2008 PHDBPD Zhai Chaoqin

Performance Modeling of Desiccant
Wheel Design and Operation
Chaoqin Zhai
School of Architecture
Carnegie Mellon University
PhD Thesis Committee:

Khee Poh Lam (Chair)
David H. Archer
Volker Hartkopf
Michael K. Sahm
John C. Fischer
Performance Modeling of Desiccant

Wheel Design and Operation
A Thesis Submitted
in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy
in Building Performance and Diagnostics
in the School of Architecture
Pittsburgh, PA
by
Chaoqin Zhai
Spring 2008
Copyright Chaoqin Zhai, 2008. All rights reserved.
Copyright Declaration
I hereby declare that I am the sole author of this thesis.
I authorize Carnegie Mellon University, Pittsburgh, Pennsylvania to lend this thesis to other
institutions or individuals for the purpose of scholarly research.
I authorize Carnegie Mellon University, Pittsburgh, Pennsylvania to reproduce this thesis by

photocopying or by other means, in total or in part, at the request of other institutions or
individuals for the purpose of scholarly research.
Copyright 2008 by Chaoqin Zhai
-i-
Acknowledgment
First and foremost, I gladly acknowledge my debt to Dr. Khee Poh Lam, chair of my thesis
committee, who has been guiding and inspiring me throughout my research. Without his constant
encouragement and advice over these years this thesis would not ever have been completed.
Im especially grateful to Dr. David H. Archer, who generously and continuously offered his time
and effort to provide me solid theoretical and technical support. His insightful thoughts and
strategic ways to problem solving during the numerous times of discussion made it possible for
me to frame the major structure of my thesis and also to detail it with systematic and logical
information.
My special thanks also go to Dr. Volker Hartkopf, Director of CBPD, for offering me the
opportunity to be a member of the fascinating IW team and for his continuous support and
encouragement during my research. I thank Dr. Michael K. Sahm from Carrier for hosting me
during my internship and offering his valuable time to communicate with me on a regular basis
regarding my research, which enabled the successful completion of my thesis in a timely fashion.
Im grateful to Mr. John C. Fischer from SEMCO for helpful discussion and quick responses to my
technical concerns every time. Its an honor of me to have them all sitting in my thesis committee.
I am very thankful to many friends and colleagues, in particular to Yi Chun Huang, Bing Dong and
Yun Gu. Many thanks also go to CBPD staff members, especially to Jim Jarrett for helping me in
my field tests with the SEMCO units.
I am grateful to my parents-in-law for taking care of my daughters, allowing me time to spend on

my research. My thanks extend to my grandparents, my dad, my brother, my brothers- and
sisters-in-law for their continuous care and support. My journey without their support would have
been unimaginable and words are just not enough to express my gratitude to all of them.
- ii -
Dedication
I dedicate this thesis to my wonderful husband Bin for all the love and support during the eight
years in the past and the many years in the future. I would also like to dedicate this thesis to my
lovely twins, Jiayu and Yujia for bringing me joy and happiness every day from the first day I
knew youre coming.
- iii -
Abstract
Desiccant coated enthalpy recovery and dehumidification devices have the potential to enhance
the dehumidification performance of HVAC systems, and thus meet the needs of improved indoor
air quality. Meanwhile, they can reduce or eliminate the energy penalty, including peak electricity
demand and overall energy consumption, associated with higher ventilation rate and better
humidity control. They make it possible for water based cooling devices such as radiant cooling
panels, water mullion and fan coils to function under humid climatic conditions, by mitigating the
impact on indoor humidity conditions.
In the past, there have been several attempts in modeling the operating performance of the
desiccant wheels. These published models are either applicable to the enthalpy recovery wheel
or the active desiccant wheel. Only a few are claimed to be applicable for both. In addition, there
has been a lack of physical understanding of the desiccant materials despite the different
moisture transport models presented in the previous publications. The practical issues
encountered in the wheel operation, such as the wheel purge, the residual moisture contained in
the desiccant materials and the impact of the wheel supporting structure, have not been
considered. Furthermore, very limited validation information has been provided for the existing
models.
This thesis presents the development of an equation based model to predict the operating
performance of both the enthalpy recovery and the active desiccant wheels, based upon
fundamental scientific and engineering principles. This model has related the desiccant wheels
performance to its design parameters and operating conditions. The moisture transfer processes
have been developed based on the physical analysis of desiccant materials. The effect of the
practical issues on the operating performance of desiccant wheels has also been considered.
- iv -
The model has been validated using the experimental data collected from a field installation of the
enthalpy recovery and the active desiccant wheels. Reasonable agreements between the
simulated and the measured performance parameters have been obtained, which indicates that
the model well represents all significant mechanisms occurring in the desiccant wheels. This
model can be used in selecting design parameters and operating variables, and in diagnosing
experimental data for the desiccant wheels.
This model has been applied to evaluate the costs and benefits provided by the enthalpy
recovery wheel. It is shown that the enthalpy recovery is an economic design feature. When
properly applied, its payback is immediate for most climatic conditions.
As a thermally activated device, the active desiccant wheel represents a good candidate for CHP
integration. This performance model has been used to develop operating strategies for the active
desiccant wheel integrated CHP system.
-v-
Table of Contents
Copyright@2008 by Chaoqin Zhai ........................................................................................ i

Acknowledgement ................................................................................................................... ii
Dedication ............................................................................................................................... iii
Abstract................................................................................................................................... iv
Chapter 1 Introduction............................................................................................................1
1.1 Background and Motivation ........................................................................................................................1
1.2 Literature Review.........................................................................................................................................2
1.2.1 Desiccant Wheels.................................................................................................................................2
1.2.2 Solid Desiccant Materials ....................................................................................................................8
1.2.3 Performance Modeling of Desiccant Wheels.......................................................................................9
1.2.4 Research Gaps in Desiccant Wheel Modeling ...................................................................................12
1.3 Research Objective ....................................................................................................................................13
1.4 Research Approach....................................................................................................................................14
1.5 Thesis Chapter Overview...........................................................................................................................15
Chapter 2 Development of the Performance Model ...........................................................17

2.1 Problem Formulation ................................................................................................................................18
2.1.1 Model Assumptions ...........................................................................................................................23
2.1.2 Governing Equations..........................................................................................................................24
2.1.3 Boundary Conditions .........................................................................................................................26
2.1.4 The Adsorption Isotherm of the Desiccant ........................................................................................26
2.2 Model Development ...................................................................................................................................28
2.2.1 Convective Heat and Mass Transfer Coefficients..............................................................................29
2.2.2 The Explicit Finite Difference Formulation.......................................................................................29
2.3 Modeling Results .......................................................................................................................................36
2.3.1 Modeling Results for the Enthalpy Recovery Wheel.........................................................................38
2.3.2 Modeling Results for the Active Desiccant Wheel ............................................................................45
2.4 Performance Indicators of Desiccant Wheels............................................................................................54
2.4.1 Performance Indicators for the Enthalpy Recovery Wheel................................................................54
2.4.2 Performance Indicators for the Active Desiccant Wheel ...................................................................56
2.4.3 Performance Indicators for the Simulated Desiccant Wheels ............................................................58
2.5 Discussion..................................................................................................................................................58
2.5.1 Step Sizes in Space and Time Domains .............................................................................................58
2.5.2 Energy and Moisture Storage in the Air.............................................................................................61
2.5.3 Heat Capacity of the Substrate...........................................................................................................63
2.5.4 Heat Conduction through the Substrate .............................................................................................64
2.5.5 Wheel Purge.......................................................................................................................................65
2.5.6 Residual Water in the Desiccant Material..........................................................................................67
2.5.7 Wheel Supporting Structure...............................................................................................................69
2.5.8 Saturation Vapor Pressure..................................................................................................................71
- vi -
2.6 Various Uses of the Performance Model ...................................................................................................72

2.7 Summary ....................................................................................................................................................74
Chapter 3 Validation of the Performance Model................................................................78

3.1 Experiment Platform..................................................................................................................................78
3.1.1 SEMCO REV 2250............................................................................................................................79
3.1.2 SEMCO FVR 2000 ............................................................................................................................82
3.2 Experiment Setup .......................................................................................................................................84
3.2.1 Experiment Setup on the Enthalpy Recovery Wheel .........................................................................85
3.2.2 Experiment Setup on the Active Desiccant Wheel ............................................................................94
3.3 Validation Results......................................................................................................................................98
3.3.1 Validation Results of the Enthalpy Recovery Wheel.........................................................................98
3.3.2 Validation Results of the Active Desiccant Wheel ..........................................................................109
3.4 Uncertainty Analysis................................................................................................................................116
3.5 Summary ..................................................................................................................................................121
Chapter 4 Integration of the Enthalpy Recovery Wheel in HVAC System Design.......124

4.1 Evaluation Procedure ..............................................................................................................................124
4.2 Discussion................................................................................................................................................133
4.3 Summary ..................................................................................................................................................135
Chapter 5 Integration of the Active Desiccant Wheel in CHP System Design ..............137
5.1 Operating Performance and Cost of the Active Desiccant Wheel ...........................................................137
5.2 Development of Operating Strategies for the Active Desiccant Integrated CHP System ........................140
5.3 Discussion................................................................................................................................................145
5.4 Summary ..................................................................................................................................................147
Chapter 6 Contribution, Conclusion and Future Work ..................................................148

6.1 Contributions ...........................................................................................................................................148
6.2 Conclusions .............................................................................................................................................153
6.3 Future Work.............................................................................................................................................155
Reference ..............................................................................................................................159
Appendix 1............................................................................................................................162
Appendix 2............................................................................................................................165
Appendix 3............................................................................................................................159
Appendix 4............................................................................................................................173
Appendix 5............................................................................................................................181
- vii -
List of Figures
Figure 1-1 Rotary Air-to-air Enthalpy Recovery Wheel .......................................................... 3

Figure 1-2 Honeycomb Structure of the Desiccant Wheels...................................................... 3
Figure 1-3 Sorption Isotherms of Various Desiccants.............................................................. 5
Figure 1-4 Desiccant Dehumidification Wheel ........................................................................ 6
Figure 1-5 Purge Section of Desiccant Wheels ........................................................................ 7
Figure 1-6 Supporting Structure of Desiccant Wheels ........................................................... 11
Figure 2-1 Schematic of the Desiccant Wheel and Its Airflow Channel Used in the Model . 19
Figure 2-2 SEM Images of the Desiccant Materials............................................................... 20
Figure 2-3 Moisture Transfer Processes in Desiccant Wheel................................................. 21
Figure 2-4 Schematic of the Finite Difference Representation of the Desiccant Wheel Model
............................................................................................................................... 30
Figure 2-5 Adsorption Isotherm of 3 Molecular Sieves and Silica Gel Used in the
Simulation ............................................................................................................. 38
Figure 2-6 Psychrometric Chart Representation of the Enthalpy Recovery Wheel Operation
............................................................................................................................... 39
Figure 2-7 Schematic of the Enthalpy Recovery Wheel Used in the Simulation................... 40
Figure 2-8 Profile of the Air and Desiccant Temperature, with the Rotation of the Enthalpy
Recovery Wheel.................................................................................................... 43
Figure 2-9 Profile of the Water Vapor Concentration in the Air and at the Interface of the Air
and the Desiccant, with the Rotation of the Enthalpy Recovery Wheel............... 44
Figure 2-10 Profile of the Moisture Loading in the Desiccant, with the Rotation of the
Enthalpy Recovery Wheel................................................................................... 44
- viii -
Figure 2-11 Psychrometric Chart Representation of the Active Desiccant Wheel Operation 46
Figure 2-12 Schematic of the Active Desiccant Wheel Used in the Simulation .................... 46
Figure 2-13 Profile of the Air and Desiccant Temperature, with the Rotation of the Active
Desiccant Wheel, rpm=0.42 ................................................................................ 50
Figure 2-14 Profile of the Water Vapor Concentration in the Air and at the Interface of the
Air and the Desiccant, with the Rotation of the Active Desiccant Wheel,
rpm=0.42 ............................................................................................................. 51
Figure 2-15 Profile of the Moisture Loading in the Desiccant, with the Rotation of the Active
Desiccant Wheel, rpm=0.42 ................................................................................ 51
Figure 2-16 Zoom in to the Water Vapor Concentration in the PA Outlet, with the Rotation
of the Active Desiccant Wheel, rpm=0.42 .......................................................... 52
Figure 2-17 Zoom in to the Water Vapor Concentration in the RgA Outlet, with the Rotation
of the Active Desiccant Wheel, rpm=0.42 .......................................................... 52
Desiccant Wheel, rpm=1.5 .................................................................................. 53
Air and the Desiccant, with the Rotation of the Active Desiccant Wheel, rpm=1.5
............................................................................................................................. 53
Figure 2-20 The Effect of the Number of Discretization in Space Domain on the Predicted
Performance of the Enthalpy Recovery Wheel ................................................... 59
Performance of the Active Desiccant Wheel ...................................................... 60
- ix -
Figure 2-22 The Effect of the Number of Discretization in Time Domain on the Predicted
Figure 2-24 The Effect of the Energy and Moisture Storage in the Air on the Predicted
Figure 2-26 The Effect of the Heat Capacity of the Substrate on the Predicted Performance of
the Enthalpy Recovery Wheel ............................................................................. 63
Figure 2-27 The Effect of the Heat Capacity of the Substrate on the Predicted Performance of
the Active Desiccant Wheel ................................................................................ 64
Figure 2-28 The Effect of the Axial Heat Conduction through the Substrate on the Predicted
Figure 2-29 Schematic of the Purge Section in the Enthalpy Recovery Wheel ..................... 66
Figure 2-30 The Effect of Purge on the Predicted Performance of the Enthalpy Recovery
Wheel................................................................................................................... 67
Figure 2-31 The Effect of Residual Water on the Predicted Performance of the Enthalpy
Recovery Wheel .................................................................................................. 68
Figure 2-32 The Effect of Residual Water on the Predicted Performance of the Active
Desiccant Wheel.................................................................................................. 69
Figure 2-33 The Impact of the Wheel Supporting Structure on the Predicted Performance of
the Enthalpy Recovery Wheel ............................................................................. 70
-x-
the Active Desiccant Wheel ................................................................................ 71
Figure 2-35 Deviation in the Saturation Vapor Pressure Calculations ................................... 72
Figure 3-1 SEMCO REV 2250 and FVR 2000 Units Installed in the IW.............................. 79
Figure 3-2 Configuration of SEMCO Revolution Unit .......................................................... 79
Figure 3-3 Flow Diagram of FVR 2000 and REV 2250 Installed in the IW.......................... 80
Figure 3-4 Psychrometric Representation of Semco FVR 2000 and REV 2250.................... 80
Figure 3-5 Solid Desiccant Dehumidification Wheel Installed in the IW .............................. 81
Figure 3-6 FVR 2000 Enthalpy Recovery Module Installed in the IW.................................. 82
Figure 3-7 The Purge Section in FVR 2000 Enthalpy Recovery Module .............................. 83
Figure 3-8 Instrumentation on the Enthalpy Recovery Module ............................................. 86
Figure 3-9 External Sensors Used to Measure the Outside Air Outlet Conditions ................ 87
Figure 3-10 External Sensors Used to Measure the Building Exhaust Air Inlet Conditions.. 88
Figure 3-11 Instrumentation on the Active Desiccant Module............................................... 95
Figure 3-12 External Sensors Used to Measure the Process Air Inlet Conditions of the Active
Desiccant Wheel.................................................................................................. 95
Figure 3-13 OA and RA Conditions during the Summer Experiment.................................... 99
Figure 3-14 Sensible Heat Balance during the Summer Experiment ................................... 100
Figure 3-15 Moisture Balance during the Summer Experiment........................................... 100
Figure 3-16 Enthalpy Balance during the Summer Experiment........................................... 101
Figure 3-17 Simulated and Measured Sensible Heat Exchange ........................................... 101
Figure 3-18 Simulated and Measured Sensible Heat Recovery Effectiveness ..................... 103
- xi -
Figure 3-19 Difference between Simulated and Measured Sensible Heat Recovery
Effectiveness ..................................................................................................... 104
Figure 3-20 Simulated and Measured Moisture Exchange................................................... 105
Figure 3-21 Simulated and Measured Moisture Recovery Effectiveness............................. 105
Figure 3-22 Difference between Simulated and Measured Moisture Recovery Effectiveness
........................................................................................................................... 106
Figure 3-23 Simulated and Measured Enthalpy Exchange................................................... 107
Figure 3-24 Simulated and Measured Enthalpy Recovery Effectiveness............................. 107
Figure 3-25 Difference between Simulated and Measured Enthalpy Recovery Effectiveness
........................................................................................................................... 108
Effectiveness during the Winter Experiment .................................................... 109
Figure 3-27 Sensible Heat Balance of the Active Desiccant Wheel during the Winter
Experiment ........................................................................................................ 110
Figure 3-28 Comparison between the Calculated and Measured Regeneration Air Inlet
Temperature of the Active Desiccant Wheel during the Winter Experiment ... 111
Effectiveness of the Active Desiccant Wheel during the Winter Experiment .. 112
Figure 3-30 Enthalpy Balance of the Active Desiccant Wheel during the Summer Experiment
........................................................................................................................... 112
Temperature of the Active Desiccant Wheel during the Summer Experiment . 113
- xii -
Figure 3-32 Moisture Balance of the Active Desiccant Wheel during the Summer Experiment
........................................................................................................................... 114
Figure 3-33 Simulated and Measured Regeneration Efficiency ........................................... 115
Figure 3-34 Simulated and Measured Heat Carryover Ratio................................................ 115
Figure 3-35 Comparison between the Simulated and Measured Sensible Heat Recovery
Effectiveness with Uncertainty Range during the Summer Experiment........... 117
Figure 3-36 Comparison between the Simulated and Measured Moisture Recovery
Figure 3-37 Comparison between the Simulated and Measured Total Heat Recovery
Effectiveness with Uncertainty Range for the Enthalpy Recovery Wheel during
the Winter Experiment ...................................................................................... 119
Effectiveness with Uncertainty Range for the Active Desiccant Wheel during the
Winter Experiment ............................................................................................ 120
Figure 3-40 Comparison between the Simulated and Measured Regeneration Efficiency with
Uncertainty Range during the Summer Experiment ......................................... 120
Figure 3-41 Comparison between the Simulated and Measured Heat Carryover Ratio with
Uncertainty Range during the Summer Experiment ......................................... 121
Figure 4-1 Hourly Load Reduction on the Cooling Coil ...................................................... 131
Figure 4-2 Hourly Load Reduction on the Regeneration Coil.............................................. 132
Figure 4-3 Predicted Annual Energy Savings at Different Location[9] ............................... 135
- xiii -
Figure 5-1 Enthalpy Removal Breakdown ........................................................................... 138

Figure 5-2 Moisture Removal Breakdown ........................................................................... 139
Figure 5-3 Operating Cost Breakdown ................................................................................. 139
Figure 5-4 Predicted Supply Air Conditions with Relation to Wheel Rotary Speed and
Bypass Ratio, Regeneration Flowrate 0.0944 m3/s............................................. 143
Bypass Ratio, Regeneration Flowrate 0.189 m3/s............................................... 143
- xiv -
List of Tables
Table 1-1 Literature Review of Desiccant Wheel Modeling .................................................. 12

Table 2-1 Design Parameters of the Desiccant Wheels Used in the Simulation .................... 37
Table 2-2 Inlet and Predicted Average Outlet Air Conditions for the Enthalpy Recovery
Wheel...................................................................................................................... 39
Table 2-3 Inlet and Predicted Average Outlet Air Conditions for the Active Desiccant Wheel
................................................................................................................................ 45
Table 2-4 Parameters of the Wheel Supporting Structure ...................................................... 70
Table 3-1 Instrumentation Specification................................................................................. 89
Table 3-2 The Data/Equation for the Enthalpy Recovery Wheel Performance Calculation .. 91
Table 3-3 The Data/Equation for the Active Desiccant Wheel Performance Calculation ..... 97
- xv -
Nomenclature
Abbreviations:
ARI
Air-Conditioning & Refrigeration Institute
ASHRAE
American Society of Heating, Refrigeration and Air-Conditioning Engineers
cfm
cubic feet per minute
CHP
Combined Heating and Power
CMU
COP
Coefficient of Performance
DOAS
Dedicated Outdoor Air System
DW
Dehumidification Wheel (Active Desiccant Wheel)
EA
Building Exhaust Air Outlet
ESS
Energy Supply System
EW
Enthalpy Recovery Wheel
HVAC
Heating, Ventilating and Air Conditioning
IW
Intelligent Workplace
LDW
Leaving Active Desiccant Wheel
LDX
Leaving DX Coil
LEW
Leaving Enthalpy Wheel
MRC
Moisture Removal Capacity
NTU
Number of Transfer Unit
OA
Outside Air
PA
Process Air
RA
Return Air
RcA
Recirculation Air
RgA
Regeneration Air
RH
Relative Humidity
rpm
revolution per minute
RSHI
Regeneration Specific Heat Input

- xvi -
SA
Supply Air
SEM
Scanning Electronic Microscopy
VFD
Variable Frequency Drive
VAV
Variable Air Volume
Parameters/Variables:
A
cross sectional area of the airflow channel, the desiccant or substrate layer, m2
Bi
Biot number
separation factor
Cp
specific heat, J/kg-K
diffusivity, m2/s
calculated variable
enthalpy or convective heat transfer coefficient, J/kg or W/m2-K
hm
convective mass transfer coefficient, kg/m2-s
heat of adsorption or vaporization, J/kg
iads
number of discretization in the adsorption section
it
indicator of the element in t domain
ix
indicator of the element in x domain
thermal conductivity, W/m-K
depth of the desiccant wheel, m
Le
Lewis number
molecular weight, kg/mole
Mass flowrate, kg/s
No_x
number of discretization in the space (x) domain
No_t
number of discretization in the time (t) domain
Nu
Nusselt number
perimeter length of the airflow channel or pressure, m or Pa
universal gas constant, R = 8.314 J/mole-K

- xvii -
time or temperature, s or oC
temperature in Kelvin, K
timestep, s
air velocity, m/s
uncertainty
distance in axial direction, m
grid size in x domain, m
moisture loading in the desiccant, kg moisture/kg dry desiccant
sensible, latent heat or enthalpy recovery effectiveness, regeneration efficiency,

heat carryover ratio
density, kg/m3
thickness of the desiccant layer, m
relative water vapor concentration
Subscripts:
amb
ambient
carryover
efan
exhaust fan
air
in
inlet
latent
desiccant matrix or mass transfer
max
maximum
min
minimum
out
outlet
pair
process air
pg
purge
regeneration
- xviii -
rair
regeneration air
sensible
sub
substrate
enthalpy
water vapor
- xix -
Chapter 1 Introduction
1.1 Background and Motivation
Previous research has associated increasing building ventilation rate with improved occupants health
and productivity[1]. Under some circumstances, increasing ventilation rate can however elevate the
indoor humidity level if the moisture in the ventilation air is not properly managed. Elevated indoor
humidity has negative health impacts such as respiratory diseases and mold growth[2]. With elevated
indoor humidity, the building occupants tend to adjust the thermostat to a lower setting in order to
achieve acceptable indoor comfort, which translates into deteriorated building energy efficiency.
Increasing ventilation rate can also be directly related to more energy consumption since the Heating,
Ventilating and Air Conditioning (HVAC) system has to cool or heat more outside air.
Therefore, building designers and mechanical engineers are faced with two issues. On one hand,
they need to meet the increasing requirement for ventilation rate, which brings in significant amount of
moisture while providing breathing air for the occupants and diluting the indoor pollutants. On the
other hand, they have to meet the needs of better indoor humidity control, and therefore provide
healthier, more comfortable and productive environment for building occupants. Moreover, they would
like to meet both requirements in an energy efficient and environment friendly manner to support the
sustainable building programs.
There are many ways to do building ventilation. Dedicated Outdoor Air System (DOAS) based on
desiccant coated enthalpy recovery and dehumidification devices provides a solution to the above
dilemma encountered by the building professionals. Desiccant coated enthalpy recovery and
dehumidification devices, either deployed together or individually, can enhance the dehumidification
performance of HVAC systems, and thus meet the needs of improved indoor air quality[3,4,5].
Meanwhile, they can reduce or eliminate the energy penalty, including the peak electricity demand
-1-
and overall energy consumption, associated with higher ventilation rate and better humidity control[6].
Desiccant coated enthalpy recovery devices fit well in current green building programs and are
required by ASHRAE 90.1-2004 for systems where the design supply air is above 2.4 m3/s (5000 cfm)
and the outside airflow is above 70% of total airflow[7] in the building. Such a provision excludes
many potential applications, such as office buildings or school facilities, where enthalpy recovery
wheels can be economically justified. In fact, the ASHRAE system capacity requirement is likely to
drop with the next version of the standard.
They make it possible for water-based cooling devices such as radiant cooling panels, water mullions
and fan coils to function effectively under humid climatic conditions, by mitigating the impact on indoor
humidity conditions. It is estimated by TIAX that radiant cooling system together with DOAS has the
potential to save 15-50% on space cooling and 20-30% on air moving (ventilation) power, compared
to conventional Variable Air Volume (VAV) system[8].
Since desiccant coated devices reduce the load on subsequent cooling/heating coils, the required
capacity of the coils can be decreased, which reduces the capital investment and helps to pay for the
deployment of desiccant devices themselves[9]. As a thermally activated device, the active desiccant
equipment provides the opportunity to utilize the rejected heat from power generation or other thermal
processes for regeneration[10], as well as the thermal energy from solar thermal receivers. By
improving the overall system energy efficiency, desiccant coated devices can also reduce the carbon
footprint of the building and its mechanical system.
1.2 Literature Review

1.2.1 Desiccant Wheels
Desiccant coated devices have different forms. A desiccant coated rotary wheel, such as the enthalpy
recovery wheel shown in Figure 1-1, is one of the most common types in commercial applications due
to its high performance as a result of its large heat and mass transfer area. This area results from its
-2-
honeycomb structure pictured in Figure 1-2, which provides numerous small, parallel channels. Air
flows through these channels, exchanging heat and moisture with the surrounding desiccant
composite structure. As indicated in Figure 1-1, the alternating ventilation and building exhaust
airflows through the channels are counter current in the two sections of the wheel, to achieve the
optimized performance of the rotating wheel. As a result of the numerous small channels, the airflows
in the channels are of laminar characteristics.
Figure 1-1 Rotary Air-to-air Enthalpy Recovery Wheel[12]
Figure 1-2 Honeycomb Structure of the Desiccant Wheels

The honeycomb structure of the desiccant wheel was developed by Carl Munter, a Swedish inventor,
in the late 1950s for low humidity applications in the defense industry[11]. The desiccant composite
usually includes the substrate and the desiccant material, although no substrate is used in some
dehumidification wheels. The substrate is normally a foil made from aluminum, stainless steel or fiber
paper, which provides structural integrity and strength. The coating process involves adhering the
particles of desiccant material to both sides of the substrate using some binder materials. Early
-3-
desiccant wheels used honeycomb paper impregnated with lithium chloride, which functioned as the
desiccant. This type of wheel was relatively easy to manufacture. However, deliquescence, loss of
the desiccant material during operation, is a major problem associated with these wheels due to the
nature of lithium chloride. More recently, silica gel and molecular sieves have been used because
they are stable and do not deliquesce[12]. Compared to lithium chloride, silica gel and molecular
sieves have lower equilibrium capacity as seen from Figure 1-3[14]. Therefore, the loading of silica
gel or molecular sieves has to be higher. The original coating method is no longer applicable. New
manufacturing techniques have emerged. One of them is to mix the desiccant material with pulp and
binder and to make desiccant paper from this mixture. The paper is then corrugated and wound into a
desiccant wheel[13]. The other technique is to form the silica gel in-situ by making a honeycomb
wheel from a glass fiber paper backbone which is first impregnated with concentrated water glass and
then reacted with an acid wash[15].
The airflow channels can have different shapes such as triangle, sinusoidal and square, but the
sinusoidal shape is preferred[16]. The height of the channel ranges from 0.5 to 2.5 millimeters. The
width ranges from 0.7 to 5 millimeters[13].
There are two types of desiccant coated rotary wheels commonly used in HVAC systems: the
enthalpy recovery wheel and the active desiccant wheel. For building ventilation purposes, the
enthalpy recovery wheel operates between the outside and the building exhaust air streams, as
shown in Figure 1-1. During the summer cooling season, the outside air is warmer and, most likely,
more humid. Heat and moisture are transferred from the outside air to the channels of the enthalpy
recovery wheel. Since the building exhaust air is cooler and less humid, heat and moisture are
transferred back from the wheel to the exhaust air. During the winter heating season, heat and
moisture are transferred from the building exhaust air to the wheel and then transferred back from the
wheel to the outside air. Heat and moisture transfer between the outside and the building exhaust air
streams are thus accomplished through the channels of the rotating wheel.
-4-
Figure 1-3 Sorption Isotherms of Various Desiccants[14]

