Journal of Energy and Power Engineering - 2016.03

Journal of Energy
and Power Engineering

Volume 10, Number 3, March 2016 (Serial Number 100)
David Publishing
David Publishing Company

www.davidpublisher.com
Publication Information:
Journal of Energy and Power Engineering is published monthly in hard copy (ISSN 1934-8975) and online (ISSN
1934-8983) by David Publishing Company located at 616 Corporate Way, Suite 2-4876, Valley Cottage, NY 10989, USA.
Aims and Scope:
Journal of Energy and Power Engineering, a monthly professional academic journal, covers all sorts of researches on
Thermal Science, Fluid Mechanics, Energy and Environment, Power System and Automation, Power Electronic, High
Voltage and Pulse Power, Sustainable Energy as well as other energy issues.
Editorial Board Members:
Prof. Ramesh K. Agarwal (USA), Prof. Hussain H. Al-Kayiem (Malaysia), Prof. Zohrab Melikyan (Armenia), Prof.
Pinakeswar Mahanta (India), Prof. Carlos J. Renedo Estbane (Spain), Prof. Mohamed Ahmed Hassan El-Sayed (Trinidad
and Tobago), Prof. Carlos Redondo Gil (Spain), Prof. Roberto Cesar Betini (Brazil), Prof. Rosrio Calado (Portugal), Prof.
Dr. Ali Hamzeh (Germany), Prof. Noor-E-Alam Ahmed (Australia), Prof. E. Ubong (USA), Prof. Shripad T. Revankar
(USA), Prof. Almoataz Youssef Abdelaziz (Egypt), Prof. Guihua Tang (China), Prof. Mohammad Rasul (Australia), Prof.
Rene Wamkeue (Canada), Prof. Ya-Ling He (China), Prof. Filippo de Monte (Italy), Prof. Masoud Rokni (Denmark), Prof.
Hosni I. Abu-Mulaweh (USA), Prof. Quan Zhang (China), Prof. Peng-Sheng Wei (Taiwan), Prof. Vinod Kumar (India), Prof.
Yuan-Kang Wu (Taiwan), Dr. Kaige Wang (USA), Dr. Fude Liu (Hong Kong), Prof. Isa Salman Hasan Qamber (Bahrain),
Dr. GM Shafiullah (Australia), Dr. Kelum Gamage (UK), Dr Isam Janajreh (UAE), Prof. Marco Casini (Italy), Prof.
Meicheng Li (China).
Manuscripts and correspondence are invited for publication. You can submit your papers via Web Submission, or E-mail to
energy@davidpublishing.com or info@davidpublishing.com. Submission guidelines and Web Submission system are
available at www.davidpublisher.com.
Editorial Office:
616 Corporate Way, Suite 2-4876, Valley Cottage, NY 10989, USA
E-mail: energy@davidpublishing.com; energy@davidpublishing.org; energy-power@hotmail.com;
energy@davidpublishing.net.
Copyright2015 by David Publishing Company and individual contributors. All rights reserved. David Publishing
Company holds the exclusive copyright of all the contents of this journal. In accordance with the international convention,
no part of this journal may be reproduced or transmitted by any media or publishing organs (including various websites)
without the written permission of the copyright holder. Otherwise, any conduct would be considered as the violation of the
copyright. The contents of this journal are available for any citation. However, all the citations should be clearly indicated
with the title of this journal, serial number and the name of the author.
Abstracted / Indexed in:
Database of EBSCO, Massachusetts, USA
Database of Cambridge Science Abstracts (CSA), USA
Chinese Database of CEPS, American Federal Computer Library Center (OCLC), USA
Ulrichs Periodicals Directory
Summon Serials Solutions
Chinese Scientific Journals Database, VIP Corporation, Chongqing, China
Chemical Abstracts Service (CAS)
ProQuest
CNKI
Google Scholar
Subscription Information:
Price:
Print $800 (per year)
Online $600 (per year)
Print and Online $880 (per year)

Tel: 1-323-984-7526, 001-323-410-1082; Fax: 1-323-984-7374; 323-908-0457
E-mail: order@davidpublishing.com
Digital Cooperative Company: www.bookan.com.cn
D DAVID PUBLISHING

www.davidpublisher.com
Journal of Energy and
Power Engineering
Volume 10, Number 3, March 2016 (Serial Number 100)
Contents
Clean and Sustainable Energy
137 Potentials of Cellular Vortex Element Modeling of Fluid Flow in Confined 2D Aquifer
Oyetunde Adeoye Adeaga and Ademola Adebukola Dare
150 Numerical Analysis of a Gravity Substructure for 5 MW Offshore Wind Turbines Due to Soil
Conditions
Min-Su Park, Youn-Ju Jeong and Young-Jun You
159 Production of Rhenium by Transmuting Tungsten Metal in Fast Reactors with Moderator
Tsugio Yokoyama, Yuki Tanoue, Atsunori Terashima and Masaki Ozawa
166 A New Generator for Tsunami Wave Generation

Tetsuya Hiraishi, Ryokei Azuma, Nobuhito Mori, Toshihiro Yasuda and Hajime Mase
173 Solid Particle Emission Abatement from Biomass-Fired Boilers at District Heating Plants
Egils Dzelzitis, Aldis Jegeris, Agnese Lickrastina and Normunds Talcis
Power and Electronic System

183 Losses Calculation of an Aerospace Retraction Wheel Motor with Regarding to
Electromagnetic-Field Analysis Investigation
Pedram Asef and Ramon Bargallo Perpina
191 Energy Efficiency of Reactive Dynamic Compensators

Rodrigo Nobis da Costa Lima, Srgio Ferreira de Paula Silva, Antnio Carlos Delaiba and Ezequiel
Junio de Lima
199 Calculation of CO Behavior in the Platform for Deeply Underground Subway Station with
Different Fire Strengths
Yong-Jun Jang, Ji-Min Ryu, Jong-Ki Kim, Min-Chul Chun and Dong-Hoe Koo
206 Redesign of Motorized and Non-motorized Transport in Cities and Sustainable Mobility
Osman Lindov, Adnan Omerhodi, Adnan Tatarevi and Samir Daferovi
Journal of Energy and Power Engineering 10 (2016) 137-149
doi: 10.17265/1934-8975/2016.03.001
D DAVID PUBLISHING
Potentials of Cellular Vortex Element Modeling of Fluid

Flow in Confined 2D Aquifer
Oyetunde Adeoye Adeaga1 and Ademola Adebukola Dare2

1. Department of Mechanical Engineering, The Ibarapa Polytechnic, Eruwa 201101, Nigeria
2. Department of Mechanical Engineering, University of Ibadan, Ibadan 200106, Nigeria
Received: August 26, 2015 / Accepted: September 25, 2015 / Published: March 31, 2016.
Abstract: Numerical methods such as finite difference, finite volume, finite element or hybrid methods have been globally used to
successfully study fluid flow in porous stratum of which aquifers are typical examples. Those methods involve mathematical
expressions which increases computation time with requirement of specific human expertise. In this paper, numerical models for
single phase flow in 1D and 2D using the conservation of mass principles, Darcys flow equation, equation of state, continuity
equation and the STB/CFB (stock tank barrel/cubic feet barrel) balance were developed. The models were then recast into pressure
vorticity equations using convectional algorithms. Derived equations were used to formulate transport equations which resemble the
conventional vorticity transport equation. Formulated numerical models were used to investigate the daily instantaneous aquifer
pressure drawdowns and pressure heads for 365 days. The developed equations were subsequently solved using cellular vortex
element technique. The developed computer program was used to investigate confined aquifer of dimensions 10 10 75 m with
single vertex image. For the aquifer rate of 0.5 m3/s, 0.1 m3/s, 0.15 m3/s, 0.2 m3/s, 0.25 m3/s, 1.0 m3/s, 2.0 m3/s, 2.5 m3/s, 3.0 m3/s,
4.0 m3/s, the respective average head drawdowns and heads were, 1.127 0.0141 m, 1.317 0.0104 m, 1.412 0.0041 m, 1.427
0.116 m,1.527 0.0141 m, 2.107 0.0171 m, 2.197 0.0191 m, 3.007 0.0171 m, 3.127 0.0041 m, 3.626 0.0121 m, and 25 kN/m2,
35 kN/m2, 33 kN/m2, 5 kN/m2, 6 kN/m2, 11 kN/m2, 25 kN/m2, 42 kN/m2, 50 kN/m2, 62 kN/m2, respectively. Cellular vortex
technique with relative little mathematics has been established to have recorded successes in numerical modeling of fluid flow in
aquifer simulation.
Key words: Numerical, Darcy, vorticity, image, subsurface porous media.
Nomenclature G0 G/Ss
H Head drawdown (m)
A Crossectional area, L2, ft2 (m2)
Ax, Ay
Crossectional area normal to x, y direction, 1. Introduction
respectively, L2, ft2 (m2)
kw Absolute permeability of water (L2) A subsurface porous media is a trap in which fluids
w Viscosity of water (poise, milli-darcy, md)
such as oil, gas and/water have accumulated over
Fluid volume formation water L3/L3,
Bw RB/SCF (reservoir barrels per standard cubic feet) millions of years by migration from source rocks,
(m3/stdm3) decays of aquatic remains etc. A porous
VR Aquifer total volume, ft3 (m3)
qw Volume flow rate of water (m3/s)
medium/material is a solid (often called frame or
w Potentials water (m/Lt2), psia (kPa) matrix) permeated by an interconnected network of
Porosity, fraction pores (voids) filled with a fluid (liquid or gas).
Circulation Usually both the solid matrix and the pore network
angular acceleration (rad/s)
(also known as the pore space) are assumed to be
Ss Storability
G Transmissivity continuous, so as to form two or more interpenetrating
continua as in a sponge. Many natural substances such
Corresponding author: Oyetunde Adeoye Adeaga, Dr., as rocks, soils, biological issues (e.g., bones, skin) and
research field: computational fluid dynamics.
man-made materials such as cements, foams and Ogundare [6]. Zheng, et al. [7] successfully applied
ceramics can be considered a porous media, A the finite element method using Biot Poro-elastic
poro-elastic medium is characterized by its porosity, approach to simulate 2D axisymmetric, 3D and full
permeability as well as the properties of its 3D reservoir in the subsistence of Venice in Italy [8].
constituents (solid matrix and fluid) [1]. Other numerical solution techniques include but not
The concept of porous media is used in many areas limited to: direct summation, vortex-in-cell, fast
of applied science and engineering; mechanics, multi-pole method, species interface/mixing, and
acoustics, geo-mechanics, soil-mechanics and density interface/ stratification [9].
rock-mechanics, engineering (reservoir engineering, Common numerical methods that are used to
petroleum engineering, construction engineering), simulate and eventually investigate fluid(s) flow in
geosciences (hydro-geology, petroleum geology, aquifers (either confined or unconfined) are finite
geophysics) biology and biophysics, material sciences element method, finite difference method, finite
etc. [2]. The numerical equations governing the flows volume method and the combination of two or more
of scientifically separable but physically immiscible of the earlier mentioned, requires relatively complex
fluids in subsurface earth (porous media) flow are mathematical composition and analysis of matrix and
inherently nonlinear, and the geometries and material matrices (i.e., inverses and transposes of sub matrices,
properties characterizing many empirical problems in factors and cofactors) occupies ample computer
under-groundwater engineering can be quite irregular. memory which consequently reduces the computation
More often than not, numerical simulation offers the speed. Also computation of these methods are iterative
only viable approach to the modeling of single-phase in nature with the iteration value carefully chosen.
and multi-phase/multi-species flows [3]. Vortex method has been found to have lighter
Ewing, et al. [4] described numerical simulation as mathematical expressions without necessarily
one which is widely used for predicting aquifer and involving expert analysis. It provides solutions at any
reservoir behaviour and forecasting its performance. location (either on-grid or off-grid) within the domain
However, the mathematical model used in the of interest unlike FEM and FDM that has rigidly grid
simulation require the knowledge of subsurface based solutions. It is not iterative in nature, hence
properties. Since aquifer is relatively inaccessible for reduces computation time while few computer
sampling, the measurable quantities at the well, memory resource is utilized. The major importance of
provide the essential information for description [5]. cellular vortex method modeling is that, each cell
Different solution techniques have been employed within the domain can be investigated in isolation or
over the years from the one that is deterministic to the as a whole depending on the concerns of the analyst.
probabilistic, all in effort to provide suitable, Also the solution of each cell is as a result of the
economical solution technique to Navier-Stokes unpredictable nature of other neighbouring cells with
equation when applied to fluid flow and heat transfer those cells closer to the yield cell experiencing highest
problems. Among these methods are: random walk influence, i.e., the farther a cell to the yielding cell, the
method, core expansion method, particle strength lesser the influence of yield and the aquifer rate on the
exchange, elliptic Gaussian blob amongst others. Not particular cell.
left out are the numerical methods of solution, these
2. Vortex Element Method
include: FDM (finite difference method), FEM (finite
element method), Monte Carlo method and recently General vortex element method is a langragian
the hybrid Monte Carlo method as reported by technique for obtaining solution to engineering
Potentials of Cellular Vortex Element Modeling of Fluid Flow in Confined 2D Aquifer 139
problems either in fluids or solids analysis and so does well. The vertex/vortices displays through image(s)
the cellular vortex method. It came to being in 1931 the intensity of vertex (mono image) or vortices
with the Rosenhead calculation of the (multiple images) diffusion, convection and
Kelvin-Helmholtz instabilities. However, in 1967, acceleration to and fro close and distant neighbouring
Batchelor obtained solutions for the differential cells taking into account their associated viscosity,
equation: mobility, media porosity, permeability, adhesive and
cohesive forces. Conventional methods for solving the
w 2 2
a s; Navier-Stokes equations, which govern fluid motion,
t x 2 y 2
(1) typically require a computational mesh of the
2
r fluid-media domain. If the geometry of the fluid
r , t exp -
4 t 4 t domain takes complicated shapes, mesh generation

often becomes the major bottleneck in the overall
Then, Chorin applied the solution obtained by simulation process. Furthermore, exterior flow
Batchelor to propose some other solution technique problems such as the flow around an airfoil require
for the Navier-Stokes equations. artificial boundary conditions or other trickery. Vortex
Summarily, the relative advantages of vortex methods are mesh-free, particle based discretization
methods over other traditional methods like FEM and which do not suffer from these disadvantages. Mesh
FDM include but not limited to the following: independence in this context means that, only the
(1) Vortex methods are practically grid independent, boundaries of the computation domainnot the
thus eliminating numerous iterations associated with domain itselfneeds to be described using a surface
RANS (Reynolds averaged Navier Stokes) simulation mesh. This greatly simplifies the meshing process and
and LES (large eddy simulation) as obtained in FEM allows infinite domains [10].
and FDM;
3.1 Vorticity Evolution Equations
(2) All problems are treated identically. No
modeling or calibration inputs are required; The complete vorticity evolution transport equation
(3) Vortex methods are grid free solution methods for incompressible fluids flow (presented in Cartesian
for dynamic systems; coordinate), written in Laplacian form is given as [11]:
(4) These methods require few mathematics with or Ad
v e c
tio
n
D
without expertise of the subject. ( u )
Dt t
Furthermore, the vortex method is synonymous
1
with fluid rotation and it is basically important in that,
(
)u

( p )
2
it closely study the seemly difficult fluid motion by V o rtex stre ch in g (2)
b a ro c lin ica l
the introduction of vortex into the porous media which
BF V isc o u
s d iffu sio n
principally contains rotational fluid stream within and
f 2
around solid matrix. e
The following assumptions were made in
3. Methodology
simplifying the complete vorticity evolution equation:
The control volume within the media is segregated (1) Fluids are considered incompressible or slightly
into vertical stratums and with horizontal cell compressible;
arrangements. Vortices (in case of multiple image (2) Gravitational body forces act significantly only
analysis) are virtually introduced into the yielding in flow of density differences;
(3) Pressure differences or relative movement of therefore required to express the 2-Dimensional form
boundaries are caused by fluid motion/dynamics. of Eq. (3), in the form that represents Eq. (4).
Therefore, a more suitable/adaptable vorticity
3.2 Basic Aquifer Numerical Equations
equation is given as:
Assumptions made were: Laminar steady flow,
Advection

D constant compressibility (geometric), constant
( u )
Dt t pressure i.e., at particular depth pressure is constant,
Viscous diffusion (3) time invariant constituents mobility,

non-compressible rock (i.e., constant).
( ) u 2

For a 1-D (x-axis flow), the derived governing
Vortexstreching
equations are [12]:
Expanding the 2D equation and rearranging in h
S - q G (6)
Cartesian and radial coordinate, respectively, and s t
assuming negligible effects of advection, vortex
S s h - -K h G (7)
stretching, we have: t
2 2 h

S K 2 h h G (8)
x 2
s t
t y 2 (4) h
2h h G (9)
u v t 0

x y x y Phase equations:
kw w Sw
r (5) Ax qw VR (10)
t r r r x w Bw x t Bw
where, is the vorticity and r is the radius. It is Eq. (11) on expansion becomes:
2
w A B k w w B q Bw V S w B V S ;
k B
Ax w w x w w w w w
w B
w x
2 x w B w x
B

w
R
t R t B w

gives
2
w A B k w w B q B V S w
k
Ax w x w x B x w w w R t B
w x 2
w w w
(11)
Introducing Laplacian compartment to obtain the general aquifer numerical model as:
M M
Ac M w 2Pw Ac Bw w Pw Ac gh M w2w Bw w Pcw M w 2Pcw Bwqw

B
w

B
w

S B P
V - w w w
R Bw Pw t

(12)
Expanding Eq. (12) and dividing through by the equivalent vortex expression/equation thus, Eq. (9)
Pw gives the mathematical model for, 2-Dimensional
coefficient of and rearranging gives the
t aquifer as thus.
K 2P 2P
A w
w w
c 2 2 K 2 2 Kw Kw
P w x y
w w B
w A gh w x B y B
c 2 2
S B y w w w w
t w

V - w w
w x
R B P
w w

2w 2w

K w K w

Pc w Pc w
Kw

x w B w y w B w
Kw
2 Pc w 2 Pc w
Ac g h
w

x2 y 2

Bw

x y

w

x2

y 2

Bw qw

Sw Bw

V -

R
Bw

Pw

(13)
K
A w
c
d (diffusive coefficient) = w (14)
S B
V - w w

R B P
w w
K
A B w
c w x B
cx (convective coefficient) = w w (15)
S B
V - w w
R B P
w w
K
A B w
c w y B
cy (convective coefficient) = w w (16)
S B
V - w w
R B P
w w
a (accelerative coefficient) =

K w

Kw
2w 2w

Kw
x w B w y w B w
Ac g h
w
Bw

x 2 y2

Pc w Pc w
(17)
2 Pc w 2 Pc w

Kw
Bw qw

x y

w

x 2 y2

Sw Bw
V
-

R

Bw

Pw

The algorithm flowchart required for the cellular average aquifer pressure head and yield drawdown.
vortex method of solution. In the following steps, it will In Fig. 6, the average aquifer pressure per layer thus
be adapted to the viscous term of the vorticity transport experienced a downward surge around the layer of yield
equation. Comparing and evaluating the equivalent and latter decreases with regular from the lowest layer
vortex coefficients as below. i.e., layer 1 to the topmost layer, i.e., layer 16. It implies
that, as one goes down the layer, the aquifer pressure
4. Summary of Input Data
drawdown and the system pressure increased. It is
An 11 11 16 discretized grid (1,936-cells) system evident that, the average aquifer pressure drawdown
was used to test the developed aquifer program. The dropped remarkably at the instance of yield
basic simulation data for the yielding well are shown commencement but due to compensation from
below. neighbouring aquifers the drawdown is compensated for
and increases with fluctuations as shown in Fig. 5, i.e.,
5. Results and Discussion
except at the yield layer when the effect of head
The coefficients s, are for 1-D and 2-D models. The variation is almost unfelt, the average head of the
fact that, the aquifer numerical equations can be aquifer reduces up the aquifer layer.
rearranged to resemble Batchelors second order partial The obtained results were validated using Fig. 7 with
differential equation of 1969 shows that, the exact those results obtained by Ertkin, et al. [13]. The same
solution obtained by him can be used to study and situation occurs to aquifer pressure head as shown in
analyze the flow of fluid in aquifer. Fig. 8.
Fig. 1 presents the flowchart of the entire simulation, The aquifer pressure head also decreases as yield rate
while Fig. 2 depicts a sub-routine to evaluate the continues with time. This is evident in Fig. 8. The
coefficients that are needed in the overall simulation. implication is that, unless there is pressure
At the yield layer, pressure changes are very compensation to the aquifer, the aquifer yield pressure
negligible while the lower yield pressure remain the may diminish with time.
same. Figs. 3 and 4 show the conditions of the aquifer Table 1 presents the summary of important users
after two weeks and twelve months of non-stop inputs for the simulation. The results obtained from this
continuous yield of 0.5 m3/s. The black portions are work was validated by making appropriate comparison
those of yield while the conventional water blue with existing work as presented in Table 2.
coloured zones shows places of water fill. The ANOVA (analysis of variance) at p = 0.05 within
It was evident in Fig. 5 that, the upper yield layer aquifer layers were presented in Table 3. This
changes considerably due to compensation to the yield eventually confirmed decreasing profile in aquifer
layer. As time passes, during simulation, the saturation pressure, head and drawdown with time and at different
of water remain constant but there is sharp drop in aquifer rate/yield.
Start Evaluating coefficients

accelerative, convective and
diffusive
s (coefficients)
Read media data
Perform accelerative, diffusive &

convective processes
Generate x, y
coordinates
Return new pressure head and

drawdown of water
Loading aquifer data

(permeability, etc.)
Write new pressure head and

drawdown
Reading data from table

(input files)
No Runs equal
preset?
Initialize time step
Yes
Reading initial Stop

parameters
Calculating fluid and

medium data
Fig. 1 Complete simulation flowchart.

