Applied Thermal Engineering 27 (2007) 223228

www.elsevier.com/locate/apthermeng

Fluid selection for the Organic Rankine Cycle (ORC)

in biomass power and heat plants
Ulli Drescher *, Dieter Bruggemann
Lehrstuhl fur Technische Thermodynamik und Transportprozesse (LTTT), Universitat Bayreuth,
Universitatsstrae 30, 95440 Bayreuth, Germany

Received 9 March 2006; accepted 21 April 2006

Available online 14 July 2006

Abstract

In small solid biomass power and heat plants, the ORC is used for cogeneration. This application shows constraints dierent from
other ORC. These constraints are described and an adapted power plant design is presented. The new design inuences the selection
criteria of working uids. A software has been developed to nd thermodynamic suitable uids for ORC in biomass power and heat
plants. Highest eciencies are found within the family of alkylbenzenes.
 2006 Elsevier Ltd. All rights reserved.

Keywords: Organic Rankine Cycle; Working uid; Plant design; Biomass; Cogeneration

1. Introduction uids are preferred, as uid mechanics leads to high turbine

In the last years, large endeavors have been made to
extend the market share of renewable energies. Power
and heat cogeneration by solid biomass is one of the most
interesting options for a sustainable and reliable energy
supply due to its high availability.
Electrical power is usually generated in processes based
on the Rankine cycle with water as a working uid. The
ORC process uses an organic working uid instead of water.
In contrast to water, the expansion in the turbine ends for
most organic uids not in the wet steam regime but in the
gas phase above condenser temperature. Thus, often an
internal heat exchanger is used to improve eciency (Fig. 1).
In comparison to water, organic uids are advantageous
when the maximum temperature is low and/or the power
plant is small. At low temperatures, organic uids lead to
a higher cycle eciency than water. In small plants, organic
*
Corresponding author. Tel.: +49 921 557161/49 921 55 7285; fax: +49
921 557165.
E-mail addresses: LTTT@uni-bayreuth.de, ulli.drescher@uni-
bayreuth.de (U. Drescher).
eciency also in partial load [1]. This is the main reason to
use ORC for biomass application. Another advantage of
ORC in small plants is a legal and economic one. Water
shows good eciency at high pressure requiring increased
safety measures which are not economically feasible for
small plants.
The ORC is not a new concept and many investigations
have been carried out [2]. Research was mainly focused on
low grade heat. Typical applications use geothermal [3] or
waste heat [4,5]. It has also been examined as a bottoming
cycle combined with gas turbines [6] or other high temper-
ature cycles.
In most of all biomass applications, octamethyltrisilox-
ane (OMTS) has been chosen as a working uid. For
OMTS, thermal as well as total heat recovery eciency is
comparatively low for a high temperature ORC process.
This is the incentive to search for uids adapted specially
to biomass application which diers from other ORC, as
we describe in detail in Section 2.
In order to identify the most suitable organic uids, sev-
eral general criterions have to be taken into consideration,
including:

1359-4311/$ - see front matter  2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.applthermaleng.2006.04.024
224 U. Drescher, D. Bruggemann / Applied Thermal Engineering 27 (2007) 223228

Nomenclature

h specic enthalpy Subscripts

p pressure c critical
T temperature con condensation
w specic work ihe internal heat exchanger
CASN Chemical Abstracts Service Number max maximum
DIPPR Design Institute for Physical Properties min minimum
EOS equation of state s isentropic
MAC maximum allowable concentration vap vaporization
OMTS Octamethyltrisiloxane 1,2,3,4,5,6,7,8,9 states of the system
ORC Organic Rankine Cycle
PP pinch point
VP vapor pressure
g eciency
D dierence