Since there is no external thermal energy input besides the limited amount of power input to rotate
the wheel, the enthalpy recovery wheel is a passive desiccant wheel, in contrast to an active
desiccant wheel, as shown in Figure 1-4. An active desiccant wheel, also known as a
dehumidification wheel, transfers moisture from the outside air supply stream to the heated
regeneration air stream. The regeneration air can be the combustion product from a direct-fired gas
burner. It can also come from a solar thermal collector, or a sensible heat exchanger which recovers
the rejected heat from the power generation process. These alternatives will be explored in Chapter 5
-5-
of this thesis. Since the temperature of this regeneration air is higher than that of the building exhaust
air, the outlet ventilation air humidity from the active desiccant wheel is not limited by the building
exhaust air conditions. The objective of the active desiccant wheel is to remove moisture from the
ventilation air stream, but the associated temperature increase needs to be counteracted for effective
system performance.
Figure 1-4 Desiccant Dehumidification Wheel[12]

For an enthalpy recovery wheel, both the heat and the mass transfer rates are important in most
applications. The desiccant material absorbs moisture and has relatively low heat capacity. The
substrate does not absorb moisture, but has a relatively high heat capacity. The mass fraction of the
desiccant material is relatively low, roughly 15 - 30%[13]. For the enthalpy wheels currently available
in the market, the thickness of the substrate is roughly 50 microns. The desiccant coating is roughly
25 microns on each side of the substrate[13]. The rotary speed of the wheel is approximately 20 - 30
rpm in order for desirable heat and mass transfer performance. The wheel split for the outside and
the building exhaust air streams is about 50/50. In order to prevent cross contamination between the
outside and the building exhaust air streams, the enthalpy wheel is built with a purge section, as
illustrated in Figure 1-5. In this section, the incoming outside air flows into the channels of the wheel
and drives the exhaust air contained in these channels into the incoming building exhaust air stream.
The required purge area is dependent on the depth of the wheel, the velocity of the purge gas flow
and the rotary speed of the wheel.
-6-
Some desiccant materials, such as 3 zeolite molecular sieves, used in the wheel can further prevent
cross contamination between the outside and the building exhaust air streams, as explained in
Chapter 3 of this thesis. As a note both the wheel purge and the selected desiccant material are only
effective to prevent the contaminants in the building exhaust air stream from entering the building
ventilation air. They have no effect on the contaminants that is contained in the outside air.
Figure 1-5 Purge Section of Desiccant Wheels[16]

For the active desiccant wheel, moisture transfer is of higher importance than heat transfer. Therefore
the mass fraction of the desiccant material is relatively high, roughly 50 - 60%, in order to increase
the moisture removal capacity. The desiccant coating is usually thicker than 25 microns[12]. The heat
capacity of the desiccant matrix is relatively low in order to minimize the heat carryover from the
regeneration side, because the regeneration air is a heated stream with a temperature of roughly
100oC. The wheel rotates at a much slower rate than the enthalpy recovery wheel, in order to provide
enough time for the moisture adsorption and to pre cool the regenerated desiccant before it can pick
up moisture again.
In order for better dehumidification capability, the adsorption capacity of the desiccant in the active
desiccant wheel is expected to be higher than that in the enthalpy recovery wheel. Due to the
different operating conditions of the active desiccant and the enthalpy recovery wheels, desired
adsorption characteristics of the desiccant materials for the two applications are different as well. The
desiccant material in the active desiccant wheel functions in a larger range in terms of temperature
and moisture concentration. It has been reported in the literature[17,18] that desiccants with Langmuir
-7-
Moderate Type 1 adsorption isotherm are desirable for the active desiccant wheel. Due to the higher
desorption rate resulting from the elevated temperature of the regeneration air, the regeneration
section of the active desiccant wheel is usually smaller than the adsorption section, resulting in a
shorter desorption period than the adsorption. The different functions of the enthalpy recovery and the
active desiccant wheels are likely to require different wheel design parameters such as the channel
size and wheel depth, and operating conditions such as the air velocity in the channel, which will be
explored in this thesis.
In addition, the building exhaust air stream is needed in the operation of the enthalpy recovery wheel.
Therefore the building outside air intake and the exhaust air outlet need to be located close to each
other.
1.2.2 Solid Desiccant Materials

The desiccant isotherm is usually used to represent the sorption characteristics of a desiccant. As
plotted in Figure 1-3, it shows the equilibrium sorption capacity of a desiccant at a certain temperature
under different humidity conditions. A set of isotherms at different temperatures is required in order to
fully characterize the sorption capacity of a desiccant. The equilibrium capacity is important in terms
of the diffusion limitation and overall heat capacity effects. The shape of the isotherm is dependent on
the dominant sorption mechanisms of the desiccants.
The heat of adsorption, which is equivalent to the heat of vaporization plus the heat of wetting, is
another parameter that affects the performance of desiccant wheels. Adsorption is an exothermic
process and low heat of wetting is desirable in order to achieve better wheel performance.
The adsorption rate is also important for desiccant wheels. It has been reported[17] that the
adsorption rate is very dependent on the desiccant thickness. The adsorption is slower for thicker
desiccant coating. The adsorption rate is also limited for salt-impregnated samples due to the pore
pluggage and the hydration of the salt.
-8-
There are some other factors that need to be considered when selecting desiccant materials:
performance degradation of the desiccant over repetitive adsorption and regeneration cycles, long
term stability and stability with combustion products when used in direct fired systems, the need for
binding material, and safety for using in air conditioning systems, etc.
1.2.3 Performance Modeling of Desiccant Wheels

The performance of desiccant wheels can be modeled by a set of equations, which represent the
conservation of energy and mass, the rates of heat and moisture transfer, and the adsorption
equilibrium. The equations have been solved either by applying the analogy between the heat and
mass transfer or numerically by using finite difference and finite volume techniques.
In early 1970s, the modeling of the enthalpy recovery wheel was based on the analogy between heat
and mass transfer. The sensible heat recovery wheel, also known as the periodic flow heat
exchanger, has been used for a long time in power plants, recovering the sensible heat from the
exhaust air to preheat the incoming combustion air. The heat transfer in these wheels has been
extensively studied. Comprehensive design theory has been presented[20]. By applying the analogy
theory, Maclaine-cross and Banks[21] developed a model for solid desiccant coated enthalpy
exchangers and dehumidifiers. By introducing new dependent variables, known as characteristic
potentials, to replace the enthalpy and the moisture content, the differential equations for coupled
heat and mass transfer were reduced to a set of uncoupled differential equations. Further
assumptions were made to convert the equations to a linear form. The solutions for sensible rotary
heat exchangers, such as the ones presented by Kays and London[20], can be used to solve the
characteristic potentials and thus the enthalpy and the moisture content in the air.
When the computational power was limited, using analogy between heat and mass transfer is
necessary because the solutions for heat transfer are readily available. However, this advantage
became less significant with the development of computers and numerical techniques[22].
-9-
Van den Bulck and Mitchell et al[23,24] developed an Effectiveness-Number of Transfer Units (-NTU)
method for the active desiccant wheel. The model was established in two steps. During the first step,
the governing equations for ideal dehumidifiers with infinite overall heat and mass transfer coefficients
were solved analytically by the method of characteristics and shock wave method. During the second
step, the performance of a silica gel coated rotary dehumidifier with finite heat and mass transfer
coefficients was empirically correlated to that of ideal dehumidifiers by using a finite difference model
MOSHMX. This model was then used to explore the impacts of different operating variables, such as
the regeneration air flowrate, regeneration temperature and wheel rotary speed, in order to maximize
the wheel performance based on the thermal and electrical energy input[25].
Zheng and Worek[26] developed a one dimensional transient model to simulate the simultaneous
heat and mass transport processes involved in the rotary desiccant dehumidifier. The governing
equations were solved using implicit finite difference method, which allowed the numerical scheme to
be unconditionally stable. The simulation results were compared with the predictions from other
programs. As applications of this model, the impacts of certain design parameters, such as the
separation factor and the maximum moisture uptake of the desiccant material and the number of
transfer units, and certain operating variables, such as the inlet temperature and humidity ratio of the
process and regeneration air, on the rotary dehumidifier performance were investigated[27,28,29].
Simonson and Besant[30,31] presented a one dimensional transient model to simulate the heat and
moisture transfer in enthalpy wheels. Different from previous publications, they assumed that part of
the heat of adsorption was conducted to the desiccant matrix and the rest was convected to the air
stream. The model was validated using experimental data and reasonable agreement between the
experiment and the model predictions was achieved. They used this model to study the sensitivity of
certain assumptions of the model and the operating variables of the enthalpy wheel. This model was
also used by Jeong and Mumma[32] to develop simple performance correlations for molecular sieves
and silica gel coated enthalpy wheels under normal operating conditions. These correlations are only
- 10 -
valid for molecular sieves and silica gel coated enthalpy wheels since the desiccant property
parameters are embedded in the development of these correlations.
The above models only included the convective heat and mass transfer resistance at the gas solid
interface. Besides the resistance at the interface, the solid side resistance for heat and mass transfer
were also considered in Majumdars model[33]. The moisture transport in the desiccant matrix was
represented by gas diffusion and surface diffusion resistances. Zhang and Niu[34]and Sphaier and
Worek[35] presented two dimensional models that considered both heat and mass transfer
resistances in both axial and thickness directions of the solid desiccant. These two models were
claimed to be applicable for both the enthalpy recovery and the active desiccant wheels. Both models
gave some insights into the heat and mass transfer processes in the desiccant matrix, which was not
provided in previous publications. However, both models are more based on abstracted mathematics
than physical understanding of the desiccant material. Whether the models truly described what is
going on in reality remains unknown. As a matter of fact, very limited validation information, if at all,
was provided in these papers.
In addition, none of the previous models considered the practical issues in the wheel operation, which
include the wheel purge shown in Figure 1-5, the residual water contained in the desiccant material
and the wheel supporting structure such as the spokes and the casing shown in Figure 1-6.
Figure 1-6 Supporting Structure of Desiccant Wheels
- 11 -
The literature review of desiccant wheel modeling is summarized in Table 1-1 as follows.
Table 1-1 Literature Review of Desiccant Wheel Modeling
Numerical
Application
Validation
Practical
issues
Maclaine-cross
and Banks [21]
No
DW
limited lab
data
No
Van den Bulck, et

al. [23,24]
No
DW
No
No
Zheng and Worek

[26]
Yes
DW
other model
No
Simonson and
Besant [30,31]
Yes
EW
field and lab

data
No
Majumdar
[ref1998]
Yes
DW
No
No
Zhang and Niu

[34]
Yes
DW, EW
No
No
Sphaier and
Worek [35]
Yes
DW, EW
limited lab
data
No
Zhai
Yes
DW, EW
field data
Yes
1.2.4 Research Gaps in Desiccant Wheel Modeling

Desiccant coated enthalpy recovery and active desiccant wheels share many similarities in wheel
structure, the use of solid desiccant materials and the combined heat and mass transfer processes. It
is expected that one performance model will be applicable for both applications, which has been
achieved by very few published models. In addition, modeling of the moisture transport processes is
one of the major tasks in the performance simulation of desiccant wheels. It requires physical
understanding of the desiccant materials, which is not provided in any of the published models.
Furthermore, the investigation of the practical issues is as important as the analysis of the combined
heat and mass transfer processes in determining the operating performance of desiccant wheels in
field installations. As pointed out before, none of the previous publications took these practical issues
into consideration.
A new performance model for both the enthalpy recovery and the active desiccant wheels, which
describes the heat and mass transfer processes based on the physical understanding of the
- 12 -
desiccant materials and considers the impact of the practical issues, is therefore needed. This model
needs to be validated through extensive field experimental data, since field experiment is the only
way to determine the desiccant wheel performance in a field installation.
In addition, there is a lack of unified performance indicators for the active desiccant wheels. Various
indicators have been used in the published literature: dehumidification effectiveness[34], specific
dehumidification power[34], coefficient of performance[27,28,29], moisture removal capacity[36],
regeneration specific heat input[36], etc. No information has been found on why these indicators have
been chosen. There is a need for unified performance parameters that can be used to make
comparative evaluations of the operating performance regardless of the design and make of the
desiccant wheels and to calculate the outlet air conditions from the desiccant wheels based on the
inlet air conditions.
Given the energy, environment and cost benefits provided by the enthalpy recovery wheels, they
have gained increasing attention from the HVAC engineers and building owners. A procedure that
helps the interested parties to evaluate the applicability and economics of the enthalpy recovery
wheels will be desirable in order for the wider deployment of this device.
As a thermally activated device, the active desiccant wheel makes a great candidate for utilizing the
rejected heat from the power generation processes, in order to improve the overall energy efficiency.
There is a need for a design procedure that integrates the active desiccant wheel in a Combined
Heating and Power (CHP) system and develops the operating strategies for the integrated system.
Such a procedure does not yet exist.
1.3 Research Objective

The objective of this thesis is to address the research gaps described in the previous section. More
specifically, the research work in this thesis aims:
to develop and implement an equation based model for the performance modeling of solid
- 13 -
desiccant wheels based on fundamental scientific and engineering principles, not model
fitting of experimental data. The major features of this model are shown in Table 1-1.
to validate the model using experimental data from field experiments
to compare and contrast the design and operation between the enthalpy recovery and the
active desiccant wheels, and to investigate the applicability of the model to both wheels
to use this model to analyze and diagnose the data collected from field experiments
to outline procedures for selecting and evaluating the costs and benefits of enthalpy
recovery wheels based on the model and experience gained in field experiments
to outline procedures for developing operating strategies for the active desiccant wheel
integrated CHP system.
1.4 Research Approach

The moisture transport model, an important aspect of the performance simulation of the desiccant
wheels, has been developed by investigating the structure of the solid desiccant materials under the
scanning electronic microscope (SEM). This model, together with heat transfer model, the energy and
material balance as well as the adsorption equilibrium equation at the air and desiccant interface,
have been assembled to form the basis of the performance simulation.
The performance model is one dimensional in the axial direction. It applies a lumped formulation in
the thickness direction of the desiccant and the substrate. The boundary conditions of this problem
represent the inlet outside/process and building exhaust/regeneration air conditions as well as the
adiabatic condition of the two ends of the desiccant composite. The solutions of this model are
iterated until the wheel reaches periodic steady state operation, which means the wheel returns to its
original condition in terms of temperature and moisture loading after one complete cycle. The
modeling
results
are
obtained
as
the
changes
of
the
outside/process
and
building
exhaust/regeneration air conditions along the wheel depth and the wheel rotation.
- 14 -
This performance model relates the wheels design parameters, such as the wheel dimension, the
channel size and the desiccant properties, and the wheels operating variables, such as the rotary
speed and the regeneration air flowrate, to its operating performance.
An active desiccant based ventilation unit, together with an enthalpy recovery module, was installed
in the Intelligent Workplace (IW)[37] at Carnegie Mellon University (CMU), as part of its energy supply
system (IWESS)[38]. This machine serves as the experimental platform for the empirical validation of
the performance model. Extensive testing data have been collected using carefully located sensors in
the machine. The testing methods used in this thesis can also be applied to other field installations of
desiccant wheels.
Applying the performance model, an evaluation procedure for the enthalpy recovery wheel and a
design procedure for the active desiccant wheel integrated CHP system have been developed, using
the installed ventilation machine in the IW as a specific example.
1.5 Thesis Chapter Overview

This thesis comprises six chapters:
Chapter 1, Introduction, introduces the motivation of this thesis work and the current research status
of the desiccant wheel modeling. It identifies the research gaps, establishes the research objectives
and approaches used in this thesis.
Chapter 2, Development of the Performance Model, presents the assumptions, governing equations
and boundary conditions used to develop the performance model of the desiccant wheels. It also
presents and interprets the modeling results for both the enthalpy recovery and the active desiccant
wheels, and investigates the impact of different parameters such as the wheel purge, substrate heat
conduction on the operating performance of both wheels. The performance indicators for both wheels
have also been developed in this chapter.
- 15 -
Chapter 3, Validation of the Performance Model, describes the experimental platform, the experiment
setup on both wheels, the data acquisition and analysis as well as the empirical validation of the
performance model. Uncertainty analysis of the experimental data has also been presented.
Chapter 4, Integration of the Enthalpy Recovery Wheel in HVAC System Design, outlines an
evaluation procedure for determining the costs and benefits provided by the enthalpy recovery wheel.
It provides guidance for HVAC engineers on the applicability and economics of the enthalpy recovery
wheel.
Chapter 5, Integration of the Active Desiccant Wheel in CHP System Design, presents the operating
energy and cost of the active desiccant wheel, based on the data obtained from the ventilation
system installed in the IW. It develops a design procedure for integrating the active desiccant wheel in
a CHP system and developing operating strategies for the integrated system for better energy
efficiency, improved indoor thermal conditions and reduced costs.
Chapter 6, Contribution, Conclusion and Future Work, summaries the major accomplishments
achieved in this thesis and conclusions drawn from the work. It also points out future areas of
research.
- 16 -
Chapter 2 Development of the Performance Model

In this chapter, an equation based model has been developed to predict the operating
performance of both the enthalpy recovery and the active desiccant wheels, based upon
fundamental scientific and engineering principles. This model relates the desiccant wheels
performance to its design parameters and operating conditions.
The design parameters include:
the wheel dimension, such as the wheel depth, the wheel diameter and the split between
adsorption and desorption sections;
the channel dimension, such as the channel shape and size;
the desiccant composite, such as the desiccant material properties and coating thickness,
as well as the substrate properties and thickness.
The operating variables include:
the rotary speed of the wheel;
the inlet outside/process air temperature, humidity and flowrate;
the building exhaust/regeneration air temperature, humidity and flowrate;
The results from the model prediction are:
the temperature and humidity of the outside/process and building exhaust/regeneration

air at any given time and location as well as the average values;
the temperature and moisture loading of the desiccant composite at any given time and
location as well as the average values.
The model can be used:
to explore different design alternatives of the enthalpy recovery and the active desiccant
wheels;
- 17 -
to predict the operating performance of the desiccant wheels once the wheel design and
operating conditions are given;
to select operating variables of the desiccant wheels for a given application.
2.1 Problem Formulation

The schematic of the desiccant wheel and a cross section of its airflow channel used in the model
development are shown in Figure 2-1. The wheel is a rotating cylinder with depth L and diameter
D. It rotates around its axis at a constant speed and the rotation of the wheel is represented by .
The wheel is split into the adsorption and desorption sections and the area of the two sections are
not necessarily equal. The wheel is sometimes built with a purge section to prevent crosscontamination between the outside/process and the building exhaust/regeneration air streams,
which is discussed in Section 2.5.5 of this chapter. The outside/process and the building
exhaust/regeneration air streams are in counter flow arrangement.
The structure and composition of the wheel are assumed homogeneous. All the channels in the
wheel are assumed identical. The wheel performance is modeled by tracking the air and
desiccant conditions in a single channel as it rotates. The channel can be of any shape, such as
sinusoidal, rectangular or circular. The channel wall is made of desiccant composite, which
consists of desiccant materials, and perhaps substrate to support the desiccants. Air flows
through the channels in the direction of the wheel axis, exchanging energy and moisture with the
desiccant composite.
Analysis of the moisture transport processes is one of the important tasks in modeling the
performance of the desiccant wheels. This task is facilitated by investigating the structure of the
desiccant materials. Figure 2-2 shows the SEM images of two desiccant materials, 3 zeolite
molecular sieves and silica gel, at various magnification levels. The first four images are for the
zeolite at 500, 2,000, 3,000 and 10,000 magnification levels, respectively. The last two are for the
silica gel at 200 and 90,000 magnification levels, respectively. Both the zeolite and the silica gel
samples were supplied by SEMCO Inc. The zeolite is coated on a thin aluminum sheet using
- 18 -
binder materials, as shown in the first two images. The appearance of this silica gel sample is
different from the dried silica gel powdered crystals found in food or electrics packages, since it is
made in-situ and not dried.
adsorption section
L
Outside /process air inlet
Outside/process air outlet
Building
exhaust/regeneration air
inlet
desorption section
Building
exhaust/regeneration air
outlet
purge section
moist air
desiccant composite
desiccant composite
r
x
moist air
desiccant composite
desiccant composite
moist air
Figure 2-1 Schematic of the Desiccant Wheel and Its Airflow Channel Used in the Model
- 19 -
3 molecular sieves (x500)
3 molecular sieves (x3,000)
Silica gel (x200)
Silica gel (x90,000)
Figure 2-2 SEM Images of the Desiccant Materials

Seem from these images, the desiccant composite is a porous matrix consisting of desiccant
particles. There are large pores between the desiccant particles, which are referred to as inter
particle or macro pores. There are also small pores residing in the desiccant particles, which are
referred to as intra particle or micro pores. The intra particle pores in the silica gel can be seen
from the last image in Figure 2-2; they are not uniform in size. The average intra particle pore size
in this silica gel is about 50 . Those in zeolite are not visible even at 10,000 magnification level
since they are only about 3 in diameter. The wall of the intra particle pores are where the
moisture is adsorbed.
The following moisture transfer model is developed by analyzing the structure of the desiccant
materials. The moisture transfers from the air flowing in the wheel channels to the intra particle
pores in two steps. In the first step, the water vapor molecules transport from the air to the
surface of the desiccant composite. This step is called the gas side transfer. In the second step,
the moisture transfers from the surface of the desiccant layer to the inter particle pores and then
- 20 -
to the intra particle pores. Finally the water molecules get adsorbed on the surface of the intra
particle pores. Heat is released at the location where the adsorption occurs. This step is called
the solid side transfer. No water is adsorbed on the surface of the inter-particle pores. Seen from
Figure 2-2, there is not a well defined separation surface between the gas side and the solid side.
Therefore, the hypothetical surface of the desiccant layer is used. These transfer processes are
illustrated in Figure 2-3.
H2O vapor
moist air
H2O vapor
(not to scale)
Figure 2-3 Moisture Transfer Processes in Desiccant Wheel

Gas side transfer is controlled by the convective mass transfer coefficient and the vapor pressure
difference between the air and the desiccant composite. There are three diffusion mechanisms,
namely ordinary diffusion, Knudsen diffusion and surface diffusion, occurring in the desiccant
composite. The solid side transfer is controlled by the diffusion coefficient, diffusion area and the
moisture concentration difference for each diffusion mechanism. The molecules of water vapor
diffuse from the hypothetical surface of the desiccant layer to the surface of the desiccant
particles through ordinary and Knudsen diffusion in the inter particle pores. Then the water
molecules diffuse from the surface of the desiccant particles into the intra particle pores through
surface diffusion of adsorbed water molecules as well as ordinary and Knudsen diffusion of water
vapor molecules in the intra particle pores. Depending on the size of the intra particle pores,
- 21 -
ordinary and Knudsen diffusion in the intra particle pores might be neglected. For example, the
intra particle pores in the 3 molecular sieves are 3 in diameter. It is just large enough to hold
the water molecule, which is 2.8 in diameter. In this case, surface diffusion of the adsorbed
water molecule is the primary means to transfer the moisture into the intra particle pores.
Despite the complexity, the controlling process for the moisture transfer in the desiccant wheels is
the moisture transport from the bulk air flowing in the channels to the hypothetical surface of the
desiccant layer, due to the small thickness of the desiccant layer.
The heat transfer Biot number is defined as[39]
Bih =
h
k
Equation 2-1
Similarly, the mass transfer Biot number can be defined as[39]
Bim =
hm
Dm
Equation 2-2
The heat transfer Biot number relates the convective heat transfer resistance between the solid
and the fluid with the resistance to thermal conduction inside the solid material. It is a measure of
the ratio of the temperature drop in the solid material and the temperature drop between the solid
and the fluid. When the heat transfer Biot number is small (Bih <0.1), most of the temperature
drop is in the fluid and the temperature of the solid can be considered uniform[39]. Similarly, the
mass transfer Biot number relates the convective mass transfer resistance between the solid and
the fluid with the resistance to mass diffusion inside the solid material. When the mass transfer
Biot number is small (Bim <0.1), most of the concentration drop is in the fluid and the
concentration in the solid can be considered uniform.
Both the heat and the mass transfer Biot numbers are small for this problem, since the desiccant
layer is very thin. For example, the coating of zeolite molecular sieves on the sample used in
- 22 -
Figure 2-2 is about 25 microns thick and the substrate is 15 microns thick. Therefore, the
temperature and moisture concentration gradient in the thickness direction (r direction in Figure
2-1) is very small. Therefore the analysis can be considered as one dimensional in the axial
direction (x direction in Figure 2-1).
The model is formulated in the same way for both adsorption and desorption sections. Depending
on the airflow direction, the air velocity is either positive or negative. Depending on the vapor
concentration in the air and the desiccant, the moisture either adsorbs onto or desorbs from the
desiccant. Depending on the temperature of the air and the desiccant, the energy either transfers
from the air to the desiccant or from the desiccant to the air.
2.1.1 Model Assumptions

The following assumptions are made in developing the model:
1. The axial heat conduction and water vapor diffusion in the air are negligible.
2. The axial water vapor and adsorbed water diffusion in the desiccant are negligible.
3. The convective heat and mass transfer rates are represented using the bulk mean air
temperature and humidity.
4. Heat conduction in the desiccant is negligible. Heat may be conducted axially through the
substrate.
5. The mid plane, indicated as dash lines in Figure 2-1 and two ends of the desiccant composite
are adiabatic and impermeable.
6. The airflow in the channel is fully developed laminar flow.
7. The heat of adsorption is released in the desiccant composite.
8. The inlet air conditions are uniform across the wheel surface, but they can vary with time.
9. Thermodynamic properties of the dry air, desiccant material, and substrate, such as density,
specific heat and heat of adsorption, remain constant during the wheel operation.
- 23 -
10. The convective heat and mass transfer coefficients remain constant during the wheel
operation. They are determined based on published coefficients between gases and solid
surfaces.
11. There is no heat or moisture storage in the wheel when it completes one rotation.
In essence, the model is one dimensional in the axial direction. It is transient, which means it
calculates the time dependent conditions of the air and the desiccant composite. It models the
simultaneous and coupled heat and mass transfer effect occurring in the wheel. It applies to both
the enthalpy recovery and the active desiccant wheels, since none of the assumptions and the
following governing equations is specific to either application. Furthermore, the wheel is in
periodic steady-state operation, which means the wheel returns to its original condition in terms of
temperature and moisture loading after one complete cycle.
2.1.2 Governing Equations

The governing equations, which describe the material and energy balance as well as the heat
and mass transfer rates, are developed based on above assumptions.
Since it is assumed that the dry air properties remain constant during the wheel operation, the air
velocity remains constant according to the mass balance of dry air. The moisture balance of the
air stream can be written as:
hm p( vg vm ) + uA
vg
x
+A
vg
t
=0
Equation 2-3
The first term in this equation represents the rate of convective mass transfer between the air and
the desiccant, which is represented by the difference in water vapor concentration between the
bulk mean air and the desiccant, and a constant convective mass transfer coefficient. The second
term represents the rate of moisture flux as a result of airflow. The third term represents the
moisture storage in the air.
- 24 -
The moisture balance of the desiccant is described as:
hm p ( vg vm ) Am m
m
=0
t
Equation 2-4
The first term in this equation represents the rate of convective mass transfer between the air and
the desiccant, corresponding to the first term in Equation 2-3. The second term represents the
moisture storage in the desiccant material.
The energy balance of the air stream is described as:
hp(tm t g ) uA g Cp g
t g
t
A g Cp g g = 0
x
t
Equation 2-5
The first term in this equation represents the rate of convective heat transfer between the air and
the desiccant composite, which is represented by the temperature difference between the bulk
mean air and the desiccant composite, and a constant convective heat transfer coefficient. The
second term represents the rate of heat flux in the air as a result of airflow, and the third term
represents the energy storage in the air. The sensible heat exchange associated with the
moisture transfer is small compared to the convective heat exchange term and it is ignored.
The energy balance of the desiccant composite is described as:
2t m
+ hm p( vg vm )H ads hp (t m t g )
x 2
t
( m AmCpm + sub AsubCpsub ) m = 0
t
k sub Asub
Equation 2-6
The first term in this equation represents the rate of heat conduction through the substrate, if it is
present in the wheel. If the substrate is not present, this term will be eliminated. The heat
conduction through the desiccant is ignored due to low heat conductivity of the material. The
second term represents the rate of heat generation as a result of moisture adsorption. The rate of
heat generation is represented by the product of the rate of moisture exchange and the heat of
adsorption. As assumed earlier, the heat of adsorption is entirely released in the desiccant
composite. The third term represents the rate of convective heat transfer between the air and the
- 25 -
desiccant composite, corresponding to the first term in Equation 2-5. The last term represents the
energy storage in the desiccant composite, including the energy stored in the desiccant and the
substrate.
2.1.3 Boundary Conditions

The boundary conditions required to solve the previous governing equations include:
the inlet outside/process and the building exhaust/regeneration air temperature and water
vapor concentration
tg
x =0
vg
tg
x =0
x=L
vg
= t pair ,in
Equation 2-7
= v , pair ,in
Equation 2-8
= t rair ,in
x= L
Equation 2-9
= v , rair ,in
Equation 2-10
the adiabatic and impermeable conditions of the two ends of the desiccant composite
t m
x
vm
x
x =0
x =0
t m
x
x=L
vm
x
=0
x=L
=0
Equation 2-11
Equation 2-12
As stated previously, the inlet air conditions can vary with time, but they are uniform across the
wheel surface. The adiabatic and impermeable conditions are required in solving the energy and
moisture balance equations of the desiccant composite.
The simulation is started with some initial conditions, but they do not influence the final modeling
results since the wheel is in periodic steady-state operation.
2.1.4 The Adsorption Isotherm of the Desiccant

There are five unknown variables in the four governing equations: temperature of the air tg,
temperature of the desiccant composite tm, water vapor concentration in the air vg, water vapor
- 26 -
concentration in equilibrium with the desiccant vm and moisture loading in the desiccant m.
Another equation is therefore needed in order to solve for the five unknown variables. The fifth
equation is the desiccant adsorption isotherm, which relates the moisture loading in the desiccant
with the relative water vapor concentration of the air that is in equilibrium with the desiccant.
A general adsorption isotherm is described as[17,18]:
m
max
1
1 c +
Equation 2-13
vm
vm, sat
Equation 2-14
The ideal gas law gives
pv =
v RT
Mv
= Rv v T
Equation 2-15
can be represented by the partial pressure of water vapor as follows.
vm RvT
Equation 2-16
p vm, sat
pvm,sat is a function of temperature only. The relationship between pvm,sat and tm can be
determined by applying the Clausius-Clapeyron equation.
ln(
H vap 1 1
p1
)=
( )
p2
R T2 T1
Equation 2-17
The application of this equation is discussed in Section 2.5.8 of this chapter.
Assuming at the standard atmospheric pressure 101,325 Pa, water boils at 373.15 K (100oC), the
saturation vapor pressure pvm,sat at temperature Tm can be calculated as
1
pvm,sat
= 1.12 *10 e
11
5196
)
Tm
Equation 2-18
where Tm = tm+ 273.15 K.

