Distributed Energy Systems

DISTRIBUTED ENERGY SYSTEMS No part of this digital document may be reproduced, stored in a retrieval system or transmitted in any form or by any means. The publisher has taken reasonable care in the preparation of this digital document, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained herein. This digital document is sold with the clear understanding that the publisher is not engaged in rendering legal, medical or any other professional services.

DISTRIBUTED ENERGY SYSTEMS

SHIN’YA OBARA

Nova Science Publishers, Inc. New York

Copyright © 2009 by Nova Science Publishers, Inc.

All rights reserved. No part of this book may be reproduced, stored in a retrieval system or transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical photocopying, recording or otherwise without the written permission of the Publisher. For permission to use material from this book please contact us: Telephone 631-231-7269; Fax 631-231-8175 Web Site: http://www.novapublishers.com NOTICE TO THE READER The Publisher has taken reasonable care in the preparation of this book, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained in this book. The Publisher shall not be liable for any special, consequential, or exemplary damages resulting, in whole or in part, from the readers’ use of, or reliance upon, this material. Independent verification should be sought for any data, advice or recommendations contained in this book. In addition, no responsibility is assumed by the publisher for any injury and/or damage to persons or property arising from any methods, products, instructions, ideas or otherwise contained in this publication. This publication is designed to provide accurate and authoritative information with regard to the subject matter covered herein. It is sold with the clear understanding that the Publisher is not engaged in rendering legal or any other professional services. If legal or any other expert assistance is required, the services of a competent person should be sought. FROM A DECLARATION OF PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE AMERICAN BAR ASSOCIATION AND A COMMITTEE OF PUBLISHERS. LIBRARY OF CONGRESS CATALOGING-IN-PUBLICATION DATA ISBN: 978-1-60741-217-5 (E-Book) Available upon request

Published by Nova Science Publishers, Inc.    New York

CONTENTS Chapter 1 Chapter 2 Chapter 3 Index

Operation Plan of a Combined Fuel Cell Cogeneration Using Genetic Algorithm

1

Route Planning of Heat Supply Piping in a Fuel Cell Energy Network

21

Load Leveling of Fuel Cell System by Oxygen Concentration Control of Cathode Gas

41 65

Chapter 1

OPERATION PLAN OF A COMBINED FUEL CELL COGENERATION USING GENETIC ALGORITHM INTRODUCTION From deregulation of energy business, and an environmental problem, the installation spread of the small-scale distribution power due to a fuel cell and a heat engine is expected. Under the objective function set up by the designer or the user, optimization planning that controls small-scale distribution power is required. In dynamic operation planning of the energy plant, the analysis method using mixed integer linear programming is developed [1, 2]. For the compound energy systems of solar modules and fuel cell cogeneration, there have been no reports of the optimization of operation planning. Therefore, there are no results showing the relationship between the objective function given to the combined system and operation planning. Such as a solar modules or wind power, greenenergy equipment is accompanied by the fluctuation of an output in many cases. Almost all green energy equipment requires backup by commercial power, fuel cells, heat engines, etc. Operation planning of the system that utilizes renewable energy differs by the objective function and power and heat load pattern. Thus, this chapter investigates the operation planning of the compound energy system composed of proton exchange membrane fuel cell cogeneration with methanol steam-reforming equipment, a solar module, geo-thermal heat pump, heat storage, water electrolysis equipment, commercial power, and a kerosene boiler. In such a complex energy system, facility cost is expensive. However, in this chapter, it investigates as a case of the independent power source for backlands with renewable energy. This chapter considers the operation planning of a system, and

2

Shin'ya Obara

the optimization of equipment capacity. The Genetic Algorithm (hereafter described as GA) applicable to a nonlinear problem with many variables is installed into the optimization calculation of the operation planning of the system [3]. In the operation analysis of a complex energy system, Mixed Integer Programming (MIP) other than GA can be used. Because the nonlinear analysis using MIP is made to approximate using a linear expression of relations, it is considered that an error is large. On the other hand, GA is applicable to the analysis of the nonlinear problem of many variables. The range of the analysis accuracy obtained by calculation with GA is understood that it can use industrially. In GA, the design variable of energy equipment is shown with many gene models. In this chapter, the objective functions given to the system were set up as (1) Minimization of error in demand-and-supply balance, (2) Minimization of the operation cost (fuel consumption) of energy equipment, (3) Minimization of the carbon dioxide gas emission accompanying operation, and (4) Minimization of the three objective functions described above. The load pattern in winter (February) and summer (August) of the average individual house in Sapporo, Japan, is used for the energy demand model shown with a case study [4]. This chapter described the operation plan of the independence energy system when installing a methanol steam-reforming type fuel cell and renewable energy into a cold region house. Such complex operation optimization of the energy system did not have a report until now. Consequently, the method of installing and analyzing the GA apply to the nonlinear problem of many variables was proposed. In points of equipment cost, it is difficult for a proposed system to spread generally. However, the installation to the area where the commercial power is not fixed is possible.

FUEL CELL, SOLAR MODULES, AND GEO-THERMAL HEAT PUMP COMBINED SYSTEM Scheme of Combined System Figure 1 shows the energy system scheme examined in this chapter. A combined system consists of a solar module (18), PEMFC-CGS (PEMFC: proton exchange membrane, fuel cell CGS: co-generation, the fuel cell is (1), the reforming equipment is (2)-(5) and (12), geo-thermal heat pump, (17), boiler, (8), commercial electric power, heat storage tank (10), and the water electrolysis equipment is (13)-(15)). Water electrolysis equipment is used to store electrical

Operation Plan of a Combined Fuel Cell Cogeneration…

3

power with hydrogen and oxygen. The arrowhead in this figure shows the substance or direction of energy flux. Each system of solar module, commercial electrical power, and PEMFC-CGS is changed with a changeover switch (6), and electrical power is supplied to the consumer. However, electrical power is not at once supplied to the demand side from two or more power systems. An electric heater (9) is installed inside the heat storage tank, and electric power is changed into heat and can be stored. Hydrogen and oxygen can be produced if electric power is supplied to an electrolysis tank (13). Hydrogen and oxygen are stored in tanks (14) and (15), respectively, and these are supplied to PEMFC and can be generated at an arbitrary time. When the heat produced by the geo-thermal heat pump exceeds the quantity demanded, surplus heat is stored in the heat storage tank. Although the exhaust heat of PEMFC and the methanol steam-reforming equipment is also supplied to the heat storage tank, when the total amount of heat exceeds the heat storage capacity, heat is radiated with a radiator (7). Tap water has heat exchanged for the heat transfer medium inside the heat storage tank, and moreover controls the temperature of this tap water by the boiler, and supplies hot water to the consumer. HEX : Heat ex changer Commercial power network Electric po wer s upply

6 18

11

16

5 14 15

13 To 10

4

1 To 10 Tap water

7 9 10 HEX

3 2 12

Exhaust 8

Hot water supply 19

P HEX

17

G.L.