thermodynamic properties, 2. Specics of ORC for cogeneration in biomass plants

stability of the uid and compatibility with materials in
contact,
safety, health and environmental aspects,
availability and costs.
This work is focused on nding an optimized working
uid for biomass ORC with respect to thermodynamic
properties. Other criteria are discussed only in qualitative
respects.
Two thermodynamic parameters have to be considered
evaluating working uids for ORC. The thermal eciency
represents the ORC itself neglecting the thermal behavior
of heat sources and sinks. Total heat recovery eciency
takes this inuence into consideration and it is the more
meaningful parameter, but more dicult to determine.
We will show that thermal eciency can be used for bio-
mass applications in order to suciently evaluate uid
selection.
6 tions at the beginning of vaporization and another pinch
Superheater
5 pinch points are the reason why thermal eciency and total
Turbine
Evaporator
4 7
Preheater Recuperator
3
8
In comparison to low grade heat applications, the ORC
for biomass combustion dierently constrains uid selec-
tion. It is obvious that temperature levels are signicantly
higher (Table 1). The condenser temperature is about
370 K to satisfy the needs of the heating network. This
means that many uids for low grade heat utilization can
not be used, because their vapor pressure is quite high even
at condensation temperature.
The turbine inlet temperature could be close to ame
temperature, but organic uids would become chemically
unstable. Therefore, maximum process temperature is lim-
ited to about 600 K [7].
The interaction of the heat ows is illustrated in temper-
atureenthalpy ow diagrams allowing to show also tur-
bine power. The extensive enthalpy ow H_ is taken
simplifying the comparison of dierent substances.
In most cases, the uid is coupled to the ue gas by ther-
mal oil to avoid local overheating and to operate the heat
exchanger in the ue gas at atmospheric pressure. This
leads to a pinch point similar to low grade heat applica-
point between thermal oil and ue gas (Fig. 2(a)). These
heat recovery do not strongly correlate. But in comparison
to low grade heat applications, the thermal oil mass ow
can be adjusted to a certain extent. Furthermore, we can
transfer the heat by more than one thermal oil cycle. Thus,
we propose a plant design with two thermal oil boilers in

Feed Pump 2 Table 1

Typical temperatures of ORC for biomass application
9
Flame temperature 1200 K
Condenser Maximum thermal oil temperature 630 K
1
Maximum ORC uid temperature 600 K
Condenser temperature 370 K
Fig. 1. Plant layout for an Organic Rankine Cycle.
 U. Drescher, D. Bruggemann / Applied Thermal Engineering 27 (2007) 223228 225

Fig. 2. T ; H_ diagram for the standard (a) and an improved (b) plant design for an ORC without superheating (state 5 and 6 are equal). PP = pinch point.
The small dierence between states 1 and 2 is neglected in the diagrams.

order to avoid the constriction of the pinch point at state 4

(Fig. 2(b)). This design allows to consider also uids with
high vaporization enthalpy at maximum process tempera-
ture. Therefore, the suggested plant design simplies uid
selection. Now thermal eciency can be used for uid eval-
uation instead of total heat recovery eciency.

3. Methods

We developed a calculation method which needs follow-

ing basic data for each substance:

molecular weight,
critical temperature,
critical pressure, Fig. 3. Typical T, S-diagram for ORC with superheating.
acentric factor,
melting point,
autoignition temperature, as well as state 6 are obtained from an isothermal change of
vapor pressure of the liquid, state starting from the 0.1 MPa isobar using PengRob-
heat of vaporization, insonEOS [10,11].
heat capacity of the liquid, The thermal eciency g is dened by
heat capacity of ideal gas.
h6  h7  jwpump j
g 1
These thermodynamic properties have been extracted h6  h 3
from the database of the Design Institute for Physical with the feed pump work
Properties (DIPPR) [8,9] which includes nearly 1800 sub- Z pmax
1
stances. The calculation method has to cover the range wpump v dp 2
from liquid phase to superheated gas (Fig. 3). Specic gpump pmin
enthalpy dierence h3  h2 is the recovered heat provided Feed pump work is calculated neglecting the uid com-
by the recuperator using h7  h8. The dierence h6  h1 pressibility which inuences enthalpy only marginally.
is input enthalpy, h7  h6 gained work wturbine and The eciency of the internal heat exchanger is
h8  h1 waste heat which may be used for space heating.
State 6 can be in the gas phase or on the dew line (then h3  h2
gihe 3
state 5 equals state 6). The maximum temperature Tmax is h7  h8
the turbine inlet temperature (state 6) and the minimum and turbine eciency is given by
temperature Tmin is the condenser temperature (state 1
h6  h7
and 9). gturbine 4
h6  h7s
The states 1, 2, 3, 4, 5 and 9 are calculated directly from
the DIPPR database. In the case of minimum process pres- In order to validate the calculation procedure, a compari-
sure below 0.1 MPa, states 7s, 7 and 8 are calculated from son with the software Fluidcal [12] which is based on an
state 9 using ideal gas heat capacity. Otherwise, these states analysis by Lemmon et al. [13] was carried out for toluene
226 U. Drescher, D. Bruggemann / Applied Thermal Engineering 27 (2007) 223228