- 27 -
The following relationship is obtained by combining Equation 2-16 and Equation 2-18.
= 4.09 *10 Tm vm e
9
5196
)
Tm
Equation 2-19
Equation 2-13 relates the moisture loading of the desiccant material with its temperature and
water vapor concentration through Equation 2-19. It supplements the governing equations shown
in Equation 2-3 through Equation 2-6 and completes the formulation of the combined heat and
mass transfer problem in the performance modeling of desiccant wheels.
In summary, the governing equation set consists of:
four partial differential equations, which describe the energy and moisture balance of the
air and the desiccant composite;
one non linear algebraic equation, which represents the desiccant adsorption isotherm.
This governing equation set is subject to the boundary conditions listed in the previous section,
which represent the inlet outside/process and building exhaust/regeneration air conditions as well
as the adiabatic condition of the two ends of the desiccant composite. The initial conditions are
not critical since the wheel is in periodic steady-state operation.
By solving this governing equation set, the change of the outside/process and the building
exhaust/regeneration air conditions along the wheel depth (x domain) and the wheel rotation (t
domain) can be obtained. The temperature and moisture loading profiles of the desiccant can
also be generated.
2.2 Model Development

In developing the model, care is exercised in determining the convective heat and mass transfer
coefficients.
- 28 -
2.2.1 Convective Heat and Mass Transfer Coefficients

The convective heat transfer coefficient is calculated based on the Nusselt number for fully
developed laminar flow in a tube with constant heat flux boundary conditions[39].
h=
k g Nu
Equation 2-20
Deff
The convective mass transfer coefficient is determined by applying the heat and mass transfer
analogy[39].
hm =
h
Cp g g Le
Equation 2-21
Assuming unity Lewis number, which is related to the mechanisms for heat and mass transfer, hm
is obtained as
hm =
h
Cp g g
Equation 2-22
2.2.2 The Explicit Finite Difference Formulation

The governing equation set is solved by numerical analysis using the explicit finite difference
method. The numbers of discretization in space and time domains have been varied to determine
the appropriate grid size and timestep for reasonable computational time and accuracy, as
discussed in Section 2.5.1. The schematic used in developing the finite difference formulation is
shown in Figure 2-4. x is defined such that the velocity of the outside/process air is positive and
that of the building exhaust/regeneration air is negative. t is defined positive with the direction of
the wheel rotation.
For the finite difference analysis, the wheel is divided into No_x elements in the space (x) domain
and No_t elements in the time (t) domain, shown as horizontal and vertical grids in Figure 2-4 The
corresponding steps in the space and time domains are x and t. The indexes for the space and
time domains are represented by ix and it. The double line in the time domain in Figure 2-4
indicates the separation between the adsorption and desorption sections, which is represented by
- 29 -
it = iads+1. The integer iads is determined based on the wheel split ratio for the adsorption and
desorption sections. The wheel adsorbs moisture from the outside/process air when it=2 - iads+1;
the wheel releases moisture into the building exhaust/regeneration air when it=iads+2 - No_t+1.
The wheel is at its initial conditions when it=1.
1/2x
1/2x
it=1
process air
it-1
it
it+1
it=iads+1
regeneration air
it=No_t+1
ix=1
ix-1
ix
ix+1
ix=No_x+1
Figure 2-4 Schematic of the Finite Difference Representation of the Desiccant Wheel Model
The five unknown variables, which are functions of x and t, can now be represented as:
tg(it,ix) temperature of the air;
tm(it,ix) temperature of the desiccant composite;
vg(it,ix) water vapor concentration in the air;
vm(it,ix) water vapor concentration at the desiccant air interface, which is in equilibrium
with the desiccant;
m(it,ix) moisture loading of the desiccant, which is a function of tm(it,ix) and vm(it,ix).
- 30 -
For the adsorption section, the inlet outside/process air conditions are known boundary conditions.
In the finite difference analysis, the calculations are carried out in the direction of the airflow. The
air and desiccant conditions at the next grid in the space domain are calculated based on the
conditions at the previous grid.
The partial derivatives are represented by the forward differences as follows.
t g
x
t g
t
=
=
t g (it , ix) t g (it , ix 1)
Equation 2-23
x
t g (it , ix) t g (it 1, ix)
Equation 2-24
t m t m (it , ix) t m (it 1, ix)

=
t
t
vg
x
vg
t
=
=
Equation 2-25
vg (it , ix) vg (it , ix 1)
Equation 2-26
vg (it , ix) vg (it 1, ix)
Equation 2-27
vm vm (it , ix) vm (it 1, ix)

=
t
t
Equation 2-28
2t m t m (it 1, ix 1) 2t m (it 1, ix) + t m (it 1, ix + 1)

=
x 2
x 2
Equation 2-29
Based on the forward differences, the finite difference form of the energy balance equation of the
air for the adsorption section is:
(hp +
u p A g Cp g
x
u p A g Cp g
x
A g Cp g
t
t g (it , ix 1) +
)t g (it , ix) hptm (it , ix)
A g Cp g
t
t g (it 1, ix)
Equation 2-30
The finite difference form of the energy balance equation of the desiccant composite:
- 31 -
hpt g (it , ix) (hp +
Am mCpm Asub subCpsub

+
)t m (it , ix) + hm pdH ads vg (it , ix)
t
t
hm pdH ads vm (it , ix) = (

+
Am m Cpm Asub sub Cpsub

+
)t m (it 1, ix)
t
t
2k sub Asub
k A
k A
t m (it 1, ix) sub 2sub t m (it 1, ix 1) sub 2sub t m (it 1, ix + 1)
2
x
x
x
Equation 2-31
The finite difference form of the moisture balance equation of the air:
(hm p +
up A
x
up A
x
A
) vg (it , ix) hm p vm (it , ix ) =
t
vg (it , ix 1) +
A
vg (it 1, ix)
t
Equation 2-32
As seen in Equation 2-4, the finite difference form of the moisture balance equation of the
desiccant requires the representation of the partial derivative of m with respect to time. This
partial derivative is obtained from the adsorption isotherm, Equation 2-13, by selecting tm and vm
as independent variables.
m m t m m vm
=
+
t
t m t vm t
Equation 2-33
The finite difference form of the moisture balance equation of the desiccant is therefore written as:
m Am m
A
t m (it , ix) hm p vg (it , ix) + (hm p + m m m ) vm (it , ix) =
t m t
vm t
Am m m
[
t m (it 1, ix) + m vm (it 1, ix)]
vm
t t m
Equation 2-34
The partial derivatives m and m are obtained from Equation 2-19 as:
t m
vm
(
m m
c max
=
=
[4.09 * 10 9 vm e
t m
t m (1 c + c ) 2 2
5196
)
Tm
(
2.13 * 10 5
vm e
Tm
5196
)
Tm
Equation 2-35
- 32 -
(
m m
c max
=
=
* 4.09 * 10 9 Tm e
vm
vm (1 c + c ) 2 2
5196
)
Tm
Equation 2-36
For the desorption section, the inlet building exhaust/regeneration air are known boundary
conditions. In the finite difference analysis, the calculations are also carried out in the direction of
the airflow. The finite difference representations of the governing equation set are similar to those
for the adsorption section, except that the air velocity is the velocity of the regeneration air and it
is now negative.
The governing equation set, which include four partial differential equations and one non linear
algebraic equation, is converted into four linear algebraic finite difference equations by
substitution. Before the finite difference scheme presented in this section is used, three other
schemes have been tried, which are detailed in Appendix 1.
The boundary conditions shown in Equation 2-7 through Equation 2-12 are also converted into
finite difference forms.
t g (it ,1) = t pair _ in
Equation 2-37
vg (it ,1) = v , pair _ in
Equation 2-38
t g (it , No _ x + 1) = trair _ in
Equation 2-39
vg (it , No _ x + 1) = v ,rair _ in
Equation 2-40
t m (it ,2) t m (it ,1)

=0
dx
2
vm (it ,2) vm (it ,1)
dx
2
=0
tm (it , No _ x + 1) tm (it , No _ x)
=0
dx
2
Equation 2-41
Equation 2-42
Equation 2-43
- 33 -
vm (it , No _ x + 1) vm (it , No _ x)
dx
2
=0
Equation 2-44
For the adsorption section, the conditions of the desiccant composite at the first (ix=1) and the
last grid (ix=No_x+1) are obtained from Equation 2-41 through Equation 2-44. The conditions of
the air at the first grid are obtained from Equation 2-37 and Equation 2-38. The conditions of the
air at the last grid can be obtained by applying the air energy and moisture balance equations for
the last grid.
2u p A g Cp g A g Cp g
+
)t g (it , No _ x + 1) = hptm (it , No _ x + 1) +
dx
dt
(hp +
A g cpg
2u p A g Cp g
t g (it , No _ x) +
t g (it 1, No _ x + 1)
dx
dt
(hm p +
+
2u p A
dx
2u p A
dx
Equation 2-45
A
) vg (it , Nu _ x + 1) = hm p vm (it , Nu _ x + 1)
dt
vg (it , Nu _ x) +
A
vg (it 1, Nu _ x + 1)
dt
Equation 2-46
The determination of the air and desiccant conditions at the first and the last grid during the
desorption section is similar except that the conditions of the air at the last grid are obtained from
Equation 2-38 and Equation 2-39 and the conditions of the air at the first grid are obtained by
applying the air energy and moisture balance equations.
(hp +
2u r A g Cp g
dx
hpt m (it ,1) +

(hm p +
A g Cp g
dt
2u r A g Cp g
dx
)t g (it ,1) =
t g (it ,2) +
A g cp g
dt
t g (it 1,1)
Equation 2-47
2u r A A
+ ) vg (it ,1) =
dx
dt
hm p vm (it ,1) +
2u r A
A
vg (it ,2) + vg (it 1,1)
dt
dx
Equation 2-48
The governing equations shown in Equation 2-30 through Equation 2-32 and Equation 2-34, and
the boundary conditions shown in Equation 2-37 through Equation 2-48 complete the finite
- 34 -
difference formulation of the combined heat and mass transfer problem for the performance
modeling of desiccant wheels. The four independent unknown variables involved in this problem
are tg(it,ix), tm(it,ix), vg(it,ix) and vm(it,ix), which can be obtained by Gaussian elimination method.
The average outlet air conditions, which indicate the operating performance of the desiccant
wheel, can be calculated by taking the average of the outside/process air outlet and the average
of the building exhaust/regeneration air outlet conditions. As stated previously, the numbers of
timesteps in the adsorption and desorption sections are iads and No_t-iads, respectively.
iads +1
t pair , out =
(it , No _ x + 1)
it = 2
iads
iads +1
v , pair , out =
vg
Equation 2-49
(it , No _ x + 1)
it = 2
iads
Equation 2-50
Nu _ t +1
trair , out =
g
it = iads + 2
(it ,1)
No _ t iads
Equation 2-51
Nu _ t +1
v , rair , out =
it = iads + 2
vg
(it ,1)
No _ t iads
Equation 2-52
According to Simonson[22], the periodic steady state operation of the wheel can be determined
either by the periodic conditions of the wheel or the conservation of energy and moisture across
the wheel. The latter is a more stringent criterion than the former at typical operating conditions.
Therefore, the conservation of energy and moisture across the wheel is used to determine the
steady state operation in this performance model. When the energy and moisture lost by one air
stream are balanced with those gained by another air stream, the accumulation of energy and
moisture in the wheel over one full rotation is zero or negligible. The wheel returns to its original
conditions over one full rotation. In developing the model, the periodic steady state operation is
defined as
- 35 -
q gain qloss
q gain
m gain mloss
m gain
3%
3%
Equation 2-53
Equation 2-54
where
.
q gain = m p (h pair ,out h pair ,in )
Equation 2-55
qloss = m r (hrair ,in hrair ,out )
Equation 2-56
m gain = m p ( w pair ,out w pair ,in )
Equation 2-57
mloss = m r ( wrair ,in wrair ,out )
Equation 2-58
The above model is programmed using Fortran 77. The temperature and water vapor
concentration of the outside/process and the building exhaust/regeneration air, the temperature
and moisture loading of the desiccant composite as functions of time and space are obtained as
the modeling results, which are illustrated in the following section.
2.3 Modeling Results

As pointed out previously, this performance model is applicable to both the enthalpy recovery and
the active desiccant wheels. The design parameters of the enthalpy recovery and active
desiccant wheels used to obtain the following simulation results are shown in Table 2-1. This
model can be used to simulate the performance of desiccant wheels with various design
parameters and operating conditions. Those shown in this section are only one example.
- 36 -
Table 2-1 Design Parameters of the Desiccant Wheels Used in the Simulation
Active Desiccant Wheel
Wheel diameter (mm)
787
787
Wheel depth (mm)
152
102
Wheel split ratio
1/2
2/3
Rotary speed (rpm)
30
0.42
Channel size (mm)
1.8*4.2
1.5*3.4
Desiccant thickness (microns)
25
65
Substrate thickness (microns)
15
75
Desiccant material
3 molecular sieves
Silica gel
Desiccant density (kg/m3)
760
700
Desiccant specific heat (J/kg-K)
1,000
1,000
Desiccant conductivity (W/m-K)
Separation factor
0.1
Maximum moisture loading (kg/kg)
0.2
0.36
Heat of adsorption (J/kg)
2,791,000
2,791,000
Substrate material
Aluminum
Glass fiber paper
Substrate density (kg/m3)
2,700
500
Substrate specific heat (J/kg-K)
900
900
Substrate conductivity (W/m-K)
237
Convective heat transfer coefficient
36.2
43.3
0.030
0.036
(W/m2-K)
Convective mass transfer coefficient
(m/s)
The adsorption isotherms of the 3 molecular sieves used in the enthalpy recovery wheel and the
silica gel used in the active desiccant wheel are plotted in Figure 2-5.
- 37 -
Figure 2-5 Adsorption Isotherm of 3 Molecular Sieves and Silica Gel Used in the
Simulation
2.3.1 Modeling Results for the Enthalpy Recovery Wheel

The inlet and predicted average outlet air conditions of the enthalpy recovery wheel are shown in
Table 2-2. The psychrometric chart representation of the enthalpy recovery wheel operation is
shown in Figure 2-6. The temperature and humidity of the outside air are reduced after going
through the wheel; those of the building exhaust air are increased. The performance of the
enthalpy recovery wheel is represented by two parallel lines on the psychrometric chart, since the
quantities of the outside and the building exhaust air are equal.
- 38 -
Table 2-2 Inlet and Predicted Average Outlet Air Conditions for the Enthalpy Recovery
Wheel
outside air
building
average outside
average building
inlet
exhaust air inlet
air outlet
exhaust air outlet
Temperature (oC)
30
23
25.1
27.8
Vapor concentration (kg/m3)
0.0150
0.00850
0.0104
0.0131
Humidity (kg/kg)
0.0125
0.00708
0.00865
0.0109
Enthalpy (J/kg)
62,119
41,147
47,325
55,898
Flowrate (m3/s)
0.472
0.472
0.472
0.472
Figure 2-6 Psychrometric Chart Representation of the Enthalpy Recovery Wheel Operation
The schematic of the enthalpy recovery wheel used in the simulation is shown in Figure 2-7. The
outside air enters the wheel from one end x=0. The building exhaust air enters it from the other
end x=L. The wheel split ratio, defined as the ratio between the face area for adsorption and the
entire face area, is 1/2. The adsorption and desorption sections are separated by brush seals.
The wheel rotates clockwise at 30 rotations per minute(rpm).
- 39 -
Outside air inlet
Outside air outlet
Building exhaust air

outlet

inlet
Figure 2-7 Schematic of the Enthalpy Recovery Wheel Used in the Simulation
The profiles of the air and desiccant conditions with respect to time, or the wheel rotation, are
plotted for two complete cycles in Figure 2-8 through Figure 2-10. Since the wheel is in periodic
steady state operation, the temperature and humidity conditions of the air and the desiccant
wheel repeat themselves over one complete rotation.
Figure 2-8 and Figure 2-9 plot the profiles of the temperature of the air and the desiccant, and the
water vapor concentration in the air and that at the air desiccant interface which is in equilibrium
with the desiccant, at x=0 and x=L. In general, the temperature and water vapor concentration of
the air and those in equilibrium with the desiccant at x=0 is higher than those at x=L, because the
warmer and more humid outside air enters the wheel from x=0. The outside and building exhaust
air outlet temperature and water vapor concentration are not uniform with the wheel rotation even
though the inlet air streams are at uniform conditions.
The changes of the outside and building exhaust air outlet temperature and water vapor
concentration with the wheel rotation are directly related to the conditions of the desiccant. Since
the outside air is warmer and more humid, heat and moisture are transferred from the outside air
to the channels of the enthalpy recovery wheel. The temperature and water vapor concentration
in the wheel increase. Therefore the wheels abilities to reduce the temperature and water vapor
- 40 -
concentration of the incoming outside air stream decrease. The outside air outlet temperature and
water vapor concentration therefore increase with the wheel rotation.
Since the building exhaust air is cooler and less humid, heat and moisture are transferred back
from the wheel to the building exhaust air stream. The temperature and water vapor
concentration in the wheel decrease. The wheels abilities to raise the temperature and water
vapor concentration of the incoming building exhaust stream decrease with the wheel rotation.
The building exhaust outlet temperature and water vapor concentration decrease accordingly.
The temperature and water vapor concentration of the desiccant appear to change linearly with
time in one section of the wheel, but non linearly in another section. This phenomenon can be
explained by the change of heat and mass transfer rates between the air and the desiccant. In
the adsorption section, the differences in temperature and water vapor concentration between the
incoming outside air and the desiccant at x=0 are at their largest values at the beginning of the
cycle, the rotation angle =0o. The differences become smaller with the wheel rotation; so do the
rates of heat and mass transfer. Therefore, the rates of change in the temperature of the
desiccant and the water vapor concentration at the interface between the air and the desiccant
become smaller. In the desorption section, the differences in temperature and water vapor
concentration between the building exhaust air outlet and the desiccant at x=0 appear constant;
so do the rates of heat and mass transfer. Therefore, the rates of change in the temperature and
water vapor concentration of the desiccant remain constant. The changes in the temperature and
water vapor concentration of the desiccant at x=L can be explained in a similar way.
When the wheel switches between the adsorption and desorption sections, the air contained in
the channels of the wheel gets carried over. The directions of the airflow are different for the
adsorption and desorption sections. Therefore, in a small region of the wheel the incoming
outside air enters the building exhaust air outlet stream and the incoming building exhaust air
enters the outside air outlet stream. This is presumably what happens when a purge section is
- 41 -
not included in the wheel design. The setup of this model inherently includes this air carryover
effect. The conditions of the wheel and the air contained in the channels of the wheel are
preserved, and the directions of the airflow are reversed when the wheel switches between the
adsorption and desorption sections.
When the wheel rotates from the building exhaust to the outside air section, for a short period of
time the air leaving the wheel is the cooler and less humid building exhaust air contained in the
channels of the wheel, which is represented by the outlet air conditions from 0 to 11.25 degree
angle in Figure 2-8 and Figure 2-9. After that period, the outside air leaves the wheel at its
characteristic outlet conditions. Its temperature and water vapor concentration increase gradually
with the wheel rotation. Similarly, the effect of the air carryover from the outside to the building
exhaust air section is represented by the warmer and more humid outlet air conditions from 180
to 191.25 degree angle. The 11.25 degree rotation angle corresponds to 0.0625 seconds for the
enthalpy recovery wheel rotating at 30 rpm. The face velocity of the outside and the building
exhaust air stream is 2.4 m/s and the depth of the wheel is 0.152 m, therefore it takes 0.0625
seconds for the air contained in the channels to be cleared out.
The two dashed lines in Figure 2-8 and Figure 2-9 indicate the average outlet outside and
building exhaust air conditions, 25.1oC dry bulb temperature, 0.0104 kg/m3 water vapor
concentration and 27.8oC, 0.0131 kg/m3, respectively. Seen from Figure 2-8 and Figure 2-9, the
average outside air outlet condition is close to the outside air outlet condition at 90o rotation angle.
The average building exhaust outlet condition is close to the building exhaust outlet condition at
270o rotation angle, 90o angle from the separation seal. This finding is consistent with previous
research, which recommended the enthalpy recovery wheel performance be calculated based on
the measured data taken at one angular position[40].
The profile of the desiccant moisture loading, corresponding to the desiccant conditions shown in
Figure 2-8 and Figure 2-9, are plotted in Figure 2-10. Seen from Figure 2-10, there is very limited
- 42 -
variation in the moisture loading in the desiccant during the operation of the enthalpy recovery
wheel. The desiccant is close to its maximum moisture loading during the wheel rotation, which is
indicated as the straight line on top of the graph. The moisture contained in the desiccant is
referred to as the residual water. Its impact on the enthalpy recovery wheel performance is
explored in Section 2.5.6 of this chapter.
(Tg: air temperature, Tm: desiccant temperature)

Figure 2-8 Profile of the Air and Desiccant Temperature, with the Rotation of the Enthalpy
Recovery Wheel
- 43 -
(Rhovg: water vapor concentration in the air, Rhovm: water vapor concentration at the interface of
the air and desiccant)
Figure 2-9 Profile of the Water Vapor Concentration in the Air and at the Interface of the Air
and the Desiccant, with the Rotation of the Enthalpy Recovery Wheel
(gamma: desiccant moisture loading)

Figure 2-10 Profile of the Moisture Loading in the Desiccant, with the Rotation of the
- 44 -
The profiles of the air and desiccant conditions with respect to space, the depth of the enthalpy
recovery wheel, are included in Appendix 2. The air conditions change roughly linearly with the
distance x.
2.3.2 Modeling Results for the Active Desiccant Wheel

The design parameters of the active desiccant wheel used to obtain the following simulation
results are shown in Table 2-1. The inlet and predicted average outlet air conditions are shown in
Table 2-3. The psychrometric chart representation of the active desiccant wheel operation is
shown in Figure 2-11. The humidity of the process air is reduced and its temperature is increased
as a result of the heat of adsorption. The temperature of the regeneration air is reduced and its
humidity is increased. This is different from the enthalpy recovery wheel operation, in which heat
and mass transfer are in the same direction. The two lines representing the air handling
processes are not parallel because the airflows are not equal. They do not follow the constant
enthalpy lines, because of the heat carried over from the desorption to the adsorption section.
Table 2-3 Inlet and Predicted Average Outlet Air Conditions for the Active Desiccant Wheel
process air
regeneration
average process
average regeneration
inlet
air inlet
air outlet
air outlet
Temperature (oC)
11
100
31.5
58.9
Vapor concentration
0.00823
0.0154
0.00239
0.0271
Humidity (kg/kg)
0.00686
0.0128
0.00199
0.0226
Enthalpy (J/kg)
28,355
135,013
36,824
118,129
Flowrate (m3/s)
0.472
0.236
0.472
0.236
(kg/m3)
The schematic of the active desiccant wheel used in the simulation is shown in Figure 2-12. The
process air enters the wheel from x=0. The regeneration air enters it from x=L. The wheel split
- 45 -
ratio is 2/3. The adsorption and desorption sections are separated by brush seals. The wheel
rotates clockwise at 0.42 rpm.
Figure 2-11 Psychrometric Chart Representation of the Active Desiccant Wheel Operation
Processed air
outlet
Processed air
inlet
Regeneration air
outlet
Regeneration air
inlet
Figure 2-12 Schematic of the Active Desiccant Wheel Used in the Simulation
The profiles of the air and desiccant conditions with respect to time, or the rotation of the active
desiccant wheel, are plotted for two complete cycles in Figure 2-13 through Figure 2-15. Similar
to the enthalpy recovery wheel, the temperature and humidity conditions of the air and the
desiccant wheel repeat themselves over one complete rotation of the active desiccant wheel,
since the wheel is in periodic steady state operation.
- 46 -
Figure 2-13 and Figure 2-14 plot the profiles of the temperature of the air and the desiccant, and
the water vapor concentration in the air and that at the air desiccant interface which is in
equilibrium with the desiccant, at x=0 and x=L. In general, the temperature of the air and the
desiccant at x=L is higher than that at x=0, because the cooler process air enters the wheel from
x=0, and the hotter regeneration air enters it from x=L. The water vapor concentration in the air
and that in equilibrium with the desiccant at x=0 is higher than those at x=L during most time of
the cycle. This is because the moisture in the process air is transferred to the wheel as it flows
through the channels, and the moisture in the wheel is transferred back to the regeneration air.
Due to the effect of the air carryover, for a very short period of time after the wheel switches from
the regeneration to the process air section, 0.1o rotation angle, the air leaving the wheel is the
hotter and more humid regeneration air contained in the channels of the wheel. Similarly for a
very short period of time after the wheel switches from the process to the regeneration air section,
the air leaving the wheel is the cooler and less humid process air contained in the channels. The
0.1o rotation angle corresponds to 0.0443 seconds for the active desiccant wheel rotating at 0.42
rpm. The face velocity of the process and regeneration air stream is 2.3 m/s and the depth of the
wheel is 0.102 m, therefore it takes 0.0443 seconds for the air contained in the channels to be
cleared out.
After clearing out the air carryover, the process air outlet temperature is still high due to the high
desiccant temperature resulting from the heat carried over from the regeneration section. The
process air outlet water vapor concentration remains high during this period, because the water
vapor concentration in the desiccant is high as a result of its high temperature. In fact, due to its
higher water vapor concentration, for a short period of time the desiccant releases moisture to the
incoming process air stream as seen from the desiccant moisture loading in Figure 2-15 and the
process air outlet vapor concentration in Figure 2-16. The desiccant gradually cools down and its
water vapor concentration decreases. Then it starts to adsorb moisture from the incoming
- 47 -
process air stream. The process outlet vapor concentration and humidity decreases and the
moisture loading in the desiccant increases. The ability of the desiccant to further adsorb
moisture is reduced as its moisture loading increases. The process air outlet vapor concentration
and humidity therefore start to increase at some point. There exists a rotation angle at which the
process air outlet vapor concentration is the lowest.
Similarly, the wheel is cool for a period of time after it switches from the process to the
regeneration air section. Its water vapor concentration is low relative to the incoming regeneration
air. Instead of releasing its moisture to the incoming regeneration air stream, the wheel actually
adsorbs moisture from it, as seen from the desiccant moisture loading in Figure 2-15 and the
regeneration air outlet vapor concentration in Figure 2-17. After that period, the wheel is heated
up by the incoming regeneration air. Its vapor concentration increases and it starts to release
moisture to the regeneration air stream. The regeneration air outlet vapor concentration increases.
The moisture loading in the wheel decreases. The ability of the desiccant to further desorb
moisture is reduced as its moisture loading decreases. The regeneration air outlet vapor
concentration therefore starts to decrease at some point. There exists a rotation angle at which
the regeneration air outlet vapor concentration is the highest.
For the active desiccant wheel, the dynamics in the air and desiccant conditions result from the
air carryover between the process and regeneration sections, and the heat carryover from the
regeneration section. This is different from the enthalpy recovery wheel, in which the dynamics
are mainly due to the air carryover effect. This difference between the active desiccant and
enthalpy recovery wheels can be explained by the fact that the active desiccant wheel operates
with a larger temperature difference between the two incoming air streams.
As stated earlier, the enthalpy recovery wheel is usually built with a purge section to clear out the
air carried over from the building exhaust air to the outside air section. In the active desiccant
wheel, the air carryover is limited due to the low rotation speed of the wheel. A purge section can
- 48 -
however be used to reduce the heat carryover from the regeneration to the process air section.
This is not commonly done in reality, but it is expected that a purge section will help improve the
operating performance of the active desiccant wheel.
In the desorption section there is a period of time during which the desiccant temperature, water
vapor concentration and moisture loading at x=L remain constant, because the desiccant and the
regeneration air at this location reach adsorption equilibrium. Because other segments of the
wheel (x<L) have not reached adsorption equilibrium with the incoming regeneration air, there is
still desorption going on and the vapor concentration of the regeneration air outlet is still higher
than that of the regeneration air inlet.
The change in the desiccant moisture loading at x=0 is larger than that at x=L. There is limited
change in the moisture loading in the desiccant at x=L. The straight line on top of Figure 2-15
indicates the maximum moisture loading of the desiccant. The impact of the residual water on the
active desiccant wheel performance is also discussed in Section 2.5.6 of this Chapter.
The two dashed lines in Figure 2-13 and Figure 2-14 indicate the average outlet process and
regeneration air conditions, 29.3oC dry bulb temperature, 0.00178 kg/m3 water vapor
concentration and 63.6oC, 0.0282 kg/m3, respectively.
These profiles of the air and desiccant conditions can give guidance on the active desiccant
wheel design and operations. For example, Figure 2-13 and Figure 2-14 show there is neither
heat nor mass transfer at x=L during the later portion of the desorption section. Figure 2-14
shows the outlet process air vapor concentration goes above its averaged value for the later
portion of the adsorption section. Both observations indicate the wheel probably rotates too slowly.
The wheel does not reach adsorption equilibrium and the outlet process air vapor concentration
does not reach its averaged value when the wheel rotation is speeded up. As shown in Figure
2-18and Figure 2-19, when the wheel rotation is increased to 1.5 rpm, none of the wheel
segments reaches equilibrium with the regeneration air and the outlet process air vapor
- 49 -
concentration does not increase to its averaged value. The average process air outlet
temperature and water vapor concentration are however increased, due to the higher heat
carryover from the regeneration section.