Methanol fuel Methanol fuel

Kerosene

1. Fuel cell s tack, 2 . Vaporizer, 3. Reformer , 4. Shifter, 5. CO oxidation , 6. Change over s witch, 7. Radiator , 8. Back-up boiler , 9. Electric heater , 10. Heat Storage tank , 11. DC/AC converter, 12 . Catalytic combustor , 13. Electrolysis tank , 14. H 2 tank, 1 5. O2 tank , 16. Chan ge o ver switch , 17. Geothermal heat pump system , 18. Solar mo dules , 19. Compress or

Figure 1. PEMFC-CGS, Heat-pump and solar module combined system for houses.

Shin'ya Obara Heat dem and ( kW)

4 20

February

15 10 5

August

Power demand [kW]

0 0.8 0.6 0.4 0.2 0

0

4

8

12

16

20

24

Time (Hour)

Energy output (kW)

Figure 2. Energy demand of Sapporo-city (Narita et al, 1996).

Thermal power Electric power

10

5

0 0

0.01

0.02

0.03

0.04

M ethanol supply (mol/s)

Figure 3. Characteristics of fuel cell stack with methanol steam reforming (Obara et al., 2003).

Methanol fuel is supplied to the reformed gas system of the methanol steamreforming equipment, and the catalytic-combustion equipment (12) installed in the evaporator (2). Kerosene fuel is supplied to the boiler (8). The energy demand pattern used for analysis is a model in February (winter) and August (summer) in the average individual house in Sapporo in Japan, and shows this in Figure 2. For

Operation Plan of a Combined Fuel Cell Cogeneration…

5

Sapporo, a cold, snowy area, the annual average temperature is 288 K, and the mean temperature in February and August is 269 K and 294 K, respectively. The operating period of a system is made into 23:00 from 0:00 of a representation day, and sampling time is expressed by t k (k = 0 , 1, 2, ....., 23) . The initial values of the capacity of each energy device set up the value used for the usual individual house. The specifications of each energy device are shown in Table 1. Table 1. Energy device specifications Solar module Area Electric energy output Fuel cell Type Fuel Reforming type Electric energy output Thermal energy output Commercial power Heat pump Type Energy source p-h diagram Thermal energy output COP Electrolysis device [8] Electrolysis efficient Accumulation of electricity Backed boiler Fuel Efficincy Thermal energy output Thermal storage tank Thermal storage capacity Heat medium temperature Thermal storage efficiency

6.0 m2 3kW (Maximum) Proton-exchange membrane fuel cell Water/Methanol=1.4/1.0 (mole ratio) Methanol steam reforming Maximum 3kW 9kW (Maximum) 5kW (Maximum) Geothermal heat source Electricity See Fig.4 5kW (Maximum) 3.0 0.85 (Constant) 180MJ Kerosene 0.85 (Constant) 40kW (Maximum) 180MJ 353K (Maximum) 0.95

Compared with the condition of the steady operation of the methanol reformer, the characteristics of a startup and a shutdown differ greatly. Cold start operation and shutdown operation require about 20 minutes, respectively. In the analysis of this section, it is assumed that the startup of the methanol reformer is always a hot start.

6

Shin'ya Obara

Relational Expression (1) Energy Output of PEMFC-CGS 3kW methanol steam reforming type PEMFC shown in Figure 3 is used for the output characteristic of the fuel cell introduced into analysis [5]. The horizontal axis of Figure 3 is divided into two or more zones, and the output characteristics are given by the analysis program by using the secondary leastsquares method approximation for each range. The electric power output at the time of supplying and generating hydrogen and oxygen stored by water electrolysis to a fuel cell is expressed by Equation (1). EFS,tk = Ic,tk ⋅ EV ,tk − ΔWFS,tk =

Qf , tk ⋅ Fd ⋅ EV ,tk − ΔWFS,tk Ec

(1)

Here, Ic,tk , EV ,tk , ΔWFS,tk , Q f , tk , Fd and Ec express current, voltage, power loss of a cell stack, hydrogen amount of supply, Faraday constant and chemical equivalent, respectively.

(2) Heat Output of Geo-thermal Heat Pump Figure 4 is a p-h diagram of Refrigerant HC-12a [6], used by the geo-thermal heat pump [7, 8]. This refrigerant is a mixed refrigerant of propane, butane, and isobutene. Although the output characteristics of the heat pump were the analysis of soil temperature TL and condensation temperature TH exactly, coefficient of performance COP t k was set to 3.0 in this section. 2000

340K

Pressu re [kPa]

1400

T H=323K 320K

1000

300K 600 280K

400 300

273.15K

TL=277K

260K 200 100

30 0

500 En thalpy [kJ/kg]

700

Figure 4. p-h diagram of Refrigerant AHC-12a (HC-TECH Inc., 1997).

Operation Plan of a Combined Fuel Cell Cogeneration…

7

(3) Characteristic Equation of Water Electrolysis Tank From sampling time t k to Δt , the electric power supplied to an electrolysis tank is expressed by E EL,tk , and the efficiency of the electrolysis tank is expressed by φ EL . In this case, hydrogen quantity QH 2 ,tk to be produced is calculated by Equation (2). Moreover, the amount of production of oxygen is similarly calculated. The efficiency of water electrolyzer refers to the results of a study [9] that used the proton exchange membrane, and φ EL is set as 0.85.