Start
Input aquifer and

fluid data
Reading grid data. Generating coordinates
Initializing time step and number of runs
Calculating coefficients
Accelerative coefficient
Diffusive coefficient
Convective coefficient
Runs equals preset?

No
Yes
Write/print results
Stop
Fig. 2 Flowchart for evaluating coefficients during simulation.
Table 1 Summary of data used for simulation.

Data from the user
Well dimensions (m)
Length 10
Breadth 10
Depth 75
Radius of well/hole 3.81 10-2
Grid dimension
Divisions in x-direction 11
Divisions in y-direction 11
Divisions in z-direction 16
Well location(m)
X 10
Y 10
Z 75
Flow rate (m3/s)
Water 0.05 m3/s
Behaviour
Simulation time step (s) 86,400
Number of preset runs 365
Number of vortex image 1
Fig. 3 3D graphical representation of aquifer condition after two months.
Fig. 4 3D graphical representation of aquifer condition after ten months.

5.0
4.5
4.0
Average pressure drawdown (kN/m2 )
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
0 2 4 6 8 10 12 14 16 18
-0.5
Aquifer layer
Fig. 5 Variation of aquifer average drawdown with layer.
45,000
40,000
Average aquifer drawdown pressure (kN/m2 )
35,000
30,000
25,000
20,000
15,000
10,000
5,000
0
0 2 4 6 8 10 12 14 16 18
Aquifer layer
Fig. 6 Variation of average aquifer drawdown pressure with layer (Current research).
8,000
7,000
Aquifer average drawdown pressure (kN/m2)
6,000
5,000
4,000
3,000
2,000
1,000
0
0 2 4 6 8 10 12 14 16 18
Aquifer layer
Fig. 7 Average aquifer drawdown pressure with layer. Existing work.
70
60
50
Aquifer head (m)
40
30
20
10
0
0 2 4 6 8 10 12 14 16 18
Aquifer layer
Fig. 8 Variation of average aquifer pressure head with layer.

Table 2 Variation of different aquifer rates/yields with average drawdown pressure after 365 days.
Aquifer rates/yield (m3/s) Present work Existing work Difference Percentage difference (%)
0.50 1.97 2.08 -0.01 -5.61
1.00 2.06 2.19 -0.03 -6.31
1.50 2.34 2.41 -0.07 -2.99
2.00 2.42 2.52 -0.17 -4.13
2.50 2.52 2.68 -0.06 -6.35
Table 3 ANOVA of results of simulation.

Standard Standard
Mean Mean Mean Standard Draw down Head Pressure
Yield deviation, deviation,
drawdown head pressure deviation, variance variance variance
(m3/s) drawdown head
(N/m2) (m) ( 104) pressure ( 10-5) ( 10-2)
( 10-3) ( 10-1)
0.5 1.977 52.02 2.923 9.97 1.82 3.42 9.94 3.31 1,169.71
1.0 2.046 50.01 2.604 8.61 1.61 3.16 8.82 3.33 1,142.63
1.5 2.336 49.02 2.432 8.30 1.55 2.92 8.32 3.22 1,141.36
2.0 2.504 46.33 2.236 7.97 1.03 2.55 7.66 3.00 1,041.33
2.5 2.634 43.24 1.967 7.62 1.01 2.37 7.21 2.91 1,001.32
6. Conclusions and Recommendations capture a little more of reality.
6.1 Conclusions References

The finite element method, finite difference method [1] Saman, P. G. 1992. Vortex Dynamics. Cambridge: Univ.
and finite volume method has successfully being used Press.
[2] Lui, L., Feng, J., Jain-ren, F., and Ke-fa, C. 2005.
to study fluid flow in aquifer. This paper has revealed
Recent Development of Vortex Methods in
the potentials of cellular vortex element modeling Incompressible Viscous Bluff Body Flows. Journal of
technique as an efficient and viable alternative Zheijang University of Science 6A (4): 283-8.
numerical tool for subsurface porous media fluid flow [3] Dyrdahl, J. 2014. Thermal Flow in Fractured Porous
Media and Operator Splitting. Master of science in
analysis with special attributes of solutions that are
mathematics thesis, Norwegian University of Science and
grid independent with little computer resource usage Technology.
and also time conserving. The method is however, [4] Ewing, R. E., Russell, T. F., and Wheeler, M. F. 1984.
both probabilistic and deterministic in solution Convergence Analysis of an Approximation of Miscible
Displacement in Porous Media by Mixed Finite Elements
evaluation.
and a Modified Method of Characteristics. Computer
6.2 Recommendations Methods in Applied Mechanics and Engineering 47 (1-2):
73-92.
It is therefore recommended that, the 3D model be [5] Shen, S. F., and Lu, Z. H. 1985. Random Walk Technique
formulated and applied to investigate the media. or Transient Heat Transfer from Flat Plate in Steady Forward
Motion. In Numerical Methods in Heat Transfer, edited
Multiple vortex images should be also investigated
by Lewis, R. W., and Morgan, K. Chistester: John Wiley.
with 1D, 2D and 3D. It is also recommended that, [6] Ogundare, M. 2006. Vortex Element Modeling of Fluid
these models be applied to other subsurface porous and Heat Flow in Vertical Channels and Ducts. Ph.D.
media to actually ascertain its generality and wider thesis, University of Ibadan.
application. The axes used to formulate the models [7] Zheng, Y., Robert, B., and Daniel, B. 2010. Reservoir
Simulation with the Finite Element Method Using Biot
were considered to be rectilinear which of course is
Poroelastic Approach. Energy Journal 4: 16-29.
not in the actual sense. It is therefore also [8] Lugt, H. J. 1983. Vortex Flow in Nature and Technology.
recommended that, the assumptions be relaxed to Malabar: Krieger Publishing Company.
[9] Ajiroba, O. A. 2006. Finite Difference Simulation of [12] Rezael, A. A., and Donaldson, U. 1987. A New
Fluid Flow in an Oil Reservoir. Unpublished B.Sc Two-Dimensional Single Phase Water Flood Model
Project, University of Ibadan. for Micro-computers. In SPE Paper 16491. Vol. 6.
[10] Mark, J. S. 2002-2007. Summary of Vortex Methods Texas: Henry L. Doherty Memorial Fund of AIME,
Literature (A Living Document Rife with Opinion). UK: 101-15.
Cambridge University, 17-26. [13] Ertekin, T., Abou-Kassem, J. H., and King, G. R. 2001.
[11] Speziale, C. G. 1987. On the Advantages of the Basic Applied Reservoir Simulation. In Society of
Vorticity-Velocity Formulation of the Equations of Fluid Petroleum Engineers Textbook Series. Vol. 10. Texas:
Dynamics. J. Comput. Phys. 73 (2): 476-80. Henry L. Doherty Memorial Fund of AIME, 64-88.
doi: 10.17265/1934-8975/2016.03.002
D DAVID PUBLISHING
Numerical Analysis of a Gravity Substructure for 5 MW

Offshore Wind Turbines Due to Soil Conditions
Min-Su Park, Youn-Ju Jeong and Young-Jun You

Structural Engineering Research Institute, Korea Institute of Civil Engineering and Building Technology, Goyang-Si 10223,
South Korea
Received: January 12, 2016 / Accepted: January 21, 2016 / Published: March 31, 2016.
Abstract: In order to increase the gross generation of wind turbines, the size of a tower and a rotor-nacelle becomes larger. In other
words, the substructure for offshore wind turbines is strongly influenced by the effect of wave forces as the size of substructure
increases. In addition, since a large offshore wind turbine has a heavy dead load, the reaction forces on the substructure become
severe, thus very firm foundations should be required. Therefore, the dynamic soil-structure interaction has to be fully considered and
the wave forces acting on substructure accurately calculated. In the present study, ANSYS AQWA is used to evaluate the wave
forces. Moreover, the substructure method is applied to evaluate the effect of soil-structure interaction. Using the wave forces and the
stiffness and damping matrices obtained from this study, the structural analysis of the gravity substructure is carried out through
ANSYS mechanical. The structural behaviors of the strength and deformation are evaluated to investigate an ultimate structural
safety and serviceability of gravity substructure for various soil conditions. Also, the modal analysis is carried out to investigate the
resonance between the wind turbine and the gravity substructure.
Key words: Offshore wind energy, gravity substructure, suction bucket foundation, substructure method, structural analysis.
1. Introduction been conducted on wind energy [1-3]. In order to

construct the offshore wind farms, the substructures
Nowadays, the main source of energy in the world
supporting offshore wind turbines have to resist loads
is fossil fuel. But the amount of fossil fuel is limited
from wind and wave. However, the size of a
and the use of it causes environmental pollution and
substructure for wind turbines is gradually increased
global warming. Thus, the studies of renewable
since the size of a tower and a rotor-nacelle becomes
energy such as hydro energy, wind energy, solar
larger with increment of gross generation. In other
energy and geothermal energy are being carried out
words, the substructure is strongly influenced by the
actively all over the world. The offshore wind energy
effect of wave forces and the safety of substructure is
has gained attention from many countries to find
decreased. Therefore, it is very important to accurately
alternative and reliable energy sources, since the
calculate the wave forces acting on substructures.
potential of offshore wind energy has been recognized
In addition, since a large offshore wind turbine has
for long and mostly associated with a nondestructive
heavy dead loads, the reaction forces on the
renewable energy. Therefore, many offshore wind
foundations become severe, thus very firm
farms are in the planning phase. Various studies have
foundations should be required. Therefore, the
dynamic soil-structure interaction has to be fully
Corresponding author: Min-Su Park, senior researcher considered. There are two main methods for the
PhD, research field: hydrodynamics of offshore structures.
It is noted that, this paper is revised edition based on analysis of dynamic structure-foundation interaction,
Proceedings of the ASME 2015 34th International Conference direct and substructure methods, as outlined by
on Ocean, Offshore and Arctic Engineering, May 31-Jun. 5,
2015, Canada. Wolf [4, 5].
Numerical Analysis of a Gravity Substructure for 5 MW Offshore Wind Turbines Due to Soil Conditions 151
In the present study, ANSYS AQWA is used to and a suction bucket foundation system as shown in
evaluate the wave forces acting on the gravity Fig. 1.
substructure for 5 MW offshore wind turbine, since As the substructure mass increases, the interaction
Morison equation cannot accurately calculate the between the superstructure and foundation system
wave forces in case of the large substructure compared intensifies, thus, the contribution to the structural
to a wave length. The substructure is significantly response of the total system increases. An impedance
influenced by the wave steepness and wave slamming function was used for analysis of the suction
effects but they are not considered in the present study, bucket-soil foundation system. Impedance coefficients
because the wave breaking is not occurred for the as damping and stiffness matrices were applied to the
design environmental wave conditions. The wave equation of motion for the foundation system [7, 8].
forces and wave run up on the substructure are Applying substructure method, the pile-soil
presented for various wave conditions. Moreover, the foundation system can be expressed with the
substructure method is applied to evaluate the effect of impedance coefficients for two-dimension such as the
soil-structure interaction with impedance coefficients horizontal direction, the rotational direction and the
as damping and stiffness matrices. Using the wave vertical direction as shown in Fig. 2.
forces and the stiffness and damping matrices
obtained from this study, the structural analysis of the
gravity substructure is carried out through ANSYS
mechanical. The structural behaviors of the strength
and deformation are evaluated to investigate an
ultimate structural safety and serviceability of gravity
substructure for various soil conditions. The first few
natural frequencies of substructure are heavily
influenced by the wind turbine. Therefore, the first
natural frequency of substructure is to be within the
soft-stiff range in between the 1 P and 3 P frequency
ranges. The rotor frequency (1 P) range lies between
0.115 Hz and 0.202 Hz, and the blade passing
frequency (3 P) range lies between 0.35 Hz and 0.61 Hz.
A safety margin of 10% on the maximum and
minimum rotor speed is adopted, which means that,
the allowable frequency is between 0.222 Hz and
0.31 Hz [6]. Therefore, the modal analysis is carried
out to investigate the resonance between the wind
turbine and the gravity substructure. It is found that,
the suggested gravity substructure can be an effective
substructure for 5 MW offshore wind turbines.
2. Formulation
Fig. 1 Geometrical definition of the gravity substructure:
The gravity substructure for 5 MW offshore wind HAT (highest astronomical tide), MSL (mean sea level),
turbine consisted of a concrete con type superstructure LAT (lowest astronomical tide).
152 Numerical Analysis of a Gravity Substructure for 5 MW Offshore Wind Turbines Due to Soil Conditions
where, the suffixs x, and z denote the component of

horizontal direction, the component of rotational
direction and the component of vertical direction,
respectively. Ep, Gs, Vs, Dp and Ip are the Young
modulus, the shear modulus of subsurface ground, the
shear wave velocity of subsurface ground, the
diameter of pile and the moment of inertia of section,
Fig. 2 Pile-soil foundation system with substructure respectively. The impedance coefficients Kp and Cp
method. denote the stiffness matrix and the damping matrix of
The impedance coefficient denotes the relation foundation system. The relationship between the shear
between the dynamic reaction force and the strength modulusEs, and the shear wave
corresponding displacements. The impedance velocityVs, is expressed as follow:
coefficient related to the suction bucket system on Gs 2.0 1 sVs (5)
relatively soft ground conditions can be represented as
follows: in which, and s denote the Poissons ratio of soil
K xx 0 K x and the density of soil, respectively.
K p 0 K zz 0 (1) In the present study, the stiffness and damping

K x 0 K coefficient of single pile, which has same thickness of
suction bucket, is firstly obtained and it is assumed
Cxx 0 C x that, the around of suction bucket is consisted of 107
C p 0 C zz 0 (2) piles to express the suction bucket-soil foundation

C x 0 C system, since the general solution to calculate the
stiffness and damping coefficient on the suction
in which:
0.75 bucket foundation system is not well known.
E p I p Gs
K xx 2.0
r 3 E p 3. Numerical Results and Discussion
0.5
E p I p Gs 3.1 Wave Force Evaluation
K x K x -1.2 2
r Ep (3) Fig. 3 shows the comparison of wave forces on
0.25
E p I p Gs gravity substructure for various water depths. The
K 1.6
r E water depth of LAT, MSL and HAT is 17.2 m, 20.48 m
p
E p A Vs and 23.75 m, respectively. In the comparison, the
K zz 50 calculated total wave forces are divided by incident
l V p
wave amplitude (H/2). The wave forces on the gravity
Dp
C xx 2.0 K xx substructure with the water depth LAT are largest
Vs
compared to the other cases. Since the wave force is
Dp
C x C x 1.5 K x closely related to the wetted surface of substructure
Vs
(4) and the water particle velocity near free surface is
Dp largest, the wave forces gradually decrease as the
C 0.5 K
Vs water depth becomes large. The peak wave force with
0.026l water depth MSL and HAT decreases about 12% and
C zz K zz
Vs 20% compared to the peak value of water depth LAT.
.
1,600 LAT
Wave force/wave amplitude (kN/m)
MSL
HAT
1,400
1,200
1,000
800
600
4 6 8 10 12 14 16 18 20
Wave period (s) (a) Panel pressure for LAT
Fig. 3 Comparison of wave forces on the gravity
substructure for various water depths.
Figs. 4 and 5 show the panel pressure and wave run

up on gravity substructure. Since the significant wave
period of southern-western sea in Korea is 12.42 s, the
comparison of those is made at the same wave period.
The panel pressure near the free surface is largest and
the pattern of wave run up around the substructure is
very similar for all cases.
3.2 Wind Turbine Model and Environmental Loads

(b) Panel pressure for MSL
In the present study, NREL (National Renewable
Energy Laboratory) 5.0 MW wind turbine model is
selected for the structural safety analysis of gravity
substructure. The details of NREL 5.0 MW wind
turbine are provided in Table 1. The total weights of
turbine and the tower model are about 350 ton
and 348 ton, respectively. The hub height is 82.72 m
from MSL and the tower length is 68.0 m [9]. The
height of gravity substructure is 32.7 m from seabed
and the water depth (MSL) is 20.48 m as shown in
Fig. 1.
At the structural safety analysis, design code IEC (c) Panel pressure for HAT
Fig. 4 Panel pressure on gravity substructure: (a) LAT, (b)
61400-3 is adopted and structure analysis is carried
MSL, (c) HAT.
out according to the ultimate design loads condition
6.1a and 6.2a presented in Table 2. DLC (design load southern-western sea of the Korea peninsular are
cases) 6.1a and 6.2a present ultimate limit states for presented in Table 3. Extreme wind and wave loads
the parked (standing still or idling) state and grid loss subjected to the gravity substructure are calculated
state, respectively. based on the wind speed and the wave period of return
Environmental loads of wind and wave for the period 50 years.
Table 2 DLC for ultimate limit state.

DLC Condition Wave Wind
Parked standing
IEC 61400-3-(6.1a) Extreme Extreme
still or idling
IEC 61400-3-(6.2a) Grid loss Extreme Extreme
Table 3 Environmental loads.

Wind Wave Limit state
Hw = 1.52 m, FLS (fatigue load
Normal 7.7 m/s
Pw = 6.2 s state)
H = 13.76 m, ULS (ultimate limit
Extreme 41.6 m/s w
(a) Wave run up for LAT Pw = 12.42 s state)
3.3 Natural Frequency and Resonance
In order to evaluate the resonance between the wind

turbine and the gravity substructure, the modal analysis
is carried out. The soil conditions are expressed by
using the Caltrans/NEHRP (national earthquake
hazard reduction program) soil profile as shown in
Table 4 [10].
(b) Wave run up for MSL Table 5 shows the stiffness and damping matrices
on suction bucket-soil foundation system using
substructure method. The natural frequencies and the
mode shapes of gravity substructure present in Table 6
and Fig 6, respectively. Since the suction bucket
stiffness increases as the shear-wave velocity of soil
increases, the natural frequency is expected to increase
with the velocity increments (Vs30). Therefore, the
natural frequency of the total system is directly
(c) Wave run up for HAT dependent on the soil conditions of the foundation.
Fig. 5 Wave run up on gravity substructure (a) LAT, (b)
The natural frequency of gravity substructure
MSL, (c) HAT.
system obtained from modal analysis is plotted against
Table 1 Details of NREL 5 MW wind turbine model.
Table 4 Caltrans/NEHRP soil profile types.
Turbine parameter Unit Value
Standard Undrained
Rating MW 5.0 Site Soil profile Vs30 penetration shear
Configuration - 3 blades class name (m/s) test N-value strength
Rotor, hub diameter m 126, 3 (bpf) (kPa)
Cut-in, rated wind speed m/s 3, 11.4 A Hard rock > 1,500 - -
Cut-in, rated rotor speed rpm 6.9, 12.1 B Rock 760 - -
Cut-out wind speed m/s 25 Very dense
C soil and soft 360 > 50 > 100
Rated tip speed m/s 80 rock
Rotor mass kg 110,000 D Stiff soil 180 15 50
Nacelle mass kg 240,000 E Soft soil < 180 < 15 < 50
Tower mass kg 347,460 Soil requiring
Coordinate location of overall center F site specific - - -
m (-0.2, 0, 64)
of mass evaluation
Table 5 Stiffness and damping matrices according to Vs30.

Stiffness coefficient Damping coefficient
Vs30
(kN, kNm) (kNs, kNms)
Kxx, Kyy 1.21E+03 Cxx, Cyy 6.70E+00
Kzz 1.12E+08 Czz 4.55E+05
90 m/s
KRxx, KRyy 6.21E+05 CRxx, CRyy 8.62E+02
KRxy, KRyx 1.83E+04 CRxy, CRyx 7.64E+01
Kzz 1.12E+08 Czz 1.52E+05
270 m/s
Kxx, Kyy 1.87E+04 Cxx, Cyy 1.67E+01 (a)
Kzz 1.12E+08 Czz 7.31E+04
560 m/s
Kzz 1.12E+08 Czz 3.62E+04
1,130 m/s
Table 6 Natural frequencies of gravity substructure for

various soil conditions (Hz).
Mode 90 m/s 270 m/s 560 m/s 1,130 m/s Fixed
1 0.1162 0.2219 0.2649 0.2813 0.3049
2 0.2436 0.2635 0.2776 0.2849 0.3062 (b)
3 0.3059 0.3875 0.5726 0.9093 2.2765 Fig. 6 Mode shapes of gravity substructure for Vs30 = 560 m/s.
4 0.8916 0.9097 1.5251 1.6144 2.2897
5 1.7226 1.9418 2.0766 2.2030 2.6899
6 2.0494 2.2414 2.2944 2.2945 3.8401
7 2.3007 2.3261 3.0837 3.0954 5.4344
8 2.9606 3.1019 3.2421 3.4220 5.4601
9 4.0152 4.1879 4.7267 4.9292 6.1734
10 5.0517 5.1409 5.2323 5.3553 9.0935
to the spectrum density of the turbine and the

environmental loads in Fig 7. It is found that, the
natural frequency of gravity substructure system with
more than Vs30 = 270 m/s is located between the rotor Fig. 7 Comparison of gravity substructure on the spectrum
density of wind turbine and wave for various soil conditions.
frequency range (1 P) and the blade passing frequency
range (3 P). Therefore, there is no possibility of wind are presented in Table 7. In the present study, the
resonance between the wind turbine and the gravity extreme values at TP for DLC 6.2a are used since they
substructure with more than Vs30 = 270 m/s. are the largest values compared to the other case. The
gravity substructure is composed of concrete with 0.5 m
3.4 Structural Results
thickness. The partial safety factor 1.35 is adapted for
The loads at TP (transition pieces) due to extreme all cases [11].
Table 7 Extreme values at TP due to extreme wind. respectively. Although the displacement and the
Fx Fy Fz Mx My Mz von-Mises stress increase as the water depth becomes
DLC
(kN) (kN) (kN) (kNm) (kNm) (kNm)
large, the bending stress decreases. The bending stress
6.1a 324 225 -6,04 -21,48 3,765 236
6.2a 1,364 -571 -6,32 39,827 82,343 -2,610 is concentrated at the connection part between circular
cylinder and concrete cone. There is also large
variation of bending stress. Therefore, this connection
part should be examined explicitly for the reliable
substructure design. The structural results are
summarized at Table 8. It is found that, the ultimate
strength of gravity substructure system satisfies ULS
design condition for various wave conditions.
Figs. 11-13 show the structural displacement, the
bending stress and the von-Mises stress with water
Fig. 8 Structural displacement of gravity substructure due depth HAT for various soil conditions, respectively.
to water depth conditions (unit: m).
Although the displacement decreases as the shear
wave velocity of soil increases, the other results have
very similar values. It means that, the displacement is
strongly influenced by soil conditions but the structural
Table 8 Structural results due to water depth conditions.