Desiccant Wheel, rpm=0.42
- 50 -
(gamma: desiccant moisture loading)

Figure 2-15 Profile of the Moisture Loading in the Desiccant, with the Rotation of the Active
- 51 -
(Rhovg: water vapor concentration in the air)

Figure 2-16 Zoom in to the Water Vapor Concentration in the PA Outlet, with the Rotation of
the Active Desiccant Wheel, rpm=0.42
(Rhovg: water vapor concentration in the air)

Figure 2-17 Zoom in to the Water Vapor Concentration in the RgA Outlet, with the Rotation
of the Active Desiccant Wheel, rpm=0.42
- 52 -

- 53 -
The profiles of the air and desiccant conditions with respect to space, the depth of the active
desiccant wheel, are included in Appendix 3.
2.4 Performance Indicators of Desiccant Wheels

Performance indicators are developed in order to compare the performance of different designs
and operations of the desiccant wheels. Two criteria are used in choosing the performance
indicators. First, comparative evaluations of the operating performance independent of the design
and make of the desiccant wheels can be made based on the performance indicators. Second,
the outlet outside/process and building exhaust/regeneration air conditions from the desiccant
wheels can be calculated based on the inlet air conditions and the performance indicators.
Besides the indicators developed in this thesis, there are other performance criteria of the
desiccant wheels, such as the wheel compactness and pressure drop. They are important for the
desiccant wheel design and operation, but they are beyond the scope of this thesis.
2.4.1 Performance Indicators for the Enthalpy Recovery Wheel

ASHRAE prescribes a set of performance indicators for enthalpy recovery wheels and another
set for active desiccant wheels used in HVAC industry. According to ASHRAE[12], sensible heat,
latent heat or moisture, and total heat recovery effectiveness, defined as the actual transfer of
sensible heat, latent heat or moisture, and total heat divided by the maximum possible transfer
between airstreams, are commonly used performance indicators for enthalpy recovery devices in
HVAC systems.
A sensible heat recovery wheel is characterized by single performance indicator, sensible heat
recovery effectiveness. The enthalpy recovery wheel is characterized by two of the three recovery
effectiveness parameters. The third can be calculated from these two indicators and the operating
conditions of the wheel.
- 54 -
s =
m OA (t OA,in t OA,out )
.
m min (t OA,in t RA,in )
m RA (t RA,out t RA,in )
.
l =
m OA ( wOA,in wOA,out )
.
m min ( wOA,in wRA,in )
t =
m OA (hOA,in hOA,out )
.
m min (hOA,in hRA,in )
Equation 2-59
m RA ( wRA,out wRA,in )
.
Equation 2-60
m RA (hRA,out hRA,in )
.
Equation 2-61
where m min is the minimum of m OA and m RA .
When the purge section is not present in the wheel, the mass flowrate of the inlet outside air
.
stream equals that of the outlet outside air stream. Both are represented by m OA . Similarly, m RA
represents the inlet and outlet mass flowrate of the building exhaust air stream. The calculation of
the performance indicators is modified when the purge section is included in the wheel, which will
be discussed in Section 2.5.5 of this chapter.
These effectiveness parameters can be used to compare different designs and operations of the
enthalpy wheel, and to calculate the outlet air conditions from the wheel. Therefore they are
chosen as the performance indicators of the enthalpy recovery wheel in this thesis.
.
t OA,out = t OA,in
s m min (t OA,in t RA,in )

.
m OA
Equation 2-62
wOA,out = wOA,in
l m min ( wOA,in wRA,in )

.
m OA
Equation 2-63
- 55 -
t m min (hOA,in hRA,in )
hOA,out = hOA,in
m OA
Equation 2-64
t RA,out = t RA,in +
s m min (tOA,in t RA,in )

.
m RA
Equation 2-65
wRA,out = wRA,in +
l m min ( wOA,in wRA,in )

.
m RA
Equation 2-66
hRA,out = hRA,in +
t m min (hOA,in hRA,in )

.
m RA
Equation 2-67
The sensible, latent and total heat recovery effectiveness determine the slope and the extent of
the lines in the psychrometric chart representation in Figure 2-6.
2.4.2 Performance Indicators for the Active Desiccant Wheel

According to ASHRAE 139-1998 Method of Testing for Rating Desiccant Dehumidifiers Utilizing
Heat for the Regeneration Process, the moisture removal capacity, MRC, and the regeneration
specific heat input, RSHI, are commonly used performance indicators for the active desiccant
wheel[36]. The MRC is defined as the mass of water vapor removed from the process air per unit
of time. The RSHI is defined as the ratio of the regeneration energy to the MRC, and the
regeneration energy is defined as the regeneration heat input to the regeneration heater.
.
MRC= m PA ( wPA,in wPA,out )
Equation 2-68
RSHI =
m RA ( hRA,in hamb )
MRC
Equation 2-69
These parameters can be used to compare the performance of different designs and operations
of the active desiccant wheel, but the outlet air conditions from the wheel can not be obtained
- 56 -
from these parameters. The following performance indicators are therefore developed for the
active desiccant wheel:
moisture removal capacity, MRC, which is defined in the same way as in ASHRAE. MRC
indicates the dehumidification capability of the active desiccant wheel for given design
and operations.
regeneration efficiency, defined as the heat required to vaporize the moisture from the
active desiccant wheel divided by the regeneration heat input to the regeneration heater.
Assuming the regeneration heater takes in the ambient air, which is the case for most
active desiccant wheels, the regeneration efficiency can be obtained as
.
r =
m PA ( wPA,in wPA,out ) H vap

.
H vap MRC
.
H vap
RSHI
Equation 2-70
heat carryover ratio, defined as the regeneration heat carried over to the process air side
divided by the regeneration heat input to the regeneration heater.
c =
hRA,in hRA,out
hRA,in hamb
m PA ( hPA,out hPA,in )
.
m RA (hRA,in hamb )
Equation 2-71
The regeneration efficiency and heat carryover ratio determine the slopes of the two lines and the
deviation from the constant enthalpy line, in the psychrometric chart representation in Figure 2-11.
The heat carryover should be limited for effective active desiccant wheel performance.
The humidity and enthalpy of the outlet air from the active desiccant wheel can be obtained from
these performance indicators. The temperature of the outlet air from the wheel can be calculated
from the psychrometric relationship based on its humidity and enthalpy.
- 57 -
wPA,out = wPA,in
r m RA ( hRA,in hamb )
.
m PA H vap
Equation 2-72
m PA
wRA,out = wRA,in +
( wPA,in wPA,out )
m RA
hRA,out = hRA,in c (hRA,in hamb )
Equation 2-73
Equation 2-74
hPA,out = hPA,in +
m RA
.
( hRA,in hRA,out )
m PA
Equation 2-75
2.4.3 Performance Indicators for the Simulated Desiccant Wheels

For the enthalpy recovery wheel modeled in Section 2.3.1, the sensible, latent and total heat
recovery effectiveness are 0.695, 0.710 and 0.705, respectively. For the active desiccant wheel
modeled in the previous section, the MRC, regeneration efficiency and the heat carryover ratio
are 0.00276 kg/s, 0.328 and 0.228, respectively.
2.5 Discussion
2.5.1 Step Sizes in Space and Time Domains
The selection of step size is a balance between the computational time and the accuracy of the
numerical solution. The numbers of discretization in space and time domains are varied in order
to determine the reasonable grid size and timestep in the finite difference analysis, as shown in
Figure 2-20 through Figure 2-23. The wheel design parameters used in these simulations are
shown in Table 2-1. The number of discretization in time domain No_t in creating Figure 2-20 and
Figure 2-21 is 10,000; the number of discretization in space domain No_x in creating Figure 2-22
and Figure 2-23 is 100.
- 58 -
Seen from Figure 2-20 and Figure 2-21, there is very little change in the predicted performance of
the enthalpy recovery and active desiccant wheel when No_x goes above 100. No_x is therefore
set as 100 in the simulation of both wheels. The resulting grid size for the enthalpy recovery and
active desiccant wheels are 0.0015 and 0.0010 m, respectively.
Performance of the Enthalpy Recovery Wheel
- 59 -
Performance of the Active Desiccant Wheel
Seen from Figure 2-23, there is little change in the predicted performance of the active desiccant
wheel when the number of discretization in time domain No_t goes above 10,000. For the
enthalpy recovery wheel, this happens when No_t is above 500, since the enthalpy recovery
wheel rotates much faster than the active desiccant wheel. In order to use the same model for
both enthalpy recovery and active desiccant wheels, No_t is set as 10,000 in the simulation of
both wheels. The resulting grid size for the enthalpy recovery and active desiccant wheels are
0.0002 and 0.0143 seconds, respectively.
Further decreasing the grid size and timestep benefits little on the computation accuracy, but it
increases the computational time significantly. For example, increasing No_x from 100 to 200 and
No_t from 10,000 to 20,000 will increase the computational time by a factor of 50.
- 60 -
The calculated numerical values of effectiveness, regeneration efficiency and heat carryover ratio
with such grid size and timestep are estimated to be within 1% of the precise solution of the
equations.
2.5.2 Energy and Moisture Storage in the Air

Seen from the governing equations presented earlier in the chapter, this performance model
considers the heat and moisture storage in the air. Figure 2-24 and Figure 2-25 show the
predicted performance of the enthalpy recovery and desiccant wheels, with and without
considering the heat and moisture storage in the air.
Seen from Figure 2-24 and Figure 2-25, the energy and moisture storage in the air has very little
effect on the predicted performance of the enthalpy recovery and active desiccant wheels. The
changes in the recovery effectiveness for the enthalpy recovery wheel are within 1%. There is
barely any change in the regeneration efficiency and heat carryover ratio for the active desiccant
wheel. This is because the energy and moisture storage in the air are negligible compared to the
- 61 -
heat and mass flux for any control volume in this problem. The heat and moisture storage terms
can therefore be eliminated from the governing equations.
- 62 -
2.5.3 Heat Capacity of the Substrate

This performance model considers the heat storage in the desiccant composite, which includes
the heat stored in the desiccant and that stored in the substrate. The amount of heat stored in the
substrate is related to its heat capacity. Figure 2-26 shows the predicted performance of the
enthalpy recovery wheel, with and without considering the heat capacity of the aluminum
substrate layer.
Figure 2-26 The Effect of the Heat Capacity of the Substrate on the Predicted Performance
of the Enthalpy Recovery Wheel
Inclusion of the substrate heat capacity improves both the sensible and the latent heat recovery
performance of the enthalpy recovery wheel, which can be explained by the fluctuation of the
desiccant conditions during the wheel operation. The higher the substrate heat capacity is, the
less the conditions fluctuate, and the closer the enthalpy recovery wheel functions to a counter
flow heat and mass exchanger, and the better its heat and moisture recovery performance is.
Figure 2-27 shows the predicted performance of the active desiccant wheel, with and without
considering the heat capacity of the glass fiber substrate layer. Inclusion of the substrate
deteriorates the performance of the active desiccant wheel. Its regeneration efficiency is reduced.
- 63 -
More heat is carried over from the regeneration to the process air section and more cooling is
therefore required to counteract this heat carryover. The substrate is dead weight for the active
desiccant wheel. It increases the weight of the wheel and thus requires more energy for the
regeneration. Ideally an active desiccant wheel should be built without substrate. In cases where
the substrate is needed for structural integrity and strength, its heat capacity should be minimized.
Figure 2-27 The Effect of the Heat Capacity of the Substrate on the Predicted Performance
of the Active Desiccant Wheel
2.5.4 Heat Conduction through the Substrate

This model assumes the axial heat conduction is through the substrate. There is no heat
conduction in the desiccant material. Figure 2-28 shows impact of the axial heat conduction
through the substrate on the predicted performance of the enthalpy recovery wheel. This axial
heat conduction deteriorates the sensible heat recovery performance of the wheel, because it
counteracts the counter flow heat exchanger. The substrate in the enthalpy wheel should
therefore be built with less conductive materials.
- 64 -
Figure 2-28 The Effect of the Axial Heat Conduction through the Substrate on the Predicted
2.5.5 Wheel Purge

The enthalpy recovery wheel is usually built with a purge section, in order to prevent cross
contamination between the building exhaust air outlet and outside air intake. The required purge
area is dependent on the depth of the wheel, the velocity of the purge airflow and wheel rotation
speed. For the enthalpy wheel simulated in Section 2.3.1, the required purge angle is calculated
as 11.25o, as stated previously.
Figure 2-29 shows the schematic of the purge section on which the model is based. The
difference between this schematic and the one shown in Figure 2-7 is that the brush seals at the
two ends of the wheel are not aligned with each other. The air leaving the purge section is
directed into the building exhaust inlet stream. It goes through the wheel one more time and ends
up in the building exhaust air outlet stream. To simplify the computational model, the air leaving
the purge section is directly put into the building exhaust air outlet stream. The outside air outlet
flowrate is reduced since a portion of the outside air inlet is used as the purge flow.
- 65 -
Outside air
outlet
Outside air inlet

inlet

outlet
Figure 2-29 Schematic of the Purge Section in the Enthalpy Recovery Wheel
The energy and moisture lost by the outside air stream is considered when calculating the wheel
performance. Those lost by the purge airflow are not included because they are not useful
transfer.
.
s =
m OA,out (t OA,in t OA,out )

.
l =
t =
(m OA,in m pg )( wOA,in wOA,out )

.
Equation 2-77

.
m OA,out (hOA,in hOA,out )
Equation 2-76
m OA,out ( wOA,in wOA,out )
( m OA,in m pg )(t OA,in t OA,out )
(m OA,in m pg )(hOA,in hOA,out )

.
Equation 2-78
The outlet outside air conditions are different from the values presented in Table 2-2, since the
outlet air in the purge section is directed into the building exhaust air outlet.
Figure 2-30 compares the predicted performance of the enthalpy recovery wheel with and without
the purge section. Inclusion of the purge reduced the wheel performance by about 5%. However,
- 66 -
the cross contamination between the building exhaust air and outside air intake is prevented and
the indoor air quality is improved.
Figure 2-30 The Effect of Purge on the Predicted Performance of the Enthalpy Recovery
Wheel
2.5.6 Residual Water in the Desiccant Material

As stated previously, there is residual water contained in the desiccant material even after the
wheel is regenerated. In the simulation, the density of the desiccant material can be adjusted to
account for this residual water. The adsorption properties of the desiccant material remain
unchanged, because the adsorption equilibrium relationship takes care of the impact of the
residual water on moisture transfer. Seen from Figure 2-10, the residual water in the enthalpy
recovery wheel is high, 18% on average. The density of the zeolite molecular sieves is increased
from 760 to 897 kg/m3 to account for this residual water. Since the process air entering the active
desiccant wheel is close to saturation condition, the residual water in the active desiccant wheel
is also high, 15% on average. The density of the silica gel is increased from 700 to 805 kg/m3 to
account for the residual water.
- 67 -
Figure 2-31 and Figure 2-32 plot the impact of the residual water on the performance of the
enthalpy recovery and active desiccant wheels, respectively. The residual water in the enthalpy
recovery wheel has very limited impact on its energy recovery performance. The residual water
increases the heat carryover ratio of the active desiccant wheel, since the wheel heat capacity is
increased. The impact on the active desiccant wheel is also limited, but larger than the enthalpy
recovery wheel, which is due to the larger ratio between the mass of the enthalpy recovery wheel
media and the mass of the air, as a result of its high rotation speed.
Figure 2-31 The Effect of Residual Water on the Predicted Performance of the Enthalpy
Recovery Wheel
- 68 -
Figure 2-32 The Effect of Residual Water on the Predicted Performance of the Active
Desiccant Wheel
2.5.7 Wheel Supporting Structure

The performance model developed in Section 2.2 only considers the core of the desiccant wheels.
In reality, the wheel also has supporting structure such as spokes and casing for structure
integrity and strength. Assuming the supporting structure goes through the same temperature
fluctuation as the substrate, it can be simulated as added thermal mass to the substrate.
For the wheels simulated in Section 2.3, the weight of the enthalpy recovery and active desiccant
wheels, including their cores and supporting structures, are assumed to be 13.6 kg (30 lbs),
which is typical for their sizes. The core weight is calculated based on the channel composition,
properties of the desiccant material and substrate and the number of channels in the wheel. The
weight of the core and supporting structure are listed in Table 2-4. The structure of the active
desiccant wheel is heavier than the enthalpy recovery wheel, because the spokes that are used
to connected the different segments of the active desiccant wheel are not present in the enthalpy
recovery wheel. The substrate density is adjusted to account for the added thermal mass of the
supporting structure.
- 69 -
Table 2-4 Parameters of the Wheel Supporting Structure

Total weight (kg)
13.6
13.6
Core weight (kg)
9.5
8.2
Structure weight (kg)
4.1
5.4
Original substrate density (kg/m3)
2700
500
Adjusted substrate density (kg/m3)
4400
1238
The impact of the supporting structure on the operating performance of the enthalpy recovery and
active desiccant wheels are plotted in Figure 2-33 and Figure 2-34. The supporting structure has
limited impact on the energy recovery performance of the enthalpy recovery wheel, because the
mass of the wheel media is relatively large compared to the mass of the air, as stated previously.
the Enthalpy Recovery Wheel
The supporting structure has significant impact on the regeneration efficiency and the heat
carryover ratio of the active desiccant wheel. For the wheel simulated in this section, the
supporting structure decreases the regeneration efficiency by 2% and increases the heat
- 70 -
carryover ratio by 10%. These observations are consistent with those from Section 2.5.3,
because the supporting structure is modeled as added thermal mass to the substrate.
the Active Desiccant Wheel
2.5.8 Saturation Vapor Pressure

The calculation of the saturation vapor pressure over the surface of liquid water is required in the
representation of the desiccant adsorption isotherm. This saturation vapor pressure is a function
of the water temperature; this function can be obtained by fitting the empirical data.
p vm , sat = e f (Tm )
Equation 2-79
where
f (Tm ) =
c1
2
3
+ c 2 + c3Tm + c 4Tm + c5Tm + c6 ln Tm
Tm
Equation 2-80
c1 = -5.8002206E3; c2 = 1.3914993; c3 = -4.8640239E-2; c4 = 4.1764768E-5; c5 = 1.4452093E-8; c6 = 6.5459673
- 71 -
The relationship between the saturation vapor pressure and the liquid temperature can also be
obtained through other analytical or empirical equations. According to ASHRAE[14], Equation 280 is a more accurate representation, but it is obviously more complicated. Significant effort is
m
m
required in obtaining the partial derivatives m and vm , which are needed in doing the finite
difference analysis. Therefore, a simpler function as shown in Equation 2-18 is used in this thesis.
The term Hvap is adjusted in order to minimize the deviation from the more accurate calculation
in Equation 2-80. When Hvap is set at 43,200 J/mole, the saturation vapor pressure determined
from Equation 2-18 agrees with the one obtained from Equation 2-80 within 5%, as shown in
Figure 2-35.
Figure 2-35 Deviation in the Saturation Vapor Pressure Calculations
2.6 Various Uses of the Performance Model

This performance model can be used in the following areas:
1.
Selection of the design parameters of desiccant wheels
- 72 -
The design parameters of the desiccant wheel include its diameter and depth, the split ratio
between the adsorption and desorption sections, the purge angle, the channel shape and size,
the properties and thickness of the desiccant material and the substrate, all of which are present
in this performance model.
The design process usually starts with setting up the design requirements, such as energy
recovery effectiveness of the enthalpy recovery wheel, the pressure drop across the wheel, its
compactness and cost. The wheel designer selects a set of initial design parameters based on
his experience, trying to meet the design requirements. He can then use this performance model
to explore different design alternatives and to optimize the required design indicators of the wheel.
For example, one might want to choose an enthalpy wheel with large diameter and depth in order
for improved energy recovery performance. However, bigger and deeper wheels require more
capital investment, occupy more space and cause a higher pressure drop in the system resulting
in higher fan power consumption. Finally, based on all these considerations, a wheel is selected
balancing its performance, size, and cost.
2.
Selection of the operating variables of desiccant wheels
The operating variables of the desiccant wheel include its rotary speed, the inlet outside/process
air temperature, humidity and flowrate, and the building exhaust/regeneration air temperature,
humidity and flowrate, all of which are present in this performance model.
The inlet outside air temperate and humidity for the enthalpy recovery wheel are dependent on
climatic conditions. Its flowrate is dependent on the number of occupants in the building and the
building function, which then determines the outside air requirement per occupant. The inlet
building exhaust air temperature and humidity are dependent on indoor conditions. Its flowrate is
dependent on the outside air flowrate and building pressure one wants to maintain. The
determination of the outside and building exhaust air conditions of the enthalpy recovery wheels
- 73 -
is discussed in more detail in Chapter 4. The inlet process air temperature, humidity and flowrate
for the active desiccant wheel are dependent on the HVAC system configuration and capacity.
The process of selecting the other operating variables is similar to the selection of the design
parameters. The wheel designer or application engineer selects a set of initial operating variables
based on his experience. He can then use this performance model to explore different operating
variables on the operating performance of the desiccant wheels. Finally the operating variables
that minimize the operating energy use or cost will be selected.
3.
Analysis of the operating data of desiccant wheels
This performance model can also be used to diagnose the operating data of the desiccant wheels.
Based on the wheel design and operating variables, its outlet air conditions can be predicted from
the model. These predictions can be compared with the measured quantities. Confidence in both
the model and the experiment can be achieved when there is reasonable agreement between the
predicted and the measured values. If they do not agree, the design parameters and operating
variable inputs to the model as well as the measured quantities should be checked. Sensitivity
studies can be performed to identity the most sensitive parameters that affect the disagreement
between the model predictions and the experiment. Close attention should be paid to these
sensitive parameters. If they still do not agree, there probably exist some mechanisms that are
not accurately accounted for in the model. Chapter 3 gives specific examples on this use case.
2.7 Summary
In this chapter, the combined heat and mass transfer problem in simulating the operating
performance of the desiccant wheels has been formulated. Based on the moisture transfer model
developed by analyzing the structure of the desiccant materials, the governing equations that
describe the energy and material balance, the heat and mass transfer rates and the adsorption
equilibrium relationship have been established. The boundary conditions have been identified.
- 74 -
The governing equation set has been solved using the explicit finite difference analysis. The
solution to this governing equation set has been iterated until a repetitive steady state solution
has been reached. This model relates the operating performance of the desiccant wheel to its
design parameters and operating conditions.
This model has been applied to predict the operating performance of both enthalpy recovery and
active desiccant wheels once the wheel design and operating conditions are given. The changes
of the air and desiccant conditions with regard to time, the rotation of the wheel, have been
predicted and explained.
The performance indicators of both enthalpy recovery and active desiccant wheels have been
developed in order to compare the performance of different designs and operations of desiccant
wheels. The performance indicators for the enthalpy recovery wheel include sensible, latent or
moisture, and total heat recovery effectiveness. The performance of the enthalpy recovery wheel
can be characterized by two of the three effectiveness parameters and the third can be calculated
based on the two and the operating conditions of the wheel. The performance indicators
developed for the active desiccant wheel include the moisture removal capacity, regeneration
efficiency and heat carryover ratio. These performance indicators can be used to make
comparative evaluations of the operating performance of different desiccant wheels. They can
also be used to calculate the outlet air conditions from the desiccant wheels.
The numbers of discretization in space and time domains are varied in order to determine the grid
size and timestep in the finite difference analysis for reasonable computational accuracy and time.
The calculated numerical values of energy recovery effectiveness, regeneration efficiency and
heat carryover ratio are within 1% of the precise solution of the governing equations.
Inclusion of the energy and moisture storage in the air has very limited impact on the predicted
performance of both enthalpy recovery and active desiccant wheels, because the energy and
- 75 -
moisture storage in the air are negligible compared to the heat and mass flux for any control
volume in this problem.
The substrate heat capacity improves both the sensible and the latent heat recovery performance
of the enthalpy recovery wheel, due to the reduced fluctuation of the desiccant conditions during
the wheel operation. The fluctuation in the wheel temperature and water vapor concentration
decreases the driving force for heat and mass transfer. The less the wheel temperature and water
vapor concentration fluctuate, the closer the enthalpy recovery wheel functions to a counter flow
heat and mass exchanger and the better its heat and moisture recovery performance is. The
substrate heat capacity deteriorates the performance of the active desiccant wheel because it
increases the weight of the wheel and thus requires more energy for the regeneration. The
substrate heat capacity should therefore be limited for improved performance of the active
desiccant wheel.
The axial heat conduction through the substrate deteriorates the sensible heat recovery
performance of the enthalpy recovery wheel. It has little impact on the latent heat recovery
performance of the wheel. The substrate in the enthalpy wheel should therefore be built with less
conductive materials.
Inclusion of a purge section in the enthalpy recovery wheel reduces the wheel performance by
about 5%. The purge section is however an important design feature since the cross
contamination between the building exhaust air and the outside air intake is prevented and the
indoor air quality is improved.
The desiccant wheels are never fully regenerated. There is residual water contained in the
desiccant material even after the regeneration process. This residual water can be modeled by
adjusting the density of the desiccant material. The residual water in the enthalpy recovery wheel
has very limited impact on its energy recovery performance. The residual water increases the
- 76 -
heat carryover ratio of the active desiccant wheel, since the wheel heat capacity is increased. The
impact on the active desiccant wheel is also limited, but larger than the enthalpy recovery wheel,
which is due to the larger ratio between the mass of the enthalpy recovery wheel media and the
mass of the air, due to its high rotation speed.
The wheel supporting structures such as spokes and casing can be modeled as added thermal
mass to the substrate, assuming they go through the same temperature fluctuation as the
substrate. The supporting structure has limited impact on the energy recovery performance of the
enthalpy recovery wheel, because the mass of the wheel media is relatively large compared to
the mass of the air. The supporting structure has significant impact on the regeneration efficiency
and the heat carryover ratio of the active desiccant wheel. For the wheel simulated in this section,
the supporting structure decreases the regeneration efficiency by 2% and increases the heat
carryover ratio by 10%.
This performance model can be used selecting the design parameters and operating variables,
and diagnosing experimental data for desiccant wheels.
- 77 -
Chapter 3 Validation of the Performance Model

In this chapter, the performance model of the desiccant wheels developed in Chapter 2 is
validated using the experimental data collected from a ventilation system installed in the IW. This
ventilation system, SEMCO REV 2250, integrates an active desiccant wheel for air
dehumidification/heating with a heat pump unit for air cooling/heating. It also includes an enthalpy
recovery module, FVR 2000, to precondition the ventilation air before it enters REV 2250.
The REV 2250 ventilation unit and the FVR 2000 enthalpy recovery module have been provided
to the IW by SEMCO and the U.S. Department of Energy. They were installed and commissioned
during the winter of 2005-2006 as part of the IWESS project. The installed system is highly
instrumented for the purpose of control and performance monitoring. The system has been
operated continuously after its installation to provide ventilation as well as cooling/heating to the
IW.
Extensive testing data have been collected and analyzed in order to characterize the operating
performance of each component and the overall system. The measured performance of the two
desiccant wheels has been compared with the predictions from the performance model for
validating the model and diagnosing the experimental data.
3.1 Experiment Platform

The SEMCO REV 2250 and FVR 2000 installed in the IW are shown in Figure 3-1.
- 78 -
Figure 3-1 SEMCO REV 2250 and FVR 2000 Units Installed in the IW
3.1.1 SEMCO REV 2250

The REV 2250 ventilation unit is illustrated in Figure 3-2. It integrates a 1.416 m3/s (3000 cfm)
circulation fan, a 50 kW air based heat pump for cooling/heating, and an active solid desiccant
wheel, which can be used for air dehumidification in summer and air heating in winter.
Figure 3-2 Configuration of SEMCO Revolution Unit (courtesy of SEMCO Inc.)