Q H 2 ,t k =

E EL,tk ⋅ E c ⋅ φ EL Fd ⋅ EV

(2)

In calculations for this case study, the hydrogen and oxygen are pressurized to 1.0 MPa, respectively. The work of the compressor is assumed to be compression work for an ideal gas. The whole compressor efficiency including an inverter controller loss and the power consumption in an electric motor, transfer loss of power, loss with insufficient air leak and cooling, and other machine losses is set up to 50%.

(4) Characteristic Equation of Heat Storage Tank and Boiler The conditional expression showing heat storage characteristics is given by Equation (3) and Equation (4) using the amount of maximum heat storage S St , max , and maximum temperature TSt ,max of the heat medium. The capacity and the specific heat of the heat storage medium are expressed by V and C p , and outside air temperature is expressed by T∞ (The heat medium is assumed to be calcium chloride). Moreover, the heat storage temperature at time t k is calculated by TSt ,t k = S St ,t k /( ρ ⋅ C p ⋅ V ) . Here, ρ express density of heat storage medium.

0 ≤ S St ,tk ≤ S St , max

(3)

T∞,tk ≤ TSt ,t k ≤ TSt , max

(4)

The characteristic equation of heat storage tank between time t k and Δt is given by Equation (5). S St ,t k − S St ,t k −1 = {H St ,in ,t k − H St ,out ,t k − φ St ⋅ ρ ⋅ C p ⋅ V ⋅ ( TSt ,t k − T∞,t k )}⋅ Δt (5)

Shin'ya Obara Thermal energy output [kW]

8

20 15 10 5 0 0

0.00025

0.00050

0.00075

0.0010

Fuel consumption [kg/s]

Figure 5. Thermal energy output of boiler.

Solar module area 6m2

Quantity of solar isolation [MJ]

5.0 4.0

August

3.0 2.0 February 1.0 0 5:00

9:00

13:00

17:00

Time [O’ clock]

Figure 6. Time change of solar module output (Nagano et al., 2002).

H St ,in,tk and H St ,out,tk express the heat input and heat output of the heat storage tank, respectively. The third term in the right-hand bracket of Equation (5) includes outside air temperature T∞,tk supposing heat storage loss is dependent on outside air temperature. However, in the analysis in this section, the efficiency of heat storage φ St is set to 0.95, and change in outside air temperature is not taken into consideration. Figure 5 shows the relationship between the fuel consumption of a boiler and hot-water-supply output. It is expressed with the calorific value of the fuel being α Boiler , the boiler efficiency being φ Boiler , and the fuel-supply

Operation Plan of a Combined Fuel Cell Cogeneration…

9

quantity of flow being FBoiler ,t k , and the characteristic equation of a boiler is given by the following equation.

H Boiler,tk = α Boiler ⋅ FBoiler,tk ⋅ φ Boiler

(6) E .S : E le ctr ic energy supply T .S : T herma l energy supply

Devices output Sa mpling time tk Electr ic ener gy stor age

System output

k −1 M

(∑

∑ S E, D , t

k

≥ EN eed ,t k )

E.S

(∑

∑ SE , D , t

k

< E Need , t k )

E.B

( ∑ ∑ S St,Di ,tk ≥ H Need, tk )

T.S

i

k =0 i = 1 k −1 M

SE ,Di ,t k

i

k =0 i =1

k

(∑

M

∑ S E,D ,t

i k

− E Need,tk )

(b)

( c)

k =0 i=1

Sst ,Di ,tk

k −1 M

k

( ∑ ∑ SS t , Di , tk < H Need, t k )

T.T

(ED i ,t k ≥ ENeed, t k )

i k

E.S

( ∑ ∑ S E,Di ,tk − E Need,tk ≤ S E,max )

k

E.B

k=0 i=1 k M

M

( ∑ ∑ S E , D i ,t k − E Need, t k > SE ,max ) M

Thermal energy output

T.T

k = 0 i =1

(∑ H Di ,tk ≥ H Need,tk ) (∑ H D i , t k + i =1

M

k

M

∑ ∑ S St, D , t

k = 0 i =1

i

k

T.T

(g)

k=0 i =1

(h) M

− H N eed , t k ≤ H S t, max )

( f)

(∑ H Di ,tk + i=1

k M

∑ ∑ SSt,D ,t

k=0 i =1

i k

− H Need ,tk )

(i)

k M

(∑ H Di ,tk + ∑ ∑ SSt,Di , tk − H Need ,tk > H St,max ) i=1

( ∑ ∑ SE ,Di ,tk −E Need, tk − S E,max )

T.S

i=1

M

( e)

k M

( ∑ ∑ S E,Di ,tk − ENeed ,tk )

k =0 i=1 k

(d)

k = 0 i =1

k M

( E D ,t < EN eed ,t )

M

( ∑ ∑ S St , Di ,t k − H Need, t k )

k = 0 i =1

Electr ic ener gy output

H Di ,tk

P ath ( a)

k=0 i=1

k−1 M

T her mal e ner gy stor age

ED i , t k

E.B : Electric energy storage T.T : Thermal energy stor age

Branch path

k=0 i=1

T.T and release (release )

(j)

M k M ⎛ ⎞ ⎜ H St,max − (∑ H D ,t + ∑ ∑ SSt,D ,t − H Need,t )⎟ ⎜ i k i k k ⎟ ⎝ i=1 k=0 i =1 ⎠

Figure 7. Energy supply path.

(5) Characteristic Equation of Solar Module Figure 6 shows the results of measurement of the production of electricity of the solar module in February and August in Sapporo [10]. However, the panel was vertically installed so that this solar module would not be covered in snow in winter. Therefore, the production of electricity decreases as shown in the results of Figure 6 at 13:00.

Energy Supply Path The energy equipment is expressed by

Di , and let subscript i

( i = 1 , 2 , 3 ,..., M , M are the number of pieces of equipment) be the equipment number. The electric power and heat that are outputted by energy device Di

10

Shin'ya Obara

follow one path (a) to (j) as shown in Figure 7. When electric energy E D i , t k generated by the system exceeds power demand E Need

,t k

, hydrogen and oxygen

are produced and stored by water electrolysis. Moreover, it is also possible to change electric power into heat with a heater, to shift time, and to supply the demand side.