Table 2
Input and constraints of the screening
Parameter Value or constraint
Melting temperature <275 K
Maximum process temperature <Autoignition temperature
Condenser pressure P5000 Pa or vapor pressure
Condenser temperature P363 K
Dryness at turbine outlet >0.9
Eciency of internal 0.95
heat exchanger
Minimum DT of internal 10 K
heat exchanger (T8T9)
Turbine eciency 0.8
Feed pump eciency 0.8
Maximum process pressure 62 MPa or vapor pressure

and isopentane. Aside from the critical point, relative dis-
crepancies were found to be less than 3% for calculated
gas and liquid heat capacity and below 1.5% for cycle
eciency.
Fig. 4. Schematic T, S-diagram illustrating the decision pattern for the
interaction between vapor pressure and process parameters. The lled
rhomb marks the turbine inlet state without superheating and the triangle
with superheating, each at 2 MPa. VP = vapor pressure, con =
condensation.

4. Screening of uids
Fig. 5 shows the number of substances which reach
Many parameters have to be considered for uid selec- given eciency values with or without internal heat
tion and ORC thermal eciency calculation (Table 2). exchanger. The implementation of a recuperator increases
The two main parameters are maximum and minimum eciency considerably. The eciency of the 100 best-suited
process temperature. The upper limit of the maximum pro- uids ranges from 25.4% to 24.3%. Thus, many uids with
cess temperature is the uid stability and material compat- similar thermal eciencies are available for considering
ibility. These are dicult to evaluate. Typical maximum further criterions.
temperature for existing biomass ORC is about 573 K. As expected, there is a strong relation between eciency
Thus, we analyzed this temperature in detail. In addition, and vaporization temperature (Fig. 6(a)). The highest e-
maximum process temperature was varied from 523 to ciencies were obtained when superheating was avoided
623 K, aware that some uids might not be adequate for and the uid was expanded directly from dew line.
this temperature range. The minimum temperature was According to Fig. 6(b), the eciency rises to approxi-
set to 363 K. If necessary, the condenser temperature was mately 25% at 1 MPa with a slight decrease for higher
raised to make the vapor pressure equal 5000 Pa, the lowest pressures.
pressure accepted for the condenser (Fig. 4). In accordance with the law of corresponding states, u-
The melting temperature should be below ambient tem- ids with low maximum process pressure at maximum
perature. Otherwise the uid may solidify during shutdown
time. For this study, the turbine eciency was xed at 80%,
which is in the typical range from 75% [14] up to 85% [15].
The maximum process pressure was limited to 2 MPa to
reduce safety measures and material expenses [16]. If the
vapor pressure at maximum process temperature was lower
than 2 MPa, the uid was expanded directly from dew line.
Otherwise, the uid was superheated (Fig. 4). In the rare case
of maximum process temperature and pressure being in the
supercritical region of a uid, the maximum process pressure
was reduced by 0.1 MPa below critical pressure and the uid
was superheated. If the adjusted process pressure was below
0.3 MPa, the uid was excluded, because the pressure ratio
in the turbine was too low for ecient power generation.