This ventilation unit can be operated either as an air conditioning system, recirculating air from
the building space and combining it with outside air for cooling/heating/ventilation in the space or
as a DOAS, processing only the ventilation air. For DOAS operation in summer, the outside air,
OA, is drawn through the direct expansion, DX, cooling coil, where it is cooled and dehumidified.
If the unit is operated as an air conditioning system, the mixture of the outside and recirculation
air, RcA, shown as the dashed line in Figure 3-3, is drawn through the DX coil. The air leaving the
DX coil, LDX, is then split into two portions: one portion goes into the active desiccant wheel
- 79 -
where it is further dehumidified, LDW; another portion bypasses around the wheel through a
modulating damper. The warm dry air from the desiccant wheel is then mixed with the cold moist,
probably saturated, air bypassing the wheel. The mixture is ducted into the IW. The bypass
damper is set. The heat pump compressor and the desiccant regeneration air temperature are
adjusted in order to obtain the operator specified supply air, SA, temperature and humidity under
different outdoor conditions. This air handling process is represented by the dashed lines in the
psychrometric chart shown in Figure 3-4 for summer and winter design days in Pittsburgh.
REV 2250
FVR 2000
Figure 3-3 Flow Diagram of FVR 2000 and REV 2250 Installed in the IW
Figure 3-4 Psychrometric Representation of Semco FVR 2000 and REV 2250 (solid lines
indicate the combined system; dashed lines, the system without enthalpy recovery module.)
- 80 -
The flow diagram shown in Figure 3-3 also includes the FVR 2000 enthalpy recovery module, and
Figure 3-4 also includes the air handling processes of the combined system, REV 2250 and FVR
2000. Both are discussed in Section 3.1.2.
The active desiccant dehumidification wheel is one of the major components in REV 2250 unit.
The other is the heat pump module. The active desiccant module includes a solid desiccant
coated rotary wheel shown in Figure 3-5, a gas burner to provide regeneration heat and a fan to
draw the regeneration air through the wheel.
Figure 3-5 Solid Desiccant Dehumidification Wheel Installed in the IW

For building ventilation, the dehumidification wheel transfers moisture from the process air stream
to the regeneration air stream. As installed in the IW, the regeneration air is a combustion
product from a direct fired gas burner. But it might also come from the air heated by a solar
thermal collector, or a sensible heat exchanger which recovers the rejected heat from the power
generation process. These possibilities will be explored in Chapter 5 of this thesis.
The dehumidification wheel in REV 2250 unit has a diameter of 0.787 m, and a depth of 0.102 m.
Silica gel is the desiccant material used on this wheel. It is formed by first impregnating the glass
fiber paper backbone with concentrated water glass (sodium silicate) and then reacting with an
acid wash. The moisture is adsorbed in the many pores residing in the silica gel mass. For air
- 81 -
dehumidification, the wheel rotates at 0.42 rpm between the process and regeneration air
streams.
In heating season, the wheel can be used as a sensible heating device. It transfers heat from the
natural gas combustion product (the regeneration air stream in the dehumidification mode) to the
process air stream. In this case, it rotates at 6 rpm to compensate for its minimal heat capacity.
3.1.2 SEMCO FVR 2000

The FVR 2000 shown in Figure 3-6 is an enthalpy recovery module which is used to pre condition
the outside air before it is brought into REV 2250. The nominal air flow through this module is
0.944 m3/s (2000 cfm). The enthalpy recovery wheel has a diameter of 0.787 m, and a depth of
0.152 m. It rotates at 30 rpm between the outside intake and the building exhaust air outlet.
OAin
OAout
RAout
RAin
Figure 3-6 FVR 2000 Enthalpy Recovery Module Installed in the IW

The purge section of the enthalpy recovery wheel is illustrated in Figure 3-7. In this case, the
purge area is about 2% of the total face area of the wheel. The resulting purge air volume under
design conditions is around 0.0378 m3/s.
- 82 -
Figure 3-7 The Purge Section in FVR 2000 Enthalpy Recovery Module
The FVR 2000 uses a molecular sieve material as the desiccant. The 3 micropores in this
desiccant are large enough to accommodate water molecules, 2.8 in diameter, but are too small
for the diameters of indoor pollutants. This desiccant, therefore, minimizes the cross
contamination from the exhaust to the fresh, outdoor air supply. Testing by Georgia Tech
Research Institute has confirmed that the SEMCO 3 wheel offers superior performance over
wheels using other desiccant materials in preventing cross contamination by common indoor
pollutants, such as isopropanol, xylene and acetaldehyde [41]. In addition, the unique adsorption
characteristics offered by the 3 molecular sieve material and the manufacturing process of
SEMCO ensure the high heat and moisture recovery effectiveness of the FVR 2000 unit.
A flow diagram of the combined system, including FVR 2000 and REV 2250 as installed in the IW,
is shown in Figure 3-3. A psychrometric chart representation of the combined system is shown as
solid lines in Figure 3-4.
The processes indicated in the upper part of the diagram represent typical summer operation.
Typical winter operation is shown in the lower part. This representation shows a DOAS operation
with the supply air temperature equal to the space return air temperature. The DOAS is only for
supplying ventilation air. The space cooling and heating load is met by other systems such as fan
coils, radiant panels or VAV systems. The difference between supply air and return air humidity is
due to the moisture generated in the space by occupants, plants and coffer makers, as well as
- 83 -
the moisture brought in by outside air infiltration. During summer, the system is controlled to
provide a constant ventilation supply air condition. During winter, if no humidification equipment is
available in the system, the indoor humidity is lower compared to the summer operation. Because
the enthalpy recovery wheel is able to recover moisture from the space exhaust air stream, the
resulting supply air and space return air humidities are higher than those for a system without
enthalpy recovery. If humidification is included in a ventilation system for winter operation,
enthalpy recovery significantly reduces the energy required for water evaporation.
The performance of the enthalpy recovery unit, the conditions of the supply and exhaust air
streams, is represented by two parallel lines on the psychrometric chart, if the quantities of the
outside air and exhaust air are equal. The slope and extent of these lines depends on the heat
and mass transfer coefficients and transfer areas provided in the enthalpy recovery wheel. The
reductions of operating energy consumption and coil capacity offered by the enthalpy recovery
module are discussed in Chapter 4.
More detailed information about the REV 2250 ventilation system, including its heat pump module,
can be found in Appendix 4.
3.2 Experiment Setup

Experiments have been conducted in order to characterize the operating performance of the
three major components in the system: the enthalpy recovery module, heat pump module and
active desiccant dehumidification module, as well as the overall system. The experimental data
on the enthalpy recovery and the active desiccant dehumidification modules serve as the
empirical validation data for the performance model developed in this thesis, and they are
reported here.
- 84 -
3.2.1 Experiment Setup on the Enthalpy Recovery Wheel

Guidelines for the performance testing of air-to-air enthalpy recovery devices are provided in
ASHRAE Standard 84-1991 Method of Testing Air-to-Air Heat Exchangers[42], where testing
methods, the required data and calculation methods, as well as testing equipment are specified.
Based on this ASHRAE Standard, ARI Standard 1060-2005 Performance Rating of Air-to-Air
Heat Exchangers for Energy Recovery Ventilation Equipment[43] establishes the reporting
requirements for equipment rating data as well as temperature and humidity conditions at which
equipment tests are to be conducted. ARI Standard 1060-2005 serves as an industry standard for
rating air-to-air heat/energy recovery devices.
While ASHRAE Standard 84-1991 can be applied to laboratory testing of enthalpy recovery
wheels, it may not be directly applicable to field testing due to financial and spatial constraints. In
addition, enthalpy recovery wheel performance in a field application may differ significantly from
laboratory testing data due to the following factors[12]:
heat and moisture exchange with the surroundings,
air leakage from/to the surroundings,
non uniform inlet air conditions,
air carryover, crossover or leakage between different air streams.
The performance testing of enthalpy recovery wheels is further complicated by the non uniform
outlet air conditions resulting from wheel rotation and, perhaps, non uniform wheel structure. The
only way to determine the enthalpy recovery wheel performance in a field installation is through
field testing.
The operating performance of FVR 2000 enthalpy recovery module has been measured for both
winter and summer operation at various outside and exhaust air flowrates. Based on the
measurements from manufacturer installed sensors and supplementary temperature and humidity
- 85 -
data loggers placed at various locations in the machine, as shown in Figure 3-8, the wheels
energy and moisture balances are confirmed, and its performance are established.
OAin
brush
seal
Exhaust fan
S
OAout H
RcA
RAout
RAin
Enthalpy
recovery
module
Supply fan
Heat pump coil
Manufacturer-installed
sensors
External temperature and

humidity sensors as data
loggers, HOBO
Purge path
Figure 3-8 Instrumentation on the Enthalpy Recovery Module

For the manufacturer installed instrumentations, temperature and relative humidity sensors are
used to measure the conditions of the outside air inlet and the building exhaust air outlet of the
enthalpy wheel. The building exhaust air inlet sensors are located in the return plenum,
underneath the main unit. The outside air outlet condition from the wheel is not directly measured
due to spatial constraints; instead, the temperature and relative humidity of the air leaving the
supply fan is recorded. Although not indicated in the diagram, pressure drops across both outside
and building exhaust air sections of the enthalpy wheel are measured by pressure transducers
located in the wheel assembly, which are used to calculate the outside and building exhaust air
volumetric flowrate. Pressure change across a piezo ring in the supply fan is also measured to
obtain the supply air flowrate.
Due to spatial limitations, some of manufacturer installed sensors are located at places where
typical representative readings of inlet and outlet air conditions may not be obtained. Therefore
- 86 -
external temperature and humidity data loggers are deployed to explore the impact of sensor
location and check the measurements obtained from manufacturer installed sensors. As shown in
Figure 3-9, four external sensors located at different positions on a metal screen at the outside air
outlet from the wheel before it enters the supply fan are used to record the outside air outlet
temperature and humidity. The average values of the four outside air outlet sensor readings are
used to represent the outside air outlet conditions in calculating the wheel performance, assuming
the airflow across the face of the wheel is even. Three external sensors are used to measure the
building exhaust air condition entering the enthalpy wheel as shown in Figure 3-10. Similarly, the
average values of the three building exhaust air inlet sensor readings are used to represent the
building exhaust air inlet conditions in calculating the wheel performance. Building exhaust air
outlet conditions are measured at the exhaust fan outlet where representative outlet air condition
readings can be obtained due to air mixing caused by the fan, provided that the temperature rise
due to exhaust fan power consumption is reasonably accounted for. Outside air inlet condition is
also recorded and checked using external sensors.
up left
down left
up right
down right
Figure 3-9 External Sensors Used to Measure the Outside Air Outlet Conditions
- 87 -
far
middle
near
Figure 3-10 External Sensors Used to Measure the Building Exhaust Air Inlet Conditions
Only temperature and humidity data loggers are used in the experiment. The air flowrate
measurements are not calibrated. The values indicated by manufacturer installed instrumentation
are used in the performance calculation.
- 88 -
Specifications of the instrumentation used in the experiment are presented in Table 3-1.
Table 3-1 Instrumentation Specification
Measurement
Measurement
Accuracy
Sensor
range
Air temperature
Thermistor
-55~150oC
0.20oC
Thermistor
-20~70oC
0.11oC
Capacitive RH sensor
0~98% RH
3.0% RH
Capacitive RH sensor
5~95% RH
2.5% RH
0-1.0 inch water
1.0% of full
column
scale
0-10.0 inch water
1.0% of full
column
scale
(Manufacturer-installed)
Air temperature
(External data logger)
Relative humidity
Relative humidity
(External data logger)
Wheel air flowrate
Pressure transducer
Supply air flowrate
Pressure transducer
2.0% of full
Electric power
Current transducer
0-10 A
scale
The mass flowrate of the outside and building exhaust air streams are calculated from the outside
and building exhaust air pressure drop measurement across the enthalpy wheel as follows. The
coefficients a, b, c and d are determined based on the testing information obtained from the
wheel manufacturers lab.
.
m OA,out = g (b PEWOA c) d
Equation 3-1
m RA,out = a g PEWRA
Equation 3-2
where a = 3625, b = 593.98, c = 70.93 and d = 1.261

- 89 -
The calculation for building exhaust air flowrate indicates the laminar flow characteristics, which is
expected given the structure of the wheel and the air flowrate. However, the calculation for
outside air flowrate does not indicate a pure laminar flow, perhaps due to flow maldistribution.
The building exhaust air flowrate entering the enthalpy wheel is obtained as the difference
between the building exhaust air outlet and the purge air flow since the purge flow is included in
the measured building exhaust air outlet flowrate, but not in the measured outside air outlet
flowrate.
.
m RA,in = m RA,out m pg
Equation 3-3
Since the building exhaust air outlet air conditions are measured after the exhaust fan, the air
temperature increase due to the fan power consumption has to be considered. The entire power
consumption of the exhaust fan ends up with the building exhaust air outlet temperature increase
due to two different mechanisms: the thermal energy going into the air stream due to motor and
fan inefficiency; the thermal effect of the air compression process in the fan.
Based on the First Law of Thermodynamics, the following equation holds for the process in the
fan:
q = h + w
Equation 3-4
where h is the air enthalpy change before and after the fan, w is the work done by the fan and q
is the overall heat exchange with surroundings. Assuming the air goes through an adiabatic
process in the fan, the work done by the fan becomes the enthalpy increase of the processed air.
w = h
Equation 3-5
Let sFan represent the total fan efficiency, which is the product of the motor efficiency and the
blade efficiency, the work done by the exhaust fan can be obtained as:
w = efan powerefan
Equation 3-6
- 90 -
efan powerefan = m RA,out t RA,work Cp g
Equation 3-7
In addition, fan inefficiency causes (1-sFan) portion of power consumed by the fan becomes heat
entering the air stream.
.
(1 efan ) powerefan = m RA,out t RA,heat Cp g
Equation 3-8
Overall, the entire power consumption of the exhaust fan ends up with the temperature increase
of the building exhaust air outlet.
t RA = t RA,heat + t RA,work
Equation 3-9
powerefan = m RA,out t RACp g
Equation 3-10
t RA,out = t RA,efan t RA
Equation 3-11
Ideally, all of the variables used to calculate the performance indicators should be directly
measured, in order to minimize the uncertainty in the experiment. It is extremely difficult to
implement in a field experiment due to financial and spatial constraints. In this experiment, the
measurements and calculations for the enthalpy recovery wheel are summarized in Table 3-2.
Table 3-2 The Data/Equation for the Enthalpy Recovery Wheel Performance Calculation
Source
Outside air inlet temperature and humidity
Direct measurements
Building exhaust air inlet temperature and
Calculated, Equation 3-11
humidity
Outside air outlet temperature and humidity
Average of four direct measurements
Building exhaust air outlet temperature and
Average of three direct measurements
humidity
Outside air flowrate
Building exhaust air flowrate
- 91 -
Once the inlet and outlet air conditions and flowrates are obtained, the energy and moisture
balance of the wheel are established, and its operating performance is determined. With
reference to the picture shown in Figure 3-6 and the schematic shown in Figure 3-8, the
maximum possible sensible heat exchange is calculated as:
.
Qs ,max = m min Cp g (tOA,in t RA,in )
Equation 3-12
The sensible heat loss of the outside air stream is calculated as:
.
Qs ,OA = m OA,out Cp g (tOA,in tOA,out )
Equation 3-13
The sensible heat loss of the purge stream is calculated as:

.
Qs , pg = m pg Cp g (tOA,in t RA,out )
Equation 3-14
The entire sensible heat loss is calculated as the sum of that lost by the outside air and the purge
streams:
Qs ,loss = Qs ,OA + Qs , pg
Equation 3-15
The sensible heat gain of the building exhaust air stream is calculated as:
.
Qs , RA = m RA,in Cp g (t RA,out t RA,in )
Equation 3-16
The maximum possible moisture exchange is calculated as:

.
Wmax = m min ( wOA,in wRA,in )
Equation 3-17
The moisture loss of the outside air stream is calculated as:

.
WOA = m OA,out ( wOA,in wOA,out )
Equation 3-18
The moisture loss of the purge stream is calculated as:

.
W pg = m pg ( wOA,in wRA,out )
Equation 3-19
The entire moisture loss is calculated as the sum of that lost by the outside air and the purge
streams:
Wloss = WOA + W pg
Equation 3-20
- 92 -
The moisture gain of the building exhaust air stream is calculated as:
.
WRA = m RA,in ( wRA,out wRA,in )
Equation 3-21
The maximum possible total heat exchange is calculated as:

.
Qt ,max = m min (hOA,in hRA,in )
Equation 3-22
The total heat loss of the outside air stream is calculated as:
.
Qt ,OA = m OA,out (hOA,in hOA,out )
Equation 3-23
The total heat loss of the purge stream is calculated as:

.
Qt , pg = m pg (hOA,in hRA,out )
Equation 3-24
The entire total heat loss is calculated as the sum of that lost by the outside air and the purge
streams:
Qt ,loss = Qt ,OA + Qt , pg
Equation 3-25
The total heat gain of the building exhaust air stream is calculated as:
.
Qt ,RA = m RA,in (hRA,out hRA,in )
Equation 3-26
The sensible heat recovery effectiveness determined from the measurements of the outside air
stream is:
.
s ,OA =
Qs ,OA
Qs ,max
.
where
m OA,out (tOA,in tOA,out )
Equation 3-27
m min (tOA,in t RA,in )

.
mmin is the minimum of m OA and m RA .
The performance parameters of the enthalpy recovery wheel are calculated based on Equation 276, Equation 2-77 and Equation 2-78 in Chapter 2.
- 93 -
3.2.2 Experiment Setup on the Active Desiccant Wheel

Guidelines for testing the moisture removal capacity and the thermal performance of active
desiccant dehumidification devices are provided in ASHRAE Standard 139-1998 Method of
Testing for Rating Desiccant Dehumidifiers Utilizing Heat for the Regeneration Process[36],
where testing methods, the required data and calculation methods, as well as testing equipment
are specified. This standard is applicable to the laboratory testing of active desiccant wheels. It
becomes extremely difficult to implement in a field testing due to financial and spatial constraints.
In addition, the field testing of the active desiccant wheel shares the same difficulties as that of
the enthalpy recovery wheel: heat and moisture exchange with the surroundings, non uniform
airflow pattern in the chambers, air carryover or leakage between different chambers, and so on.
Furthermore, the outlet air conditions from the active desiccant wheel vary more significantly with
the wheel rotation than the enthalpy recovery wheel as seen from Chapter 2, which makes the
field performance testing of the active desiccant wheel even more difficult.
The operating performance of the active desiccant wheel during both winter sensible heating and
summer dehumidification has been tested using manufacturer installed sensors and external
temperature and humidity data loggers, as shown in Figure 3-11.
There is a manufacturer installed temperature sensor located after the heat pump coil to measure
the temperature of the process air inlet of the active desiccant wheel. No manufacture installed
humidity sensor is available to measure the humidity of the process air inlet. Therefore, external
temperature and humidity data loggers are used to measure the humidity of the process air inlet
and to confirm the temperature measurement of the manufacturer installed sensor. Four data
loggers are placed at different locations in the process air section, upstream of the active
desiccant wheel as shown in Figure 3-12, in order to capture the non uniformity of the inlet air
conditions. The average values of these four data loggers are calculated to represent the inlet
process air conditions of the wheel.
- 94 -
RgAout
regeneration
fan
PAout
S
SA
brush
seal
S
Burner
Active desiccant wheel
Heat pump coil
PAin
ambient
RgAin
S
Manufacturer-installed
sensors
External temperature and

humidity sensors as data
loggers, HOBO
Figure 3-11 Instrumentation on the Active Desiccant Module
Figure 3-12 External Sensors Used to Measure the Process Air Inlet Conditions of the
The process air outlet conditions are also measured with four external data loggers. As seen from
Chapter 2, in certain region there is rapid change in the outlet air temperature and water vapor
concentration. It is difficult for find a proper location to capture the average values of the outlet air
conditions. The average of the four data loggers does not represent the average of the entire
process air section. The process air outlet condition are therefore calculated from the measured
supply air and leaving heat pump coil air conditions, based on the energy and moisture balance
of air mixing occurring after the active desiccant wheel and the bypass damper.
- 95 -
m PA hPA,out + m bypass hPA,in = m SA hSA

.
Equation 3-28
m PA wPA,out + m bypass wPA,in = m SA wSA
Equation 3-29
The regeneration air inlet temperature is measured by the manufacturer installed sensor.
However, it is not reliable due to the non uniform air conditions and spatial limitations, as
explained in Section 3.3.2. In determining the active desiccant wheel performance, the
regeneration air inlet temperature is calculated assuming energy balance across the wheel. The
regeneration air inlet humidity is not directly measured due to spatial constraints. Instead, it is
calculated based on the outdoor humidity, considering the added moisture during the combustion
process occurring in the gas burner, shown in Equation 3-33. If a natural gas flow meter was
present in the system, the amount of added moisture could be determined from the natural gas
flowrate. Due to spatial constraints, the regeneration air inlet conditions are not confirmed using
external data loggers.
A manufacturer installed temperature sensor is available to measure the regeneration air outlet
temperature from the wheel. An external temperature and humidity data logger is placed in the
regeneration air outlet hood to measure the air humidity and to confirm the readings from the
manufacturer installed temperature sensor. Since both sensors are placed after the regeneration
fan, typical readings of the regeneration air outlet conditions can be obtained. The air temperature
increase due to the power consumption of the regeneration fan is accounted for in a similar
fashion to the building exhaust air outlet condition from the enthalpy recovery wheel.
Similar to the enthalpy recovery wheel, the process and regeneration air flowrate through the
active desiccant wheel is determined from the pressure drop across the wheel based on the
equations derived from the manufacturers lab.
.
m PA = e g PPA APA
Equation 3-30
- 96 -
m RgA = e g PRgA ARgA
Equation 3-31
where e = 1457
The calculation for the process and regeneration air flowrate indicates the laminar flow
characteristics, which is expected given the structure of the wheel and the air flowrate.
The total supply air flowrate is computed from the delta pressure measurement across the piezo
ring in the supply fan. Similarly, the coefficient f is determined based on the testing information
obtained from the manufacturers lab. The calculation is consistent with the flowrate
measurement with a Venturi flow meter.
.
m SA = f g
PSA
Equation 3-32
Aring
where f = 753.06
The measurements and calculations for the active desiccant wheel are summarized in Table 3-3.
Table 3-3 The Data/Equation for the Active Desiccant Wheel Performance Calculation
Source
Process air inlet temperature and humidity
Average of four direct measurements
Regeneration air inlet temperature
Calculated, energy balance across the wheel
Regeneration air inlet humidity
Process air outlet temperature and humidity
Calculated, Equation 3-28, Equation 3-29
Regeneration air outlet temperature and
Direct measurement
humidity
Process air flowrate
Regeneration air flowrate
- 97 -
The specifications of the instrumentation are the same as shown in Table 3-1. The performance
parameters of the active desiccant wheel are calculated based on Equation 2-70 and Equation 271.
3.3 Validation Results

3.3.1 Validation Results of the Enthalpy Recovery Wheel
The operating performance of the enthalpy recovery wheel during summer operation has been
tested under two flowrate settings: OA 0.64 m3/s (1360 cfm), RA 0.48 m3/s (1015 cfm) and OA
0.41 m3/s (870 cfm), RA 0.38 m3/s (800 cfm). During the winter testing, the OA and RA flowrates
were maintained at 0.64 m3/s (1300 cfm) and 0.51 m3/s (1080 cfm), respectively. For each test,
the conditions of inlet and outlet air streams were monitored continuously for several days, with a
data sampling interval of 5 minutes.
The FVR 2000 unit installed in the IW incorporates a stop/jog economizer control, which is used
to stop the wheel under moderate outdoor conditions for better energy efficiency. The jog function
is used to rotate the wheel periodically for self cleaning. For performance analysis, the data points
collected during this economizer operation were eliminated.
3.3.1.1 Summer Experiment

The inlet outside and building exhaust air conditions during the summer experiment with higher
flowrate setting are shown in Figure 3-13.
- 98 -
Figure 3-13 OA and RA Conditions during the Summer Experiment

Figure 3-13 indicates that the outside air humidity is higher than the building exhaust air humidity
for all data points. The outside air temperature is higher than the building exhaust air temperature
for most data points. In addition, the experiment provides some points where outside air
temperature is lower, as is observed on a cool, humid summer day. No data are shown when
outdoor temperature is around 23.5oC, due to the stop/jog economizer control as mentioned
earlier.
The sensible heat, moisture and total heat balances between the outside air supply and the
building space return air streams are established in the enthalpy recovery calculations, as shown
in Figure 3-14, Figure 3-15 and Figure 3-16, respectively. The sensible heat balance in Figure
3-14 shows a linear equivalence relation slightly displaced corresponding to a small constant heat
loss from the wheel. Figure 3-14 shows a few outliers in the measured data. They occur when the
sensible heat exchange is large, indicating the outdoor temperature is much higher than building
space temperature. From an inspection of the original experimental data, it is found that the
measured exhaust air temperature exceeds the outdoor temperature for these data points and it
happens during the early afternoon. Since the exhaust air temperature sensor was placed inside
the metal exhaust air hood, solar radiation absorbed by the hood might result in higher sensor
readings. A radiation shield over the exhaust air temperature sensor might be used to reduce the
measurement error.
- 99 -
Figure 3-14 Sensible Heat Balance during the Summer Experiment

The moisture balance in Figure 3-15 shows greater scatter due to the lesser accuracy of moisture
measurements. The linear equivalence relation is also displaced corresponding to a constant
moisture gain in the operation of the wheel. The total heat balance in Figure 3-16 shows a linear
equivalence relation displaced corresponding to a constant total heat gain in the operation of the
wheel, which in essence reflects the sensible heat and moisture balance.
Figure 3-15 Moisture Balance during the Summer Experiment

- 100 -
Figure 3-16 Enthalpy Balance during the Summer Experiment

Once the sensible, latent and total heat balances are confirmed, the wheel performance is
established by its sensible, latent and total heat recovery effectiveness. Measured and simulated
sensible heat loss of the outside air stream are plotted against the maximum possible sensible
heat exchange in Figure 3-17.
Figure 3-17 Simulated and Measured Sensible Heat Exchange

- 101 -
The plot indicates a linear relationship between the measured or simulated sensible heat
exchange and the maximum possible exchange. The slopes of the regression lines represent the
sensible heat recovery effectiveness. The intercepts represents the heat gain or loss in the
operation of the enthalpy recovery wheel. There appears to be some sort of sensible heat loss in
the simulation due to two reasons. Firstly, the results are obtained numerically and the number of
iterations is limited for reasonable computational time. Secondly, the results scatter resulting from
the different inlet air conditions.
Simulated and measured sensible heat recovery effectiveness are plotted against the
temperature difference between outdoor and the building space in Figure 3-18. The range of the y
axis in Figure 3-18 is limited and a few data points are eliminated in order to have a clearer view
of the data. The sensible heat recovery effectiveness in Figure 3-18 is nearly constant, except
when the outdoor temperature is close to the space temperature. When the outdoor temperature
approaches the indoor temperature, the denominator in the sensible heat recovery effectiveness
calculation approaches zero. Eventually the numerator also approaches zero, but it appears that
the denominator approaches zero at a faster rate. The sensible heat recovery effectiveness
approaches infinity. In addition, the effectiveness goes to either positive or negative infinity
depending on the direction from which the outdoor temperatures approach space temperatures,
indicating some sort of heat gain/loss in the operation of the wheel. Disturbances in the field
experiment such as solar radiation absorbed by the wheel enclosure, air leakage between
different chambers of the wheel or uneven air distribution in the chambers, which are negligible
for large temperature differences between outdoor and the building space, can cause significant
error in the measurement when the temperature differences are small. It is difficult to attribute the
heat gain/loss to any specific reason due to the field experiment nature of these measurements.
However, whatever the reasons might be, the sensible heat recovery effectiveness under such
conditions is not of significant concern, because the sensible heat exchange is small, as shown in
Figure 3-17.
- 102 -
Figure 3-18 Simulated and Measured Sensible Heat Recovery Effectiveness

Figure 3-19 plots the difference between simulated and measured sensible heat recovery
effectiveness during this experiment. Compared with the field experiment results, the simulation
underpredicts the sensible heat recovery effectiveness by about 5%. In general, the deviation
between the simulated and measured sensible heat recovery effectiveness increases with the
difference between the outdoor and indoor temperature, which is probably due to the heat
exchange between the wheel assembly and the outdoor surroundings.
- 103 -
Effectiveness
Simulated and measured moisture loss of the OA stream are plotted against the maximum
possible moisture exchange in Figure 3-20. Similar to the sensible heat exchange, Figure 3-20
shows a linear relationship between the measured or simulated moisture exchange and the
maximum possible exchange. The slopes of the regression lines indicate the moisture or latent
heat recovery effectiveness. The intercepts represents the moisture gain or loss in the operation
of the wheel.
Simulated and measured latent heat recovery effectiveness are plotted against the humidity
difference between the outdoor and the building space in Figure 3-21. The latent heat recovery
effectiveness is nearly constant for different outdoor humidities. The latent heat recovery
effectiveness shows larger scatter than the sensible heat recovery effectiveness due to the lesser
accuracy of the moisture measurements. In addition, there is larger scatter in the measured latent
heat recovery effectiveness when the outdoor humidity is closer to the indoor humidity. When the
amount of moisture transfer is small, small uncertainty in the measurements can cause large
uncertainty in the calculation of the latent heat recovery effectiveness.
- 104 -
Figure 3-20 Simulated and Measured Moisture Exchange
Figure 3-21 Simulated and Measured Moisture Recovery Effectiveness