LAT MSL HAT
Displacement Max 0.1219 0.1238 0.1256
(m) Min 0.0563 0.0571 0.0586
Bending stress Max 18.054 17.969 17.842
Fig. 9 Structural bending stress of gravity substructure (MPa) Min 1.4E-15 1.7E-15 1.0E-16
due to water depth conditions (unit: Pa). Max 11.790 11.806 11.832
Von-Mises stress
(MPa) Min 5.5E-14 7.6E-14 5.6E-14
Normal stress Max 3.592 3.560 3.528
(x axis, MPa) Min -2.902 -2.930 -2.941
(y axis, MPa) Min -5.310 -5.350 -5.401
(z axis, MPa) Min -5.702 -5.836 -5.925
Fig. 10 Structural von-Mises stress of gravity substructure

due to water depth conditions (unit: Pa).
Using the wave forces from ANSYS AQWA and

the wind loads at TP, the structural analysis of gravity
substructure is carried out through ANSYS
mechanical. For various water depths the structural
displacement, the bending stress and the von-Mises Fig. 11 Structural displacement of gravity substructure
stress with Vs30 = 560 m/s are plotted at Figs. 8-10, due to soil conditions (unit: m).
4. Conclusions
The gravity substructure with suction bucket
foundation system for 5 MW offshore wind turbine is
suggested to examine soil-structure interaction. Using
the wave forces and the stiffness and damping
matrices obtained from the substructure method, the
structural analysis of the gravity substructure is
carried out through ANSYS mechanical for various
Fig. 12 Structural bending stress of gravity substructure
due to soil conditions (unit: Pa). soil conditions. The results obtained from the present
study are summarized as follows:
(1) Since the wave force is closely related to the
wetted surface of substructure and the water particle
velocity near free surface is largest, the total wave
forces on gravity substructure gradually decrease as
the water depth increases. Moreover, although the
structural displacement is strongly influenced by soil
conditions, the structural stress is slightly influenced.
(2) From the structural analysis of gravity
Fig. 13 Structural von-Mises stress of gravity substructure
due to soil conditions (unit: Pa).
substructure, it is found that, the gravity substructure
system satisfy structural safety in respect of ULS.
Table 9 Structural results due to soil conditions. Also, the resonance between the wind turbine and the
270 560 1,130
m/s m/s m/s
Fixed gravity substructure is not occurred.
Displacement Max 0.266 0.125 0.074 0.026 (3) Consequently, the suggested gravity
(m) Min 0.183 0.058 0.019 0.000 substructure with suction bucket foundation system
Bending stress Max 17.84 17.84 17.84 17.84 can be an effective substructure for reducing
(MPa) Min 1.E-16 1.E-16 9.E-17 0.000 hydrodynamic effects and construction costs in the
Von-Mises Max 11.83 11.83 11.83 11.83 southern-western sea of the Korea peninsular.
stress
(MPa) Min 1.E-13 5.E-14 2.E-14 0.000 (4) In the present study, the static structural analysis
Normal stress Max 3.528 3.528 3.528 3.528 is only carried out. So the linearized initial stiffness is
(x axis, MPa) Min -2.941 -2.941 -2.941 -2.942 used for the analysis. However, next step the dynamic
Normal stress Max 6.769 6.769 6.769 6.769
structural analysis will be performed with the
(y axis, MPa) Min -5.401 -5.401 -5.401 -5.401
nonlinear foundation stiffness.
Normal stress Max 7.625 7.625 7.625 7.615
(z axis, MPa) Min -5.925 -5.925 -5.925 -5.935 Acknowledgments
stress is slightly influenced by soil conditions. The This research was supported by Korea Institute of
maximum bending stress and the von-Mises occurs at Marine Science and Technology Promotion through
the connection part between TP and circular cylinder. the research project Safety evaluation of concrete
The structural results are summarized at Table 9. For substructure systems for offshore wind power
various soil conditions, the ultimate strength of (D11206012H370000130) and Korea Institute of
gravity substructure system also satisfies ULS design Energy Technology Evaluation and Planning through
condition. the research project Development of hybrid
substructure systems for offshore wind power Structures). Stuttgart: Institute of Aircraft Design
Universitt Stuttgart.
(SUBJID0000000014637).
[7] Yamada, Y., Kawano, K., Iemura, H., and
Venkataramana, K. 1988. Wave and Earthquake
References
Response of Offshore Structure with Soil-Structure
[1] Jan, W., Marte, R., Christian, S., and Edgar, G. H. 2009. Interaction. Proceeding Structure Eng./Earthquake Eng.
Life Cycle Assessment of a Floating Offshore Wind 5 (2): 361-70.
Turbine. Renew Energy 34 (3): 742-7. [8] Park, M. S., Koo, W. C., and Kawano, K. 2011.
[2] Jonkman, J. M. 2009. Dynamics of Offshore Floating Dynamic Response Analysis of an Offshore Platform
Wind Turbines-Model Development and Verification. Due to Seismic Motions. Engineering Structures 33 (5):
Wind Energy 12 (5): 459-92. 1607-16.
[3] Lozano-Minguez, E., Kolios, A. J., and Brennan, F. P. [9] Jonkman, J., Butterfield, S., Musial, W., and Scott, G.
2011. Multi-criteria Assessment of Offshore Wind 2009. Definition of a 5 MW Reference Wind Turbine
Turbine Support Structures. Renew Energy 36 (11): for Offshore System Development. National Renewable
2381-7. Energy Laboratory.
[4] Wolf, J. P. 1985. Dynamic Soil Structure Interaction. [10] Bernard, R. W., Jason, T. D., and Thomas, S. 2012.
New Jersey: Prentice Hall, Inc. Guidelines for Estimation of Shear Wave Velocity Profile.
[5] Wolf, J. P. 1988. Soil Structure Interaction Analysis in PEER (Pacific Earthquake Engineering Research) report
Time Domain. New Jersey: Prentice Hall, Inc. 2012/08.
[6] Fischer, T., De Vries, W., and Schmidt, B. 2010. Upwind [11] DNV (Det Norske Veritas). 2013. DNV-OS-J101:
Design Basis (WP4: Offshore Foundations and Support Design of Offshore Wind Turbine Structures. DNV.
doi: 10.17265/1934-8975/2016.03.003
D DAVID PUBLISHING
Production of Rhenium by Transmuting Tungsten Metal

in Fast Reactors with Moderator
Tsugio Yokoyama1, Yuki Tanoue2, Atsunori Terashima2 and Masaki Ozawa1

1. Research Laboratory for Nuclear Reactors, Tokyo Institute of Technology, Tokyo 152-8550, Japan
2. Department of Nuclear Engineering, Graduate School of Science and Engineering, Tokyo Institute of Technology, Tokyo 152-8550,
Japan
Received: January 20, 2016 / Accepted: February 02, 2016 / Published: March 31, 2016.
Abstract: The feasibility of rhenium (Re) production by irradiating tungsten (W) metal in a medium size fast reactor was evaluated by
using a Monte Carlo code. The fast reactor can produce about 50 kilograms of Re per every 3 years, which corresponds 10% of
Japanese domestic production. The specific activity of Re can be reduced below the exemption level or even the natural Re level if W
and osmium is separated after the irradiation. The use of ZrD1.7 moderator reduces the specific activity by half compared to that of
ZrH1.7 case, and even the no moderator case is permissible to produce the production of Re which has lower specific reactivity than that
of natural Re.
Key words: Rhenium, tungsten, fast reactor, transmutation, MVP, ORIGEN2, specific activity.
1. Introduction The neutron energy spectrum of FRs (fast reactors)

shows faster or harder than that or LWRs (light water
In recent years, the demand for rhenium (Re) is
reactors) due to the lack of neutron moderators. The
growing as an additive to super alloys for jet engines.
average number of emitted neutrons of FRs per
However, rhenium is one of the rarest elements in earth.
absorption, where the main fissile nuclide is
The abundance of Re is much smaller: 10-9 times of Si
plutonium-239, is about 2.5, while that of LWRs is 2.1.
and 10-4 times of tungsten (W). The price of Re is
Furthermore, the neutron loss of parasitic absorption
about 500 times that of W at present. Therefore, active
of structure and coolant per fission neutron is 0.1 in
technology for creation of Re that we advocate has a
FRs, while that of LWRs is 0.5. Thus more excess
great value.
neutrons are expected to utilize for transmutation if
Aprs ORIENT research program, newly initiated in
the system is designed appropriately.
2011, includes transmutation of stable elements to create
In our study, we focused two points for setting the
rare metals/RE (rare earth) elements by (n, ) reaction
transmutation system in FRs, amount of produced Re
with subsequent beta decays in the reactors [1-3]. Re
for practical use and specific activity of Re for
can be produced by nuclear transmutation where W
radiation protection. The exemption level of specific
captures neutrons and disintegrates to Re [4].
activity, that is a radioactivity level under that market
The objective of this study is to clarify the
use is approved, is defined by IAEA for each
feasibility of production of Re from W metal in fast
radioactive nuclide as Bq/gram-element [5].
reactors by using neutron moderator.
2. Candidates for Neutron Moderator in FRs
Corresponding author: Masaki Ozawa, professor, research
fields: spent nuclear fuel reprocessing, PUREX process Fig. 1 illustrates the transmutation of W to Re with
chemistry, partitioning and transmutation, actinides and fission
product separation chemistry. other related reactions in reactors. The isotope
abundance of W indicates that, two main paths for Re deuteride are employed because zirconium can retain
production are: much hydrogen (1.7 times Zr or more at 600 C) and
(1) W-184 (abundance: 31%) to W-185 by neutron have some experiences for the reactor materials in
capture to Re-185 by beta disintegration; FRs.
(2) W-186 (abundance: 28%) to W-187 by neutron
3. Analysis Method
capture to Re-187 by beta disintegration.
As shown in Fig. 2, lower energy neutrons react The code system used in the analysis is shown in
with more W isotopes. Fig. 3 shows the scattering and Fig. 4. A continuous energy 3-dimensional Monte
absorption cross section of the typical neutron Carlo code, MVP [6] with its burn-up calculation
moderator nuclides, proton and deuteron. Proton has a routine MVP-BURN [7] and JENDL-4.0 library [8]
larger scattering cross section than deuteron by 1 order where the neutron energy spectra of each zone of
while deuteron has much smaller absorption cross reactors are exactly simulated as well as detailed
section. We have selected, thus, materials containing configurations is used for evaluating the transmutation
proton or deuteron as candidates for effective rate of W, the effective neutron cross sections and
moderator, and zirconium hydride or zirconium neutron flux for the use of ORIGEN2.
1.34 3.36 1.25 3.32
Beta disintegration Os-186
(barn)
Os-187
(barn)
Os-188
(barn)
Os-189
(barn)
Os-190
3.7d 17h 24h

Neutron
Betacapture reaction
disintegration
410 10 y
4.03 0.015 0.0021

Re-185 Re-186 Re-187 Re-188 Re-189
Neutron capture reaction (barn) (barn) (barn)
75d 24h 70d

W-180 1.19 W-182 0.81 W-183 1.58 W-184 0.50 W-186 0.597 6.43
W-181 W-185 W-187 W-188
0.12% (barn) 26.50% (barn) 14.31 (barn) 30.64% (barn) 28.43% (barn) (barn)
121d 115d 5.1d
2.1b 2.33b
Ta-181 Ta-182 Ta-183 Ex.note
Ex.note
0.81 W-183 : Nuclide
(barn) 14.31% : Natural abundance for stable isotopes
1gr. cross section for fast reactor blanket (ORLIBJ4.0)
Fig. 1 Nuclear transmutation scheme of tungsten (W) to rhenium (Re).
1.E+04
1.E+03
1.E+03 Scattering
Capture cross section (barn)
W186 (proton)
1.E+02 1.E+01 1 order
Cross section (barn)
Scattering
1.E+01 (deuteron)
1.E-01
1.E+00
1.E-03
3 orders
Capture
1.E-01 W184 (proton)
1.E-05
1.E-02 Capture
(deuteron)
1.E-03 1.E-07
1.E-04 1.E-02 1.E+00 1.E+02 1.E+04 1.E+06 1.E-05 1.E-03 1.E-01 1.E+01 1.E+03 1.E+05 1.E+07
Neutron energy (eV) Neutron energy (eV)
Fig. 2 Capture cross section of W isotopes. Fig. 3 Cross section of hydrogen and deuteron.
Production of Rhenium by Transmuting Tungsten Metal in Fast Reactors with Moderator 161
power of the core is 710 MWt, the coolant is liquid

JENDL4.0
sodium and the active core size is 1.8 m in diameter
and 0.9 m in height.
MVP The number of the target assemblies is 54 located at
(MVPBURN)
the peripheral of the core where the impact of the target
on core performance can be minimized.
Effective The pin layout of the target assembly is shown at
WtoRe cross
production section
the right hand in Fig. 5. Outer 43 pins contain tungsten
(W) pellets, which total height is 0.9 m.
ORIGEN2 Inner 18 pins contain zirconium hydride (ZrH1.7) or
zirconium deuteride (ZrD1.7) pellets, which face the
active core.
Specific We have set the no-moderator case for the comparison,
activity
where all of 61 pins contain W pellets only.
Fig. 4 Capture cross section of W isotopes.
The one-point model burn-up code ORIGEN2 with
5.1 Cross Section and Neutron Flux at W Region
one group cross section system [9] is used for
evaluating the specific activity of each isotope by As shown in Table 1, the one group effective
applying the effective cross section and flux evaluated capture cross section and neutron flux at W region are
by the MVP code. evaluated by MVP for the use of ORIGEN2, where
specific reactivity is calculated. The effective cross
4. Core and Assembly for Transmutation
sections of moderator applied cases increase by a few
The fast reactor core layout and the assembly for times that of no moderator case (Case 5), while the
the transmutation are shown in Fig. 5. The thermal fluxes reduce by about 0.6-0.9 time only.
1st row : 9 pins

(ZrH1.7 or ZrD1.7)
2nd row : 9 pins

Radial blanket (ZrH1.7 or ZrD1.7)
Outer core 43 tungsten pins

(W pellet diameter : 10.4 mm)
Inner core
0.1 m
1.8 m
Target assembly
Core layout (pin heterogeneous model) (total 54 assemblies)
Fig. 5 Core layout and target assembly model for MVP calculation.
Table 1 One group effective capture cross section and neutron flux at W region by MVP code.
Case Number of pins Pin map W-182 W-183 W-184 W-186 Re-185 Re-187 Neutron flux
(barn) (barn) (barn) (barn) (barn) (barn) (n/cm2/s)
ZrH1.7 18 pins
Case 1 0.574 1.203 0.237 0.580 9.348 4.823 7.57E+14
W 43 pins
ZrD1.7 18 pins
Case 2 0.268 0.792 0.230 0.176 2.355 2.361 1.09E+15
W 43 pins
ZrH1.7 9 pins
Case 3 ZrD1.7 9 pins 0.418 1.053 0.235 0.365 6.255 4.138 8.79E+14
W 43 pins
ZrD1.7 9 pins
Case 4 ZrH1.7 9 pins 0.442 1.115 0.24 0.399 6.728 4.39 8.65E+14
W 43 pins
W 61 pins
Case 5 0.187 0.459 0.169 0.124 1.093 1.008 1.21E+15
(no moderator)
Original ORLIBJ40
0.808 1.583 0.500 0.597 4.027 0.015 -
(infinite dilute model)
Table 2 Production weight of Re after 3 years irradiation.

Re-187
Case Number of pins Pin map W load Re weight Re/W Re-185 Re-187
/Re-185
(kg) (kg) (%) (kg) (kg) -
ZrH1.7 18 pins
Case 1 3,520 51.2 1.45 14.1 37.1 2.63
W 43 pins
ZrD1.7 18 pins
Case 2 3,520 40.3 1.14 23.6 12.7 0.54
W 43 pins
ZrH1.7 9 pins
Case 3 ZrD1.7 9 pins 3,520 44.4 1.26 17.6 27.8 1.58
W 43 pins
ZrD1.7 9 pins
Case 4 ZrH1.7 9 pins 3,520 45.6 1.30 17.3 27.2 1.57
W 43 pins
W 61 pins
Case 5 5,010 49.1 0.98 29.1 19.9 0.68
(no moderator)
5.2 Production of Re after 3 Years Irradiation The Re production rate of moderator applied cases is
The production weight of Re after 3 years 1.5 times that of no-moderator case at maximum.
irradiation in the core has been evaluated by The weight of Re-187 is larger than that of Re-185 in
MVP-BURN code. Table 2 lists the load weight of W, Case 1 (ZrH moderator), while Re-185 is larger than
production weight of Re and the ratio with isotropic Re-187 in Case 2 (ZrD moderator).
fractions. This relation is investigated by analyzing neutron
The produced Re weight at 3 years irradiation will be spectra as shown in Fig. 6. The softer neutron
more than 50 kg in prototype reactor, which spectrum by ZrH1.7 in Case 1 produces more Re-187
corresponds to about 10% of Japanese total production due to the higher W-186 cross section in lower energy.
of Re at present. The harder spectrum by ZrD1.7 in Case 2 produces less
ZrH1.7 moderator ZrD1.7 moderator

(Case 1) (Case 2)
6.E+04 5.E+04
Total weight of isotopes (g)

5.E+04
Irradiation Cooling Irradiation Cooling Re-total
4.E+04
4.E+04
Re-187
3.E+04 Re-185
Change in 3.E+04
2.E+04 Os-total
Re isotopes 2.E+04
Os-188
during irradiation 1.E+04 1.E+04
Os-186
0.E+00 0.E+00
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Irradiation and cooling time (year) Irradiation and cooling time (year)
1.E-04 1.E-04
Neutron flux (1/cm2/s/source)
Core region
1.E-05 1.E-05
1.E-06 1.E-06
W region
1.E-07 1.E-07
Neutron spectrum
1.E-08 1.E-08
in W region
1.E-09 1.E-09
1.E-10 1.E-10
1.E-04 1.E-02 1.E+00 1.E+02 1.E+04 1.E+06 1.E-04 1.E-02 1.E+00 1.E+02 1.E+04 1.E+06
1.E+04 1.E+04
Capture cross section (barn)
1.E+03 1.E+03
W186 1.E+02
1.E+02
1.E+01
Average of
1.E+01 Average of
Spectrum averaged W184 W-184
1.E+00 W-186 1.E+00
capture cross section 1.E-01 1.E-01
Average of Average of
1.E-02 W-184 1.E-02 W-186
1.E-03 1.E-03
1.E-04 1.E-02 1.E+00 1.E+02 1.E+04 1.E+06 1.E-04 1.E-02 1.E+00 1.E+02 1.E+04 1.E+06
Fig. 6 Comparison of ZrH1.7 (Case 1) and ZrD1.7 (Case 2) moderator.
Table 3 Effective cross section of W isotopes. exemption level of Re specific activity is achieved in a
Case (moderator) W-184 W-186 half year cooling if W and Re is separated after the
Case 1 (ZrH) 0.237 b 0.580 b irradiation as shown in Fig. 8.
Case 2 (ZrD) 0.230 b 0.176 b
1.E+14
Irradiation Cooling W-total
Re-187. The effective cross section of W isotopes in W-185
Specific activity of isotopes (Bq/g-element)
W-187
the target region is shown in Table 3. 1.E+12 W-188
Re-total
Re-186
5.3 Specific Activity of Re Re-187
1.E+10 Re-188
Os-total
The specific activity of Re during irradiation and Os-185
Os-191
cooling has been evaluated by ORIGEN2 with the 1.E+08
Re-188
cross section and flux shown in Table 1.
Fig. 7 shows the specific activities of the isotopes 1.E+06
6 orders
of W, Re, and Os (osmium) in Case 1 where the higher
1.E+04 than
weight of Re production is maximum among the study exe. level
cases. The exemption levels of Re isotopes determined
1.E+02
by IAEA [5] are listed in Table 4. Re-188 exemption level
In Fig. 7, the main activity of Re is that of Re-188,
1.E+00
the daughter of W-188, during the cooling time, if W 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
and Re are mixed when the specific activity is 6 orders Irradiation and cooling time (year)
higher or more than the exemption level. Thus the Fig. 7 Specific activity when Re is not separated from W.
1.E+14
Table 4 Exemption level of Re isotopes. Irradiation Cooling
Isotope (T1/2) Bq/g