5. Results

About 700 substances of the DIPPR database passed the

preselection described above and were included in the sub- Fig. 5. Number of substances exceeding the eciency g in ORC processes
sequent comparison. with Tmax = 573 K and Tmin = 363 K.
 U. Drescher, D. Bruggemann / Applied Thermal Engineering 27 (2007) 223228 227

. .

. .

. .

. .

. .

. .
. . . . . .

Fig. 6. Relation between eciency and vaporization temperature (a) and maximum process pressure (b) with Tmax = 573 K and Tmin = 363 K. Triangles
symbolize uids with lower vaporization temperature and maximum process pressure of 2 MPa. Filled rhombs stand for uids with vaporization at
maximum process temperature and corresponding vapor pressure as maximum process pressure.

process temperature have a vapor pressure at condenser

temperature, which is lower than the specied 5000 Pa.
Thus, the calculation method raises condensation tempera-
ture to the temperature correlating with a vapor pressure of
5000 Pa. The resulting smaller temperature dierence
between heat input and output leads to low eciencies.
At the above mentioned maximum process pressure of
about 1 MPa, some uids have a vapor pressure at con-
denser temperature exceeding 5000 Pa. Thus, the full tem-
perature dierence can be used for the cycle. The slight
decrease can be explained by the needed work of the feed
pump (see Eq. (1)). This means that there is an optimal
maximum process pressure.
In order to demonstrate these eects, a detailed analysis
for typical uids has been carried out. For the uid selec-
tion, also the toxicity was taken into account by using a
maximum allowable concentration (MAC) of 50 ml/m3.
Fig. 7 shows thermal eciency Eq. (1) for the homologous Fig. 7. Eciency of ORC with recuperator depending on temperature for
series from toluene to n-butylbenzene and the widely used selected uids. Tmax varying from 523 to 623 K, Tmin = 363 K and
octamethyltrisiloxane (OMTS). Results and typical uid OMTS = Octamethyltrisiloxane.
properties can be taken from Table 3. To illustrate the
inuence of vaporization enthalpy (h5  h4), its ratio to OMTS vaporizes at higher temperature as toluene, but
input enthalpy (h6  h3) is given. its vaporization enthalpy is signicantly lower. It has the
Toluene has the highest vaporization enthalpy, but at a lowest eciency of the selected uids.
low temperature level. Thus, toluene shows worst eciency As expected, eciency rises with maximum process tem-
of the alkylbenzenes. Butylbenzene shows highest e- perature. Butylbenzene gains 6.5% points and OMTS 4.7
ciency. It has the lowest maximum and minimum process percent points from 523 to 623 K. In general, eciency dif-
pressure, vaporizes at maximum process temperature and ference between the uids rises with higher maximum pro-
its condenser temperature is only slightly lifted. cess temperature.