Figure 3-22 plots the difference between simulated and measured moisture recovery
effectiveness during this experiment. The simulation results agree with the field measurements
within 10% for most of the data points.
- 105 -
Figure 3-22 Difference between Simulated and Measured Moisture Recovery Effectiveness
Figure 3-23 plots the simulated and measured enthalpy loss of the outside air stream against the
maximum possible enthalpy exchange. Figure 3-23 shows a linear relationship between the
measured or simulated enthalpy exchange and the maximum possible exchange. The slopes of
the regression lines indicate the enthalpy recovery effectiveness. The intercepts represents the
enthalpy gain or loss in the operation of the wheel.
Simulated and measured enthalpy recovery effectiveness are plotted against the outside air
temperature in Figure 3-24. For simulated data, the enthalpy recovery effectiveness is nearly
constant for different outdoor temperatures. For measured data, the enthalpy recovery
effectiveness appears to increase with the outdoor temperature, which essentially reflects the
sensible heat recovery performance shown in Figure 3-19. This is probably due to the heat
exchange between the wheel assembly and the outdoor surroundings, as explained earlier.
The average measured enthalpy recovery effectiveness during this testing is about 70%, as
indicated in Figure 3-24. No data are shown when outdoor temperature is around 23.5oC, due to
the stop/jog economizer control.
- 106 -
Figure 3-23 Simulated and Measured Enthalpy Exchange
Figure 3-24 Simulated and Measured Enthalpy Recovery Effectiveness

Figure 3-25 plots the difference between simulated and measured enthalpy recovery
effectiveness during this experiment. The simulation results agree with the field measurements
within 10%. The deviation is approximately linear with the outside air temperature, which is
related to the portion of sensible and latent heat exchange. When the outdoor temperature is
- 107 -
higher, there is a larger portion of sensible heat exchange in the enthalpy exchange. The
measured enthalpy recovery effectiveness is also affected more by the heat exchange between
the wheel and the ambient.
Figure 3-25 Difference between Simulated and Measured Enthalpy Recovery Effectiveness
The validation results for the summer operation with reduced flowrate are presented in Appendix
5. The deviation between the simulated and the measured effectiveness parameters are larger
than the previous experiment, which are probably related to the non uniform airflow patterns in
the wheel assembly or the dependence of heat and mass transfer coefficients on the airflow rate.
3.3.1.2 Winter Experiment

In the analysis of measured data obtained during the heating season in the Pittsburgh climate,
the moisture exchange in the wheel is not considered due to the lack of driving potential as a
result of negligible humidity difference between the outside and building exhaust air streams. In
essence, only sensible heat exchange performance is examined in the winter operation of the
enthalpy recovery wheel. The measured sensible heat recovery effectiveness is about 75%
during the winter operation of the enthalpy recovery wheel.
- 108 -
The difference between the simulated and measured sensible heat recovery effectiveness during
the winter experiment is shown in Figure 3-26. The plot indicates that the simulation agrees with
the measured results within -4~1%.
Effectiveness during the Winter Experiment
3.3.2 Validation Results of the Active Desiccant Wheel

Both the winter heating and the summer dehumidification performance of the active desiccant
wheel have been tested and compared with the simulation results from the performance model.
3.3.2.1 Winter Experiment

During the winter heating experiment, the process and regeneration air flowrates are maintained
at 0.47 m3/s (1000 cfm) and 0.25 m3/s (540 cfm), respectively. The sensible heat balance of the
active desiccant wheel functioning as a sensible heating device during winter operation is shown
in Figure 3-27. There appears to be a large sensible heat loss in the wheel assembly. One of the
reasons is suspected to be the inaccurate temperature measurement in the regeneration air inlet
leaving the gas burner based on two considerations. Firstly, the temperatures of the process air
- 109 -
inlet and the regeneration air outlet are directly measured and confirmed between the
manufacturer installed sensors and external data loggers. The supply air temperature, on which
the calculated process air outlet temperature is based, is also confirmed between the
manufacturer installed sensor and external data logger. Secondly, in the direct fired gas burner,
the natural gas flame temperature is approximately 1000oC. The combustion product has to be
mixed with much air in order to bring down its temperature to approximately 70-100oC, which is
required for regenerating the active desiccant wheel. There is limited space in the gas burner to
allow for the air mixing. Therefore, it is likely that the manufacturer installed sensor does not give
a typical reading of the regeneration air inlet temperature.
Figure 3-27 Sensible Heat Balance of the Active Desiccant Wheel during the Winter
Experiment
The other reason is suspected to be the air leakage between the process and regeneration air
chambers. From Figure 3-11, it is estimated that the pressure in the inlet regeneration air
chamber is less than the outlet process air chamber. Therefore the process air outlet could be
leaked into the regeneration air inlet. If this leakage exists, it occurs after the gas burner where
the regeneration air inlet temperature measurement has been made. Essentially, the regeneration
- 110 -
air entering the active desiccant wheel will have a lower temperature than what the measurement
indicates.
The regeneration air inlet temperature is calculated assuming the sensible heat lost by the
regeneration air is balanced with that gained by the process air. The calculated regeneration air
inlet temperature is compared with the measured value in Figure 3-28. On average the calculated
value is 39oC lower than the measurement.
Temperature of the Active Desiccant Wheel during the Winter Experiment
The sensible heat recovery effectiveness of the wheel is computed based on the measured
process air inlet and regeneration air outlet temperatures, and the calculated process air outlet
and regeneration air inlet temperatures, as indicated in Table 3-3. The measured sensible heat
recovery effectiveness during the active desiccant winter heating operation is around 90%. The
wheel performance is simulated based on the measured process air inlet and calculated
regeneration air inlet temperatures. The difference between the simulated and measured results
is plotted in Figure 3-29. The plot indicates that the simulation agrees with the measurements
within 8%.
- 111 -
Effectiveness of the Active Desiccant Wheel during the Winter Experiment
3.3.2.2 Summer Experiment

In the summer dehumidification operation, there also exists large deviation between the enthalpy
gained by the process air and the enthalpy lost by the regeneration air, as shown in Figure 3-30.
Figure 3-30 Enthalpy Balance of the Active Desiccant Wheel during the Summer
Experiment
- 112 -
Similar to the winter operation, the regeneration air inlet temperature is calculated based on the
enthalpy balance of the active desiccant wheel, as shown in Figure 3-31. Compared to the winter
experiment, the calculated regeneration air inlet temperature is higher and less uniform in the
summer experiment.
Temperature of the Active Desiccant Wheel during the Summer Experiment
As stated previously, the regeneration air inlet humidity is calculated based on the outdoor
humidity, considering the added moisture during the combustion process in the gas burner. The
amount of moisture added during the combustion process is estimated from the energy and
material balance of combustion. The chemical reaction occurring in the burner can be written as:
4 N 2 + O2 + xCH 4 = xCO2 + 2 xH 2 O + 4 N 2 + (1 2 x)O2
Equation 3-33
Assuming the lower heating value of natural gas 800 kJ/mole, the combustion process adds
around 2g/kg of moisture into the regeneration air inlet when the outdoor temperature is 20oC and
the regeneration air inlet temperature is 70oC.
- 113 -
The moisture balance, which is based on the measured process air inlet and regeneration air
outlet humidities, and the calculated process air outlet and regeneration air inlet humidities, is
established in Figure 3-32. The plot shows large scatter in the moisture balance due to the
uncertainty in the moisture measurement.
Figure 3-32 Moisture Balance of the Active Desiccant Wheel during the Summer
Experiment
The simulated and measured regeneration efficiency is plotted in Figure 3-33. The heat carryover
ratio is plotted in Figure 3-34. The simulation over predicts the regeneration efficiency and under
predicts the heat carryover ratio. The possible explanations include: 1) the parameters, such as
the heat and mass transfer coefficients, the heat of adsorption, used in the performance model
differ from the reality. The published heat and mass transfer coefficients from the textbook can
deviate from the reality by 20%. Under different operating conditions, the heat of adsorption can
also differ by 10%. 2) some mechanisms occurring in the active desiccant wheel operation, such
as non uniform inlet air conditions, uneven desiccant loading and air carryover, crossover or
leakage between air streams, are not properly captured.
- 114 -
When the uncertainties in the field measurement and the model prediction are considered, the
simulated regeneration efficiency and heat carryover ratio show reasonable agreement with the
measured values, as discussed in Section 3.4.
Figure 3-33 Simulated and Measured Regeneration Efficiency
Figure 3-34 Simulated and Measured Heat Carryover Ratio
- 115 -
3.4 Uncertainty Analysis

Two types of uncertainty are relevant to the work performed in this thesis. Firstly, due to
experiment setup and instrumentation uncertainty, there is uncertainty in the calculated
performance indicators. Secondly, due to the uncertainty in the parameters and inlet air
conditions used in the simulation, there is uncertainty in the predicted performance indicators.
Both uncertainties can be estimated using standard error propagation methods, such as the one
recommended by NIST[44]:
U F = U xi
i xi
Equation 3-34
When applying Equation 3-34 to estimate the uncertainty associated with the experiment setup
and instrumentation error, x represents the directly measured quantities. Partial derivative
F
can be obtained from the equations used to calculate the performance indicators based on x.
x
When estimating the uncertainty associated with the model prediction, x represents the
parameters or the inlet air conditions used in the simulation. Partial derivative
F
represents the
x
sensitivity of the model prediction to the variable x, which can be obtained from the model. The
uncertainty associated with the instrumentation error is presented in this section.
Based on the instrumentation specifications shown in Table 3-1, the uncertainties in the
measured sensible heat, moisture and total heat recovery effectiveness parameters for the
summer experiments are calculated. They are represented by Us, Ul and Ut respectively. The
possible range of the measured sensible, latent heat and enthalpy recovery effectiveness can
then be represented by [s-Us, s+Us], [l-Ul, l+Ul] and [t-Ut, t+Ut].
- 116 -
The comparison between the simulated effectiveness parameters and the measured quantities
together with their uncertainties are shown in Figure 3-35 through Figure 3-37 for the summer
experiment with higher flowrate. The plots for the summer experiment with reduced flowrate are
presented in Appendix 5. The uncertainty in the sensible heat recovery effectiveness for winter
operation is presented in Figure 3-38. The sensible heat recovery effectiveness shown in Figure
3-35 varies widely, because it approaches either negative or positive infinity when the outdoor
temperature is close the space return air temperature as explained earlier. The moisture recovery
effectiveness has larger uncertainty than the sensible heat recovery effectiveness, since the air
humidity is calculated from the measured air temperature and relative humidity.
For the summer experiment with higher flowrate, the simulated sensible heat, moisture and total
heat recovery effectiveness lies within the possible range of the measured quantities, which
indicates good agreement between the simulation and the measurements. The same is also true
for the winter experiment.
Effectiveness with Uncertainty Range during the Summer Experiment
- 117 -
Figure 3-36 Comparison between the Simulated and Measured Moisture Recovery
Figure 3-37 Comparison between the Simulated and Measured Total Heat Recovery
- 118 -
Effectiveness with Uncertainty Range for the Enthalpy Recovery Wheel during the Winter
Experiment
The comparison between the simulated sensible heat recovery effectiveness and the measured
quantities together with their uncertainties for the active desiccant wheel operating as a sensible
heating device in winter is shown in Figure 3-39. The plot indicates reasonable agreement
between the simulation and the field experiment.
The comparison between the simulated regeneration efficiency and the measured quantities
together with their uncertainties for the active desiccant wheel operating as a dehumidification
device in summer is shown in Figure 3-40. The plot for the heat carryover ratio is shown in Figure
3-41. For both performance indicators, the predictions lie on the edge of the uncertainty range.
- 119 -
Effectiveness with Uncertainty Range for the Active Desiccant Wheel during the Winter
Experiment
Figure 3-40 Comparison between the Simulated and Measured Regeneration Efficiency
with Uncertainty Range during the Summer Experiment
- 120 -
Figure 3-41 Comparison between the Simulated and Measured Heat Carryover Ratio with
Uncertainty Range during the Summer Experiment
The performance model developed in this thesis is based on fundamental scientific and
engineering principles. The convective heat transfer coefficient is determined from the Nusselt
number for fully developed laminar flow in a tube with constant heat flux boundary condition. The
convective mass transfer coefficient is determined by applying the heat and mass transfer
analogy and assuming a unity Lewis number. The adsorption and thermodynamic properties of
the desiccants are taken from public literature. The published heat and mass transfer coefficients
from the textbook can deviate from the reality by 20%. Under different operating conditions, the
heat of adsorption of the desiccant can also differ by 10%. When the uncertainty in model
prediction is considered, Figure 3-40 and Figure 3-41 indicate reasonable agreement between
the model prediction and the field measurements.
3.5 Summary
This chapter deals with the experimental validation of the performance model for the enthalpy
recovery and active desiccant wheels. The experimental platform, the field experiment setup
including the instrumentations and calculations involved in the experiments, and the validation
- 121 -
results have been reported. Two steps have been taken in analyzing the experimental data:
establishing the energy and material balance across the wheel, and validating the model
prediction against the field measurements.
The only way to determine the operating performance of the enthalpy recovery and active
desiccant wheels in a field application is through field testing. The field testing is complicated by
the heat and/or moisture exchange with the surroundings, air leakage from/to the surroundings,
the non uniform inlet air conditions or wheel construction, and the air carryover or leakage
between different air streams. The field testing is further complicated by the non uniform outlet air
conditions resulting from the wheel rotation. This non uniformity is more significant for the active
desiccant wheel, which makes its performance testing even more difficult. For the active
desiccant wheel installed in the IW, the not well mixed combustion product entering the wheel
regeneration side makes the situation more complicated.
The experiments have been conducted on the SEMCO REV 2250 active desiccant integrated
ventilation system and the FVR 2000 enthalpy recovery module. The sensible heat recovery
performance of the enthalpy recovery and the active desiccant wheels during winter operation,
the sensible and latent heat recovery performance of the enthalpy recovery wheel during summer
operation, and the regeneration performance and heat carryover effect of the active desiccant
wheel during summer operation have been tested. The tests have been carried out using
manufacturer installed sensors and external data loggers placed at various locations in the
machine.
Uncertainties associated with the measured performance parameters have also been estimated
based on the specifications of the instrumentations. Reasonable agreements between the
simulated and the measured performance parameters have been obtained for both the enthalpy
recovery and the active desiccant wheels, during their winter and summer operations. This
- 122 -
agreement indicates that the model well represents all significant mechanisms occurring in the
desiccant wheels.
- 123 -
Chapter 4 Integration of the Enthalpy Recovery Wheel in

HVAC System Design
When considering the inclusion of an enthalpy recovery wheel in a ventilation system, HVAC
designers need to consider both the installed cost of this wheel and its benefits in terms of
reduced costs for
cooling and heating equipment due to the reductions in their required capacities;
operation in providing cooling and heating of the ventilation air.
The most important factors in determining these cost benefits provided by an enthalpy recovery
wheel are ventilation air flowrate, climatic conditions, enthalpy recovery wheel performance,
operation schedule and control strategies. A detailed hourly building energy simulation that
considers all of these factors is needed in order to provide accurate estimates of the costs;
however, in the preliminary design phase of an HVAC system a relatively simple calculation can
provide useful cost benefit estimates of reasonable accuracy.
4.1 Evaluation Procedure

The following steps can be taken in evaluating the cost benefits of an enthalpy recovery wheel.
The system shown in Figure 3-3 and Figure 3-4 is used to illustrate how this procedure is
followed in a specific application. This system is assumed to operate as a DOAS that provides for
both the ventilation needs and the entire dehumidification requirements for a space, the IW in
Pittsburgh, PA, with its maximum occupancy of 60 people. The DOAS supplies air at space
neutral temperature with a low dew point as needed and space cooling/heating is provided by
hydronic, water-based, cooling/heating units.
1. Locate the outdoor design condition, OA, and the desired indoor condition, RA, on a
psychrometric chart.
- 124 -
ASHRAE[14] provides a set of outdoor design conditions for different cities across the
United States. The 0.4% wet bulb cooling design condition for Pittsburgh is 29.2oC dry
bulb temperature and 63.3% relative humidity, RH. For indoor conditions, ASHRAE[45]
specifies a comfort zone based on indoor air temperature, indoor humidity and radiant
temperature. The desired indoor condition in summer selected in this example is 23.3oC
dry bulb temperature and 50% RH.
The 99.6% heating design dry bulb temperature for Pittsburgh is -16.7oC. The desired
indoor dry bulb temperature in winter is 23.3oC. This value assumes that no
humidification equipment is available in the system and that the indoor humidity during
winter will vary depending on outdoor humidity and the moisture generation inside the
space.
2. Determine the required supply air, SA, flow, temperature and humidity.
A dedicated outdoor air system, DOAS, provides the entire ventilation load for the space.
ASHRAE 62.1-2004[46] prescribes ventilation air requirement for different building types
in order to achieve acceptable indoor air quality. For an office space, the ventilation air
requirement is 9.44 L/s-person. The DOAS for the IW, with its maximum occupancy of 60
people, needs an airflow of 0.566 m3/s under design conditions.
Supply air temperature is determined based on the amount of sensible load the DOAS
needs to carry. For the IW, the DOAS supplies air at space neutral temperature, 23.3oC.
The DOAS does not bear any space sensible load. On the other hand, the DOAS needs
to meet the entire latent load of a space. Based on this requirement, the supply air
humidity can be calculated from the desired space humidity, the supply air flowrate, the
moisture generated in the space, and the moisture brought in by outside air infiltration.
- 125 -
Consequently, the supply air conditions determined in this example are 23.3oC dry bulb
temperature and 35% RH in summer, and 23.3oC dry bulb temperature in winter. As
explained earlier, the supply air humidity during winter is allowed to vary with outdoor
humidity.
3. Determine the outlet air conditions from the enthalpy recovery wheel, LEW and EA.
From the manufactures catalog, an enthalpy recovery wheel is selected based on the
flow rate, directly related to the face velocity, and the desired enthalpy recovery
performance as well as the financial and spatial constraints. Total heat recovery
effectiveness increases with decreasing wheel face velocity in addition with increasing
wheel depth.
For a given outside air flowrate, the wheel performance also varies with the exhaust and
outside air flow ratio. The outside air flowrate is determined by the ventilation air
requirement in the space. Generally speaking, the exhaust air flowrate is less than the
outside air flow in order to maintain a positive pressure in the space. For simplicity, the
exhaust flow selected in this example is the same as the outside air flow, 0.566 m3/s.
Ideally, a wheel is chosen that provides the lowest possible face velocity and greatest
depth. However, bigger and deeper wheels require more capital investment, occupy more
space and cause a higher pressure drop in the system resulting in a higher fan power
consumption. Based on all these considerations, a wheel is selected optimizing its
performance, size, and cost. The wheel selected in this example has a sensible and total
heat recovery effectiveness of 75% at the design outside and exhaust air flowrates. The
air conditions leaving an enthalpy recovery wheel can be determined from the following
equations:
- 126 -
t LEW = t OA
s m min (t OA t RA )
.
Equation 4-1
m OA
.
hLEW = hOA
t m min (hOA hRA )

.
Equation 4-2
m OA
.
t EA = t RA +
s m min (tOA t RA )
.
Equation 4-3
m RA
.
hEA = hRA +
t m min (hOA hRA )

.
Equation 4-4
m RA
As shown in Figure 3-3, the outside air, OA, and the return air, RA, are in counter current
flow through the enthalpy recovery wheel at equal rates. The two lines in Figure 3-4
representing the OA and RA in the enthalpy wheel are parallel in this example.
4. Illustrate the overall air handling processes on the psychrometric chart.
The air handling processes can be illustrated on the psychrometric chart based on the
overall system configuration. The overall air handling processes for this example, shown
in Figure 3-3 are illustrate in Figure 3-4, with the solid and dashed lines representing the
system with and without an enthalpy recovery wheel, respectively. For a system without
an enthalpy recovery wheel, the process starts from OA. With enthalpy recovery, the
process starts from LEW. In either case, the air needs to be conditioned to the desired
supply air condition, SA.
Since the enthalpy recovery wheel recovers moisture from the building exhaust air, the
winter supply air humidity and space humidity are slightly higher than those without
enthalpy recovery wheel. The presence of an enthalpy recovery wheel does not change
the summer supply air humidity and space humidity, because the humidity is controlled.
When an enthalpy recovery wheel is present, the control of the system is adjusted to
- 127 -
obtain the same desired supply air conditions. For the system illustrated in Figure 3-3 and
Figure 3-4, this adjustment means a proper leaving DX coil air condition and a proper
leaving active desiccant wheel air condition.
Under some operating conditions, there is no need to adjust the control of the system
when an enthalpy recovery wheel is included. For example, for the summer design
condition shown in Figure 3-4, the leaving DX coil air condition with and without the
enthalpy recovery wheel is the same. Consequently the leaving active desiccant wheel
air condition LDW and LDW* are also the same.
The determination of LDX and LDX* assumes a bypass ratio of 60% and a 20% enthalpy
increase of the process air after it goes through the active desiccant wheel as a result of
heat carryover from the regeneration section.
5. Determine the reductions in the system cooling/heating capacities provided by the

enthalpy recovery wheel and consequently in the capital investments.
The calculations required in this step are different for different overall ventilation system
configurations. For the system illustrated in Figure 3-3 and Figure 3-4 in summer, the
enthalpy change from OA to LDX indicates the load on the cooling coil without an
enthalpy recovery wheel; that from LEW to LDX*, the loads with the enthalpy recovery
wheel.
The calculation of desiccant regeneration heating loads assumes that the amount of heat
required to regenerate the desiccant equals the amount of heat to vaporize the adsorbed
water and that the efficiency of the regeneration system is 40%. The corresponding
humidity changes from LDX to SA and from LDX* to SA indicate the summer load on the
- 128 -
regeneration air heating coil and its heat supply without and with an enthalpy recovery
wheel.
In winter, the enthalpy change from OA to SA indicates the load on the heating coil
without an enthalpy recovery wheel; from LEW to SA*, the load on the heating coil with
the enthalpy recovery wheel.
The design load calculations for the cooling/heating and the regeneration coils and for the
requisite cooling/heating system based on these conditions illustrated in Figure 3-4 are
the load on summer cooling coil without an enthalpy recovery wheel: 22.8 kW;
the load on summer cooling coil with an enthalpy recovery wheel: 11.1 kW;
the regeneration heating load without an enthalpy recovery wheel: 10.0 kW;
the regeneration heating load with an enthalpy recovery wheel: 10.0 kW;
the load on winter heating coil without an enthalpy recovery wheel: 26.6 kW;
the load on winter heating coil with an enthalpy recovery wheel: 6.6 kW.
According to Besant and Simonson[47], the capital cost of cooling and heating equipment
varies from $100 to $350 per kW. Based on the average value, $225 per kW, calculations
in this example indicate:
the saving on summer cooling equipment cost: $2,638;
the saving on summer regeneration equipment cost is $0;
the saving on winter heating equipment cost: $4,483.
In current market, the capital cost of the enthalpy recovery wheel is about $5 per L/s[47].
For the wheel selected in this example, the capital investment is $2,830.
As mentioned earlier, the system in this example uses a heat pump unit for summer
cooling and winter heating. Since the heating load is greater than the cooling load, the
- 129 -
heat pump unit is sized based on the heating load. Therefore, the saving on capital
investment due to the reduced capacity of the heat pump is $4,483. Considering the
capital cost of the enthalpy recovery wheel, the saving on capital investment of the entire
system is $1,653.
6. Estimate the annual operating costs savings.
Hourly weather data for ambient air temperature and humidity, and utility prices for
electrical energy and natural gas are required for this estimation. The website of the
Department of Energy provides hourly weather files for various cities in the United States
and selected cities in other countries. The utility price used in this calculation is $0.07 per
kWh for electricity and $10.50 per thousand ft3 for natural gas.
In general, the savings on summer operating costs depend on the air handling processes
involved. The annual cooling load reduction can be obtained by summing the hourly
cooling load reduction through the entire cooling season.
The hourly cooling load reduction for the system illustrated in Figure 3-3 and Figure 3-4 is
plotted in Figure 4-1. Figure 4-1 indicates that for some hours the load reduction on the
cooling coil is negative; the enthalpy recovery wheel actually increases the cooling coil
load. This situation occurs when outdoor enthalpy is lower than indoor enthalpy, but the
HVAC system is in a cooling mode due to solar radiation and heat generation in the
space. If the enthalpy recovery wheel is properly controlled, as assumed in this example,
its operation will be suspended or the wheel bypassed under such conditions. The annual
load reduction on the cooling coil is calculated as 7,660 kWh, after removing the negative
reductions. Savings of $134 are achieved in the electricity costs for summer cooling with
a heat pump unit that has a COP of 4 for cooling.
- 130 -
Figure 4-1 Hourly Load Reduction on the Cooling Coil

The hourly load reduction on the regeneration coil is calculated based on the humidity
difference between LDX and LDX*. As seen from Figure 4-2, the enthalpy recovery wheel
reduces the load on the regeneration coil except for a few hours. The annual
regeneration load reduction is obtained as 16,319 MJ by summing the hourly
regeneration load reduction through the entire cooling season. The regeneration gas cost
saving is calculated as $162.
- 131 -
Figure 4-2 Hourly Load Reduction on the Regeneration Coil

In winter, the enthalpy recovery wheel functions primarily as a sensible heat recovery
wheel since the moisture recovery is negligible. The estimation of the savings on winter
heating energy consumption is therefore straightforward. An enthalpy recovery wheel
with sensible heat recovery effectiveness s can reduce the load on the winter heating
coil by s in percent, as illustrated in Figure 3-4. For the wheel selected in this example,
the load on winter heating coil is reduced by 75%.
For the system without the enthalpy recovery wheel, the hourly heating load can be
calculated based on the difference between hourly outdoor temperature and the desired
supply air temperature, and the ventilation air flowrate:
.
Q = m OA Cp (t SA t OA )
Equation 4-5
The hourly heating load reduction provided by the enthalpy recovery wheel is Q*s. The
annual heating load reduction can be obtained by summing the hourly heating load
reduction through the entire heating season. For the heating season from Jan 1 to Apr 30
and Oct 16 to Dec 31, the annual heating load reduction provided by the enthalpy
- 132 -
recovery wheel is calculated to be 47,557 kWh in this example. If the heat pump unit has
a COP of 5 for winter heating, savings of $666 on electricity costs for winter heating are
achieved.
7. Estimate the payback period of the enthalpy recovery wheel.
The payback period of an enthalpy recovery wheel can be calculated based on the
incremental capital cost, the annual operating cost savings and discount rate. In this
example, net savings of $1,653 on capital investment and savings of $962 annually on
operating cost of the heat pump unit and the gas boiler are expected. The payback on an
enthalpy recovery wheel is immediate.
4.2 Discussion
This example calculation deals with an enthalpy recovery wheel integrated with a DOAS system.
Enthalpy recovery wheels can also be applied to VAV systems to precondition the outside air.
The calculation procedure for a VAV system is similar to that for a DOAS, except that the air
handled by the VAV system is the mixture of the outside air and the recirculation air from the
space, and the amount of air is much greater. With reference to Figure 3-4, the air entering the
DX coil will be the mixture of OA and RA when an enthalpy recovery wheel is not present; the
mixture of LEW and RA, when an enthalpy recovery wheel is present. The mixed air condition lies
on the line connecting RA and OA or LEW. It will be closer to RA, since the amount of RA is
greater than that of OA or LEW. The reductions on the cooling/heating equipment capacities and
the annual operating cost for a VAV system remain the same as those for a DOAS. An enthalpy
recovery wheel is still an economic design option.
This example assumes a heat pump unit for summer cooling and winter heating, and a gas
burner for summer regeneration of the active desiccant wheel. If a DX coil is used for summer
cooling and a gas boiler is used for winter heating, this winter heat supply can serve for summer
- 133 -
regeneration, and the regeneration heating equipment cost can be eliminated. In this case, the
saving on capital investment of the entire system is $4,291, and the reduction on the annual
operating cost for electricity and natural gas is $2,001. The payback for an enthalpy recovery
wheel is also immediate.
In addition, this example assumes no winter humidification equipment incorporated in the system.
If humidification equipment is present, the difference between indoor and outdoor humidity will be
substantial and moisture recovery in an enthalpy recovery wheel will provide a significant benefit
in reducing the energy and cost requirements of ventilation air humidification. The savings in
humidification costs will be a function of the moisture recovery effectiveness of the enthalpy
recovery wheel.
Furthermore, this example calculation indicates the potential energy and cost savings in
Pittsburgh, PA. In fact, an enthalpy recovery wheel can be expected to provide substantial energy
savings under most climatic conditions. The results of a detailed energy simulation in EnergyPlus,
by Lam et al[9] predict the energy saving potential of enthalpy recovery at different geographical
locations, shown in Figure 4-3. While the plot does not provide precise numbers, it does give
HVAC designers a starting point when considering the integration of enthalpy recovery in system
designs under different climatic conditions.
- 134 -
Figure 4-3 Predicted Annual Energy Savings at Different Location[9]

Besides the savings on capital investments and operating costs, enthalpy recovery wheels are
also expected to enhance the dehumidification performance of HVAC systems and thus to
improve the indoor humidity conditions[4], which are important to occupants health, comfort and
productivity.
4.3 Summary
Based on the knowledge gained from field testing and computer modeling, a procedure has been
developed to estimate the capital cost reduction and the annual operating cost savings provided
by the enthalpy recovery wheel. This chapter provides guidance for HVAC engineers on the
applicability and economics of an enthalpy recovery wheel.
The most important factors in determining the benefits provided by an enthalpy recovery wheel
are climatic conditions, ventilation air flowrate, enthalpy recovery wheel performance, operation
schedule and control strategies. When properly applied, an enthalpy recovery wheel provides
reduced capital investments in the cooling and heating equipment, which help pay for the
- 135 -
enthalpy recovery wheel itself. It also reduces the operating energy consumption and peak
electricity demand. The payback on an enthalpy recovery wheel is immediate for most climatic
conditions. Besides the benefits on capital cost and operating energy usage, an enthalpy
recovery wheel can also improve indoor humidity conditions, essential to occupants health,
comfort and productivity.
- 136 -
Chapter 5 Integration of the Active Desiccant Wheel in