Re-186 (3.8 d) 1 103 1.E+12 Re-total
Re-187 (5 1010 y) 1 106
Re-188 (17 h) 1 102
1.E+10
5.4 Specific Activity for Various Moderators Re-186

1.E+08 Re-188
Fig. 9 illustrates the specific reactivity of Re for
various types of moderators when Re is separated
1.E+06
from W and Os after the irradiation. The use of ZrD1.7 Re-187 exemption level
reduces the specific activity by half compared to that 1.E+04 Re-187

of ZrH1.7 or no moderator due to the decrease of the 1.E+03
Re-186 exemption level
fraction of Re-187 (T1/2 = 4.33 1010 years). 1.E+02
Re-188 exemption level
The specific activity of produced Re is lower than
that of natural Re (2,610 Bq/g, Re-187 62.6%) for all 1.E+00
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
cases in Fig. 9, which suggests the no moderator case
Irradiation and cooling time (year)
is permissible for the production of Re in the fast Fig. 8 Specific activity when Re is separated from W and
reactor shown above. Os after irradiation.
1.E+14 Re-total 1.E+14 1.E+14

Irradiation Cooling Irradiation Cooling Irradiation Cooling
1.E+12 1.E+12 1.E+12
1.E+10 1.E+10 1.E+10

Re-186
1.E+08 1.E+08 1.E+08
Re-188
1.E+06 1.E+06 1.E+06

Re-187 exemption level Re-187 exemption level Re-187 exemption level
1.E+04 Re-187 1.E+04 1.E+04
1.E+02 1150 (Bq/g) 1.E+02 658 (Bq/g)

1.E+02
938 (Bq/g)
1.E+00 1.E+00 1.E+00

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Irradiation and cooling time (year) Irradiation and cooling time (year) Irradiation and cooling time (year)
ZrH1.7 moderator (Case 1) ZrD1.7 moderator (Case 2) No moderator (Case 3)
(a) (b) (c)
Fig. 9 Specific activity of Re vs. types of moderator.
6. Conclusions of Fission Products. Nuclear Science and Techniques 26

(1): S010311.
The feasibility of rhenium (Re) production by [4] Yokoyama, T., Terashima, A., and Ozawa, M. 2014.
Evaluation of Rhenium Production Rates in Tungsten
irradiating tungsten (W) metal in a medium size fast
Irradiated in Fast Reactors by Using Continuous Energy
reactor was evaluated by using the Monte Carlo Monte Carlo Code MVP. In Proceedings of the 8th
calculation code MVP and ORIGEN2. The prototype International Symposium on Technetium and Rhenium:
fast reactor can produce about 50 kilograms of Re per Science and Utilization (ISTR 2014), 526-32.
[5] IAEA (International Atomic Energy Agency). 2014.
every 3 years, which corresponds to about 10% of
Radiation Protection and Safety of Radiation Sources:
Japanese domestic production. The specific activity of International Basic Safety Standards. IAEA Safety
Re can be reduced below the exemption level or even Standard Series No. GSR Part 3.
the natural Re level if W and Os is separated after the [6] Nagaya, Y., Okumura, K., Mori, T., and Nakagawa, M.
2005. MVP/GMVP II: General Purpose Monte Carlo
irradiation. The use of ZrD1.7 moderator reduces the
Codes for Neutron and Photon Transport Calculations
specific activity by half compared to that of ZrH1.7 Based on Continuous Energy and Multigroup Methods.
case, and even the no moderator case is permissible to JAERI (Japan Atomic Energy Research Institute), JAERI
produce the production of Re which has lower specific 1348.
[7] Okumura, K., Mori, T., Nakagawa, M., and Kaneko, K.
reactivity than that of natural Re.
2000. Validation of a Continuous-Energy Monte Carlo
Burn-Up Code MVP-BURN and Its Application to
References
Analysis of Post Irradiation Experiment. Journal of
[1] Han, C. Y., Ozawa, M., and Saito, M. 2012. Resource Nuclear Science and Technology 37 (2): 128-38.
Ability on Nuclear Fuel Cycle by Transmutation [8] Shibata, K., Iwamoto, O., Nakagawa, T., Iwamoto, N.,
Approach. Science China Chemistry 55 (9): 1746-1751. and Ichihara, A. 2011. JENDL-4.0: A New Library for
[2] Ozawa, M., Han, C. Y., Kaneshiki, T., and Nilsson, M. Nuclear Science and Engineering. Journal of Nuclear
2013. Aprs ORIENT, a P & T-Based New Resource Science and Technology 48 (1): 1-30.
Strategy in Nuclear Fuel Cycle. In Proceedings of the [9] Okumura, K., Sugino, K., Kojima, K., Jin, T., Okamoto,
International Nuclear Fuel Cycle Conference (GLOBAL T., and Katakura, J. 2013. A Set of ORIGEN2 Cross
2013), 1499-503. Section Libraries Based on JENDL-4.0: ORLIBJ40.
[3] Terashima, A., and Ozawa, M. 2015. Numerical JAEA (Japan Atomic Energy Agency), JAEA-Data/Code
Analysis on Element Creation by Nuclear Transmutation 2012-32.
doi: 10.17265/1934-8975/2016.03.004
D DAVID PUBLISHING
A New Generator for Tsunami Wave Generation
Tetsuya Hiraishi1, Ryokei Azuma1, Nobuhito Mori2, Toshihiro Yasuda2 and Hajime Mase2
1. Disaster Prevention Research Institute, Kyoto University, Kyoto 612-8235, Japan
2. Disaster Prevention Researtch Institute, Kyoto University, Uji 611-0011, Japan
Abstract: A middle size experiental wave generator has been implemented is the Ujikawa Open Laboratory, Disaster Prevention
Research Institute, Kyoto University. The generator is composed of a pistontype wave maker, a head strage water tank and a current
generator to mainly reproduce long waves like tsunami and storm surge. The paper desribes several experimental series to predict the
applicability of the generator to model tests. The three operating sysemes are capable to be controlled in one operating sysytem and
start time is contorolled separately according with the target tsunami and storm surge profiles. A sharp tsunami profile is reproduced
in adjusting the start timing of piston type wave maker and opening gates of head storage tunk. Any type of tsunami waves are
reproduced in the generator and it becomes a storong tool to predict the effective of resiliency of hardwares.
Key words: Tsunami, storm surge, experimental generator tool, wave maker.
1. Introduction like friction factors. The verification of numerical

modelling can be carried out by using the
The 2011 Tohoku Pacific off great earthquake
experimental data.
tsunami caused the worst hazard in the Northern
A middle size experimental wave flume has been
Japanese coastline. The various numerical simulation
implemented in the Ujikawa Laboratory, Disaster
and field survey have revealed the tsunami height
Prevention Research Institute, Kyoto University. Fig. 1
distribution at each devastated areas. However, the
shows a bird-eye view of the flume. In the flume, the
detail mechanism of several building collapses and
following three generators are equipped to reproduce
turning down of coastal tidal walls has not yet clearly
the tsunami and storm surge waves in shallow water
discussed mainly because of the unknown
and after run-up:
transformation of tsunami in shallow waters and
(1) flume size: 45 m 4 m 0.8 m (side wall
run-up grounds.
height is 2 m) with water tank (1 m deep);
In 2013, typhoon No.13 also caused the large storm
(2) piston type wave generator with 2.5 m stroke;
surge mainly in the middle parts of Philippines. The
(3) current generator (the maximum 0.8 m3/s);
run-up storm surge waves are transformed to bore
(4) water storage tank over wave generator (4 m3).
and give impulsive force to the coastal structures.
We conducted several fundamental tests using the
However the detail profile of the impulsive force is
new flume and generated an impulsive wave with a
not revealed. The physical study using experimental
sharp peak. The newly developed wave flume in our
instruments often become the useful tools to study on
laboratory will be demonstrated in the final paper.
the tsunami run-up phenomena, impulsive force due to
bore and surging wave force induced in storm surges. 2. Experimental Method
Meanwhile, the experimental results are necessary to
The eight wave gages are implemented to measure
determine the several coefficients in numerical models
the variation of wave profile. The position of wave
Corresponding author: Tetsuya Hiraishi, professor, gages is expressed in Fig. 2. Table 1 shows the
research field: coastal sedimentation.
A New Generator for Tsunami Wave Generation 167
1/20 slope
1/10 slope
Head storage tank
Wave maker
Current generator
Water tank
Fig. 1 Bird view of new experimental basin.
WG8
WG3
WG2
WG1
Fig. 2 Position of wave gage.
Table 1 Experimental case for piston wave maker.

Exp. no. Wave height (cm) Water depth (cm)
Trial 1 10 80
Trial 2 20 80
Trial 3 30 80
Trial 4 40 80
Trial 5 50 80
Photograph 1 Generation tsunami (soliton) wave.
experimental case using the wave maker. As shown in period high water level is reproduced in the current
Table 1, the water depth is constant and 80 cm. Target generator. The head storage tank is also employed to
wave height varies from 10 cm to 55 cm. In each case, reproduce the bore wave in the basin.
two trial was carried out to check the uniformity of At the first, we studied the efficiency of piston type
experimental results. Photograph 1 shows the wave generator. Table 1 shows the experimental case for
generation of one peaky wave generated in the piston piston type wave maker test. The measurement of water
type wave generator. The wave paddle is removed to level was carried out at three times for each case.
the backside slowly as not to generate waves in the
wave basin. A large wave is generated in removing the
paddle to the front direction in a short time.
Fig. 3 shows a tsunami wave profile obtained at the
offshore observation point [1] 20 km far from the
coast line. High peaky wave is observed at the initial
Water level (m)
time. This profile is obtained in the offshore side, so

the peak may become more representative in shallow
water areas. The low peak with long wave succeeds.
The time duration of long period tsunami wave
profiles are very long. Therefore several operation
modes are inevitable to reproduce such tsunami waves.
High peak wave is represented in soliton wave in
0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000
the experimental basin. In our basin, the piston type
Time after shock (s)
wave maker represents to the generator of soliton
wave at initial stage of tsunami. The succeeding long Fig. 3 Tsunami profile obtained in the offshore.
The soliton wave is generated in such operation

mode. The soliton wave profile is expressed as
follows:
3
asech sech 1 sech
4
(1)
3 5
1
4 8
Photograph 2 Water fall from the head storage tank.

1 3
1 wave height in the different volumes stored in the
2 20
where, awave height, hwater depth, ttime. head storage tank. The maximum volume of head
Table 2 shows the experimental case for tsunami storage tank is 4 m3. In the experiment, the volume
wave only generated by current generator. The system was varied with constant height of opening and
is controlled by imputing the water volume per minute. opening speed. Photograph 2 shows the water fall in
The duration is the time of pump working. In the front of the storage tank.
experiment, we used the two time duration (120 s and
3. Experimental Results
240 s). However the water level becomes stable in 120 s,
so the cases with 120 s duration are introduced in the At the first, profile of initial tsunami wave
paper. generated by the piston maker is expressed. Fig. 4
Table 3 shows the case to compare the reproduced shows the theoretical soliton and practically generated
wave profile. The comparison of the both profiles
Table 2 Experimental case for current generator.
shows the experimental wave profile agrees well to
Exp. no. Constant flow (m3/min) Time duration (s)
the theoretical ones. Therefore the peaky sharp
Trial 6 0.1 120
Trial 7 0.2 120 tsunami wave is capable to be reproduced a soliton
Trial 8 0.3 120 wave in a piston wave maker.
Trial 9 0.35 120 Fig. 5 shows the variation of tsunami height
Trial 10 0.1 240 measured at No.1 gage for input tsunami height. For
Trial 11 0.2 240
each case, three measurements were done and the
Trial 12 0.3 240
Trial 13 0.35 240 averaged related line was indicated. Target tsunami
height is almost reproduced in inputting the same
Table 3 Experimental case for head storage water tank.
height to operation system.
Tank height Tank volume Open height
Exp. no. Fig. 6 shows the wave height distribution along the
(cm) (m3) (cm)
Trial 14 50 0.5 50 longitude line of basin. In the flat part, in the basin
Trial 15 50 1.0 50 tsunami height (Hmax) is almost uniform and similar.
Trial 16 50 1.5 50
The height becomes small at the most far point on the
Trial 17 50 2.0 50
Trial 18 50 2.5 50 slope because the wave is broken on the shallow
Trial 19 50 3.0 50 slope.
Trial 20 50 3.5 50 Fig. 7 shows the variation the tsunami wave height
Trial 21 50 4.0 50
(No.1) related for the input volume data to current
Theoretical
Wave height (cm) value
Experiment
TRY1
Experiment
TRY2
Time (s)
Fig. 4 Comparison of theoretical and experimental soliton wave profile.
Output wave height (cm)
Input wave height (cm)
Fig. 5 Relation between the input and measured soliton wave height.
Hmax
Hmax
Position of wave gages
Fig. 6 Tsunami height distribution along the longitude direction in basin.
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40

3
Input water volume (m )
Fig. 7 Relation of generated tsunami height and storage water volume.
generator. The tsunami height linearly as the input to the heights. The initial volume of the tank was main
water volume increases. The maximum tsunami height parameter to determine the generated wave height.
becomes about 30 cm. Fig. 7 demonstrates the current The characteristic of the newly developed basin is
generator is effective to reproduce tsunami wave as capability to operate the three generators (piston
well the piston wave maker. maker, current generator, head tank). Fig. 9a shows
Fig. 8 shows the tsunami height variation when the the one example of generated tsunami profile using
head storage tank is opened. The water volume stored the three operation modes. The initial two peaks the
in the tank is shows in Table 3. The maximum water sharp tsunami profiles generated the piston maker and
volume inside the tank is 4 m3. Fig. 8 demonstrates head tank. The operation timing is also controlled and
the tsunami wave height increase as the input water the peaks of tsunami are capable to be occurred at the
volume increases. The maximum wave height is about same time. Fig. 9b shows the tsunami profile when the
20 cm when the maximum water volume is stored. tsunami peak generated wave maker and that
The relation is approximately linear and the generated generated in head tank appear at the same time. The
tsunami height is controlled in the input water tank peak height becomes much higher than that induced
volume. by one operation tool. The composing of the three
The height of the tank and opening gate height are operation modes is applicable to any tsunami profile
controlled, but the both parameters had no influences observed and simulated in the target area.
Input volume tank (m3)
Fig. 8 Relation of tsunami height and water volume in head tank.

50
40
Head tank
Water level (cm)
30
Piston type
wave maker
20 Current
generator
10
0
0 50
-10
Time (s)
Time (s)
(a) Separate generation (b) Timing adjustment operation

Fig. 9 Applicability of generation for any type of tsunami profile.
4. Conclusions storage tank determine the reproduced tsunami height;

(4) The higher peak tsunami wave profile can be
The following conclusions are derived in the paper:
reproduced by adjusting the operation timing.
(1) A new tsunami and storm surge reproduction
basin has been developed with the three operation Reference
mode;
[1] Takahashi, S. 2011. Urgent Survey for 2011 Great East
(2) Soliton wave expressing the peaky tsunami
Japan Earthquake and Tsunami Disaster in Ports and
profile is reproduced in the piston wave maker; Coasts. Technical Note of the Port and Airport Research
(3) Current volume and water volume in the head Institute.
doi: 10.17265/1934-8975/2016.03.005
D DAVID PUBLISHING
Solid Particle Emission Abatement from Biomass-Fired

Boilers at District Heating Plants
Egils Dzelzitis1, Aldis Jegeris2, Agnese Lickrastina1 and Normunds Talcis2

1. Heat, Gas and Water Technology Institute, Riga Technical University, Riga, LV-1048, Latvia
2. District Heating Company Rigas Siltums, Riga, LV-1012, Latvia
Received: August 04, 2015 / Accepted: February 15, 2016 / Published: March 31, 2016.
Abstract: The district heating company Rigas siltums operates biomass fuelled boiler in Riga city. Three systems consisting of biomass
boilers having a comparatively similar heat capacity and particle abatement units like multicyclons, electrostatic precipitators and
flue gas condensers are compared. The main goal of the study is to evaluate the boiler plant as a system where solid particles are both
emitted and caught. The results show that, the particulate matter can be efficiently trapped from flue gases by the particle abatement
technologies, and the electrostatic precipitator with sufficiently large collection surfaces is able to provide appropriate flue gas
treatment of the particulate matter in the biomass boilers also without pre-cleaning of the flue gas in multicyclons.
Key words: Biomass, district heating plant, particulate formation, particulate abatement.
1. Introduction Three different types of wood chips boilers located

at district heating plants are analyzed in this article.
The strong environmental requirements and the EU
The boiler plants have comparatively similar loads
decision to decrease the share of the imported fuels in
and are connected to the same district heating network.
the Europes energy balance promote the utilization of
The boilers have different combustion conditions and
the local energy resources in combination with the
utilize different particle abatement technologies.
best practice of the environmental protection
There are usually three different types of biomass
technologies [1].
ashes considered in the plantsbottom ash, coarse fly
Wood and wood residues compose the greatest share
ash and fine fly ash. Bottom ash constitutes the major
among local renewable fuels in Latvia thus becoming
fraction and is characterized by the presence of sand,
a very attractive and sustainable alternative also in
stones and soil. These fractions decrease the melting
large district heating systems. The main disadvantage
point thus causing slag formation. Cyclone ash is finer
of the biomass combustion is ash related problems.
and consists of inorganic ash particles. This ash
For medium to large scale applications, the particulate
fraction is mainly composed of coarse fly ash. The
abatement at the back of the combustion chamber is
filters, electrostatic precipitators and gas condensation
needed for the further decrease of the particulate
units collect the finest fly ash particles which
matter amount [2]. In order to reach the particle
normally consist of aerosols [3].
emission requirements, the heat production unit
According to the Latvian regulations, the
should be considered as a system where all parts
permissible emission limit for particulate matter
are influencing each other and each of them is
emissions from the boiler plants with the installed heat
important.
capacity from 10 MW up to 50 MW fuelled by solid
Corresponding author: Agnese Lickrastina, Dr.sc.ing, fuels or wood chips must be below 500 mg/m3 with
docent, research fields: district heating, environmental issues
the 6% oxygen content in the flue gases [4].
and biomass combustion.
The chemical properties of biomass vary that, hence, Table 1 Mass percentage of ash from biomass combustion.
causes variations in chemical composition of biomass Particle size (m) Fuel ash Bark Chipping
> 50 Coarse 65-85 60-90
ashes. The ash concentration for wood chips varies
< 10 Cyclone ash 10-25 50-70
depending on the wood type and on the presence of <1 Fine ash 2-10 10-20
bark and inorganic materials like sand and stones. It is
attention in biomass combustion is paid to the
known that, wood bark has the highest ash content
formation of PM10 characterizing particles with the
directly influencing the ash concentration of biomass
aerodynamic diameter less than 10 m. These would
fuel. The major part of wood ash is typically
include both coarse and fine particles. Attention
composed of calcium (Ca) and potassium (K), part of
should be also paid to PM2.5, i.e., to particles with the
which volatilizes during the combustion and forms
diameter less than 2.5 m that are removed from the
fine particles [5]. According to the data in Ref. [6], the
combustion zone by flue gases.
mass concentration of fly ash particles behind the
There are several particle abatement technologies
boiler is in the range of 60-3,000 mg/nm3.
most commonly used for flue gases cleaning, such as
Wood bark typically has also a comparatively high
cyclones, electrostatic precipitators and wet scrubbers
concentration of heavy metals (Zn and Cd) which are
which are often used as flue gas condensers [9].
mostly recovered in fly ash as they vaporize during
Relatively large particles can be easily trapped by
the combustion. They often are found on the surface
means of multicyclones. Multicyclones are often used
of fine fly ash particles or in aerosols [3]. It is found in
for pre-cleaning of the flue gas in front of the flue gas
Ref. [7] that, the fine ash fraction from a woody
condensation unit [5]. Multicyclones are known as
biomass is typically formed by potassium sulphates,
simple-design units with low to moderate gas side
chlorides and carbonates.
pressure drop, inexpensive and working in a wide
The data from Ref. [8] give a general overview
range of temperatures. The main disadvantages of
about the distribution among biomass ashes (Table 1). multicyclones are the low efficiency of small particle
It shows that, more than 60% of the total amount of collection and the lower total operation efficiency
ash is composed of coarse ash with the particle under part-load conditions [8, 10]. The typical
diameter more than 50 m, whereas fine ash efficiency of cyclones is 55%-99% for particles from
constitutes 2%-20%. There are large variations for the 1.0 m to 100 m, whereas the highest efficiency
same type of biomass fuel characterizing the huge (90%) is reached for particles larger than 10 m.
diversity of the biomass constituents. The ESP (electrostatic precipitator) uses
During the complete combustion, all inorganic electrostatic charges to remove the particulates from
particles originated from the minerals present in the the flue gas flow. The electrostatic precipitator is very
fuel are divided into two large groups. The large efficient for very small particles. The disadvantage is
particles form bottom ash and are carried away from its sensitivity to the varying flow rates and amount of
the combustion zone by the ash conveyor. The smaller particles, as well as the decreased efficiency during its
particles called fly ash are taken by flue gases. The operation [9]. Larger collective surface areas and
latter need particular attention in the operation of the lower flue gas flows through the ESP lead to a higher
biomass combustion units. particle removal efficiency as dust particles have more
Fly ash is defined as PM (particulate matter) and is time and surface to collect. The efficiency of large
characterized by the size of its particles. The fly ash ESP units can reach more than 99% [2]. Due to that,
from the combustion of wood is set for the total today, the ESP is one of the most widely used
amount of solid particles, for PM10 and PM2.5. Particular technologies for modern heat and power generation
Solid Particle Emission Abatement from Biomass-Fired Boilers at District Heating Plants 175
plants [11]. licensed companies and only under the actual