ENERGY BALANCE AND OBJECTIVE FUNCTION Objective Function of System The objective function given to the system is given by (1) Minimization of error in demand-and-supply balance, (2) Minimization of the operation cost (fuel consumption) of the energy equipment, (3) Minimization of the carbon dioxide gas emission accompanying operation, and (4) Minimization of the three objective functions described above. Equation (7) and Equation (8) are energy balance equations of electric power and heat, respectively. E FS ,tk + EUtility,tk + E Stp,tk = E System,tk + ΔE EL,tk + ΔE HP,tk +ΔE CPH,tk +ΔE CPO,tk + E H ,tk

(7)

αFS ⋅ FFS,tk ⋅φFS + αBoiler ⋅ FBoiler,tk ⋅φBoiler + HHP,tk + HSt,tk = HSystem,tk + HRad,tk + ΔHSt,tk (8) The left-hand side in Equation (7) and Equation (8) is the amount of energy inputted into the system, and the right-hand side expresses the amount of energy outputted from the system. Here, E FS ,tk , EUtility,tk , E Stp,tk and E System,tk express electric power of fuel cell stack, commercial power and power storage, respectively. ΔE EL,tk , ΔE HP,tk , ΔE CPH ,tk , ΔE CPO,tk and E H ,tk express power consumption of electrolyzer, heat pump, hydrogen compressor, oxygen compressor and heater, respectively. αFS and αBoiler express calorific value of fuel of fuel cell stack and boiler. φFS and φBoiler express efficiency of fuel cell stack and boiler. FFS,tk and FBoiler,tk express fuel quantity of flow of fuel cell stack and boiler. HHP,tk , H St,tk , H System,t k , HRad,tk and ΔHSt,tk express heat of heat pump, heat storage tank, system, radiator and heat storage loss, respectively.

Operation Plan of a Combined Fuel Cell Cogeneration…

11

Objective function (1) described in Introduction is an operating pattern when the difference in input-output of energy balance Equation (7) and Equation (8) serves as the minimum. Objective function (2) is an operating pattern when fuel cost and commercial power cost serve as the minimum. The operation cost of equipment Di between time t k and Δt is calculated from fuel flow rate FDi ,t k and unit fuel price C fuel , Di that are supplied to the equipment. Therefore, the operation cost of the whole system is calculated by Equation (9). Here, CUtility,tk

express

commercial power cost. M

CSystem,t k = ∑ (C fuel, Di ⋅ FDi ,t k ⋅ Δt ) + CUtility,tk

(9)

i =1

Objective function (3) expresses the operation pattern whose amount of greenhouse gas discharge calculated from the fuel consumption is the minimum. Amount of emission WSystem,t k of greenhouse gases is calculated by Equation (10). However, the number of gas compositions that contribute to greenhouse gas discharged by equipment Di is expressed by S .

WSystem,tk = ∑∑ (GDi , EX j ⋅ ε Di , EX j ,tk ⋅ FDi ,tk ⋅ Δ t ) M

S

i =1 j =1

Table 2. Energy cost and greenhouse-warming coefficient (Japanese Environment Agency, 2000) Kerosene fuel

Methanol fuel Commercial power

0.01097 Dollar/J 3.099 kg?CO2 /kg 2.026 kg/Dollar 0.01772 Dollar/J 1.379 kg?CO2 /kg 0.0647 Dollar/J (9:00-21:00) 0.01515 Dollar/J (22:00-8:00) 0.000099167 kg?CO2 /kJ

(10)

12

Shin'ya Obara Table 3. Calculation result of each purposes of February (Kerosene (kg)) operation cos t

Minimization of Minimization of operation cost

14.72 (15.36)

the error of demand-and -supply balance 0 .439

greenh ouse g as 13.76 (13.35)

Minimization of the error of demand-and-s upply balance

22.40

0.0170

18.82

Minimization of the amount of green -ho use gas discharge

15.66

0 .426

13.16

Table 4. Calculation result of each purposes of August (Kerosene (kg)) the error of operation greenhouse demand-and cost gas discharge -supply balance Minimization of operation cost

3.61 (4.28)

0.795

3.42 (2.60)

Minimization of the error of demand-and -supply balance

6.85

0.0247

4.174

Minimization of the amount of green-house gas discharge

5.27

0.199

2.55

Here, GDi , EX j expresses a global-warming factor per unit weight of fuel,

ε Di , EX j ,t k being the weight concentration of EX j , and FDi ,t k being the amount of fuel supply to equipment Di . Table 2 shows fuel cost and a global-warming factor [11], and is analyzed using these values in an analysis case.

Multi-objective Optimization As shown in Equation (11), the operation pattern that minimizes the sum that multiplies each objective function by weight is a multiple-objective optimal solution.

Operation Plan of a Combined Fuel Cell Cogeneration…

⎛ Period N p minimize ⎜ ∑ ∑ w j ⋅ f j ( xt k ) ⎜ t =1 j =1 ⎝ k

⎞ ⎟ ⎟ ⎠

13

(11)

In order for the same level to compare the amount of energy loss, operation cost, and amount of greenhouse gas emission, each value was replaced to the amount of kerosene and evaluated. w j in Equation (11) is given beforehand and the value of this equation searches for the minimum solution using GA. Here, w j , f j ( x t k ) and N p express weight of an objective function, objective function and number of objective functions, respectively.

ANALYSIS RESULTS Results of Optimization Tables 3 and 4 are the calculation results when optimizing under each objective function using the energy demand pattern of a representative day in February and August, and are converting all values into a kerosene weight. The number in ( ) of the table is the value of the conventional energy system (using commercial power and a kerosene boiler). If this system is optimized by operation cost minimization, compared with the cost of the conventional system, there will be a maximum of 4% and 16% reduction in February and August, respectively. Reduction rates differ every month because the energy demanded and the solar module output are different.

Equipment Capacity The analysis results of the ratio of maximum output to equipment capacity at the time of planning operation with each objective function (this value is described as RMC below) are shown in Table 5 and Table 6. If the value of RMC is lower than 1, a decrease in the initialized equipment capacity (Table 1) is possible. On the other hand, equipment with a larger value of RMC than 1 has insufficient capacity. The analysis results of RMC of heat storage and power storage (storage of hydrogen and oxygen by water electrolysis) when operating the system under each objective function are shown in Figure 8. From the results of Figure 8, the capacity reduction ratio of a heat storage tank and power storage

14

Shin'ya Obara

equipment can be seen. From the method described above, the optimization of the capacity of each piece of equipment that composes a system can be designed.