CHP System Design
In Chapter 2 and Chapter 3, the performance model of desiccant wheels has been developed and
validated. Using the IWESS project as a specific example, this chapter illustrates how this model
can be applied in developing operating strategies for the active desiccant wheel integrated CHP
system.
5.1 Operating Performance and Cost of the Active Desiccant

Wheel
For the system illustrated in Figure 3-3 and Figure 3-4, the required cooling and moisture removal
are achieved by three components, namely the enthalpy recovery module, the heat pump unit
and the active desiccant module. The enthalpy removal, which includes both sensible and latent
cooling, accomplished by each of the three components at a typical summer operating condition
is plotted in Figure 5-1. The control settings corresponding to Figure 5-1 are
OA condition: T = 28.9 oC, RH = 52.0%, w = 0.0139 kg/kg, v = 0.63 m3/s
RA condition: T = 23.9 oC, RH = 50.7%, w = 0.0090 kg/kg, v = 0.27 m3/s
SA condition: T = 19.1 oC, RH = 53.5%, w = 0.0073 kg/kg, v = 0.89 m3/s
Total heat removal: 22.0 kW
Total moisture removal: 0.0048 kg/s
- 137 -
Figure 5-1 Enthalpy Removal Breakdown

It is seen that about one fifth of the enthalpy reduction is achieved by the FVR enthalpy recovery
module and the rest by the heat pump module. As expected, the active desiccant module made
very limited contribution to enthalpy reduction, since the air temperature rises as its humidity is
reduced. But enthalpy reduction is only part of the story in HVAC operation; moisture removal
has also to be considered since it affects occupants thermal comfort and the operation of the
sensible cooling devices such as radiant panels and fan coils. The moisture removal
accomplished by each component is plotted in Figure 5-2. The active desiccant wheel provides
more than half of the moisture removal; the rest is provided by the cooling coil and the enthalpy
recovery module.
The operating cost of the system is shown in Figure 5-3. Among the three components in the
system, the enthalpy recovery module has the lowest operating cost. The enthalpy recovery
module only requires power input to operate the exhaust fan and to rotate the wheel, a very
limited amount. The high price of natural gas in current fuel market is responsible for the high
operating cost of the active desiccant wheel.
- 138 -
Figure 5-2 Moisture Removal Breakdown
Figure 5-3 Operating Cost Breakdown

If the regeneration energy for the active desiccant wheel can be obtained for free or at
significantly reduced cost from renewable energy sources or as the rejected heat from power
generation systems, the active desiccant wheel will become very attractive since relatively small
amount of power is needed to run the regeneration fan and to rotate the wheel. This alternative is
explored in this chapter, using the IWESS as a specific example.
- 139 -
5.2 Development of Operating Strategies for the Active

Desiccant Integrated CHP System
The following questions need to be answered when integrating the active desiccant wheel in a
CHP system:
Are the desired supply air conditions achievable with the available regeneration air
temperature and quantity?
How should the variable control factors be adjusted in order to achieve the desired supply
air conditions?
The following steps can be taken in developing operating strategies for the integrated system:
1. Determine the temperature and quantity of the rejected heat available from the upstream
equipment.
For the IWESS project, a bio diesel engine generator has been installed as the prime
mover. 20 kW thermal energy can be made available from the engine coolant, which is a
mixture of glycol and water. It has been proposed to utilize this rejected heat from the
generator to regenerate the active desiccant wheel in the cooling season. Due to spatial
constraints, the engine generator is located in the basement of the building, six levels
down from the ventilation system. It has been considered to make this rejected heat
available to the ventilation system through two heat exchangers: a coolant-water heat
exchanger located next to the engine generator and a water-air heat exchanger located
inside the ventilation system. The heated water is used as the energy transfer medium
since it is more energy effective in transferring thermal energy from the basement to the
ventilation system.
- 140 -
The quantity of the rejected heat from the engine coolant matches the regeneration
energy requirement of the active desiccant wheel. According to the manufacture of the
engine generator, the coolant temperature can go to 93oC without affecting the engine
performance. The approximation between the regeneration air temperature and the
coolant temperature depends on the heat loss in the system and the design of the heat
exchangers. With proper insulation on the piping and reasonable heat transfer areas in
the heat exchangers, the regeneration air temperature can reach 70oC.
2. Determine the indoor, outdoor design conditions and the supply air requirements from the
active desiccant wheel integrated HVAC system.
The ventilation system illustrated in Figure 3-3 and Figure 3-4 is assumed to operate as a
DOAS that provides for both the ventilation needs and the entire dehumidification
requirements for a space, the IW in Pittsburgh, PA, with its maximum occupancy of 60
people. The DOAS supplies air at space neutral temperature and space cooling/heating
is provided by hydronic, water-based, cooling/heating units.
The outdoor and indoor design conditions as well as the desired supply air conditions
have been provided in Chapter 4:
Outdoor design condition: dry bulb temperature 29.2oC, RH 63.3%, humidity ratio
0.0161 kg/kg;
Indoor design condition: 23.3oC, RH 50%, 0.00893 kg/kg;
Desired supply air condition: 23.3oC, RH 35%, 0.00622 kg/kg, flowrate 0.566
m3/s
There are five variable control factors in this system: the regeneration air temperature,
the regeneration air flowrate, the rotary speed of the active desiccant wheel, the bypass
ratio around the active desiccant wheel and the leaving DX coil air conditions.
- 141 -
The regeneration air temperature and flowrate are obviously restricted by the upstream
equipment, the engine generator. The maximum regeneration air temperature is limited
by the coolant temperature. It is set at 70oC in this example. The maximum regeneration
air flowrate is limited by the amount of rejected heat available. It is determined to be
0.378 m3/s.
3. Develop performance maps for the proposed integrated system.
In order to answer previous questions, performance maps that relate the supply air
conditions and different settings of the variable control factors need to be developed. The
supply air conditions are determined from the leaving DX coil air conditions, the leaving
desiccant wheel air conditions and the bypass ratio based on the enthalpy and moisture
balance of air mixing as seen from Figure 3-3. The validated performance model is used
to predict the leaving desiccant wheel air conditions based on the leaving DX coil air
conditions, the regeneration air conditions and the process air flowrate, which is a
function of the bypass ratio.
Figure 5-4 through Figure 5-7 plot the predicted supply air conditions with relation to the
wheel rotary speed and the bypass ratio around the active desiccant wheel, at various
regeneration air flowrates. These plots assume a fixed leaving DX coil air condition: dry
bulb temperature 12.4oC and humidity ratio 0.00896 kg/kg, and a fixed regeneration air
temperature 70 oC. The solid lines in these plots represent the supply air humidity and the
dashed lines represent the supply air temperature. Different lines in the plots indicate
different bypass ratio around the active desiccant wheel.
- 142 -
Bypass Ratio, Regeneration Flowrate 0.0944 m3/s
- 143 -
4. Determine whether the desired supply air conditions are achievable and how the various
control factors should be adjusted.
- 144 -
As seen from Figure 5-4 through Figure 5-7, the supply air humidity decreases and the
supply air temperature increases with decreasing bypass ratio. In addition, the supply air
humidity decreases and the supply air temperature increases with increasing
regeneration air flowrate. The desired supply air condition is not achievable with any
bypass ratio when the regeneration flowrate is 0.0944 m3/s. The desired supply air
humidity is achievable with bypass ratio 0.2, 0.3 or 0.4, when the regeneration flowrate is
0.189 m3/s. It is even obtainable with bypass ratio 0.5, when the regeneration flowrate is
0.283 or 0.378 m3/s. The desired supply air humidity can not be achieved with any
regeneration air flowrate when the bypass ratio is set at 0.6. For smaller bypass ratio and
higher regeneration air flowrate, for example bypass ratio 0.3 and regeneration flow
0.378 m3/s, the supply air temperature is lower than the desired value when the supply
air humidity reaches the desired level. The lower supply air temperature means that the
DOAS is now carrying some space sensible load, which provides some freedom or safety
factor in the overall system operation.
5.3 Discussion
The previous performance maps are developed based on the selected leaving DX coil air
condition and regeneration air temperature. The supply air conditions are solved as functions of
the regeneration air flowrate, the wheel rotation speed and the bypass ratio. In a similar fashion,
the supply air conditions can be solved as functions of the regeneration air temperature, the
regeneration airflow and the leaving DX air condition, given a certain wheel rotation speed and
bypass ratio.
Seen from the performance maps, the answer to the second question stated at the beginning of
Section 5.2 is not unique. The desired supply air condition can be obtained from different settings
of the various control factors. A logical extension to the question would be: which settings will
result in the lowest operating energy consumption or operating cost? It then becomes an
- 145 -
optimization problem, with the operating energy consumption or cost as the objective function.
This problem can be approached by defining the operating energy consumption or cost
associated with the change of each operating variable and minimizing the objective function by
varying the settings of the control variables.
Taking the ventilation system shown in Figure 3-3 and Figure 3-4 as a specific example, the
optimization function can be described as:
Minimize (system operating energy consumption) = f(regeneration air flowrate, regeneration air
temperature, leaving DX air condition, wheel rotation speed, bypass ratio)
The regeneration air flowrate is related to objective function through fan power required to
push/draw the regeneration air through the active desiccant wheel. The regeneration air
temperature is related to the objective function through the upstream equipment, the engine
generator, and the loss when transferring the thermal energy from the engine generator to the
ventilation system. In this system, the leaving DX air condition is controlled by adjusting the
frequency of the compressor. Therefore the leaving DX air condition is related to the objective
function through the energy use of the vapor compression system. The wheel rotation speed is
controlled by adjusting the variable frequency drive (VFD) on the wheel motor. The wheel rotation
speed is therefore related to the objective function through the power use of the wheel motor,
which is limited. The bypass ratio is related to the objective function through the fan power
required to push/draw the process air through the active desiccant wheel.
The sensitivity of the system operating energy consumption with regard to the different control
variables can be obtained analytically or empirically from the system operation. The optimization
can be then conducted based on the sensitivity study.
- 146 -
5.4 Summary
In this chapter, a procedure has been outlined for developing the operating strategies for the
active desiccant wheel integrated CHP system. The validated performance model has been
applied in predicting the supply air conditions from the integrated system under different settings
of the control variables.
Performance maps that relate the supply air conditions and the different settings of the control
variables have been constructed for the proposed system in the IWESS project. These
performance maps show that the desired supply air condition can be achieved with certain
settings of the control variables, namely the regeneration air temperature, the regeneration air
flowrate, the leaving DX air condition, the wheel rotation speed, and the bypass ratio. The
settings of the control variables to achieve the desired supply air condition are not unique. It is an
optimization problem to determine the control settings to achieve the lowest system operating
energy consumption or cost, in which the system operating energy consumption or cost is the
objective function, and different control factors are the variables.
- 147 -
Chapter 6 Contribution, Conclusion and Future Work

This thesis presents the development and validation of an equation based model to predict the
operating performance of both enthalpy recovery and active desiccant wheels. This model has
been applied to evaluate the costs and benefits provided by the enthalpy recovery wheel, and to
develop operating strategies for the active desiccant wheel integrated CHP system.
6.1 Contributions
1. Development of a moisture transfer model based on physical analysis of the desiccant
materials
There are inter particle and intra particle pores in the desiccant materials and the moisture is
adsorbed on the surface of the intra particle pores. The moisture transfers from the air flowing
in the wheel channels to the intra particle pores in two steps: gas side transfer, during which
the water vapor molecules transport from the air to the surface of the desiccant composite,
and solid side transfer, during which the moisture transfers from the surface of the desiccant
layer to the inter particle pores and then to the intra particle pores. Heat is released at the
location where the adsorption occurs.
Gas side transfer is controlled by the convective mass transfer coefficient as well as the
vapor pressure difference between the air and the desiccant composite. There are three
diffusion mechanisms, namely ordinary diffusion, Knudsen diffusion and surface diffusion,
occurring in the desiccant composite. The solid side transfer is controlled by the diffusion
coefficient, the diffusion area and the moisture concentration difference for each diffusion
mechanism.
Despite the complexity, the controlling process for the moisture transfer in the desiccant
wheels is the moisture transport from the bulk air flowing in the channels to the hypothetical
- 148 -
surface of the desiccant layer, due to the small thickness of the desiccant layer. The
predictions from the performance model developed based on this understanding reasonably
agree with the field experimental data, which indicates that the model well captures all of the
significant mechanisms occurring in the wheel operation.
2. Determination of the effect of the practical issues, including the wheel purge, the residual
water and the wheel supporting structure, on the operating performance of the desiccant
wheels
The enthalpy recovery wheel is usually built with a purge section, in order to prevent cross
contamination between the building exhaust air outlet and outside air intake. The air leaving
the purge section is directed into the building exhaust air inlet stream. It goes through the
wheel one more time and ends up in the return air outlet stream. The required purge area is
dependent on the depth of the wheel, the velocity of the purge gas flow and wheel rotation
speed. Inclusion of a purge section in the enthalpy recovery wheel reduces the wheel
performance by about 5%. The purge section is however an important design feature since
the cross contamination between the outside and the building exhaust air streams is
prevented and the indoor air quality is improved.
In the active desiccant wheel, the air carryover is limited due to the low rotation speed of the
wheel. A purge section can however be used to reduce the heat carryover from the
regeneration to the process air section. This is not commonly done in reality, but it is
expected that a purge section will help improve the operating performance of the active
desiccant wheel.
The desiccant wheels are never fully regenerated. There is residual water contained in the
desiccant material even after the regeneration process. This residual water can be modeled
by adjusting the density of the desiccant material. The residual water in the enthalpy recovery
- 149 -
wheel has very limited impact on its energy recovery performance. The residual water
increases the heat carryover ratio of the active desiccant wheel, since the wheel heat
capacity is increased. The impact on the active desiccant wheel is also limited, but larger
than the enthalpy recovery wheel, which is due to the larger ratio between the mass of the
enthalpy recovery wheel media and the mass of the air resulting from its high rotation speed.
The wheel supporting structures such as spokes and casing can be modeled as added
thermal mass to the substrate, assuming they go through the same temperature fluctuation
as the substrate. The supporting structure has limited impact on the energy recovery
performance of the enthalpy recovery wheel, because the mass of the wheel media is
relatively large compared to the mass of the air. The supporting structure has significant
impact on the regeneration efficiency and the heat carryover ratio of the active desiccant
wheel. For the wheel simulated in this thesis, the supporting structure decreases the
regeneration efficiency by 2% and increases the heat carryover ratio by 10%.
3. Construction of an equation based performance model for both the enthalpy recovery and
active desiccant wheels based on fundamental scientific and engineering principles, rather
than curving fitting of experimental data
The combined heat and mass transfer problem in simulating the operating performance of the
desiccant wheels has been formulated and implemented. This model relates the operating
performance of the desiccant wheels to their design parameters and operating conditions.
This performance model is applicable to both the enthalpy recovery and active desiccant
wheels. This performance model can be used selecting the design parameters and operating
variables of the desiccant wheels a certain application, in order for better indoor thermal
conditions, improved building energy efficiency and reduced costs. It can also be used to
diagnose experimental data for the desiccant wheels.
- 150 -
4. Empirical validation of the performance model using data collected from field experiments
The only way to determine the operating performance of the enthalpy recovery and active
desiccant wheels in a field application is through field testing. The field testing is complicated
by the heat and/or moisture exchange with the surroundings, air leakage from/to the
surroundings, the non uniform inlet air conditions or wheel construction, and the air carryover
or leakage between different air streams. The field testing is further complicated by the non
uniform outlet air conditions resulting from the wheel rotation. This non uniformity is more
significant for the active desiccant wheel, which makes its performance testing even more
difficult.
Through careful experiment design and data acquisition, the operating performance of the
SEMCO REV 2250 active desiccant integrated ventilation system and the FVR 2000 enthalpy
recovery module installed in the IW has been tested. The experimental data have been used
to validate the performance model developed in this thesis. Guidance has also been provided
on locating the sensors for representative readings of inlet and outlet air conditions.
5. Systematic comparison between the enthalpy recovery and active desiccant wheels
The enthalpy recovery and active desiccant wheels are two types of solid desiccant based
equipment commonly used in HVAC systems. They share similarities in the wheel structure,
the use of solid desiccant materials and the combined heat and mass transfer processes.
However they have significant difference in terms of the design intention, the design
parameter selection and the operating conditions, which have been systematically compared
in this thesis. The temperature and moisture concentration characteristics with time, the
wheel rotation, and distance, the wheel depth, have been plotted and justified for both wheels.
- 151 -
6. Development of performance indicators for both the enthalpy recovery and active desiccant
wheels
The performance indicators for the enthalpy recovery wheel include the sensible, latent or
moisture, and total heat or enthalpy recovery effectiveness. The performance of the enthalpy
recovery wheel can be characterized by two of the three effectiveness parameters. The third
can be calculated based on the two and the operating conditions of the wheel. The
performance indicators developed for the active desiccant wheel include the moisture
removal capacity, regeneration efficiency and heat carryover ratio. These performance
indicators can be used to make comparative evaluations of the operating performance of
different desiccant wheels. They can also be used to calculate the outlet air conditions from
the desiccant wheels.
7. Development of an evaluation procedure to estimate the costs and benefits provided by the
enthalpy recovery wheel
Based on the knowledge gained from field testing and computer modeling, a procedure has
been developed to estimate the capital cost reduction and the annual operating cost savings
provided by the enthalpy recovery wheel. Guidance on the applicability and economics of an
enthalpy recovery wheel has been provided for HVAC engineers.
8. Development of a procedure for developing control strategies for the active desiccant wheel
integrated CHP system
A procedure has been outlined to develop operating strategies for the active desiccant wheel
integrated CHP system. The validated performance model has been applied in predicting the
supply air conditions from the integrated system under different settings of the control
variables. Performance maps that relate the supply air conditions and different settings of the
- 152 -
control variables have been constructed for the proposed system in the IWESS project.
These performance maps show that the desired supply air condition can be achieved with
certain settings of the control variables, namely the regeneration air temperature, the
regeneration air flowrate, the leaving DX air condition, the wheel rotation speed, and the
bypass ratio. The settings of the control variables to achieve the desired supply air condition
are not unique. It is an optimization problem to determine the control settings to achieve the
lowest system operating energy consumption or cost, in which the system operating energy
consumption or cost is the objective function, and different control factors are the variables.
6.2 Conclusions
The following conclusions have been drawn from the work accomplished in this thesis.
1. The combined heat and mass transfer processes are similar for both the enthalpy recovery
and active desiccant wheels. It is therefore possible to have a performance model that is
applicable to both wheels.
2. The controlling process for the moisture transfer in the desiccant wheels is the moisture
transport from the bulk air flowing in the channels to the hypothetical surface of the desiccant
layer, due to the small thickness of the desiccant layer. The combined heat and mass transfer
problem in the desiccant wheels can be formulated as one dimensional since both heat and
mass transfer Biot numbers are small and the temperature and moisture concentration
gradient in the thickness direction are negligible.
3. Inclusion of the energy and moisture storage in the air has very limited impact on the
predicted performance of both enthalpy recovery and active desiccant wheels. Inclusion of
the substrate heat capacity improves both the sensible and the latent heat recovery
performance of the enthalpy recovery wheel. Inclusion of a substrate layer deteriorates the
performance of the active desiccant wheel. The axial heat conduction through the substrate
- 153 -
deteriorates the sensible heat recovery performance of the enthalpy recovery wheel and it
has little impact on the latent heat recovery performance of the wheel.
4. Inclusion of a purge section in the enthalpy recovery wheel reduces the wheel performance
by about 5%. The residual water in the enthalpy recovery wheel has very limited impact on its
energy recovery performance. The impact of the residual water on the active desiccant wheel
is also limited, but larger than the enthalpy recovery wheel. The wheel supporting structure
has limited impact on the energy recovery performance of the enthalpy recovery wheel, but it
has significant impact on the regeneration efficiency and the heat carryover ratio of the active
desiccant wheel. The supporting structure decreases the regeneration efficiency by 2% and
increases the heat carryover ratio by 10%.
5. Reasonable agreements between the simulated and the measured performance parameters
have been obtained for the enthalpy recovery and active desiccant wheels, during both winter
and summer operations. This agreement indicates that the model well represents all
significant mechanisms occurring in the desiccant wheels.
6. The most important factors in determining the benefits provided by an enthalpy recovery
wheel are climatic conditions, ventilation air flowrate, enthalpy recovery wheel performance,
operation schedule and control strategies. When properly applied, an enthalpy recovery
wheel provides reduced capital investments in the cooling and heating equipment, which help
pay for the enthalpy recovery wheel itself. It also reduces the operating energy consumption
and peak electricity demand. The payback on an enthalpy recovery wheel is immediate for
most climatic conditions.
7. The active desiccant wheel represents a good application for utilizing the rejected heat from
the power generation processes. Using the IWESS project as a specific example, it is found
that the desired supply air condition can be achieved with different settings of the control
- 154 -
variables, namely the regeneration air temperature, regeneration air flowrate, leaving DX air
condition, wheel rotation speed and bypass ratio.
6.3 Future Work

1. Investigation of the impact of the active desiccant wheel on the indoor thermal conditions and
other air conditioning units
The active desiccant wheel is expected to provide better indoor humidity control and
therefore improved energy efficiency and effectiveness for the indoor sensible cooling
devices such as radiant panels, fan coils or water mullions. With improved indoor humidity
control, the thermostat can have a higher setting, still achieving the same level of thermal
comfort. The chilled water temperature entering the sensible cooling devices can be lower,
which results in more effective cooling. Using the IW as a specific example, the impact of the
active desiccant wheel on the indoor thermal conditions and other air conditioning units are to
be studied.
Based on the experience gained in the IW, investigations are to be made on how to design
and operate active desiccant dehumidification equipment together with sensible cooling
devices in an energy efficient and cost effective manner in order to provide occupants with
healthy and comfortable indoor environment.
2. Integration of the active desiccant wheel with the engine generator
The performance maps presented in Chapter 5 relate the supply air condition to the
regeneration air flowrate, the wheel rotation speed and the bypass ratio around the wheel,
assuming a fixed regeneration air temperature and leaving DX air condition. More
performance maps are to be developed to explore the impact of the regeneration air
temperature and leaving DX air condition on the supply air condition and overall system
- 155 -
energy performance. Since the setting of the control variables to achieve the desired the
supply air conditions are not unique, optimization will be conducted to determine the settings
of the control variables that result in the lowest system operating energy consumption or cost.
Finally, the integrated system is to be designed, installed, tested and evaluated.
3. Replacement of the DX coil with a chilled/hot water coil in the SEMCO ventilation system
The SEMCO ventilation system uses a DX coil for air cooling and dehumidifying in summer
and air heating in winter. To better fit the idea of system integration, replacement of the DX
coil with a chilled/hot water coil, which makes use of the chilled water from the absorption
chiller and the hot water from the solar thermal receivers, is planned. Generally speaking, a
chilled water coil has less dehumidification capability than a DX coil, which means the air
entering the subsequent active desiccant wheel will be more humid. The impact on the wheel
performance is to be investigated. The integrated system is to be designed, installed, tested
and evaluated.
4. Development and implement of more intelligent operating strategies for the ventilation system
During the testing period, the ventilation system in the IW has been operated continuously,
regardless of the indoor and outdoor conditions. The system is supplied with sophisticated
control logics from SEMCO, which determines the settings of its control variables, such as
the compressor frequency, supply fan frequency, natural gas valve, etc, based on user
specified supply air flowrate, temperature and humidity setpoints. Currently, these setpoints
are manual inputs. In order for better energy efficiency and cost effectiveness, more
intelligent operating strategies, which decide the on/off of the system or the different
components in the system and the various setpoint inputs to the current control logic, based
on the occupancy, the indoor and outdoor thermal conditions, are to be developed,
implemented, tested and evaluated.
- 156 -
5. Systematically comparison of the different approaches to do building ventilation and indoor