Flue gas condensers are often designed as wet operating conditions. The obtained results are
scrubbers having double effect on the operation of the evaluated in this study. Particular attention was paid to
biomass boiler plant. They act as a particulate the efficiency of the particulate abatement units
abatement technology ensuring the capturing of fine installed in the boiler plants. The main goal of the
fly ash particles [12] and recovering the latent heat study is to evaluate the boiler plant as a system where
from the flue gas flow. During the operation, the solid particles are both emitted and caught. Three
process water is injected into the flue gas condenser. systems including biomass boilers and particle
The surface of small droplets absorbs the fine fly ash abatement units like multicyclons, electrostatic
particles and even the gases from the flue gas flow precipitators and flue gas condensers are compared in
simultaneously absorbing also the heat from the gas this study.
flow. The total surface area of all the droplets is large
2. Evaluation of the Efficiency of Particulate
enough to ensure the high efficiency of the absorption
Abatement Technologies at the District
process. The higher is the moisture content in the fuel,
Heating Plants in Riga
the larger amount of heat is carried by the flue gases [8].
Recovering the latent heat significantly increases the The district heating company Rigas siltums has
boiler efficiency up to 30%. The wood biomass gradually enlarged the existing boiler plants with the
moisture could reach even 55%-60% of the moisture wood chips boilers. Thus the biomass has become the
content thus making the flue gas condensers an main heat source providing particular care of the flue
attractive option for the flue gas treatment in the gas treatment issues. From the environmental
biomass boiler plants [13]. protection point of view, the main difference between
The measurements at three biomass fuelled boiler the formation of gaseous emissions from the natural
plants were made only during the full load gas combustion and the biomass combustion is
operation in order to avoid the influence of the determined by the formation of fly ash in the flue gas.
part-load conditions and incomplete burning of the The special regulations by the Latvian environmental
biomass. This ensures that, the particle formation protection authority set forth the maximum
mainly originates from the minerals initially present in permissible calculated limits for solid particle
the biomass. The boiler plants have different emissions for each boiler [14-16].
combustion technologies. It is assumed that, the The particle formation should be initially
combustion technology has an effect on the amount of considered from the combustion process point of view
solid particles originating during the biomass as the ash formation is directly influenced by fuel
combustion. properties, combustion conditions, appliances, excess
As all boiler plants are fuelled with wood chips air and boiler load [1].
delivered by the same biomass fuel production In order to avoid the influence of the incomplete
company, significant variations due to the variations combustion on the ash formation, the full load
in wood chips fuel composition are not expected. operation with the optimal air supply is ensured
Measurements are made behind the boiler and behind (Table 2). The variations of the fuel composition
each of the flue gas cleaning units, i.e., behind the could be considered insignificant as one and the same
multicyclone, behind the electrostatic precipitator and biomass fuel production company simultaneously
behind the flue gas condensing unit. Measurements supplies fuel to the three investigated boiler plants.
were made during the last heating season by the With the complete and efficient combustion provided,
Table 2 Fixed parameters of the study. chips on dry and ash free basis is the following:
Fuel Wood chips delivered by the same company carbon around 50%, hydrogen around 6% and oxygen
Boiler load Full load about 44%.
Optimal conditions according to the
Combustion The combustion technology directly influences the
technology
biomass gasification and char combustion process.
Table 3 Variable parameters of the study.
There are two main technologies for the biomass
District District District
heating plant heating plant heating plant combustion: fixed bed or fluidized bed. During the
Vecmilgravis Zasulauks Ziepniekkalns fixed bed combustion, the primary combustion air is
Installed
heating 14 MW 20 MW 22 MW supplied to the fuel from the bottom. It initiates drying
capacity and gasification of the biomass with the subsequent
Bubbling Rotating
Combustion Moving grate combustion of the char. To ensure complete and
fluidized-bed water-cooled
technology pre-furnace
furnace grates reliable fuel combustion, the air is typically supplied
Multicyclone
Installed Electrostatic also as secondary and tertiary air into two or three
Multicyclones Electrostatic
particulate precipitator
Flue gas precipitator locations above the fuel bed. The fixed bed
abatement Flue gas
condenser Flue gas
technologies condenser combustion is provided at the district heating plants
condenser
Vecmilgravis and Ziepniekkalns. At the DHP
Table 4 Wood chips characteristics.
(district heating plant) Vecmilgravis, the
Moisture content 35%-40%
combustion is ensured by the moving grate
Volatiles 60%-80%
Ash content 2% pre-furnace, whereas the pre-furnace at the DHP
Ziepniekkalns is designed with rotating
the changeable parts are the different combustion water-cooled grates. The fixed bed systems are
technologies and the particulate abatement units designed for biomass with the variable moisture
installed behind the boiler (Table 3). content and variable fuel particle size [17].
The particulate amount in the flue gases is first A bubbling fluidized bed furnace is installed at the
measured directly behind the boiler. Then the DHP Zasulauks. In this furnace, the biomass is
measurements are made repeatedly behind each combusted as a suspension of gas, sand, fuel and ashes.
particulate abatement unit, i.e., behind the Also, in this case, the primary air is supplied through
multicyclones, the electrostatic precipitators and the bottom of the bed. The fluidized bed systems are
behind the flue gas condensers. The last measurement flexible with regard to the fuel material, moisture
behind the flue gas condensers shows the amount of content and at the same time they are sensitive to the
particulate matter blown into the atmosphere. fuel particle size [17, 18].
The efficiency of the flue gas condensers is
influenced by the flue gas water content originated 2.1 District Heating Plant Vecmilgravis
from the wood biomass humidity. The moisture In 2010, wood-chip fired water heating boilers were
content of wood chips could vary from 25% to 60%. installed at the district heating boiler plant
During the study, it was in the range from 35% to Vecmilgravis. In 2013, they were equipped with a
40% (Table 4). The content of volatiles generally is flue gas condenser in order to increase the capacity of
high for wood chips and could reach 60%-80% the boiler plant and decrease the particulate matter
depending of the type of wood species composing the emitted through the stack.
fuel. The ash content varies slightly around 2%. Now there are two hot water boilers AK-7000 with
For the industrial application cases discussed in this the total installed heating capacity 14 MW (Fig. 1).
article, the most typical composition of the wood Both boilers are equipped with the moving grate
Fig. 1 Flue gas circulation scheme at the DHP Vecmilgravis.
pre-furnace and hot water boilers with the flue gas condenser, where about 7% of the particles
multicyclones. The multicyclones are of MC-7000 initially originated during the combustion are trapped.
type and they are produced by the local company Due to that, the total amount of particulate matter
Komforts. The total size of each unit is 2,300 emitted into the atmosphere consists of 68 mg/nm3, or
2,845 4,550 mm with the DN590 inlet and outlet 22% of the particles coming out from the boilers (Fig. 2).
ports. The capacity of the multicyclon is 7.4 m3/s. The The maximum permitted amount of particulate
flue gas condenser is located behind the multicyclone emissions leaving the boiler plant could reach 77 mg/nm3
for the further decrease of the particulate emissions. for PM10 and 42 mg/nm3 for PM2.5 [14].
The measurements were made behind the boiler in
order to evaluate the amount of particulates originated 100.0%
during the combustion. The next measuring point was 90.0% 22.0%
80.0%
located behind the multicyclone and the third one 7.2%
70.0%
behind the flue gas condenser. The flow of dry gas 60.0%
through the particulate abatement units reaches 50.0%
2.05-3.72 Nm/s. 40.0%
70.8%
30.0%
According to the measurements data, the total
20.0%
averaged particulate emissions are 309 mg/m3 directly
10.0%
behind the boilers. The multicyclone entraps more 0.0%
than 70% of the particles generated during the
Trapped in multicyclons Trapped in flue gas condenser
combustion in the pre-furnace (Fig. 2). The amount of
Blown into the atmosphere
particles behind the multicyclone reduces from
Fig. 2 Trapping distribution of the particulate matter
309 mg/nm3 down to 90 mg/nm3. originated in the biomass boilers at the DHP
The particulate matter is further decreased in the Vecmilgravis in the flue gas cleaning devices.
The comparison between the actual emissions and of the boiler show that, the amount of particles
the maximum allowable emissions shows that, the originated during the combustion in the fluidized bed
boiler emits 88% of the maximum allowable PM10 reaches 297 mg/nm3. After leaving the boiler, the flue
particulate matter. Assuming that, only coarse fly ash gases enter the electrostatic precipitator, where the
is trapped in the multicyclone, the proportion of fine particulate matter amount was decreased by 85%,
fly ash from the grate fired boiler is about 30%. from 297 mg/nm3 to 42 mg/nm3 (Fig. 4).
After leaving the electrostatic precipitator, the
2.2 District Heating Plant Zasulauks
flue gases enter the flue gas condenser primarily
A water heating boiler with the nominal capacity designed to improve the boiler plant efficiency. Only
20 MW equipped with a bubbling fluidized-bed a small amount of about 3.3% of the initially
furnace was installed in 2013 at the district heating originated amount of particles is trapped there. The
plant Zasulauks. final amount of particles leaving the boiler plant is
The water heating boiler is equipped with an 32.4 mg/nm3, or 10.9% (Fig. 4). The dry gas flow
electrostatic precipitator and a flue gas condenser during the measurements varied from 6.3 Nm3/s to
(Fig. 3). The electrostatic precipitator has the 9.81 Nm3/s.
dimensions 16,064 5,695 18,669 mm and the The maximum limit for the solid particle emission
dimensions 1,300 1,300 mm of the inlet and outlet from the district heating plants Zasulauks has been
ports. The gas flow through the unit reaches 16 Nm3/s. calculated as 84 mg/m3. The maximum permissible
The measurements to evaluate the particle amount for the particulate emissions PM10 is set to
abatement from this boiler were made in three 29 mg/nm 3 , whereas the maximum for PM 2.5 is
placesdirectly behind the boiler, at the exit of the 18 mg/nm 3 [15].
electrostatic precipitator and between the flue gas In fact, the total amount of particulates leaving the
condenser and the stack. boiler plant is less than 40% of the maximum
The measurements made behind the convective part permissible particulate matter emissions.
Fig. 3 Flue gas circulation scheme at the DHP Zasulauks.

100% plant efficiency by preheating the water and air

10.9%
90% 3.3% supplied for combustion.
80%
After passing the economizers, the flue gases enter
70%
the gas cleaning system involving several steps.
60%
50% Firstly, the gases pass through the multicyclone and
85.8%
40% then through the electrostatic precipitator. The flue
30%
gas condenser is installed between the electrostatic
20%
precipitator and the stack (Fig. 5). The multicyclone
10%
0% wellon has been added as a component at the boiler
supply. The dimensions of the unit are 2,090 2,446
Trapped in ESP Trapped in flue gas condenser
4,040 mm. The capacity of the multicyclone
Blown into the atmosphere
reaches 13.61 m3/s for the flue gases with the
Fig. 4 Trapping distribution of the particulate matter temperature below 320 C, or 6.73 Nm3/s. The gas
originated in the biomass boilers at the DHP Zasulauks
in the flue gas cleaning devices. flow velocity through the multicyclone ranges
4.2-4.5 m/s. The dimensions of the electrostatic
2.3 District Heating Plant Ziepniekkalns
precipitator are 10,680 6,700 14,815 mm, with the
The biomass cogeneration steam boiler plant with cross-section area of 3,500 6,000 mm for the inlet
the installed heat capacity up to 22 MW and electrical and outlet ports. The consumption of the flue gases is
capacity 4 MW is in operation since 2013. The 25.28 m3/s at the gas temperature 163 C,
pre-furnace for biomass combustion is designed with correspondingly, 16.99 Nm3/s.
rotating water-cooled grates. Tangentially supplied The measurements were made at the exit of the
secondary and tertiary air provides a stable and boiler, behind the multicyclone, behind the
complete combustion process. The boiler is equipped electrostatic precipitator and between the flue gas
with two economizers to ensure the higher total boiler condenser and the stack.
Fig. 5 Flue gas circulation scheme at DHP Ziepniekkalns.

100%
12.3%
3. Conclusions
90%
8.9%
80% Three systems with different wood chips
70% 16.0%
combustion technologies and slightly different
60%
particulate abatement units have been evaluated in this
50%
40% study. Considering the biomass plants as a system
30% 62.8% consisting of separate units allows selecting an
20% optimal combination under the given circumstances.
10%
During the investigation, the influence of variations
0%
of the biomass and combustion conditions has been
Trapped in multicyclons Trapped in ESP
eliminated. The same biomass producer supplied
Trapped in flue gas condenser Blown into the atmosphere
wood chips simultaneously to all three plants. The
Fig. 6 Trapping distribution of the particulate matter optimal combustion conditions according to the
originated in the biomass boilers at the DHP combustion technology have been provided in all
Ziepniekkalns in the flue gas cleaning devices.
biomass pre-furnaces. The amounts of fly ash particles
The total amount of particles originated from the from the different types of biomass boilers operated at
fuel in the combustion zone was 293 mg/m3. The the district heating company Rigas siltums were
amount of particulate matter was gradually decreasing measured under full load and stable operation during
when passing through the flue gas treatment devices. the heating season. The measurements were made
After leaving the boiler, the coarse dust particles were behind the boiler and behind each of the flue gas
separated from the flue gas in the multicyclone. About treatment unit.
184 mg/nm3, or 63% of the particulate matter The measurement data show the comparatively
originated during the combustion were trapped there. similar solid particle amount originated during the
Then flue gases entered the electrostatic precipitator, combustion independently on the furnace design. It
where the amount of particulate matter was decreased allows us to conclude that, the combustion technology
to 62 mg/nm3. Upon cleaning, the flue gases were is not a determining factor for the solid particle
condensed in the flue gas condenser, where about formation if the optimal combustion conditions are
26 mg/nm3 of the solid particles were trapped. The provided in the furnace and if there are no significant
dry gas flow during the measurements varied from variations of the wood biomass composition (Table 5).
7.3-7.6 Nm3/s. The particulate matter can be efficiently trapped
As a result, the total amount of particulate matter from the flue gases by the particle abatement
emitted into the atmosphere is 36 mg/nm3, or only technologies. The flue gas condensers are an optimal
12% of the emissions initially originated from the solution for the biomass boiler plants which ensures
biomass combustion (Fig. 6). For the biomass boiler the recovering of latent heat and the increase of the
located at the district heating plant Ziepniekkalns, efficiency with the simultaneous decrease of
the limit for the total amount of particulates is set to particulates in the flue gas.
83 mg/nm3. The limits for PM10 and PM2.5 are set to The comparison between multicyclons and
58 mg/nm3 and 42 mg/nm3, respectively [13]. The electrostatic precipitators has revealed that, the latter
particles emitted into the atmosphere constitute only could be a good solution for the biomass plants. The
43% of the limits set by the environmental protection results show that, the electrostatic precipitator with
authority. sufficiently large collection surfaces is also able to
Table 5 Particulate emissions from the district heating plants (mg/m3).

DHP Vecmilgravis DHP Zasulauks DHP Ziepniekkalns
Behind the boiler 309 297 293
Behind the multicyclons 90 - 184
Behind the ESP - 42 62
Behind the flue gas condensers/blown into the atmosphere 68 32.4 36
Percentage captured by particulate abatement units 78% 89% 88%
provide an appropriate flue gas treatment of Emissions from Small-Scale Biomass Combustion.
Thesis for a degree of licentiate of engineering, Chalmers
particulate matter in the biomass boilers avoiding the
Univeristy of Technology.
flue gas pre-cleaning in the multicyclons. [7] Sippula, O., Hytonen, K., Tissari, J., Raunemaa, T., and
The actual emissions during the measurements Jokiniemi, J. 2007. Effect of Wood Fuel on the
constitute about 40% of the limits set forth by the Emissions from a Top-Feed Stove. Energy & Fuels 21
(2): 1151-60.
environmental protection authority and about 9% of
[8] Ebert, F. 2001. Particle Separation for Biomass
the limits stated by the laws of the Republic of Latvia [4]. Combustion, Aerosols from Biomass Combustion.
This allows to conclude that, the selection of an Presented at the International Seminar Organised on
appropriate particle abatement technology for the Behalf of IEA (International Energy Agency)
Bioenergy Task 32: Biomass Combustion and Cofiring,
biomass boilers cannot be underestimated, and in
and the Swiss Federal Office of Energy, Zurich,
order to decrease the particulate matter emitted into Switzerland.
the atmosphere from the industrial-scale boiler, the [9] AEA Technology Plc. 2010. The Assessment of Flue Gas
proper flue gas treatment of the particles should be Particulate Abatement in Wood Burning Boilers. Report
for Forestry Commission Scotland.
considered and appropriate equipment should be
[10] Brunner, T., Obernberger, I., Jller, M., Arich, A., and
installed. Plt, P. 2001. Behavoiur of Ash Forming Compounds in
Biomass Furnaces-Measurement and Analyses of
References Aerosols Formed during Fixed-Bed Biomass
[1] Obaidullah, M., Bram, S., Verma, V. K., and Ruyck, J. D. Combustion. Presented at the International Seminar
2012. A Review on Particle Emissions from Small Scale Organised on Behalf of IEA Bioenergy Task 32: Biomass
Biomass Combustion. International Journal of Combustion and Cofiring, and the Swiss Federal Office
Renewable Energy Research 2 (1): 147-59. of Energy, Zurich, Switzerland.
[2] Nussbaumer, T. T. 2010. Overview on Technologies for [11] Mikael, O., Jorma, J., Jouni, H., Pasi, M., and Jarkko, T.
Biomass Combustion and Emission Levels of Particulate 2006. Views and Conclusions from the FINE
Matter. Zrich: FOEN (Swiss Federal Office for the ParticlesTechnology, Environment and Health
Environment), 1-79. Technology Programme. In Combating Particulate
[3] Biedermann, F., and Obernberger, I. 2005. Ash-Related Emissions in Energy Generation and Industry, edited by
Problems during Biomass Combustion and Possibilities Peter, H. Finland: Tekes.
for a Sustainable Ash Utilisation. Presented at the [12] Strand, M., Pagels, J., Szpila, A., Gudmundsson, A.,
International Conference World Renewable Energy Swietlicki, E., Bohgrd, M., and Sanati, A. 2002. Fly Ash
Congress (WREC), Aberdeen, Scotland. Penetration through Electrostatic Precipitator and Flue
[4] Regulations of the Cabinet of Ministers Nr. 187. 2013. Gas Condenser in a 6 MW Biomass Fired Boiler.
The Procedure to Prevent, Restrict and Control the Energy & Fuels 16 (6): 1499-506.
Emission of the Air Polluting Substances from [13] Neuenschwander, P., Good, J., and Nussbaumer, T. 1998.
Combustion Units. Cabinet of Ministers Nr. 187. Combustion Efficiency in Biomass Furnaces with Flue
[5] Sippula, O., Hokkinen, J., Puustinen, H., Yli-Pirila, P., Gas Condensation. Presented at the Biomass for Energy
and Jokiniemi, J. 2009. Particle Emissions from Small and Industry, 10th European Conference and Technology
Wood-Fired Distrcit Heating Units. Energy & Fuels 23 Exhibition, Wrzburg, Germany. Accessed August 17,
(6): 2974-82. 1998. http://www.etsi.org.
[6] Johansson, L. 2002. Characterisation of Particle [14] Environmental Office of Riga Region, State
Environmental Office, Ministry of Environmental [16] Environmental Office of Riga Region, State
Protection and Regional Development of the Republic of Environmental Office, Ministry of Environmental
Latvia. 2010. Permission of Category A Polluting Protection and Regional Development of the Republic of
Activities Nr. RI 10 IA 0005. Environmental Office of Latvia. 2012. Permission of Category A Polluting Nr.
Riga Region, State Environmental Office, Ministry of RI12IA0006. Environmental Office of Riga Region,
Environmental Protection and Regional Development of State Environmental Office, Ministry of Environmental
the Republic of Latvia. Protection and Regional Development of the Republic of
[15] Environmental Office of Riga Region, State Latvia.
Environmental Office, Ministry of Environmental [17] Obernberger, I. 2009. Reached Developments of
Protection and Regional Development of the Republic of Biomass Combustion Technologies and Future Outlook.
Latvia. 2012. Permission of Category A Polluting Presented at the 17th European Biomass Conference and
Activities RI12IA0001. Environmental Office of Riga Exhibition, Hamburg, Germany.
Region, State Environmental Office, Ministry of [18] Sippula, O. 2010. Fine Particle Formation and Emissions
Environmental Protection and Regional Development of in Biomass Combustion, Academic Dissertation. Kuopio:
the Republic of Latvia. University of Eastern Finland.
doi: 10.17265/1934-8975/2016.03.006
D DAVID PUBLISHING
Losses Calculation of an Aerospace Retraction Wheel

Motor with Regarding to Electromagnetic-Field Analysis
Investigation
Pedram Asef and Ramon Bargallo Perpina

Department of Electrical Engineering, Polytechnic University of Catalonia-BarcelonaTech, Barcelona, CAT 08036, Spain
Received: December 23, 2015 / Accepted: January 06, 2016 / Published: March 31, 2016.
Abstract: A 3-D FEA (finite element analysis) transient and steady-state design proposal for high-speed with Nd-Fe-Br (reversible)
magnets in aerospace application will be examined under design considerations of n = 12,000 rpm, short-duty, sinusoidal drive, low
cogging, high efficiency at peak torque, and etc. for an ARWM (aerospace retraction wheel motor). In construction, the PMs
(permanent magnets) fixed on the rotor core which is surface-mounted magnets retained by a carbon-fiber bandage. Redundant
windings, resistant to fault propagation have accounted. Besides, an axial water-jacket housing without end-cap cooling has involved.
All performed characteristic performances of the correlated ARWM will verify by comparison through 2-D and 3-D FEA results. In
this paper, design process has dealing with determination of various kinds of losses such as electromagnetic and mechanical losses. In
terms of both classified losses, copper, stator back iron, stator tooth, PM, rotor back iron, air-friction and sleeve losses were calculated.
The 3-D end-winding effects were included in the modeled ARWM by the authors.
Key words: ARWM, high-speed PM motor, redundant winding, sleeve, end-winding, short duty, air-friction losses.
Nomenclature Number of coils per pole per phase