RMC [ % ]

80

40

0

RMC [ % ]

(a) February

80

Thermal storage Accumulation of electricity

Maximum

Average 40

0 Mu lt i ob j ec t io

n

Mi Mi n theimiza op eni mi z t e d r at a ti o i e r o - su ma r or n o io n n p pl nd - of f c os of y b a nd t a la nce

Min i gr eth e am izat en h mou ion discouse nt o of har -ga s f g e es

(b) August

Figure 8. Energy storage result.

Objective Function and Characteristics of Operation Plan Figures 9 and 10 show the balance results of electric power and heat when optimizing an operation plan under each objective function using the energy demand pattern of representative February and August days. "System energy output" in these figures is the characteristic of electric power and heat outputted by the system configuration equipment. However, power storage output and heat storage output are not included in this characteristic, but are separately shown as "Storage". Moreover, the energy demanded is shown as "Demand". In Figure 9 and Figure 10, although the characteristic of "Demand" is flat, there ate times when the characteristic of "System energy output" is extremely

Operation Plan of a Combined Fuel Cell Cogeneration…

15

large. The reason for this is that it stores energy for a short period in the system, when the stored energy is released in large quantities. Figure 11 shows the ratio of energy outputted by each piece of equipment. The operation plan when optimizing the energy system of Figure 1 under different objective functions from the analysis results described above has the characteristics described below. Table 5. RMC of February (RMC=Maximum output/Cevice capacity) Commercial Fuel power cell

Solar Heat module pump Boiler Electricity Heat use use 0.338 0.662 0.980 0.964

Minimization of the error of demand-and-supply balance

0.90

0.906

Operation cost minimization

0.57

0

0.361

0.639 0.982 0.990

Greenhouse gas minimization

0.82

0.899

0.207

0.793 1.00

1.00

Table 6. RMC of August

Commercial Fuel power cell 0.87

Solar Heat module pump Boiler Electricity Heat use use 0.714 0.159 0.841 0.604 0.255

Operation cost minimization

0.81

0.631 0.566

0.434 0.614 0.564

Greenhouse gas minimization

0.48

0.503 0.444

0.556 0.890 0.139

(1) Operation Plan of the Minimization of Error in Demand-and-Supply Balance As Figure 9 (a) and Figure 10 (a) show, the stored electric power and heat is mostly used up at 23:00. Moreover, as shown in Table 5 and Table 6, both months are high-output months, and the operating point of a fuel cell is planned so that partial-load operation is avoided. The operation of a fuel cell with a high value of RMC (Maximum output / Cevice capacity) avoids low-efficiency, partial-load operation.

16

Shin'ya Obara Powe r Storage

System e nergy output

20 10

20

Dema nd

10 0

0 Syste m en ergy output-De mand - 20

2

5

8

11 14 17 Time [H our]

180

20

-20

23

2

11 14 17 Tim e [ Hour]

20

23

Hea t

Heat [MJ]

45

90 45

0

0

- 45

-45 2

5

8

11 14 17 Time [ Hour]

20 23

2

Power

30

20

20

Powe r [M J]

30

10

5

8

11 14 17 Tim e [Hour]

20

23

(b) Minimiza tion of opera tion cost

(a) M inim ization of the error of dem and-a nd-supply ba lance

P owe r [M J]

8

135

90

Power

10 0

0 -20

5

180

Hea t

135 Hea t [MJ]

Powe r

30

Power [MJ]

P ow er [M J]

30

2

5

8

11 14 17 Time [Hour]

20

-20

23

2

5

8

11 14 17 Tim e [ Hour]

20

23

180 He at

Hea t 180 135

90

He at [M J]

Hea t [MJ]

135

45 0

90 45 0

-45 2

5

8

11 14 17 Time [Hour]

20

23

-45

2

5

8

11 14 17 Tim e [Hour]

20 23

(d) M ultiple-pu rpo se operational p la nning result

Figure 9. Operational planning result of February.

Operation Plan of a Combined Fuel Cell Cogeneration… Power

30

17

Powe r

30

Storage

10 De mand

Powe r [M J]

Power [M J]

20

System e nergy output

0.0

10 0

System e nergy output- Dem and

-20

2

5

8

11 14 17 Time [Hour]

20

-2 0

23

Heat

60

2

Heat [ MJ]

20

- 15

11 14 17 Time [Hour]

20 23

He at

40 20

-1 5 2

5

8

11 14 17 Time [Hour]

20

23

2

( a) Minimiz ation of the error of de mand- andsupply ba lance

8

Power [ MJ]

10

11 14 17 Time [Hour]

20

23

Powe r

30

20

0 -20

5

(b) M inimization of ope ration cost

Power

30 Powe r [M J]

8

0

0

20 10 0

2

5

8

11 14 17 Time [H our]

-20

20 23

Hea t

60

2

5

8

40 20 0

11 14 17 Time [H our]

20 23

Heat

60

Heat [ MJ]

Hea t [M J]

5

60

40

He at [ MJ ]

20

40 20 0

-15

-15 2

5

8

11 14 17 Time [Hour]

20

23

( c) M inim ization of the a mount of green-house gas discharge

Figure 10. Operational planning result of August.

2

5

8

1 1 14 17 Tim e [Hour]

20

23

(d) M ultiple-purpose opera tiona l planning result

18

Shin'ya Obara 100

Ratio [ % ]

80

Solar module 128.7kWh

124.3kWh

16 8.5kWh

122.2kWh

60 40 20 0

13.9kWh 14.0kWh 22.5kWh 7.9kWh 5.7kWh 13.0kWh 20.6kWh 11.5 kWh

1 2.9kWh 5.4kWh 11.6kWh

(a) February repres entative day Solar modu le 100

8 .5kWh

8.9kWh

Boiler

Ratio [ % ]

80 23.7kWh

20

1.8kWh

1 0.3kWh

5.6kWh

3.0kWh

Heat pump 4.5kWh

22.6kWh

60 40

8.1kWh

7.9k Wh

Fuel cell 7.3kWh 9.6kWh

4.7kWh

Commercial power

9.6kWh 10.6 kWh

8.1kWh

8.7kWh

8.9kWh

0 M M in i opin i m erami za tio Mu t he i zat t io n n o l ti p d io e r e -su m a ro r n o f l e-p u c o st f pp ly n d- a o f rp o s bal nd es anc e

Mi n t im g reeh e a miza ti o n n o d ishcou s eu-nt of o f h arg g ase e s

(b) August representative day

Figure 11. Result of Energy output component.