humidity control
In the past, there have been at least three ways to do building ventilation and indoor humidity
control, as shown in Appendix 4. A systematic comparison between the three system
configurations, including its installed equipment capacity, capital cost, operating energy and
cost needs to be conducted.
6. Development of the design procedures for the desiccant wheels
A design procedure is to be developed for the desiccant wheels. The procedure will outline
the steps taken to go from the design requirements to the selection of the various design
parameters. The procedure will also outline how the performance model can be used to
facilitate the wheel design process. The procedure will also provide guidance on the selection
of various operating variables, optimizing the capital and operating cost, compactness and
pressure drop of the desiccant wheels.
7. Plan for additional instrumentation for the SEMCO ventilation system
Based on the experience gained during the field experiments, additional instrumentation is to
be planned for the performance monitoring and data acquisition of the SEMCO ventilation
system. For example, the humidity of the leaving DX air is not measured. Since the air
leaving the DX coil is close to saturation, capacitance RH sensors currently used in HVAC
systems will not work. Other types of RH sensors that are designed for high humidity
conditions need to be investigated and installed.
- 157 -
In addition, a natural gas flow meter will have to be installed to measure the natural gas
consumption of the system. This natural gas flow measurement can be used to calculate the
regeneration air inlet temperature leaving the gas burner and confirm the readings from the
manufacturer installed temperature sensor.
8. Investigation of the impact of wheel purge on the active desiccant wheel performance
A purge section in the active desiccant wheel is expected to reduce the amount of heat
carryover from the regeneration to the process air section. Presumably it can improve the
operating performance of the active desiccant wheel. The impact of the wheel purge on the
active desiccant wheel performance is to be explored using the performance model.
9. Uncertainty analysis of the model predictions
As pointed out in Chapter 3, there are uncertainties associated with the model prediction due
to the uncertainties in the input parameters and inlet air conditions. The sensitivity of the
model prediction with regard to these input parameters and inlet air conditions are to be
studied. The uncertainties in model prediction are to be quantified.
10. Empirical validation of the predicted temperature and humidity profiles of the desiccant
material and air
The predicted average leaving air conditions from the desiccant wheels have been validated
in this thesis. The temperature and humidity profiles of the air and the temperature and
moisture loading profiles of the desiccant material with respect to time and distance are to be
measured and validated.
- 158 -
Reference
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
Fisk, W. (2002). How IEQ Affects Health, Productivity. ASHRAE Journal 44(5): 56-60.
Bayer, C. W., S. A. Crow and J. C. Fischer (2000). Causes of Indoor Air Quality Problems
in Schools. Technical Report Prepared for the Department of Energy
Sand, J. R. and J. C. Fischer (2003). Active Desiccant Integration with Packaged Rooftop
HVAC Equipment. 21st International Congress of Refrigeration. Washington DC
TIAX (2003). Matching the Sensible Heat Ratio of Air Conditioning Equipment with the
Building Load SHR.
Zhai, C., X. Liao, et al. (2005). Simulation Analysis of Building Humidity Control and
Energy Consumption in Summer for Different System Configurations. 9th International
Building Performance Simulation Association (IBPSA) Conference, Montreal, Canada.
Mumma, S. A. and K. Shank (2001). Achieving Dry Outside Air in an Energy Efficient
Manner. ASHRAE Transations 107(1): 553-561
ASHRAE (2004). ANSI/ASHRAE Standard 90.1-2004, Energy Standard for Buildings
Except Low-rise Residential Buildings, Atlanta, ASHRAE.
TIAX (2002). Energy Consumption Characteristics of Commercial Building HVAC Systems
Volume III: Energy Savings Potential
Lam, K. P., S. Lee, et al. (2005). Simulation of the Effect of an Energy Recovery Ventilator
on Indoor Thermal Conditions and System Performance. 9th International Building
Performance Simulation Association (IBPSA) Conference, Montreal, Canada.
Radermacher, R. and X. Liao (2003). The Impact of Waste Heat Activated Desiccant
Systems on Buildings. International Conference on Building Systems and Facilities
Management, Singapore.
Jeong, J. W. (2004). Simplified Ceiling Radiant Cooling Panel and Enthalpy Wheel Models
for Dedicated Outdoor Air System Design. Department of Architectural Engineering. State
College, Pennsylvania State University. PhD Thesis.
ASHRAE (2000). Handbook of HVAC Systems and Equipment. Atlanta, ASHRAE.
Fischer, J. C. and T. L. Thomas (1994). Langmuir Moderate Type 1 Desiccant Mixture for
Air Treatment. U. S. P. a. T. Office. United States, SEMCO Incorporated: 1-28.
ASHRAE (2001). Handbook of Fundamentals. Atlanta, ASHRAE.
Fischer, J. C. (2006). Personcal Communication.
Slayzak, S. J. and J. P. Ryan (2000). Desiccant Dehumidification Wheel Test Guide,
National Renewable Energy Laboratory.
Belding, W. A., W. M. Worek, et al. (1991). "Desiccant Development for Gas-fired
Desiccant Cooling Systems." ASHRAE Transactions 1991, Part 1: 587-594.
Collier, R. K., B. M. Cohen, et al. (1992). Desiccant Properties and Their Effects on the
Performance of Desiccant Cooling Systems. Desiccant Cooling and Dehumidification. L.
Harriman. Atlanta, ASHRAE: 75-81.
Ruthven, D. M., (1984). Principles of Adsorption and Adsorption Processes, Wiley,
Toronto.
Kays, W. M. and A. L. London (1984). Compact Heat Exchangers. Third Edition, New York,
McGraw-Hill Book Company.
Maclaine-cross, I. L. and P. J. Banks (1972). Coupled Heat and Mass Transfer in
Regenerators Prediction Using an Analogy with Heat Transfer. International Journal of
Heat and Mass Transfer 15: 1225-1242.
Simonson, C. J. (1998). Heat and Moisture Transfer in Energy Wheels. Department of
Mechanical Engineering. Saskatoon, University of Saskatchewan. PhD Thesis.
Van den Bulck, E., J. W. Mitchell, et al. (1985a). "Design Theory for Rotary Heat and Mass
Exchangers - I. Wave Analysis of Rotary Heat and Mass Exchangers with Infinite Transfer
Coefficients." International Journal of Heat and Mass Transfer 28(8): 1575-1586.
Van den Bulck, E., J. W. Mitchell, et al. (1985b). "Design Theory for Rotary Heat and Mass
Exchangers - II. Effectiveness-Number-of-Transfer-Units Method for Rotary Heat and mass
Exchangers." International Journal of Heat and Mass Transfer 28(8): 1587-1595.
- 159 -
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
Van den Bulck, E., J. W. Mitchell, et al. (1992). The Design of Dehumidifiers for Use in
Desiccant Cooling and Dehumidification Systems. Desiccant Cooling and Dehumidification.
L. Harriman. Atlanta, American Society of Heating, Refrigerating and Air-conditioning
Engineers (ASHRAE): 118-126.
Zheng, W. and W. Worek (1993a). "Numerical Simulation of Combined Heat and Mass
Transfer Processes in a Rotary Dehumidifier." Numerical Heat Transfer, Part A 23: 211232.
Zheng, W., W. M. Worek, et al. (1993b). "Control and Optimization of Rotational Speeds for
Rotary Dehumidifiers." ASHRAE Transactions 1993, Part 1: 825-833.
Zheng, W., W. M. Worek, et al. (1995a). "Effect of Operating Conditions on Optimal
Performance of Rotary Dehumidifiers." Journal of Energy Resources Technology 117(1):
62-66.
Zheng, W., W. M. Worek, et al. (1995b). "Performance Optimization of Rotary
Dehumidifiers." Journal of Solar Energy Engineering 117(1): 40-44.
Simonson, C. J. and R. W. Besant (1997a). "Heat and Moisture Transfer in Desiccant
Coated Rotary Energy Exchangers: Part I. Numerical Model." HVAC&R Research 3(4):
325-350.
Simonson, C. J. and R. W. Besant (1997b). "Heat and Moisture Transfer in Desiccant
Coated Rotary Energy Exchangers: Part II. Numerical Model." HVAC&R Research 3(4):
351-368.
Jeong, J. W. and S. A. Mumma (2005). "Practical Thermal Performance Correlations for
Molecular Sieve and Silica Gel Loaded Enthalpy Wheels." Applied Thermal Engineering
25: 719-740.
Majumdar, P. (1998). "Heat and Mass Transfer in Composite Desiccant Pore Structures for
Dehumidification." Solar Energy 62(1): 1-10.
Zhang, L. Z. and J. L. Niu (2002). "Performance Comparison of Desiccant Wheels for Air
Dehumidification and Enthalpy Recovery." Applied Thermal Engineering 22: 1347-1367.
Sphaier, L. A. and W. M. Worek (2004). Analysis of Heat and Mass Transfer in Porous
Sorbents Used in Rotary Regenerators. International Journal of Heat and Mass Transfer
47: 3415-3430.
ASHRAE (1998). Standard 139-1998 Method of Testing for Rating Desiccant
Dehumidifiers Utilizing Heat for the Regeneration Process. Atlanta, ASHRAE.
http://www.arc.cmu.edu/cbpd/index.html
http://www.cmu.edu/iwess/
Incropera, F. P. and D. P. Dewitt (1996). Fundamentals of Heat and Mass Transfer, Fourth
Edition, Wiley, Toronto.
Shang, W. and Besant, R. W., 2001, Energy Wheel Effectiveness Evaluation, Part II:
Testing and Monitoring Energy Wheels in HVAC Applications, ASHRAE Transaction,
107(2).
Georgia Institute of Technology (1998). Independent Performance Verification of SEMCOs
Total Energy Recovery Wheels, Part 2b-Test Results for FV Series of Total Energy
Recovery Wheels, Technical Report Prepared for SEMCO Incorporated.
ANSI/ASHRAE Standard 84-1991, Method of Testing Air-to-Air Heat Exchangers,
ASHRAE, Atlanta.
ARI Standard 1060-2005, Performance Rating of Air-to-Air Heat Exchangers for Energy
Recovery Ventilation Equipment, ARI, Arlington.
Taylor B. N. and C. E. Kuyatt (1994). Guidelines for Evaluating and Expressing the
Uncertainty of NIST Measurement Results. NIST Technical Note 1297. 1994 Edition.
Gaithersburg, MD.
ASHRAE (2004). ANSI/ASHRAE Standard 55-2004, Thermal Environmental Conditions for
Human Occupancy, Atlanta, ASHRAE.
ASHRAE (2004). ANSI/ASHRAE Standard 62.1-2004, Ventilation for Acceptable Indoor Air
Quality, Atlanta, ASHRAE.
Besant, R. W. and C. J. Simonson (2000). Air-To-Air Energy Recovery, ASHARE Journal,
May 2000: 31-38
- 160 -
Appendix 1 Finite Difference Analysis of the Adsorption Isotherm

Appendix 2 Profiles of the Air and Desiccant Conditions with Distance for the
Appendix 3 Profiles of the Air and Desiccant Conditions with Distance for the
Appendix 4 More on SEMCO REV 2250
Appendix 5 Summer Experiment of the Enthalpy Recovery Wheel, Reduced
Flow
- 161 -
Appendix 1 Finite Difference Analysis of the Adsorption

Isotherm
The adsorption isotherm of the desiccant shown in Equation 2-13 is a non linear algebraic
equation. It has to be linearized, converted into the finite difference form and solved together with
the other four governing equations. There are different approaches to linearize this equation and
to convert it into the finite difference form. Some approaches work and the other do not. Before
the one presented in Section 2.2.2 is taken, three other approaches have been tried.
The first approach tried in this thesis is to linearize the adsorption isotherm shown in Equation 213 using Taylors theorem.
m =
m
t m
t m + m vm
t m
vm
t (it , ix)
it 1,ix m
m
t m
m
vm
t (it 1, ix)
it 1,ix m
Equation A1-1
it 1,ix
m
vm
vm (it , ix ) + m (it , ix)
it 1,ix
vm (it 1, ix) + m (it 1, ix)
Equation A1-2
The partial derivative m in Equation 2-4 is represented in the finite difference form as
t
m m (it , ix) m (it 1, ix)

=
t
t
Equation A1-3
The finite different forms of other governing equations remain the same as Equation 2-30,
Equation 31, Equation 2-32 and Equation 2-34. The governing equation set are therefore
converted into five linear algebraic finite difference equations. There are five unknown variables in
- 162 -
this equation set; besides tg(it,ix), tm(it,ix), vg(it,ix) and vm(it,ix), m(it,ix) becomes another
unknown variable in this equation set.
This model is programmed and used to simulate the operating performance of the enthalpy
recovery wheel. The temperature and vapor concentration of the air and the desiccant converge
after a certain number of iterations. However the moisture loading of the desiccant increases with
increasing number of iterations. The wheel reaches steady state operation except for the
desiccant loading. In each rotation, some moisture is deposited in the wheel. The isotherm
relationship is violated.
Then the second approach is tried. Instead of using Equation 2-13 which represents the moisture
loading in the desiccant to its temperature and water vapor concentration, the water vapor
concentration in the desiccant is represented by its temperature and moisture loading.
vm =
p vm, sat c
Rv Tm ( max + c 1)
m
Equation
A1Error! No text of specified style in document.-4
vm =
vm
t m + vm m
t m
m
Equation A1-5
The same problem as encountered in the first approach still exists. Comparing the modeling
results from the first and the second approaches, the predicted temperature and water vapor
concentration of the air and the desiccant are the same. But the predicted desiccant loadings are
different.
The change of the moisture loading in the desiccant predicted by the second approach is not
reasonable. However, this does not seem to affect the predicted air conditions.
- 163 -
It is suspected that the problem was resulted from the truncation errors in the Taylor expansion
shown in Equation A1-1 and Equation A1-5. Therefore the third approach, which does not apply
the Taylor theorem, is tried. The algorithm used in the third approach is as follows:
the mass transfer rate is estimated at the beginning of each timestep;
the governing equations are solved for tg(it,ix), tm(it,ix), vg(it,ix) and m(it,ix), based on the
estimated mass transfer rate;
vm(it,ix) is determined from the adsorption isotherm;
the rate of mass transfer is updated based on the newly obtained vg(it,ix) and vm(it,ix);
this process is repeated until a converged solution is obtained;
the calculation is then moved on to the next timestep and the same process is followed
until the periodic steady-state operation of the wheel is reached.
The third approach works well for simulating both the enthalpy recovery and the active desiccant
wheels. The problem encountered in the first two approaches is eliminated. However, this
approach takes significant computational time since a number of iterations have to be made
before a converged solution is reached.
Finally the approach described in Section 2.2.2, which expresses the partial derivative m using
t
m
m
partial derivatives t m and vm , and therefore eliminates the unknown variable m(it,ix) from the
equation set, is applied. This approach does not have the problem associated with the first and
second approaches, and it needs less computational time than the third approach.
- 164 -
Appendix 2 Profiles of the Air and Desiccant Conditions

with Distance, the Enthalpy Recovery Wheel
The profiles of the air and desiccant conditions with respect to space, the depth of the enthalpy
recovery wheel are plotted in Figure A2-1 through Figure A2-5. The plots indicate the air and
desiccant conditions for six rotation angles, three in the adsorption section and three in the
desorptoin section. The adsorption section covers the rotation angle =0o+~180o-. The desorption
section covers =180o+~0o-. The plots also include the average desiccant conditions during the
adsorption and the desorption sections.
Seen from Figure A2-1 and Figure A2-2, the temperature and water vapor concentration of the
outside air decrease as it flows through the channels in the wheel, because energy and moisture
are transferred from the outside air to the channels of the wheel; the temperature and water vapor
concentration of the building exhaust air increase as it flows through the channels in the wheel,
because energy and moisture are transferred back from the wheel to the building exhaust air.
The air carryover effect mentioned earlier is more clearly shown in Figure A2-1 and Figure A2-2.
At the ration angle =0o+ where the wheel switches from the building exhaust air to the outside
air section, the warmer and more humid outside air clears out the cooler and less humid building
exhaust air contained in the channels of the wheel. Similarly the cooler and less humid building
exhaust air clears out the warmer and more humid outside air contained in the channels of the
wheel at =180o+, where the wheel switches from the outside air to the building exhaust air
section. Except for these small regions, the changes in the temperature and water vapor
concentration of the air appear to be linear with the depth of the wheel.
The dashed lines in Figure A2-1 and Figure A2-2 indicate the average outside air and building
exhaust air outlet conditions. The plots clearly show that the average outlet air conditions are
- 165 -
close to those at the 90o angle from the separation seals, when the wheel rotates half way
through the adsorption and desorption sections.
Figure A2-1 Profile of the Air Temperature, with the Depth of the Enthalpy Recovery Wheel
Figure A2-2 Profile of the Water Vapor Concentration in the Air, with the Depth of the
- 166 -
The desiccant conditions for the six rotation angles, together with the average desiccant
conditions for the adsorption and the desorption sections, are plotted in Figure A2-3 through
Figure A2-5. In general, the temperature, water vapor concentration and the moisture loading of
the desiccant decrease with the depth of the wheel, since the warmer and more humid outside air
enters the wheel from the top and the cooler and less humid building exhaust air enters it from
the bottom. Seen from Figure A2-3, the desiccant temperature swings back and forth as the
wheel switches between the outside air and building exhaust air sections; so does the water
vapor concentration shown in Figure A2-4, and the moisture loading in the desiccant shown in
Figure A2-5. The average desiccant conditions for the adsorption section are close to those for
the desorption section; the average desiccant conditions for both sections are close to those at
the 90o rotation angle from the separation seals.
Figure A2-3 Profile of the Desiccant Temperature, with the Depth of the Enthalpy Recovery
Wheel
- 167 -
Figure A2-4 Profile of the Water Vapor Concentration in the Desiccant, with the Depth of
the Enthalpy Recovery Wheel
Figure A2-5 Profile of the Moisture Loading in the Desiccant, with the Depth of the
- 168 -
Appendix 3 Profiles of the Air and Desiccant Conditions

with Distance, the Active Desiccant Wheel
The profiles of the air and desiccant conditions with respect to space, the depth of the active
desiccant wheel, are plotted in Figure A3-1 through Figure A3-5. The plots indicate the air and
desiccant conditions for six rotation angles, three in the adsorption section and three in the
desorptoin section. The adsorption section covers the rotation angle =0o+~240o-. The desorption
section, =240o+~0o-. The plots also include the average desiccant conditions during the
adsorption and the desorption sections.
Seen from Figure A3-1 and Figure A3-2, the water vapor concentration of the process air
decreases as it flows through the channels in the wheel from the top to the bottom, because
moisture is transferred from the process air to the channels of the wheel. Its temperature
increases as a result of the heat of adsorption. The water vapor concentration of the regeneration
air increases as it flows through the channels from the bottom to the top, because moisture is
transferred back from the wheel to the regeneration air. Its temperature decreases since heat is
transferred from the regeneration air to the wheel.
Similar to the enthalpy recovery wheel, the air carryover effect is more clearly shown in Figure
A3-1 and Figure A3-2. At the ration angle =0o+ where the wheel switches from the regeneration
air to the process air section, the warmer and more humid regeneration air clears out the cooler
and less humid process air contained in the channels of the wheel. The cooler and less humid
process air clears out the warmer and more humid regeneration air contained in the channels of
the wheel at =240o+, where the wheel switches from the process air to the regeneration air
section.
- 169 -
Figure A3-1 Profile of the Air Temperature, with the Depth of the Active Desiccant Wheel
Figure A3-2 Profile of the Water Vapor Concentration in the Air, with the Depth of the Active
Desiccant Wheel
The desiccant conditions for the six rotation angles, together with the average desiccant
conditions for the adsorption and the desorption sections, are plotted in Figure A3-3 through
- 170 -
Figure A3-5. The desiccant temperature increases, and its water vapor concentration and
moisture loading decrease with the depth of the wheel, because the process air entering the top
of the wheel has a lower temperature and higher relative water vapor concentration.
Unlike the enthalpy recovery wheel, the average desiccant conditions for the adsorption section
are different from those for the desorption section. The desiccant conditions in the active
desiccant wheel vary more significantly than those in the enthalpy recovery wheel.
Figure A3-3 Profile of the Desiccant Temperature, with the Depth of the Active Desiccant
Wheel
- 171 -
Figure A3-4 Profile of the Water Vapor Concentration in the Desiccant, with the Depth of
the Active Desiccant Wheel
Figure A3-5 Profile of the Moisture Loading in the Desiccant, with the Depth of the Active
Desiccant Wheel
- 172 -
Appendix 4 More on SEMCO REV 2250

Comparison between the SEMCO and Traditional Approaches for
Dehumidification
In a conventional rooftop air conditioning unit, the air flow is cooled below its dew point
temperature to reduce its moisture content and then is reheated to the desired supply
temperature, as indicated by the dashed lines in Figure A4-1. In prior desiccant integrated
systems, all of the air flow first flows through the desiccant wheel to achieve the desired supply
humidity, approximately a constant enthalpy process. A large amount of moisture is adsorbed,
and the air temperature leaving the desiccant wheel is high, as shown by the dotted lines in
Figure A4-1.
Figure A4-1 Psychrometric Representation of Different Air Handling Processes for

Pittsburgh Summer Design Day
(solid lines indicate the SEMCO approach; dashed lines, the conventional rooftop air conditioning
approach; and dotted lines, the prior desiccant integrated approach.)
As seen from Figure A4-1, the installed ventilation unit offers the following advantages over
conventional rooftop air conditioning units and previous desiccant integrated systems:
- 173 -
higher outlet cooling coil temperature than conventional rooftop unit. It reduces the
probability of frozen coil and other equipment or control problems associated with that.
lower cooling coil capacity than a conventional rooftop unit, comparable to previous
desiccant integrated systems. Calculations indicate a 35% reduction in cooling load for a
Pittsburgh summer design day compared with a conventional rooftop unit, shown in
Figure A4-1. The cooling coil can operate at full load or close to full load conditions for a
greater portion of the time, and the coil performance is improved.
better humidity control at part load conditions. In conventional rooftop unit, the moisture
condensed on the cooling coil can re-evaporate and enter the space when the
compressor cycles off and the supply fan continues to operate, resulting in elevated
indoor humidity level during part load conditions. In this installed ventilation unit, the reevaporated moisture is captured by the desiccant wheel, thus providing improved indoor
humidity conditions.
less heat required for the installed desiccant wheel regeneration than for conventional
rooftop reheating and previous desiccant integrated system regeneration. The amount of
reheat and regeneration energy is illustrated in Figure A4-1. For the SEMCO approach,
regeneration energy can be calculated from the humidity difference between leaving coil
air and leaving desiccant wheel air and the heat of vaporization of water, by assuming the
required amount of regeneration heat equals the amount of heat required to vaporize the
adsorbed water. Similarly the regeneration energy required for previous desiccant
integrated approach can be calculated from the humidity difference between outside air
and leaving desiccant wheel air. For conventional rooftop approach, reheat energy can
be calculated from the temperature difference between the supply air and the leaving
cooling coil air, and the specific heat of air. Consideration has to be given to the amount
of air being processed in different systems. Assuming 1/3 of total supply air going through
the desiccant wheel, which is a normal setting for Revolution unit, the regeneration
energy required by the installed system is calculated to be 58% of the rooftop reheat
- 174 -
energy and 55% of the previous desiccant wheel regeneration energy for Pittsburgh
summer design day indicated in Figure A4-1.
The heat pump unit in REV 2250 has the capability of providing cooling in summer and heating in
winter. A schematic of the heat pump module is provided in Figure A4-2, except that there is only
one compressor and therefore one refrigerant circuit in REV 2250. The switch between cooling
and heating mode is achieved by the reversing valve, as illustrated in Figure A4-3 and Figure A44.
Figure A4-2 Schematic of the Heat Pump Module in Semco Ventilation Unit (courtesy of
SEMCO)
- 175 -
Figure A4-3 Operation of the Heat Pump Module in Cooling Mode
Figure A4-4 Operation of the Heat Pump Module in Heating Mode
Installation and Commissioning of the System

Installation of the combined system was finished within one day on 10 November 2005. It
replaced a previous ventilation unit which was used as the ventilation unit for the IW. The FVR
2000 and REV 2250 were provided as a self contained system, greatly simplifying the installation.
The installation crew put in a curb adaptor, placed the machine, and then connected the gas line,
electrical supply, air duct, and the internet connection for data acquisition and control. A curb
adaptor was needed to match the air supply and return ducts to the IW.
- 176 -
System commissioning followed directly after the installation. It was done by a crew of two from
SEMCO. During the commissioning, the outside and building exhaust air flowrates through the
FRV 2000 unit were adjusted by dampers. The total air flowrate of the machine was set by
adjusting the fan frequency and verified in the air duct to the IW. The bypass damper around the
desiccant wheel was adjusted. The system performance was then continuously monitored for 12
hours; the operational data were archived and analyzed to ensure the system was operating
optimally.
Control and Data Acquisition of the System

A comprehensive set of instruments has been installed for both system control and performance
monitoring. Both control and data acquisition of the system are accomplished through WebCTRL,
as illustrated in Figure A4-5.
Figure A4-5 Control and Data Acquisition of the FVR 2000 and REV 2250
External adjustments of the control system include:
system on/off;
supply air (or space) temperature and humidity through control set points;
supply air flowrate through a control set point;

- 177 -
outside air and exhaust air flowrates through dampers;
heating source: heat pump or gas;
bypass damper around the desiccant wheel.
Internal adjustments include:
supply fan variable frequency drive, VFD;
compressor VFD;
condenser fan VFD;
desiccant wheel VFD;
regeneration fan VFD;
regeneration natural gas valve;
heat pump in heating or cooling mode;
switch from heat pump to gas heating.
Once the external adjustments are set by the user, the internal adjustments are made based on
the control algorithms programmed in WebCTRL in order to achieve user specified settings. The
supply air flowrate controls the supply fan frequency; the temperature, the compressor frequency;
the supply air humidity, the gas flow. Defrost and equipment safety modes are separately
programmed.
Measurements from the sensors as well as indications of various control signals are recorded at
user specified time intervals. The data are stored in the control server and can be traced back at
anytime. As an example, Figure A4-6 shows the outside, supply and return air temperature on 25
February 2007. These data can also be reported in spreadsheets to facilitate data analysis.
- 178 -
Figure A4-6 Sample Measurement Data of the FVR 2000 and REV 2250
Operation of the System

Since its installation, the system has been operated in two different modes: as an air conditioning
system, which provides ventilation air and a portion of cooling/heating for the IW through
recirculation air; and as a dedicated outdoor air system, which delivers ventilation air at a user
selected temperature and humidity while cooling/heating is provided by other systems such as
radiant panels and mullion pipes supplied with chilled or heated water. Supply air from the system
is delivered to the IW through floor mounted air diffusers that are located at each workstation.
According to guidelines on high performance HVAC system design and operation, developed by
the researchers in the IW, provision of ventilation air should be separated from space
cooling/heating. Furthermore, space cooling/heating should be achieved by water based units
whenever possible. Therefore, ideally the ventilation system should be operated to provide
adequate amount of fresh, outside, air for the occupants at comfortable temperature and humidity
conditions and at a humidity sufficiently low to allow for the effective operation of the water based
cooling units. The system is therefore a DOAS. Given the normal occupancy rate in the IW,
0.283 m3/s (30 occupants at a 0.00944 m3/s-person ventilation rate) of ventilation air is required
for reasonable indoor air quality. However, in order to take advantage of the cooling and heating
- 179 -
capacity offered by the REV 2250, the machine was operated as an air conditioning system for
most of the time. The supply air flow was set to 0.944~1.416 m3/s. In addition, due to the
operation of the active desiccant wheel, the system was able to deliver air dry enough to allow for
effective operation of the radiant panels and fan coils in the summer. As a result, comfortable
indoor conditions were maintained in the IW during both heating and cooling seasons and
positive feedback was given by most of the occupants.
- 180 -
Appendix 5 Summer Experiment of the Enthalpy

Recovery Wheel, Reduced Flow
The inlet outside and building exhaust air conditions during the summer experiment with the
reduced flowrate setting are shown in Figure A5-1. Similar to the previous experiment, Figure
A5-1 does not show any data when outdoor temperature is around 23.5oC, due to the stop/job
economizer control.
Figure A5-1 OA and RA Conditions during the Summer Experiment, Reduced Flow
The sensible heat, moisture and enthalpy balances between the outside air supply and the space
return air are established in Figure A5-2, Figure A5-3 and Figure A5-4. Similar to the previous
experiment, the sensible heat, moisture and enthalpy balances show linear equivalence relation
slightly displaced corresponding to a small sensible heat, moisture or enthalpy gain, which is
probably due to the heat transfer or air leakage from the outdoor surroundings to the wheel
assembly. The deviation in sensible heat balance in Figure A5-2 increases with the outdoor
temperature, which might be a result of the heat transfer from the surroundings and/or the solar
radiation absorbed by the exhaust air hood as explained earlier. The moisture balance in Figure
A5-3 shows larger scatter than the sensible heat balance due to the less accuracy of moisture
measurements. The enthalpy balance in Figure A5-4 also indicates greater scatter, which
essentially reflects the moisture balance.
- 181 -
Figure A5-2 Sensible Heat Balance during the Summer Experiment, Reduced Flow
Figure A5-3 Moisture Balance during the Summer Experiment, Reduced Flow
- 182 -
Figure A5-4 Enthalpy Balance during the Summer Experiment, Reduced Flow
The measured and simulated sensible heat, moisture, and enthalpy loss of the outside air stream
are plotted against the maximum possible sensible heat, moisture and enthalpy exchange in
Figure A5-5, Figure A5-6 and Figure A5-7, respectively. The plots indicate linear relationships
between the measured or simulated and the maximum possible quantities. The sensible, latent
heat and enthalpy recovery effectiveness, represented by the slopes of the regression lines, are
smaller compared to the previous experiment. The difference between the simulated and the
measured effectiveness parameters are larger than the previous experiment, which are shown
more clearly in Figure A5-8 through Figure A5-10. These larger deviations are probably related to
the non uniform airflow patterns in the wheel assembly or the dependence of heat and mass
transfer coefficients on the airflow rate.
- 183 -
Figure A5-5 Simulated and Measured Sensible Heat Exchange, Reduced Flow
Figure A5-6 Simulated and Measured Moisture Exchange, Reduced Flow
- 184 -
Figure A5-7 Simulated and Measured Enthalpy Exchange, Reduced Flow
Figure A5-8 Difference between Simulated and Measured Sensible Heat Recovery
Effectiveness, Reduced Flow
- 185 -
Figure A5-9 Difference between Simulated and Measured Moisture Recovery Effectiveness,
Reduced Flow
Figure A5-10 Difference between Simulated and Measured Total Heat Recovery
Effectiveness, Reduced Flow
As seen from Figure A5-11through Figure A5-13, the simulation slightly overpredicts the sensible
heat, moisture and total heat recovery effectiveness for the experiment with reduced flow.
- 186 -
Considering the uncertainty associated with the model prediction, these plots indicate reasonable
agreement between the simulation and the measurements.
Figure A5-11 Comparison between the Simulated and Measured Sensible Heat Recovery
Effectiveness with Uncertainty Range during the Summer Experiment, Reduced Flow
Figure A5-12 Comparison between the Simulated and Measured Moisture Recovery
- 187 -
Figure A5-13 Comparison between the Simulated and Measured Total Heat Recovery
- 188 -

2008 PHDBPD Zhai Chaoqin

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

2008 PHDBPD Zhai Chaoqin

Încărcat de

Drepturi de autor:

Formate disponibile

Performance Modeling of Desiccant

Wheel Design and Operation

PhD Thesis Committee:

Performance Modeling of Desiccant

Copyright Chaoqin Zhai, 2008. All rights reserved.

I hereby declare that I am the sole author of this thesis.

I authorize Carnegie Mellon University, Pittsburgh, Pennsylvania to reproduce this thesis by

Copyright 2008 by Chaoqin Zhai

I am grateful to my parents-in-law for taking care of my daughters, allowing me time to spend on

Copyright@2008 by Chaoqin Zhai ........................................................................................ i

Chapter 2 Development of the Performance Model ...........................................................17

2.6 Various Uses of the Performance Model ...................................................................................................72

Chapter 3 Validation of the Performance Model................................................................78

Chapter 4 Integration of the Enthalpy Recovery Wheel in HVAC System Design.......124

Chapter 6 Contribution, Conclusion and Future Work ..................................................148

Figure 1-1 Rotary Air-to-air Enthalpy Recovery Wheel .......................................................... 3

Figure 5-1 Enthalpy Removal Breakdown ........................................................................... 138

Table 1-1 Literature Review of Desiccant Wheel Modeling .................................................. 12

Air-Conditioning & Refrigeration Institute

American Society of Heating, Refrigeration and Air-Conditioning Engineers

cubic feet per minute

Combined Heating and Power

Carnegie Mellon University

Dedicated Outdoor Air System

Dehumidification Wheel (Active Desiccant Wheel)

Building Exhaust Air Outlet

Energy Supply System

Enthalpy Recovery Wheel

Heating, Ventilating and Air Conditioning

Leaving Active Desiccant Wheel

Leaving Enthalpy Wheel

Moisture Removal Capacity

Number of Transfer Unit

revolution per minute

Regeneration Specific Heat Input

Scanning Electronic Microscopy

Variable Frequency Drive

Variable Air Volume

specific heat, J/kg-K

enthalpy or convective heat transfer coefficient, J/kg or W/m2-K

convective mass transfer coefficient, kg/m2-s

heat of adsorption or vaporization, J/kg

number of discretization in the adsorption section

indicator of the element in t domain

indicator of the element in x domain

thermal conductivity, W/m-K

depth of the desiccant wheel, m

molecular weight, kg/mole

Mass flowrate, kg/s

number of discretization in the space (x) domain

number of discretization in the time (t) domain

perimeter length of the airflow channel or pressure, m or Pa

universal gas constant, R = 8.314 J/mole-K

air velocity, m/s

distance in axial direction, m

grid size in x domain, m

moisture loading in the desiccant, kg moisture/kg dry desiccant

sensible, latent heat or enthalpy recovery effectiveness, regeneration efficiency,

thickness of the desiccant layer, m

relative water vapor concentration

desiccant matrix or mass transfer

1.2 Literature Review

Figure 1-1 Rotary Air-to-air Enthalpy Recovery Wheel[12]