PF Power factor
Electrical cycle
Rated power
Density
Rated speed
Conductivity
Fs Fundamental frequency
Supply voltage to the controller
Short-circuit current
Peak tooth flux-density
Rms (root mean square) phase current
Peak yoke flux-density
/ Rms current density in phase windings
Peak flux density in stator tooth on open-circuit
Angular circumferential around the air-gap
Flux density in the stator core
Magnet energy-production
Intrinsic coercivity
Winding factor
Density of magnet material
Lamination space factor of the stator-core
Shaft diameter
Peak fundamental air-gap flux-density
Dse Stator outer diameter
ls Active stack stator length
Rotor diameter Factor to include the effect of slot-leakage flux on
Stator core yoke saturation face factor
Cross-sectional area of a stator slot Stator equivalent length including effect of fringing
Length of the path of one pole-pitch as measured in
Effective slot-depth
the mid-point of the stator-core
PM (permanent magnet) arc Ls Leakage inductance/phase
m Stator number of phases
Corresponding author: Pedram Asef, PhD, research fields: N Number of turns per phase
electrical machinery and its drives.
p Number of pole pairs In this research, baseline of the motor topography

Total number of the stator slots will be discussed by authors. Besides, shrinking motor
Width of phase separator (excluding legs) dimensions until desired temperature rise has reached.
Turns in series per phase
In case of such a highly demanding application,
Air-gap length
HSPMSM (high-speed surface-mounted pm
Torque angle
synchronous motor) topology proposed with proper
Thermal conductivity
Specific heat results in terms of electromagnetic analysis and losses
Reynolds number calculation by the authors.
LFV Local fluid velocity For the PMs fixed on the rotor core, which is
Axial velocity of the average surface-mounted magnets retained by a carbon-fiber
% Percentage of the efficiency bandage because improved tensile modulus and adhesion
properties over T700S type, where applications of this
1. Introduction
never twisted fiber include aircraft/airplane and high
In the application area of aerospace, highly engineered performance sporting goods, where demanding
with smallest motor and sensor types are at harsh conditions require superior composite properties.
environment. High performance and precision which As Fig. 2 shows the construction, an axial
provided through embedded 3-D FEA (finite element water-jacket housing without end-cap cooling with dn
analysis). Whereas the following clues have derived in = 34, 85 10-5 m3/s of fluid volume flow rate, =
the ARWMs (aerospace retraction wheel motors) design 0.125 W/(mC), = 1,000 kg/m3, 1.88 kJ/(kgC),
such as short-duty, low cogging, high efficiency at peak 1.55E+04 and 53.37 m/s have derived
torque, banded rotor/large air-gap, redundant windings in order to abound the temperature along minimizing
that occurs resistant to fault propagation [1], react time losses.
almost 10 s, must withstand by two consecutive In order to have the highest accuracy for the
extend/retract cycles, which is flooded with the simulation, the maximum accessible number of
ARWM fuel, must supply 16 kW of mechanical output elements of a competent mesh has noticed by 991,253
power at 12,000 rpm [2]. Through below schematic, elements and 4,561,205 of nodes which can be seen by
better understanding of the application represented by Fig. 3.
Fig. 1 [3-5]. In addition, providing a mechanical gearbox
to match motor speed to the required output speed
operates advantageously due to very high relative
torque density mechanical gearing can achieve [6]. The
mechanical advantages synonymous with EMAs
(electromechanical actuators) result in small, compact
actuators ideal for the aerospace industry, where
torque/force density, fault tolerance and reliability are
of vital importance [7]. The higher torque density, the
simplification and maintenance are some of their most
Operation time (t > 0) Standby time (t = 0)
crucial advantages leading PM (permanent magnet) Fig. 1 Schematic of the studied ARWM for airplane
machines to be noted as an attractive option [8]. application.
Losses Calculation of an Aerospace Retraction Wheel Motor with Regarding to 185
(a) (b)
Fig. 2 Designed prototype of the ARWM in (a) 2-D model and (b) 3-D with 1/4 of whole model and 1/2 length.
of the speed. Besides, for ignoring excessive iron losses,

the flux should be confined to do not alternate in a
high-frequency. Redundant windings clear-cut due to
attain a sinusoidal open-circuit back-EMF
(electromotive force) waveform have been accounted.
According to the back-EMF and the rotor-position
sensing, an AC control with sinusoidal phase current,
180 conduction for each inverter leg should be used
by a PWM (pulse-width-modulation) inverter with a
position encoder [4, 5, 9]. In this study, design of the
machine geometry data and its FEA results are held by
Table 1 and Table 2, respectively.
Fig. 3 Generated mesh for the whole model Table 1 Design of the machine geometry.
Variable Linear current density fixed
2. Electromagnetic Field Analysis through
80 mm
Embedded 3-D and 2-D FEA 30 mm
64 mm
In terms of the chosen topology, radial-flux strategy
106 mm
because of the fabrication and slotted stator by standard 130 mm2
round radial laminations, has been elected in order to g 2 mm
maximize the electrical loading. As usually AC motors 130 mm
13 mm
have more than 8-poles owing to be enable to have
12
fractional-slot windings.
Slots/pole 1.5
Here the ARWM via a 8-poles (surface-mounted P 4
PMs) in order to decrease the costs of the motor will be m 3
presented. Although, the number of poles is a function 3 mm
Table 2 FEA design data.

Variable Value Unit
Vs 690 V
@ 20 1.3 T
1,592 kA/m
7,500 kg/m3
800 Hz
15 A/mm2
34 A
55 Fig. 4 Compared TVW and MSI methods to calculate
69 A cogging torque.
0.34 mH
0.8
210 kJ/m3
1.6 T
12,000 rpm
23 kW
PF 0.9
CPP 0.5
% 95.5%
As in the open-circuit test, the variation of cogging Fig. 5 Comparison between MSI and skewed-MSI results.
torque caused by the interaction of the magnetized
rotor and the slotted stator. Here the cogging torque
was calculated through:
(1)
The approach used in this text is somewhat the

reverse of this latter method. However, two different
methods have been applied in order to have extreme
precision results. One is base on the local co-energy Fig. 6 Comparison between skewed-normalized and
non-skewed normalized harmonics at the cogging torque.
derivative which has been rated in the air-gap TVW
(time-varying weight) method. Second one is the MSI peak yoke flux-densities under no-load condition has
(maxwell surface integral) method that is also applied been showed.
in the air-gap. Both can be seen in Fig. 4. In order to calculate the flux-density in the stator
By doing skew, here the cogging has core (tooth and yoke), Eq. (2) can be given [10]:
straightforwardly been reduced as Fig. 5. Their
(2)
normalized maximum harmonics have also been
presented in compare with non-skewed one by Fig. 6. The flux-density and the EMF/turn on the tooth of
A 3-D transient magnetic flux density distribution of the stator have been showed by comparison between
the ARWM under rated load condition and 2-D 2-D and 3-D results which are soundly packed nearby
steady-state magnetic flux density distribution under themselves in below Fig. 10.
no-load condition have been illustrated in Figs. 7 and 8. Besides, the flux-density distribution on the air-gap
In addition, the flux paths have also been indicated in region via a comparison between 2-D and 3-D has been
Fig. 8. By Fig. 9, a comparison between peak tooth and represented in the below graph Fig. 11.
(a) (b)
(c) (d)
Fig. 7 A 3-D transient magnetic flux density distribution of the ARWM under rated load condition with 1/2 stack length: (a)
@ t = t = 1.465E-04 s, (b) @ t = t = 2.930E-04 s, (c) @ t = t = 5.859E-04 s, (d) @ t = t = 1.143E-03 s.
Fig. 8 2-D steady-state magnetic flux density distribution Fig. 10 Comparison between the flux-density on the stator
under no-load condition. tooth, and tooth EMF/turn in 2D and 3D.
Fig. 9 Comparison between peak tooth and peak yoke Fig. 11 2-D and 3-D results on the flux-density distribution
flux-densities. in the air-gap region.
In the following discussion as always, windage

(air-friction) consideration must be accounted in the
Fig. 12 The rated torque of the ARWM. high-speed condition due to its higher windage loss and
its effect on temperature rise of the motor. Thus, the
Which becomes from Eq. (3) [10]: paper argues the way that, windage losses have
cos (3) scrutiny been calculated in the rotor surface and the
both rotor ends through results by 3-D FEA
As Fig. 12 shows that, the average torque is 18.5 Nm. electromagnetic field analysis. In this calculation, the
In this section, the paper has been evaluated the windage loss for overcoming drag resistance of the
electromagnetic analysis of the studied motor through rotating cylinder can be calculated by Jaaris method as
three different FEA software in order to achieve better below [16]:
results from simulation. As the results present, the k (7)
electromagnetic manner of the ARWM was matched where, is the friction coefficient that depends on
with what was expected in terms of characteristics. the structure of the surface, k is a constant coefficient
3. Losses Calculation of the ARWM which was chose 1.7, is angular velocity which is
1,130 rad/s, and r is the radius of the rotor (0.032 m). In
In such applications with extreme synchronous speed
order to calculate Cf, first through the
and power density, thus, in this case, the proportion of
Couette-Reynolds number ( ) for accounting the
iron loss is larger than copper loss. In this paper, iron
enclosure and the Axial-Reynolds number ( ) for
losses included eddy-current (on the cores, PMs, and
accounting the axial-flow that must be calculated via
sleeve) and hysteresis losses (on the cores) what can be
below expressions [16].
both calculated by below Eqs. (4) and (5) [11, 12]:
(8)
1 (4)

(9)
d (5)
After that, the can be calculated by Eq. (10) as:
where, , N, d, k and are maximum flux-density,
. .
the number of minor loops, thickness of the lamination . 1 (10)
material, a constant which will be chosen between
Then, by having all above parameters, the windage
k = 0.6-0.7, and the flux reversal associated with a
loss at rotor surface will be possible. However, the
given minor loop, respectively [13, 14]. Rotor
windage in the both rotor ends ( , ) also can be
eddy-current loss becomes the dominant loss
important to evaluate through Eq. (11):
mechanism, contributing 83% of the total loss [15].
, 0.5 , (11)
On the other hand, Copper losses and Friction
where, , for the both rotor ends has been defined
losses which both can be calculated through Eqs. (6)
as:
and (7) [6]:
.
(6) , . (12)
where, , , and are the nth current
The tip-Reynolds number under the condition of
density harmonic, copper conduction resistivity,
cylinders rotating at a free space should be given as:
copper volume along each slot, and end-winding
copper volume, respectively. (13)
Table 3 Losses calculation results. power density on the winding has been found similarity
Eddy-current Hysteresis Windage Total to sum of this value on the stator core plus PMs.
Loss (W)/part
loss loss loss (W)
Winding 550 - - 550 4. Conclusions
Stator
110 94 - 200
back-iron In this paper, the various aspects of the economy
Stator tooth 450 160 - 610
PMs 160 - - 160 problem have been considered in terms of active
Shaft 4 6 - 10 materials and sizing issues. The ARWM model under
Rotor-surface - - 2 2 technical conditions for electromagnetic-field analysis
Rotor-ends - - 0.1 0.1 has been analyzed in 3-D (transient-case) and 2-D
Sleeve 1 - - 1
(steady-state-case), with its notable results. Where
Total (W) 1,275 260 2.1 1,533.1
technical conditions relied on short-duty, low cogging,
high efficiency at peak torque, banded rotor/large
air-gap, Redundant windings that occurs resistant to
fault propagation, react time almost 10 s, must
withstand by two consecutive extend/retract cycles.
Besides, through a highly engineered design in terms
of skewing, the harmonics at the cogging torque has
straightforwardly been reduced in compare to the one
without skewing. Thus, the cogging torque was
minimized perfectly.
Fig. 13 The percentage of total eddy-current and hysteresis
losses.
All the simulation results have been developed by a
comparison between 3-D and 2-D results as a
validation of the electromagnetic-field analysis.
With regarding to the presented
electromagnetic-field analysis results, the eddy-current,
hysteresis and windage losses of the model have been
calculated and compared.
Acknowledgment
The authors gratefully acknowledge the contribution
of Dr. D. Staton from Motor Design Ltd., UK.
Fig. 14 Comparison of the power density by each
individual part. References
The following table (Table 3) summarized the [1] Zhao, W., Cheng, M., Hua, W., and Jia, H. 2008. A
Redundant Flux-Switching Permanent Magnet Motor
calculated losses and Fig. 13 presents the percentage of
Drive for Fault-Tolerant Applications. In Proceedings of
total eddy-current and hysteresis losses. the IEEE International Conference on Vehicle Power and
Subsequent presented Fig. 14, focused on the power Propulsion Conference, 1-2.
density value (as a compulsory value to determine the [2] Mecrow, B. C., Jack, A. G., Atkinson, D. J., Green, S. R.,
Atkinson, G. J., King, A., and Green, B. 2004. Design
temperature distribution) in the whole motor
and Testing of a Four-Phase Fault-Tolerant
comparability, where the power density value on the Permanent-Magnet Machine for an Engine Fuel Pump.
PMs and stator core aligned. Meanwhile, value of the IEEE Trans. on Energy Conversion 19 (4): 671-4.
[3] Cao, W., Mecrow, B. C., Atkinson, G. J., Bennett, J. W., Madison, WI: University of Wisconsin-Madison, 120-7.
and Atkinson, D. J. 2012. Overview of Electric Motor [11] Zhao1, N., and Liu, W. 2015. Loss Calculation and
Technologies Used for More Electric Aircraft (MEA). Thermal Analysis of Surface-Mounted PM Motor and
IEEE Trans. on Industrial Electronics 59 (9): 3226-7. Interior PM Motor. IEEE Trans. on Magnetics 51 (11):
[4] Alastair, F. 2008. Practical Application of CAD in a High 1-2.
Power Density Motor for a Very Short Duty Aerospace [12] Asef, P. 2014. Design, Characteristic Analysis of PM
Actuator. UK: Electro-Kinetic Designs Ltd. Wind Generator Based on SMC Material for Small
[5] Robinson, J. I. 2007. An Integrated Approach to Direct-Drive Wind Energy Conversion System. In
Aerospace Motor Design. IEEE Trans. on Aerospace 1 Proceedings of the International Conference on
(2): 847-50. Renewable Energy Research and Application (ICRERA),
[6] Gerada, C., Bradley, K., Whitley, C., and Towers, G. 2007. 44-6.
High Torque Density PM Machines for High [13] Gao, N., and Yu, L. 2013. Research on Loss and
Performance Operation. In Proceedings of the IEEE Electromagnetic Heat Coupling of High Speed Permanent
Industrial Electronics Society (IECON), 210-3. Magnet Synchronous Motor. In Proceedings of the IEEE
[7] Galea, M., Gerada, C., Raminosoa, T., and Wheeler, P. W. International Conference on Mechatronics and
2011. Considerations for the Design of a Tubular Motor Automation (ICMA), 85.
for an Aerospace Application. In Proceedings of the [14] Gilson, G. M., Raminosoa, T., Pickering, S. J., Gerada, C.,
IEEE Electrical Machines and Systems (ICEMS), 1-2. and Hann, D. B. 2010. A Combined Electromagnetic and
[8] Kakosimos, P. E., Tsampouris, E. M., Kladas, A. G., and Thermal Optimisation of an Aerospace Electric Motor. In
Gerada, C. 2012. Aerospace Actuator Design: A Proceedings of the IEEE Electrical Machines (ICEM),
Comparative Analysis of Permanent Magnet and 1-4.
Induction Motor Configurations. In Proceedings of the [15] Atkinson, G. J., Mecrow, B. C., Jack, A. G., Atkinson, D.
IEEE Electrical Machines and Systems (ICEMs), 2538-41. J., Sangha, P., and Benarou, M. 2006. The Analysis of
[9] Dorrell, D. G., Hsieh, M., Popescu, M., Evans, L., Staton, Losses in High-Power Fault-Tolerant Machines for
D. A., and Grout, V. 2011. A Review of the Design Issues Aerospace Applications. IEEE Trans. on Industry
and Techniques for Radial-Flux Brushless Surface and Applications 42 (5): 1162-8.
Internal Rare-Earth Permanent-Magnet Motors. IEEE [16] Aglen, O. 2003. Loss Calculation and Thermal Analysis
Trans. on Industrial Electronics 58 (9): 3741-6. of a High-Speed Generator. In Proceedings of the IEEE
[10] Lipo, T. A. 2011. Introduction to AC Machine Design. International Electric Machines and Drives Conference,
In Library of Congress Cataloging in Publication Data. 1118.
doi: 10.17265/1934-8975/2016.03.007
D DAVID PUBLISHING
Energy Efficiency of Reactive Dynamic Compensators
Rodrigo Nobis da Costa Lima1, Srgio Ferreira de Paula Silva1, Antnio Carlos Delaiba1 and Ezequiel Junio de
Lima2
1. Energy Rationalization and Power Quality Group, Faculdade de Engenharia Eltrica, Universidade Federal de Uberlndia,
Uberlndia 38.400-902, Brazil
2. Instituto Federal de Educao, Cincia e Tecnologia do Sul de Minas Gerais, Poos de Caldas 37713-100, Brazil
Abstract: A parameter that allows an evaluation of power quality transmitted, or distributed, between energy source and the final
user is electric system power factor. Among other aspects, a bigger power factor, close to unit value, relieves operational conditions
of lines and cables, besides, it improves feeders voltage behavior. Due to load variation along the day, the dynamic compensation of
power factor allows maintaining this parameter close to the ideal. This paper brings a study about a reactive dynamic compensator
based on the voltage control in a capacitive element, varying the reactive energy in accordance with the system demand, everything
from the energy efficiency point of view. In distribution systems, the losses due to this variable compensation can be lower than in
other compensation methods and also the voltage presents a better behavior, justifying its application.
Key words: Reactive dynamic compensator, power factor correction, losses, energy efficiency, tap changing.
1. Introduction registered in 2001. The only reason Brazil is not

facing an energy rationing like in the beginning of the
The electrical energy demand is growing, just as its
century is because a great part of the energy consumed
cost. This situation supports and favors the energy
in the country is being provided by thermoelectric
efficiency surveys applied to the electrical system.
plants. In this scenario, any measure taken to improve
According to the Brazilian National Energy Balance
energy efficiency is greatly welcomed, since it
of 2014, hydraulic plants provided 70.6% of the
relieves the network and allows better use of
electrical energy generated in Brazil [1]. Even though
generation resources.
the advance of the other energy sources, mainly wind
In the energy transmission between the load and the
power, the predominance and consequently, the
source, active power must be bigger than the reactive
dependence of hydraulic sources is quite big. This
power. Reactive power demanded by the loads must
kind of generation is strongly weather dependent and,
be produced near the load preferably. This condition
when there is a lack of rain, the energy rationing issue
reduces feeders line current and, as consequence,
shows up, contributing to the rise of the energy cost.
reactive power set up an important role in voltage
In order to illustrate this risk, Fig. 1 [2] compares the
regulation and system reliability [3]. The parameter
reservoirs levels of the Brazilian southeast and center
that quantifies the rate between power and total
west regions in the 2014 and 2015 years to the 2001
(apparent) power is called power factor.
levels, year of the greatest energy crisis in Brazil.
The basic idea of power factor correction is
As it can be seen, the registered values in the
producing some of the reactive power demanded by
beginning of this year were lower than the ones
the load near itself. The most common, and probably
Corresponding author: Rodrigo Nobis da Costa Lima, the oldest manner of supplying this reactive power is
master, research fields: power systems with emphasis in energy by capacitor banks.
efficiency and power quality.
2. Working Principle of the DRC

As said before, the power demanded by a
distribution system, both reactive and active, varies
along the day. Thus, the ideal situation would be if
the reactive power compensation followed this
variation. Instead of what is obtained with the
Fig. 1 Reservoirs levels of the Brazilian southeast and installation of fixed capacitor banks, the DRC
center west regions.
provides reactive power according to the network
Considering that a power systems load along a day requirement. The change in the reactive injected in the
varies [4, 5], the demand for reactive power is also system is made varying the voltage applied on the
variable. This load variation can be noticed by the capacitor bank.
load curves for different consuming classes The secondary side winding of the equipment is
(commercial, residential or industrial) [6]. divided into taps, allowing a voltage variation since a
Urban and country distribution feeders have load minimum value until the reactive elements rated
curves with big difference between the power voltage. As the capacitor is not disconnected and the
demanded in peak hours and in the remaining. The use voltage taps are changed gradually, the switching
of fixed capacitor banks in these cases can lead to an transients are highly reduced when compared to the
undesired increase in voltage and even in losses at insertion banks made by breakers or mechanic
some periods of the day. A solution could be using switches. This sentence is based on the utilization of a
switched banks, however, the switching devices cost significant amount of taps, higher than 10, resulting in
and the transient caused by the switching itself may voltage steps lower than 10%. Fig. 2 shows the
complicate the utilization of these equipment. equipments simplified architecture.
An alternative solution is the variation of the It is important to notice that, the reactive
voltage applied to the capacitor bank. Whereas the compensation does not need to be made by capacitors,
reactive power generated by a capacitor varies directly if the system presents an excess of reactive power, the
with the square of the voltage on it, the amount of capacitor bank can be replaced by a reactor. In this
energy can be controlled by the voltage variation. An paper, the compensation will be made only by
on load tap changing device [7] is able to perform this capacitor banks.
voltage control. Along this paper, the three one-phase reactive
The DRC (dynamic reactive compensator) is compensators are going to have the following
nothing but a transformer with the secondary winding parameters:
divided into taps feeding a capacitor bank. The transformer rated power: 200 kVA;
purpose of this paper is to analyze the impact the capacitor bank rated power: 200 kVAr;
installation of the DRC will cause in a distribution transformer percentual impedance (tap + 16):
feeder. This equipment was developed and patented 11.26%;
by a brazilian company called ITB (Indstria de no load losses: 513 W;
Transformadores Birigui) equipamentos eltricos [8]. short circuit losses: 2,532 W;
A real network is compensated by DRCs in Ref. [9]. magnetizing current: 0.53 A.
This paper focuses on monitoring the power factor For this setup, the reactive power injected on the
behavior of the network for 11 months and the power network for each tap, assuming rated voltage applied
factor improvement is quite satisfactory. on the equipment, is shown in Fig. 3.
Energy Efficiency of Reactive Dynamic Compensators 193
Fig. 2 Simplified DRCs squeme.
Fig. 4 Commercial load curve.