(2) Operation Plan of the Minimization of Equipment Operation Cost (Fuel Consumption) From the results of Table 5 and Table 6, the electric power generated by the solar module is supplied to the demand side with electric power rather than heat compared with other objective functions. This is because the cost of this method of producing is high compared with the amount of heat. As the results of Table 5 show, the output of commercial power lowers and there is no fuel cell operation. In optimization with this objective function, operation that suspends operation, converts power into heat, stores power is planned.

Operation Plan of a Combined Fuel Cell Cogeneration…

19

(3) Operation Plan of the Minimization of the Greenhouse Gas Emission From the results of Table 5, operation of a heat pump and boiler is planned with a maximum-output point on representative February days with great heat demand. Moreover, application is planned with a value with a high supply rate by the heat of the solar module, and high RMC of the fuel cell. On the other hand, heat supply, rather than the boiler, is mainly concerned with the heat pump according to the results of representative August days with little heat demanded as shown in Table 6 and Figure 11 (b). (4) Operation Plan of the Minimization of Multi-Objective Functions When operating the system under a multi-objective function, the result of Figure 8 to RMC is a value near the average value adding each objective function. As shown in Figure 9 (d) and Figure 10 (d), the operation plan under this objective function includes the characteristic of the operation plan of other objective functions.

CONCLUSIONS How the operation optimization of the combined system of PEMFC-CGS, a solar module, a geo-thermal heat pump, heat storage and power storage equipment, a commercial electric power, and a boiler using GA should be analyzed was described. The capacity optimization of equipment that composes a system was considered from the analysis results. If the capacity of each energy device changes with the objective functions given to the system and examination of equipment cost is added to the results in this section, it can be utilized in real design. Moreover, the characteristics when planning the operation of a system under each objective function were investigated. As a result, for example, the minimization objective of operation cost, operation that suspends operation and converts power into heat, power storage operation is planned. If the objective function concerning minimization of greenhouse gas discharge is given to the system, there will be many opportunities to use the electric power of a solar module as heat; moreover, heat supply mainly concerned with the heat pump rather than the boiler is planned. In the design and operation plan of a combined system containing renewable energy equipment, care should be taken concerning the composition and operating method that change greatly according to objectives given to the system.

Chapter 2

ROUTE PLANNING OF HEAT SUPPLY PIPING IN A FUEL CELL ENERGY NETWORK INTRODUCTION Small-scale fuel cells are distributed in houses, apartment houses, small-scale stores, etc, and power and heat are supplied to them. High facility cost, efficiency fall at the time of partial load, and mismatch of heat output to power generation are issues of this technology. Therefore, in this section, a technology where power and heat demand of tens of or buildings or less correspond by the centralization system and by the distributed system of fuel cells is investigated. The fuel cell centralization system supplies energy to each building from a fuel cell and auxiliary machinery that have been prepared in one arbitrarily selected building. On the other hand, the fuel cell distribution system supplies energy by placing a fuel cell in each building. Until now, improvement in efficiency with cooperation system control [12], peak cut of the power load using a water electrolyzer [13], and estimation of the amount of heat radiation of the hot-water piping for exhaust heat supply [14] are discussed regarding the fuel cell distribution system. When supplying energy using the central system, a heat energy system (hot-water piping) and a power system (power transmission line) are connected to each house, composing an energy network. In the operation of the distributed system, a fuel system (reformed gas piping) is added to these networks, and each network follows an objective function in cooperation with other. Because the energy network examined in this section has a short electric transmission distance, there is little energy loss of power transmission compared with the heat transport. However, it is more advantageous regarding energy cost to have improved power

22

Shin'ya Obara

transmission loss rather than heat transport, because energy unit prices differ. On the other hand, in an energy network with large heat demand of buildings in a cold region, such as houses and apartments, office buildings and hospitals, heat transport loss has large effect on energy cost and system efficiency. Until now, the distributed installation planning of cogeneration using a genetic algorithm has been investigated [15]. Also in this report, the cost of heat transport and the problem of efficiency are pointed out. It is considered that a lot of heat transport using hot-water piping requires route planning of the piping taking heat loss into account. Then, an exploratory program of the piping route that has the minimum amount of heat radiation was developed concerning the hot-water piping of a fuel cell network [14]. With this program, the optimum piping route of the central system and distributed system can be sought. In the last report, it turned out that the heat release of the hot-water piping route of the optimized distributed system is greatly reduced compared with the central system. Therefore, this section considers the route planning of hot-water piping that takes into consideration the load fluctuation of each house connected to the fuel cell network as the next step. Furthermore, this section also considers the influence of the route planning of hotwater piping when connecting solar module equipment to the output changes of the fuel cell network. Moreover, a PEM (proton exchange membrane) fuel cell network is installed into an urban area model in Sapporo, and the optimum hotwater piping route is analyzed. The optimum piping route in winter, summer and mid-term is analyzed, and heat releases are also investigated.

FUEL CELL NETWORK AND ENERGY BALANCE Fuel Cell Network Figure 12 shows the equipment model of the distributed system that installs a fuel cell in each building and supplies power and heat to each building through an energy network. On the other hand, a fuel cell and common auxiliary machinery are installed in a machinery room, and the method of supplying electric power and heat through the energy network is described as the central system. In the case of the central system, the heat energy system (hot-water piping) and power system (power transmission line) of each building are connected. In the case of the distributed system, in addition to these, the fuel system (reformed gas piping) of a fuel cell is also connected. A heat transfer medium is flowed for the hot-water network of both systems, and waste heat recovery of the fuel cells and heat supply