Table 3
Thermodynamic properties and results of selected uids
h5 h4
Working uid CASN Tc (K) pc (MPa) pmax (MPa) Tvap (K) h6 h3 (%) pmin (kPa) Tcon (K) g (%)
OMTS 107-51-7 564 1.44 1.34 560 15 13.8 363 22.5
Toluene 108-88-3 592 4.11 2.00 536 42 54.1 363 23.2
Ethylbenzene 100-41-4 617 3.61 2.00 570 36 24.3 363 24.3
Propylbenzene 103-65-1 638 3.20 1.41 573 40 11.4 363 24.9
Butylbenzene 104-51-8 661 2.89 0.92 573 43 5.0 364 25.3
228 U. Drescher, D. Bruggemann / Applied Thermal Engineering 27 (2007) 223228
6. Conclusions
This study presents a procedure to calculate ORC e-
ciency with sucient accuracy based on the DIPPR data-
base and the PengRobinsonEOS. This method is used
to nd thermodynamic suitable uids for ORC in biomass
power plants. In comparison to most other ORC applica-
tions, the pinch point at the start of vaporization can be
avoided by an adapted plant design. Eciency correlates
with a low minimum temperature, high vaporization tem-
perature and a high amount of vaporization enthalpy to
input enthalpy. If these parameters are optimal, low maxi-
mum and minimum process pressure leads to a slight
improvement of eciency. For biomass plants operating
with a maximum process temperature of 573 K, a maxi-
mum pressure between 0.9 and 1.5 MPa is best. The family
of alkylbenzenes shows highest eciencies.
A biomass ORC can be improved by raising maximum
temperature. If the maximum temperature could be shifted
from 573 K to 623 K, eciency would gain about 2.5%
points.
Other optimization strategies like air preheating may
improve the power to heat ratio, a fact entailing higher
protability. To evaluate such options, investigations con-
cerning complete system analysis are necessary, including
energy evaluation, detailed pinch point analysis, coupled
cooling processes etc.
The overall objective, a cost ecient and sustainable
energy supply is an incentive for further research work in
this eld.
References
[1] S.K. Ray, G. Moss, Fluorochemicals as working uids for small
Rankine cycle power units, Advanced Energy Conversion 6 (1966)
89102.
[2] S. Devotta, F.A. Holland, Comparison of theoretical Rankine power
cycle performance data for 24 working uids, Heat Recovery Systems
5 (6) (1985) 503510.
[3] R. DiPippo, Geothermal power plants: Principles, Applications and
Case Studies, Elsevier Advanced Technology, Oxford, 2005.
[4] V. Maizza, A. Maizza, Unconventional working uids in organic
Rankine-cycles for waste energy recovery systems, Applied Thermal
Engineering 21 (2001) 381390.
[5] B.-T. Liu, K.-H. Chien, C.-C. Wang, Eect of working uids on
organic Rankine cycle for waste heat recovery, Energy 29 (2004)
12071217.
[6] Y.S.H. Najjar, Ecient use of energy by utilizing gas turbine
combined systems, Applied Thermal Engineering 21 (2001) 407
438.
[7] G. Angelino, C. Invernizzi, Cyclic methylsiloxanes as working uids
for space power cycles, Journal of Solar Energy Engineering,
Transactions of the ASME 115 (1993) 130137.
[8] American Institute of Chemical Engineers, Design Institute for
Physical Properties (DIPPR), project 801, evaluated process design
data. Brigham Young University, USA, 2004. Available from:
<http://dippr.byu.edu/>.
[9] V.W. Wilding, R.L. Rowley, J.L. Oscarson, DIPPR Project 801
evaluated process design data, Fluid Phase Equilibria 150151 (1998)
413420.
[10] S.I. Sandler, Chemical and Engineering Thermodynamics, John Wiley
and Sons Inc., New York, 1999.
[11] B.E. Poling, J.M. Prausnitz, J.P. OConnell, The Properties of Gases
and Liquids, McGraw-Hill, New York, 2001.
[12] W. Wagner, Fluidcal, Chair of Thermodynamics, Ruhruniversitat
Bochum, Germany, 2004. Available from: <http://www.ruhr-uni-
bochum.de/thermo/Software /Seiten/Fluidcal-eng.htm>.
[13] E.W. Lemmon, R. Span, New equations of state for technical
applications, in: Paper at the 15th Symposium on Thermophysical
Properties, Boulder, USA, 2003.
[14] G. Angelino, P. Colonna di Paliano, Multicomponent working uids
for organic Rankine cycles (ORCs), Energy 23 (1998) 449463.
[15] G. Angelino, P. Colonna di Paliano, Organic Rankine Cycles (ORCs)
for energy recovery of molten carbonate fuel cells, in: 35th Interso-
ciety Energy Conversion Engineering, Las Vegas, USA, 2000.
[16] M.J. Lee, D.L. Tien, C.T. Shao, Thermophysical capability of ozone-
safe working uids for an organic Rankine cycle system, Heat
Recovery Systems & CHP 13 (5) (1993) 409418.