Tap
Fig. 3 Reactive power supplied by DRC in different taps,

when submitted to rated voltage.
Fig. 5 Residential load curve.
3. Study Case
load curves are for three-phase power. Thereby, apart
In order to better comprehend the DRCs from reactive compensation, as the DRC controls the
functionalities, some computational simulations [10] power factor in each phase independently, it is
were performed in a hypothetical distribution feeder. possible to analyze the current and voltage unbalance.
The loads were divided into two types: commercial Three distincts operational situations were
and residential. The load curves were made using simulated: system without reactive compensation,
Refs. [6, 11] as references. These curves are shown in compensation done by fixed capacitors and
Figs. 4 and 5. compensation done by the DRCs devices.
The simulated feeders rated voltage is 13.8 kV and In all the above situations, 24 cases, representing
it feeds 44 transformers connected in /Y grounded the 24 h of the implemented load curves were
(Dyt), which attend 44 constant current type loads. Fig. 6 simulated.
shows the unifilar diagram of the simulated circuit.
3.1 No Compensation
The loads were specified by phase and their power
values were randomly chosen between 90% and 110% The first of the three simulated cases was the
of the three-phase power. The one-phase power factor system without compensation, i.e., only the conductors,
were also chosen randomly, the possible values are the loads and their respective transformers were
between 0.80 and 0.90. The purpose of this process is implemented. In this case, the source supplied about
to obtain an unbalanced system, since the values in the 5.6 MVA (pf = 0.841) and 0.9 MVA (pf = 0.852)
Bus I
Bus II
Capacitors
Bus III
DRCs
Fig. 6 Studied feeder unifilar diagram.
in the highest and in the lowest demand situation, capacitor banks can be seen in Fig. 6. In this
respectively. operational condition, the maximum and minimum
demands were 5 MVA (pf = 0.929) and 1.2 MVA
3.2 Compensation by Fixed Capacitors
(pf = -0.691), respectively. A reduction of 600 KVAr
In the second simulation, three fixed capacitor and a power factor improvement on the highest
banks had been installed (two banks of 210 kVAr and loading period can be seen due to the compensation
one bank of 900 kVAr of rated three-phase power). To done by the capacitors. On the other hand, on the low
elucidate, the three banks have been installed in three loading period, the reactive power excess caused an
different buses and the biggest bank has been installed increase in the value of the demand and a decrease in
close to the load center. The position of these three the power factor, becoming capacitive in part of the
time. The DRCs purpose is to actuate on the low compensated systems were higher than the
loading period in order to reduce the excessive non-compensated systems. However, the utilization of
reactive flowing on the network and improve the fixed capacitors led the power factor to values lower
power factor. than -0.5 at the low loading moments. This condition
is favorable to voltage and losses enhance. Fig. 8
3.3 Compensation by DRC
shows the power factor behavior in the three situations,
In the third simulation, three fixed capacitor banks seen by the source.
were operating as well (two banks of 210 kVAr and Also from the sources point of view, the power
one bank of 300 kVAr of rated three-phase power), factor variation was much smaller when the time that
besides three one-phase DRCs of 200 kVAr feeding the system was compensated by the DRC.
200 kVAr of capacitor banks (with identical As it has been mentioned before, at the low loading
parameters described in Section 2). Important periods, there was an excess of reactive power flowing
reminder: The DRCs were installed on the same bus on the system when it was compensated only by fixed
that the 300 kVAr bank was, which was also the bus capacitors, leading the power factor smaller values. As
where the 900 kVAr bank was installed in the second the demanded reactive power varied along the day, the
case (fixed capacitors only). The other two smaller DRC changed its secondary winding tap in order to
banks location and rated power are the same in both change the amount of reactive injected on the network.
situations. In this case, the maximum and minimum Fig. 9 shows the tap switchs position of the three
demands were approximately 5 MVA (pf = 0.93) and equipments (three-phase bank made by one-phase
0.8 MVA (pf = -0.96). In the highest demand case, the equipment) during the simulated 24 h, it can be
control of the DRC performed the higher possible observed that, the diagrams outline is similar to the
compensation, according to its rated power (operating
as a fixed bank), in the low loading situation, the
control set the voltage applied on the bank to zero, not
allowing the equipment to inject excessive reactive
power.
4. Results
Following, there are going to be presented the
obtained results for the three considered operational
conditions described above. The analyses were Fig. 7 DRCs bus power factor for each case.
focused on three parameters: power factor, electric
losses and voltage behavior.
4.1 Power Factor
Fig. 7 presents the power factor variation for each

of the three cases on the bus that the DRC was
installed. It is worth to remember that, the equipment
was configured so that it would keep the power factor
on its respective bus between -0.98 and 0.98.
As expected, the power factor values for the Fig. 8 Feeders power factor, seen by the source.
more advantageous at the low demand periods.

Fig. 13 shows the difference between the losses
obtained in situation 2 (fixed capacitors only) and 3
(DRC), for each simulated hour.
It can be observed that, the system with the DRC is
more efficient until hour 10 (positive inclination),
after the rise of the demand, the curve changes its
inclination, meaning that, the system with fixed
capacitors starts to show better results from the losses
Fig. 9 DRCs tap switcher position. point of view. The positive inclination happens
load curves, once the reactive power demand has a

behavior similar to the active power demand.
Once the system loads are unbalanced, each phase
DRC is going to act in order to keep the one-phase
power factor within the stipulated limits. As a
consequence of this independence, the tap position of
each equipment can be different for the same situation,
as can be seen in Fig. 9. This way, the DRC
installation can decrease the unbalance rate.
Fig. 10 Losses in the case without reactive compensation.
4.2 Technical Losses
The losses caused by the current flow through the

feeder conductors are called technical losses. Figs. 10-12
show the losses in the compensation devices
(capacitors and DRCs), in transformers and
conductors, and the total losses of the feeder,
considering the 24 h simulated.
By varying the amount of reactive injected on the
system, the DRC limits the excessive reactive in the
low loading periods, what is confirmed by the results
Fig. 11 Losses in the DRCs case.
shown in the charts above.
An important information obtained by the diagrams
above is the difference between the conductor losses
observed for the DRC and fixed capacitors situations.
In the first one, the cable losses were 97 kWh lower.
The DRC disadvantage is its inner transformer,
more specifically the losses caused by this element.
This can be clearly noticed in Fig. 11, where the
DRCs consumed more than 100 kWh during the
simulation period. However, even with these high
losses, the system as a whole was more efficient, been Fig. 12 Losses in the fixed capacitors case.
Fig. 13 Accumulated losses difference between the DRC

and the fixed capacitors cases.
Fig. 14 Voltage at bus I.
because there is an excessive reactive power in the
low loading period when the system is compensated
only with fixed capacitors, leading to higher Joule losses
in the cables, different from what happens in the DRCs
case, once their tap switches are in low positions (Fig. 9)
and there is not excessive production of reactive power.
Meanwhile, the negative inclination happens because
the system demands a larger amount of reactive power
and that makes the injection of reactive in both
systems very close, however, the DRCs system
Fig. 15 Voltage at bus II.
causes higher losses due to their inner transformers.
Nevertheless, at the end of a day, the DRC system
resulted in a reduction of 61.66 kWh on the losses
when compared to the fixed capacitors system.
4.3 Voltage Regulation
The amount of reactive power flowing through an

electric system is deeply linked to the voltage levels
along the circuit. This way, in situations with an
excess of reactive power, sometimes high voltage
levels are observed.
Fig. 16 Voltage at bus III.
The diagrams of Figs. 14-16 show the voltage
variation in three medium voltage buses. The bus When the amount of reactive power injected in the
number I is the closest to the source and the bus system is varied, the DRC does the secondary control
number III is the further, which is also the one where of voltage, since its control parameter is the power
the 900 kVAr bank or the DRCs and the 300 kVAr factor.
bank are installed, depending on the simulated In the DRC case, as a result of the control of the
situation. The upper and lower limits demarcated in reactive power produced, the voltage also presented
the diagrams are the ones described in the Brazilian the best behavior, mainly on the low loading period
standard [12]. The names and locations of the buses (0-8 h). The voltage difference between the two cases
are shown in Fig. 5, by the colors. was small due to the compensator size.
5. Conclusions Ministrio de Minas e Energia. Balano Energtico

Nacional 2014: Relatrio Final. Accessed May 06, 2015.
Due to the load variation along the day, a variable https://ben.epe.gov.br/downloads/Relatorio_Final_BEN_
reactive compensation may be necessary. 2014.pdf.
[2] Operador Nacional do Sistema Eltrico (Brazil). 2015.
The compensator presented in this paper adjusts the
Energia Armazenada 2015. Operador Nacional do
amount of reactive power injected on the system Sistema Eltrico. Accessed May 06, 2015. http://www.
according to its bus power factor. ons.org.br/historico/energia_armazenada_out.aspx.
The computational simulation of the equipment was [3] Al-mhanna, T. H. 2009. Power Quality Improvement
Using Dynamic Compensation of Reactive Power to Control
highly satisfactory. For the simulated systems, the
the Voltage Values. The Iraqi Journal for Mechanical
dynamic compensator had a better performance in all and Material Engineering, Special Issue 0: 839-49.
the analyzed parameters when compared to the fixed [4] Chis, M., Salama, M. M. A., and Jayaram, S. 1997.
capacitor banks (power factor, losses and voltage Capacitor Placement in Distribution Systems Using
Heuristic Search Strategies. IEE Proceedings-Generation,
regulation in steady state). However, each system
Transmission and Distribution 144 (3): 225-30.
must be studied in detail before the installation of the [5] Deng, Y., Ren, X., Zhao, C., and Zhao, D. 2002. A
equipment. Heuristic and Algorithmic Combined Approach for
It can be easily seen that, the DRC presents more Reactive Power Optimization with Time-Varying Load
Demand in Distribution Systems. IEEE Transactions on
advantages over the fixed capacitor banks in systems
Power Systems 17 (4): 1068-72.
with a variable demand of reactive power. This way, [6] Jardini, J. A., Tahan, C., Gouvea, M. R., Ahn, S. U., and
in systems where the demand remains steady, the Figueiredo, F. M. 2000. Daily Load Profiles for
trade of fixed capacitors to the dynamic compensator Residential, Commercial and Industrial Low Voltage
Consumers. IEEE Transactions on Power Delivery 15
may not be so advantageous due to the latter inner
(1): 375-80.
losses. [7] Gao, D., Lu, Q., and Luo, J. 2002. A New Scheme for
It is worth noticing that, the compensator power is On-Load Tap-Changer of Transformers. In Proceedings
going to be another determinant factor in the obtained of the PowerCon 2002 (International Conference on
Power System Technology), 1016-20.
results with its installation. In this paper, it was
[8] ITB Equipamentos Eltricos Ltda. 2010. Regulador
installed a small bank of capacitors along with the reativo. Brazil Patent 1002561-8 A2, filed July 7, 2010,
DRC, assigning to the compensator only the fine and issued May 20, 2012.
tuning of the reactive power injected. However, if the [9] Takayanagi, H., and Iwamoto, D. 2014. Compensador
Reativo Monofsico Para Correo do Fator de
power of each equipment was 300 kVAr instead of
Potncia. CIDEL (Congreso Internacional de
200 kVAr, the voltage regulation, mainly in the Dirtibuicin Eltrica).
moments of lower reactive demand, would be more [10] Quality Engenharia e Sistemas Ltda. 2012. Simulador de
satisfying, but the losses in the DRC transformer Sistemas Eltricos de Potncia, SSEPUser Guide.
Quality Engenharia e Sistemas Ltda. Accesssed March 12,
would be larger. In view of this situation, before the
2012. http://www.qes.com.br.
installation of the equipment, a study should be made [11] Francisquini, A. A. 2006. Estimao de Curvas de Carga
taking in account the system to be compensated as em Pontos de Consumo e em Transformadores de
much as the desired results. Distribuio. In Dissertao (Mestrado em Engenharia
Eltrica)-Faculdade de Engenharia. Ilha Solteira:
For future works, the transient analysis of the tap
Universidade Estadual PaulistaUNESP.
switching and the actuation of the DRC as a harmonic [12] ANEEL (Agncia Nacional de Energia Eltrica). 2015.
passive filter are listed. PRODISTProcedimentos de Distribuio de Energia
Eltrica no Sistema Eltrico Nacional. ANEEL. Accessed
References May 22, 2015. http://www.aneel.gov.br/visualizar_texto.
[1] Empresa de Pesquisa Energtica (Brazil). 2014. cfm?idtxt=18.
doi: 10.17265/1934-8975/2016.03.008
D DAVID PUBLISHING
Calculation of CO Behavior in the Platform for Deeply

Underground Subway Station with Different Fire
Strengths
Yong-Jun Jang, Ji-Min Ryu, Jong-Ki Kim, Min-Chul Chun and Dong-Hoe Koo
Railroad Safety Center, Korea Railroad Research Institute, Gyeonggi-do 437-757, South Korea
Abstract: Effect of different fire strengths on the smoke distribution in the subway station is investigated. Shin-Gum-Ho station (line
#5) in Seoul is selected as a case study for variation of CO (carbon monoxide) distribution caused by the fire in the platform. The
ventilation in the station is set to be an air supply mod in the lobby and an air exhaustion mod in the platform. One-side main tunnel
ventilation (7,000 m/min) is applied to operate in the tunnel. The fire is assumed to break out in the middle of train parked in the
platform tunnel. Two kinds of fire strength are used. One is 10 MW and the other is 20 MW. Ventilation diffusers in the station are
modeled as 317 square shapes & four rectangular shapes in the lobby and platform. The total of 7.5 million grids is generated and
whole domain is divided to 22 blocks for parallel computation. Large eddy simulation method is applied to solve the momentum
equation. The behavior of CO is calculated according to different fire strengths and compared with each other.
Key words: Carbon monoxide, subway station, fire strength, main tunnel ventilation, LES.
1. Introduction showed that LES produced better results than RANS.

Fletcher, et al. [4] experimentally investigated the
Korea government has steadily supported the
tunnel fire using a pool fire. They also calculated the
fire-accident prevention research since fire disaster in
same fire-driven flow as their experiments using k-
Daegu subway station which occurred in 2003. It has
model and compared the results. Gao, et al. [5]
enforced to establish the several laws for fire safety
theoretically studied the tunnel fire flow which was
standard in railway system in Korea. However, the
experimentally done by Fletcher, et al. [4]. They
fire-safety measures for subway station and subway
employed LES method and compared the results with
tunnel have not been much arranged yet in Korea.
experimental data. Jang, et al. [6] experimentally
Some researchers have studied the fire-driven flow
investigated the ordinary and emergency ventilation
analysis in tunnel. Hwang and Edwards [1] studied the
system in deeply underground subway station to study
hot stratified flow and ventilated flow in tunnel.
the exact ventilation capacity of subway station. Jang,
Hwang and Wargo [2] showed the thermally generated
et al. [7] analyzed the smoke distribution in the
reverse stratified layers in fire tunnel using
subway station with various main tunnel ventilations.
experimental method. Jang, et al. [3] studied the
They showed the smoke density was reduced in the
back-layer flow phenomena in the straight tunnel
platform of station with main tunnel ventilation.
using LES (large eddy simulation) and RANS
Effect of fire strength on the smoke distribution in
(Reynolds-averaged Navier-Stokes). They compared
the subway station is investigated in this research.
the simulated results with experimental data and
Shin-Gum-Ho subway station (depth: 46 m), which is
Corresponding author: Yong-Jun Jang, principal the same station as studied by Jang, et al. [7], is
researcher, research field: thermo-fluid dynamics in railway.
200 Calculation of CO Behavior in the Platform for Deeply Underground Subway
Station with Different Fire Strengths
Table 1 Case study for different fire strengths. Table 2 Ventilation rate in the lobby and platform of
Shin-Gum-Ho station for emergency state.
Tunnel ventilation Fire
Cases Descriptions
capacity strength Flow rate (m3/h)
No main tunnel Case
1 0 10 MW Supply air Exhaustion
ventilation
Emergency Lobby 42,126 -22,561
2 One side exhaustion -7,000 m/min 10 MW
state Platform - -276,094
No main tunnel
3 0 20 MW
ventilation
4 One side exhaustion -7,000 m/min 20 MW
selected as a case study for variation of CO (carbon

monoxide) distribution caused by the different fire
strengths. The CO density is most concerned when (a) No main tunnel ventilation
fire break out in subway station.
Fire is assumed to break out in the middle of train
which is parked in platform. The two kinds of fire
strength are used. One is 10 MW and the other is 20 MW.
(b) Main tunnel ventilation (7,000 m/min)
Currently, ventilation operation mode in
Fig. 2 Application of main tunnel ventilation.
Shin-Gum-Ho station is set up as a full exhaustion
mode in the platform in case of platform-fire emergency. condition for emergency state in Shin-Gum-Ho station.
Furthermore, main tunnel ventilation is also applied to The pressurized air is supplied in the first and second
be operated as prescribed in Table 1 for helping of floor lobby. The contaminated air by smoke in the
smoke exhaustion in platform. The LES method in platform is exhausted through the ventilation. Table 2 [6]
FDS (fire dynamics simulator) code is used to simulate shows the ventilation rate in the lobby and platform of
the fire-driven flow in this station. Parallel computation Shin-Gum-Ho station for emergency state.
is employed to reduce the calculation time. Fig. 2 shows the conditions for main tunnel ventilation
and capacity which is well described in Table 1. Fig. 3
2. Modeling of Station
shows the ventilation distribution in the first and
The total number of floors in Shin-Gum-Ho station second floor underground. Ventilation diffusers are
is 8. The first and second floor undergrounds are modeled as 95 square shapes of 0.6 m (x) 0.6 m (y)
lobby and the eighth floor underground is platform. in the lobby (Fig. 3) and as 222 square shapes of
Fig. 1 shows the overall view and ventilation 0.6 m (x) 0.6 m (y) and four large rectangular
shapes of 1.2 m (x) 0.8 m (y) in the platform (Fig. 4).
3. Governing Equation and Numerical

Methods
The FDS code solves numerically a form of
Navier-Stokes equations, and the core algorithm for
momentum equations is an explicit predictor-corrector
scheme that uses second order accurate finite-difference
approximation. The flow variables are updated in time
using an explicit second-order Runge-Kutta scheme.
Fig. 1 Ventilation fan modes in Shin-Gum-Ho station for
Turbulence is treated by means of the Smagorinsky
emergency state. form of LES.
Calculation of CO Behavior in the Platform for Deeply Underground Subway 201
(a) (b)
Fig. 3 Diffuser arrangement of (a) the first floor underground and (b) the second floor underground.
Fig. 4 Diffuser arrangement in the platform of Shin-Gum-Ho station.
The governing equations [8] for LES in FDS are as Where t should be modeled, Smagorinsky
follows: model [9] is employed in FDS code.
DU i 1 P U i U j t (Cs ) 2 S (5)
- ( ) ij (1)
Dt xi x j x j xi
S (2Sij Sij )1/2 (6)
ij U iU j U i U j (2)
(xyz )1/3 (7)
1
ij ij kk -2 t Sij (3)
3
where, Cs is an empirical constant and is 0.2. In Eq. (7),
1 U U j (4)
Sij ( i ) x, y, z denote grid distances in x, y, z coordinates,
2 x j xi
respectively.
The grid size is then selected to be 0.2 m and the
In Eqs. (1)-(4), , P , Ui , ij , Sij and t ,
total grid numbers are 7.5 million (Fig. 5). The whole
respectively denotes density, mean static pressure, domain is divided to 22 blocks and computed on 10
mean velocity, turbulent fluctuation tensor, strain rate CPU (central processing unit) in parallel using MPI
tensor and turbulent viscosity. (message passing interface).
Stairway (Fig. 1 4 ) entrance

region
Y X
Reference vector (1 m/s)
Fig. 5 Grid modeling for Shin-Gum-Ho subway station. (a)

Stairway (Fig. 1 4 ) & middle floor (Fig. 1 5 )