Route Planning of Heat Supply Piping…

23

for each building are attained. Although the route setting of hot-water piping can be arbitrarily planned in this section, the flow of the heat transfer medium is one way. As shown in Figure 12 (b), a city gas reformer (7), a water separator removing water from reformed gas (8), a reformed gas compressor (9), a cylinder (10), a heat storage tank (11), and a back-up boiler (12) are installed in the machinery room of the distributed system. The fuel cell exhaust heat of each building is first supplied to the same building in the distributed system. When there is excess or insufficient heat, it corresponds by operating a heat storage tank and an auxiliary boiler through the network. On one side, the equipment installed in the machinery room of the central system is a fuel cell other than a reformer, a heat storage tank, and a back-up boiler. A header ((4) and (5)) is installed in each building of the central system and the distributed system. Space heating and hot water supply in a building are performed through a radiator and heat exchanger (3) connected to the header. E

G

F

l FE

Hydrogen cylinder (10) Reformer (7)

lDG l EC l AD

D

C

Compress or (9)

l GB

l BF

Water separator (8)

A

l AC

B

Thermal storage tank (11)

B

(a) Fuel cell network model

Heat exchanger

Header (4)

Fuel cell s tack (1) Header (5)

Back-up boiler (12)

(b) Auxiliary machine ins talled in u tility room

Hot water inlet Hydrogen inlet Oxygen inlet Electric power inlet Electric power outlet

Radiators (3)

Town-gas supply (6)

DC-AC converter (2) Hot water outlet

(c) Fuel cell unit installed in each building

Figure 12. Distributed system of the fuel cell network.

24

Shin'ya Obara

Heat Release Model of Hot-Water Piping Figure 13 shows the model of (a) the hot-water piping route, (b) capacity of the fuel cell placed in each Building, (c) change of hot water temperature and (d) piping heat release per unit length of both systems of the fuel cell network. In these models, both systems install a machinery room in Building A, and a heat transfer medium flows in the order of Building A, B, C, D, E, F, G and A as shown in Figure 13 (a). As shown in Figure 13 (b), a fuel cell is installed in Building A in the central system. In addition, a fuel cell of arbitrary capacity is installed in Buildings A to G in the distributed system. As shown in Figure 13 (c), in the central system, hot water is inputted into Building A with inlet temperature T A,in,t . Because hot water has input from the fuel cell exhaust heat, heat storage tank, and auxiliary boiler, it rises to temperature T A,out ,t and is released from Building A. The hot water temperature of the central system drops to T A,in,t due to the heat demand of Buildings B-G as well as the heat release of the piping. Moreover, hot water returns to the machinery room in Building A. In the central system, since the hot water temperature falls in the order of Buildings A to G, the difference in the temperature of the outside air and the hot water becomes small. Therefore, as shown in Figure 13 (d), the heat release per unit length of the piping becomes small. Compared with this, in the distributed system, the outlet hot water temperature of each building is decided by the balance of hot water, input quantity of heat to each building, the amount of exhaust heat of the fuel cell installed in the building, and the heat demand of the building. Therefore, the outlet hot water temperature of each building fluctuates as shown in Figure 13 (c). As a result, the heat release per unit length of piping also fluctuates as shown in Figure 13 (d). Therefore, the sum of heat release changes with the building order that the hotwater piping passes.

Output Characteristics of the Fuel Cell The model of the PEM fuel cell load factor, and the heat output and power load are shown in Figure 14. However, the power output is a value of the AC-DC converter outlet, and the heat output is a value of the fuel cell outlet. In the analysis case described in after Section, the fuel cell capacity of the central system and the distributed system is determined to be 1.2 times the maximum power load. The load factor of the fuel cell at an arbitrary time is calculable from the power generation capacity described above, and the power load. In addition, if the load

Route Planning of Heat Supply Piping…

25

factor of the fuel cell is given to Figure 14, the heat output ( H f ,m,t of Equation (13) in the following section) of the fuel cell is calculable. Buildings A

(a) Route of hot water supply B

C

Capacity of each fuel cell unit [W]

l AB

l BC

D l CD

E l DE

F l EF

G l FG

l GA

(b) Capacity of each fuel cell unit

Central sys tem Distributed s ystem

(c) Hot water temperature

Heat loss per unit Hot water l ength [W/m] t emperature [K]

T A ,out,t Central system TB ,out, t TB ,in ,t

Distributed s ystem

T A,i n ,t (d) Heat loss of each hot water piping

Central system

Distributed s ystem A

B

C D E Location of buildings

F

G

Heat energy output / Power load [-]

Figure 13. Arrangement plan of fuel cell units.

12

8 y = 13.93x −0 .536

4

0

0

Figure 14. Cell performance.

20

40 60 Load factor [%]

80

100

26

Shin'ya Obara

Energy Balance Equation The number of buildings connected to a network was set at N , and a total of M f fuel cells generated electricity in sampling time t ( N ≥ M f and M f =1 in the central system). At this time, the power balance is expressed with Equation (12) and the heat balance is expressed with Equation (13). Mf

N

V

m =1

n =1

v =1

∑ E f ,m,t = ∑ Eneed ,n,t + ∑ ΔEsub,v,t

(12)

Mf

N

N

m=1

n=1

n=1

∑ H f ,m,t + H st,t + Hbo,t = ∑ H need,n,t + ∑ ΔH hw,nn′,t Area3 AP4

F18 S10 D12 F17 F16

0

Area2 S7 F6 AP2 S6 F5

Area1 DH3 SO1

AP1

(13)

D1 S1 S2 D2 S3 F1 F2

DH5 D7 DH7 S9 AP3 F3 F7 D5 DH2 F15 F14 D11 F13 F8 S4 D3 DH1 F10 F9 S5 F4 F11 D10 DH6 D9 D8 D6 DH4 D4 F12 S8 CS1 F21 D14 F25 F19 F20 AP5 S12 F2 2 SO3 D13 S11 D15 DH10 D16 DH8 AP6 Area5 F23 F24 CS2 SO2 DH9 Area4 50m

S : Single-person household D : Family household (2 persons) F : Family household (3~ 4 persons) DH : Two household house (5 or more persons) SO : Small office CS : Convenience store AP : Apartment house

?AP(1) : AP2, AP4 6 houses (single person? 1, two persons? 2, three persons?2, four persons?1) ?AP(2) : AP3, AP6 8 houses (single persons?3, two persons? 2, three persons?2, four persons?1) ?AP(3) : AP1, AP5 10 houses (single persons?10)

Figure 15. Urban area model.