Fig. 6 shows velocity vectors (maximum 0.97 m/s)
in the entrance region ( in Fig. 1) and middle floor
region ( in Fig. 1) of the stairway for fire strength
10 MW without main tunnel ventilation at Z
1,300-1,500 s. The fresh air in the lobby is moving Y X

down through the stairway to the platform. It is
(b)
observed that, the hot smoke air due to the fire in the Fig. 6 Velocity vectors (every other vector) in the stairway:
platform cannot overcome the incoming flow from the (a) entrance region and (b) middle floor region for 10 MW
lobby. Therefore, CO cannot move up to the lobby without tunnel ventilation.
with fire strength 10 MW in the platform. Fig. 7 Stairway (Fig. 1 4 ) entrance

region
shows velocity vectors in the same locations as in
Fig. 6 with main tunnel ventilation (7,000 m/min).
The velocity vectors (maximum 2.2 m/s) in the
stairway become stronger than without main tunnel
ventilation. The reason is that, the pressure in the
platform is decreased due to the operation of tunnel Z
ventilation. Large vortices are found in the stairway Y X

and its size is about 1.5 m in height. Fig 8 shows
(a)
velocity vectors (maximum 1.12 m/s) in the same Stairway (Fig. 1 4 ) & middle floor (Fig. 1 5 )
locations as in Fig. 6 for fire strength 20 MW without
main tunnel ventilation. The magnitude of velocity is
similar with case of 10 MW. The back flow is slightly
generated near the top of middle floor because of
buoyancy force of smoke caused by stronger fire
strength in the platform. Fig 9 shows velocity vectors Z
(maximum 2.3 m/s) in the same locations as in Fig. 6 Y X

with fire strength 20 MW and with main tunnel
(b)
ventilation (7,000 m/min). The magnitude of velocity
Fig. 7 Velocity vectors (every other vector) in the stairway:
is increased compared with no main tunnel ventilation, (a) entrance region and (b) middle floor region for 10 MW
however, is similar with Case 2. with tunnel ventilation.
Stairway (Fig. 1 4 ) entrance Fig. 10 shows the soot (smoke) distribution in the
region
platform. Large amounts of soot occupy the platform
without main tunnel ventilation (Figs. 10a and 10c). In
Y X

(a)
Stairway (Fig. 1 4 ) & middle floor (Fig. 1 5 )
(a) Case 1 (10 MW)
Y X

(b)
Fig. 8 Velocity vectors (every other vector) in the stairway:
(a) entrance region and (b) middle floor region for 20 MW
without tunnel ventilation.
Stairway (Fig. 1 4 ) entrance (b) Case 2 (10 MW)

region
Y X

(a)
Stairway (Fig. 1 4 ) & middle floor (Fig. 1 5 ) (c) Case 3 (20 MW)
Y X

(b)
Fig. 9 Velocity vectors (every other vector) in the stairway: (d) Case 4 (20 MW)
(a) entrance region and (b) middle floor region for 20 MW Fig. 10 Soot distributions in the platform with different
with tunnel ventilation. fire strengths.
500
1 1 case of operation of main tunnel ventilation, the
450 2 2
400
3
4
3
4
smoke is not much propagated into the platform due to
5 5
350 6 6
(10 MW)
the tunnel ventilation (Figs. 10c and 10d). However,
7 7
CO (ppm)
300 the soot is a little more propagated into the platform

250
4 4 with 20 MW than with 10 MW. It is clearly seen that,
200 4 4 4 4 4 4
150 4 the main tunnel ventilation plays an important role to
4
100
3
eliminate the smoke in the platform and help the
3 3 3
50
3 63 6 6 passenger escape to safety zone.
07
10
2
3
4
5
6 3 63 7
13
2
5 17 6
2 1000
5 157
2
500 , ,
1500
Time (s)
Fig. 11 shows CO distribution in the platform with
(a) Case 1 (10 MW) different fire strengths. The CO densities are
500
1 1 calculated at seven points in the platform. The soot
2 2
400
3
4
3
4
density is 50% increased with stronger fire strength
5 5 (10 MW)
6 6 20 MW in the platform. With main tunnel ventilation,
7 7
CO (ppm)
300
however, the soot is decreased by 65%-80%.
200
Interesting thing is the soot density has become steady
state after 1,000 s since fire occurrence in both 10 MW
100
4 4 4 4
and 20 MW case. It seems that, the steady state
4 4 4
circumstance is dependent on total ventilation capacity
0
0
3
1 500
2
5
6
7 3
7
1
2
5
6 ,
1000
3
7
1
2
5
6 ,
1500
Time (s) in the station and tunnel.
(b) Case 2 (10 MW)
500
1 1
4. Conclusions and Summary
2 2
3 3
400 4 4 CO distributions generated by the fire in the
5 5
(20 MW)
6
platform with different fire strengths in the subway
6
7 7
4
CO (ppm)
300
4 4 4 4 4 4 4
4 4 4
station are calculated. The Shin-Gum-Ho station in the
200 Seoul is modeled. The smoke flow due to the fire is
4
analyzed and then the distribution of CO is calculated.
100 3
3 3 3 3 3 LES and parallel computing is employed to simulate
36 36 3 3
3 3 6 3 6 the fire-driven flow in the whole station.
07
10
2
3
4
5
6 61 7
2 500 1 7
5 2 1000
5 1 7 2
5 11500
, ,6
Time (s) The flow in the stairway and middle floor goes
(c) Case3 (20 MW) down to the platform due to the low pressure caused
500
1 1
by strong ceiling ventilation in the platform. Main
2 2
3 3
400 4
5
4
5
tunnel ventilation induces much increased going-down
(20 MW)
6 6
7 7 flow in the stairway. With fire strength 10 MW in the
CO (ppm)
300
platform, the velocity in the stairway is about 0.9 m/s
200 or less. With fire strength 20 MW in the platform, the
4 velocity in the stairway is about 1.1 m/s which is
100
4
4
similar with 10 MW case. The back flow is slightly
4 4 4 4
07
1
2
3
4
5
6 3 500
7
1
2
5
6 37
1
2
5
6 37
1
2
5
6 generated near the top of middle floor because of
0 ,
1000 ,
1500
Time (s) buoyancy force with fire strength 20 MW. The
(d) Case 4 (20 MW) magnitude of velocity in the stairway with main
Fig. 11 Calculated CO in the platform with different fire
strengths. tunnel ventilation is two times larger than without
tunnel ventilation. The soot density is 50% increased [4] Fletcher, D. F., Kent, J. H., and Apte, V. B. 1994.
Numerical Simulations of Smoke Movement from a Pool
with stronger fire strength 20 MW in the platform.
Fire in a Ventilated Tunnel. Fire Safety Journal 23 (4):
With main tunnel ventilation, however, the soot is 305-25.
decreased by 65%-80%. [5] Gao, P. Z., Liu, S. L., Chow, W. K., and Fong, N. K.
2004. Large Eddy Simulation for Studying Tunnel
Acknowledgment Smoke Ventilation. Tunneling and Underground Space
Technology 19 (6): 577-86.
This research is supported by a grant (PK1606C9) [6] Jang, Y. J., Lee, H. S., and Park, D. S. 2012.
from Korea Railroad Research Institute. Experimental Study for the Capacity of Ordinary and
Emergency Ventilation System in Deeply Underground
References Subway Station. Journal of the Korean Society for
Railway 15 (6): 579-87.
[1] Hwang, C. C., and Edwards, J. C. 2005. The Critical [7] Jang, Y. J., Ryu, J. M., Park, S. H., and Koo, D. H. 2015.
Ventilation Velocity in Tunnel FiresA Computer Analysis of Smoke Distribution in the Subway Station
Simulation. Fire Safety Journal 40 (3): 213-44. with Various Main Tunnel Ventilations. Journal of
[2] Hwang, C. C., and Wargo, J. D. 1986. Experimental Energy and Power Engineering 9 (4): 405-10.
Study of Thermally Generated Reverse Stratified Layers [8] McGrattan, K., McDermontt, R., Hostikka, S., and Floyd,
in a Fire Tunnel. Combustion and Flame 66 (2): 171-80. J. 2010. Fire Dynamics Simulator (Ver. 5) Users Guide.
[3] Jang, Y. J., Kim, H. B., Kim J. H., and Han, H. Y. 2009. Washington: NIST (National Institute of Standards and
Comparative Study on the Numerical Simulation for the Technology).
Bake-Layer of The Tunnel Fire-Driven Flow with LES [9] Smagorinsky, J. 1963. General Circulation Experiments
and RANS. Transactions of the KSME (B) 33 (3): with the Primitive EquationI. The Basic Experiment.
156-63. Monthly Weather Review 91 (3): 99-164.
doi: 10.17265/1934-8975/2016.03.009
D DAVID PUBLISHING
Redesign of Motorized and Non-motorized Transport in

Cities and Sustainable Mobility
Osman Lindov1, Adnan Omerhodi1, Adnan Tatarevi2 and Samir Daferovi3

1. Department of Traffic, Faculty of Traffic and Communications, University of Sarajevo, Sarajevo 71000, Bosnia and Herzegovina
2. Department Approval of the Vehicle Type, Institute of Traffic and Communications, Sarajevo 71000, Bosnia and Herzegovina
3. Department for Road and Rail Transport, Ministry of Communications and Transport, Sarajevo 71000, Bosnia and Herzegovina
Received: December 21, 2015 / Accepted: January 06, 2016 / Published: March 31, 2016.
Abstract: Redesign of motorized and non-motorized transport in cities and sustainable mobility possibilities and utilization of
already occupied areas for motorized and non-motorized transport in cities. Reserved area for public transport may be the redesign
much more to take advantage of without taking up new space in cities. Redesigned solutions shown in the work of public mass
transport and the redesign of non-motorized transport, bicycle and pedestrian paths point to the improved use and safety of movement
of passengers, cyclists and pedestrians. This paper presents five redesigned concept designs as improving existing forms of transport
and movement of cyclists and pedestrians in cities. Redesigned conceptual designs of motorized and non-motorized transport in cities
should serve as ideas for the growing problems of urban development in the segment of insufficient surface for pedestrians and
cyclists, as well as sustainable mobility transport people in the city.
Key words: Redesign, motorized transport, non-motorized transport, sustainable mobility.
1. Introduction priority. Given the stated, it would be too long to wait

for another next 30 years to implement the plans
Final goal of global directions of the European
stated in strategic documents. It is necessary to react
cities development, including Sarajevo, in the period
adequately and promptly, and that it possible only
untill 2023 is a qualitative, comfortable, safe and
through the innovative and fresh solutions that have
reliable transport, fully adjusted to the demands of
to be valorized in the cities [1]. Most of developed
passengers. According to stated plans, planned
European cities have action plans and established
development of the public passenger transport system
development strategies for sustainable transport, and
shall be based on the inclusion of railway into the
they intensively work on informing and educating
passenger transport system, as well as the
residents on necessity of transformation of the existing
transformation of the existing tramway passenger
mode of transport into the sustainable one [2].
service into the system of LRT (light rail transport).
Experiences of the cities that have already passed the
These plans have existed for the last 30 years and are
initial phases of project implementation show that,
planned for the next 30 years. However, from the
education of residents has to be the first step in
security and ecologic aspect, cities become
fulfilling such goals. Some of the advantages of
problematic because certain plans are not fully
sustainable transport are: healthier life of residents,
realized as planned. Those plans did not include the
less noise, less pollution and nicer city. And if the
segment of sustainable development that is today the
residents are not familiarized with these facts, and if
they do not recognize them as their own, and there is
Corresponding author: Osman Lindov, Ph.D., research no since if residents are not familiar with these facts,
fields: traffic safety and environmental traffic.
Redesign of Motorized and Non-motorized Transport in Cities and Sustainable Mobility 207
and if they do not recognize them [3].
2. Redesign of Road ProfileInnovative

Solutions for Motor, Pedestrian and Bicycle
Paths
In the last 30 years, there were no significant
changes in the width of road profile. Both pedestrian
and motor traffic, as well as bicycle paths are built at
the same level. Innovative solutions propose the Fig. 1 Redesign of road profile (motor, bicycle and
pedestrian paths).
construction of the road above the level, in other
words, to construct the surface for motor transport at
the separate level, and to elevate the cycling surface,
and then to construct the pedestrian paths above that
with pellucid panels above the pedestrian and cycling
lane [4]. Such projecting and construction will provide
significantly increased road safety (Fig. 1).
3. Redesign of the Miljacka River (a)

BedInnovative Solutions for
Pedestrian/Bicycle Paths
Alternative modes of transport in the cities will be
required, primarily cycling connected and linked with
adequate pedestrian paths, and more detailed and
precise planning of free space that will be used for
(b)
such alternative modes of transport [5]. Redesigning Fig. 2 Redesignedinnovative solution for Miljacka River
of the Miljacka River bed that stretches from the bed.
central part of the city towards the periphery in the
total length of 10 km, implies putting into operation
4. Redesign of Rail Tramway
the space occupied by the river itself. The redesign
BeltInnovative Solutions for Tramway
itself of the Miljacka River bed will be characterized
Transport
by the construction of the terraces on the both sides of Occupied belt for tramway transport in Sarajevo is
the river with pedestrian and cycling paths (Fig. 2). a designed and occupied space used for lower and
The aim of the redesign is to provide new look of the upper rail machines transport. It currently covers a
bed with modern and useful paths, without any huge belt and already in the city central parts creates
damage to the river and its importance and role for significant problems due to noise emission [7].
urban area. Innovative solutions provide solution for tramway
Redesign of the Miljacka River bed will contribute transport that will be pleasant and adequate for the
to greater safety of pedestrians and cyclists because it environment, where the route of the rail will partly be
will divide physically separated terraces [6]. underground and above the platform there will be
(a) Before (b) After

Fig. 3 Redesignedinnovative solution for tramway belt.
useful multi-purpose space (bicycles, etc.). Tramway spatial and environmental pollution issues. Travelators
transport in Sarajevo longitudinally stretches for some (moving pavement) will be placed along the space of
14 km with the belt width from 13 m to 25 m. Descent the existing tramway rails and it would have secured
of the rail in the ground to the level of tram windows space, physical divided from the vehicle traffic, with
will provide visual contact with the area, avoiding the pellucid panels visually and reasonably in harmony
unpleasant feel of the metro and other systems that with the travelors lane. In the crossroads, in other
are completely situated underground. In some parts of words, in parts where motor traffic intersects the
the crossroads, this system will be completely situated streets with travelors, adequate crossings have to be
under the ground and then again to the level of tram put in place in order to secure pleasant crossing to the
windows. This type of solution on the crossroads will other part of the trailor, and thus ensure nice,
secure constant flow of tramway transport, without comfortable and safe transport along the entire length
stopping, which would lead to an increased transport of the rail (Fig. 4).
flow and transport safety. The platform above the
6. Pedestrian Zone RedesignInnovative
route may be designed and used for multi-purpose
Solutions for Pedestrian Routes
usage, such as pedestrian paths, cycling lanes, as well
as the space for emergency services (Fig. 3). Redesigning in pedestrian zonespedestrian routes
innovative solutions, as initial and final solution,
5. Innovative Type of Transport in the Old
envisage an increase in time that pedestrians spend in
Part of Urban AreaAssembly Line
the pedestrians zones due to different weather
Since rail vehicles (tramways) inevitably make conditions (sunshine, rain, snow, etc.). This innovative
noise in the areas of their passage, their usage is solution includes roofing the pedestrian zones. That
limited, particularly in the old parts of urban areas [8]. would include of pleasant transparent panels that will
The city of Sarajevo has had, and still has, one of the be used as protection against sun, rain and other
first tramway system that passes through the old part weather conditions. These panels will also be movable
of the city, and if adequate noise protection standards and placed above the pedestrian zone in the bad
would be applied, its usage will be prohibited in that weather conditions (sun, rain or snow). This would
part of the city. The alternatives are trolleybus or bus affect the increased flow of pedestrians daily and
networks. Beside the noise, the cities also face with throughout the year, that would lead to a decrease in
Redesign of Motorized and Non-motorized Transport in Cities and Sustainable Mobility 209
(a) Before (b) After

Fig. 4 Innovative solution for transport in old part of citymoving walkway.
movement and use of motor vehicles in that part of the

city. Day which will, again, decrease the movement
and usage of motor vehicles in that part of the city. At
the same time, cleaner, nicer, more pleasant, more
human and safer city would be provided (Fig. 5).
In the central city zone, where pedestrian paths are
situated alongside the main city road, there is a
Fig. 6 Redesigninnovative solution for pedestrian path

alongside the road.
(a)
possibility to build adequately designed domes or
arches above the pedestrian routes that would be adjusted
to the buildings the arches would be connected to. The
building with which the arches will be connected
(Fig. 6).
(b) 7. Conclusions
Redesigned, or in other words, innovative solutions
for traffic and transport flows, are impracticable in the
short run, mostly as result of slow changes of
standards in designing and construction. Redesigned
and innovative solutions presented in this paper,
authors incentives to consider them as possible
(c)
Fig. 5 Redesigninnovative solution for pedestrian path. solutions in urban areas from the aspect of sustainable
development of cities worldwide. Stockholm: Royal Institute of Technology.

[3] International Engineering and Consulting Group
Most of transport solutions today are classic and
SYSTRA (Systmes de Transport, Transport Systems).
expensive, and there is almost no place for innovative 2009. A Study of Optimization of the City Transport
and non-standard solutions. Network and Making Long-Term Plan for Transport
Certainly, each of solutions described in this work Canton Sarajevo. International Engineering and
Consulting Group SYSTRA.
is not new and the once that cannot be used elsewhere.
[4] Lindov, O., and Omerhodi, A. 2013. Analysis of
Also, technologies presented in this work for certain Efficiency of Road Traffic Improvement MeasuresLocal
solutions are not unknown nowadays, they are only Activities and Best International Practice. Suvremeni
being applied in other fields and purposes. Authors are PrometModern Traffic 33 (1-2): 262-7.
[5] Stoki, D. 2009. Education of CitizensThe Main
aware that, each of the presented solutions may be
Precondition for Sustainable Transport. Faculty of Civil
improved and further developed along with the Engineering and Architecture in Nis, Science + Practice
technological development. 12 (1): 224.
[6] Lindov, O., Omerhodi, A., Tatarevi, A., and
Reference Daferovi, A. 2015. Redesign of Motorized and
Non-motorized Transport in Cities and Sustainable
[1] Organisation for Economic Co-operation and Mobility. Presented at the International Conference
Development. 2002. Implementing Sustainable Urban Towards a Human City, Vojvodina, Novi Sad.
Travel Policies, Key Messages for Governments. In [7] Lindov, O., Tatarevi, A., and Daferovi, S. 2013.
Proceedings of the ECMT (European Conference of Modern Methods of Road Audits in Order to Increase
Ministers of Transport), 13-8. Traffic Safety. Suvremeni PrometModern Traffic 33
[2] Archer, J. 2005. Indicators for Traffic Safety Assessment (1-2): 277-82.
and Prediction and Their Application in Micro-simulation [8] Lindov, O. 2012. Traffic Culture. Faculty of Traffic
Modelling: A Study of Urban and Suburban Intersections. and Communication of the University of Sarajevo.

Journal of Energy and Power Engineering - 2016.03

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

Journal of Energy and Power Engineering - 2016.03

Încărcat de

Drepturi de autor:

Formate disponibile

Journal of Energy

and Power Engineering

David Publishing Company

David Publishing Company

David Publishing Company

166 A New Generator for Tsunami Wave Generation

Power and Electronic System

191 Energy Efficiency of Reactive Dynamic Compensators

Potentials of Cellular Vortex Element Modeling of Fluid

Oyetunde Adeoye Adeaga1 and Ademola Adebukola Dare2

Key words: Numerical, Darcy, vorticity, image, subsurface porous media.

Start Evaluating coefficients

Read media data

Perform accelerative, diffusive &

Return new pressure head and

Loading aquifer data

Write new pressure head and

Reading data from table

Initialize time step

Reading initial Stop

Calculating fluid and

Fig. 1 Complete simulation flowchart.

Input aquifer and

Reading grid data. Generating coordinates

Initializing time step and number of runs

Runs equals preset?

Fig. 2 Flowchart for evaluating coefficients during simulation.

Table 1 Summary of data used for simulation.

Fig. 3 3D graphical representation of aquifer condition after two months.

Fig. 4 3D graphical representation of aquifer condition after ten months.

Fig. 5 Variation of aquifer average drawdown with layer.

Fig. 7 Average aquifer drawdown pressure with layer. Existing work.

Fig. 8 Variation of average aquifer pressure head with layer.

Table 3 ANOVA of results of simulation.

6. Conclusions and Recommendations capture a little more of reality.

6.1 Conclusions References

Numerical Analysis of a Gravity Substructure for 5 MW

Min-Su Park, Youn-Ju Jeong and Young-Jun You

1. Introduction been conducted on wind energy [1-3]. In order to

where, the suffixs x, and z denote the component of

Figs. 4 and 5 show the panel pressure and wave run

3.2 Wind Turbine Model and Environmental Loads

Table 2 DLC for ultimate limit state.

Table 3 Environmental loads.

3.3 Natural Frequency and Resonance

In order to evaluate the resonance between the wind

Table 5 Stiffness and damping matrices according to Vs30.

Table 6 Natural frequencies of gravity substructure for

to the spectrum density of the turbine and the

Table 8 Structural results due to water depth conditions.

Fig. 10 Structural von-Mises stress of gravity substructure

Using the wave forces from ANSYS AQWA and

Production of Rhenium by Transmuting Tungsten Metal

Tsugio Yokoyama1, Yuki Tanoue2, Atsunori Terashima2 and Masaki Ozawa1

1. Introduction The neutron energy spectrum of FRs (fast reactors)

3.7d 17h 24h

4.03 0.015 0.0021

75d 24h 70d

Fig. 1 Nuclear transmutation scheme of tungsten (W) to rhenium (Re).

power of the core is 710 MWt, the coolant is liquid

1st row : 9 pins

2nd row : 9 pins

Outer core 43 tungsten pins