The left-hand side of Equation (12) expresses the power in the AC-DC converter outlet of a total of M f fuel cells in the generating mode. Moreover, the

Route Planning of Heat Supply Piping…

27

2nd term of the right-hand side expresses the power consumption of V number of auxiliary machines V (reformed gas compressor, hot-water circulating pump, etc.). The 1st term of the left-hand side of Equation (13) expresses the exhaust heat of fuel cells, and the 2nd term and the 3rd term express the heat output of the heat storage tank and back-up boiler, respectively. Moreover, the 2nd term of the right-hand side expresses the heat release of the hot-water piping surface that connects each building. ΔH hw, nn′,t expresses the heat release from hot-water piping that connects Building n to Building n ′ , and is calculated from Equation (14). This section assumes the installation of hot-water piping on the ground. However, underground piping etc. is analyzable by the same calculation. In the analysis case in after Section, in order to make the hot-water flow rate in the piping 1 m/s or less, the piping inside diameter was 60 mm. Around the hot-water piping is equipped with a 40-mm-thick polystyrene-foam heat insulating mold. Moreover, the overall heat transfer coefficient ( h in Equation (14)) between the hot water in the piping and outside air is calculated as 8.0 W/m2 K.

ΔH hw,nn ',t = h ⋅ π ⋅ D p ⋅ lnn ' ⋅ (Tn,out ,t − Tatm,t )

(14)

Urban Area Model and Energy Demand Pattern Figure 15 shows an urban area model in Sapporo assuming application of the fuel cell network. The number of the buildings of this urban area model is 74. This is separated into 16 buildings in Area 1, 23 buildings in Area 2, 13 buildings in Area 3, 12 buildings in Area 4, and 10 buildings in Area 5. The energy network is constructed for each Area. The usages of a building are a single-person household (S: the symbols correspond to Figure 15), a two-person household (D), a three- to four-person household (F), two households living together comprising five or more persons (DH), a small office (SO), a 24-hour convenience store (CS) and an apartment house (AP). The apartment house is a scheme of six households (AP2 and AP4 in Figure 15), eight households (AP3 and AP6), and ten households (AP1 and AP5). The power and heat demand pattern of each building in winter, summer, and mid-term are shown in Figure 16 [16-18]. Note that the number of household members differs in the apartment house in Figures 16 (e) to (g). Figure 17 shows the total amount of power and heat demand on each representative day in Area 1 to Area 5.

0.4 0.0

1

5

9 13 17 Tim e [Hou r]

0.8

16 8

1.2 0.8

1

5

9 1 3 17 Time [Hou r]

21

20 10 0

8 .0 6 .0 4 .0

1

5

9 1 3 17 21 Time [H our] (c) Fam ily h ous eho ld ( 3~ 4pers ons ) 6 ho uses (sin gle p erso n? 1, two pers ons? 2, three pers ons ?2, fo ur pers ons? 1)

2 .0 0 .0

1

5

9 13 1 7 Tim e [Ho ur]

21

8 6

1

5

9 1 3 17 2 1 Time [Hou r] (e) Ap artm ent (AP-1 )

10 h ous es (sing le p ers ons ?1 0)

4 2 0

1

5

9 13 1 7 Time [H our]

21

80 40 0

1

30

5 9 13 17 21 Time [H our] (g) Ap artment (AP-3)

20 10 0

5

9 13 17 Tim e [Ho ur]

21

16 8 0

1 5 9 13 17 21 Time [Ho ur] (b ) Family hou sehold ( 2 pers ons )

1.2 0.8 0.4 0.0

1

5

9 1 3 17 Time [H our]

21

20 10 0

1 5 9 13 17 21 Time [H our] (d) Fam ily h ous eho ld ( 5 or more p ers ons )

8 h ous es (sin gle pers ons ? 3, 16 tw o pers ons ?2, three pers ons ? 2, four p erso ns? 1) 8 0

21

1

5

9 13 17 Time [Ho ur]

0

1

5

9 13 17 21 Time [Hour] (f) A partmen t (A P-2 )

8 6 4 2 0

1

5

9 13 17 Tim e [Hou r]

1

5

9 13 17 Tim e [Hour] (h ) Office

21

1

5

9 13 17 Time [Hou r]

21

1 00

40 0

1

2 00

80

1

5

9 13 1 7 Ti me [Ho ur]

Figure 16. Energy demand patterns.

21

Consumpti on of electric power [kW]

0.4 0.0

Cosumpt ion of Consumption of heat energy [kW] el ectric power [kW]

9 13 17 21 Time [H our] (a) Family hous eh old ( Sin gle pers on )

Consumption of electric power [kW]

5

Cosumption of heat energy [kW]

1

Cosumption of heat energy [kW]

0

0.4

0.0

21

Cosumpti on of heat energy [kW]

Cosumption of Consumption of heat energy [kW] electric power [kW] Consum ption of elect ric power [kW] Cosumption of heat energy [kW] Consumption of el ectric power [kW] Cosumpt ion of heat energy [kW] Consumption of el ectric power [kW]

M arks February M ay Au gus t

0.8

Cosumption of Consumption of heat energy [kW] electric power [kW]

Shin'ya Obara Cosumpt ion of Consumption of heat energy [kW] electric power [kW]

28

21

80 60 40 20 0

8 4 0

(i) Convenience s tore

Energy needs ?103 [M J/Day]

Route Planning of Heat Supply Piping…

29

Area 1 Area 2 Area 3 Area 4 Area 5

15.0 10.0 5.0 0.0

E need ,d ay

H need , day

February

E need , day

H need , da y

May

En eed ,d ay

H need , day

August

Figure 17. Energy needs for each area.

ROUTE PLANNING METHOD OF PIPING Route Planning Method of Piping Using the TSP (Traveling Salesman Problem [19]) The chromosome model indicating the piping route is installed into the general genetic algorithm, and gene manipulation, such as crossover and mutation, is added. In this case, many routes that pass the same building two or more times, and route solutions that do not pass some buildings are obtained. Since such routes cannot be managed, all of these chromosome models will be canceled and their analysis efficiency is very low. So, in this section, the view of route order expression by Dewdney is installed [20]. According to this view, although each gene model expresses the building number of each building, this number is not an actual number but the listed number of a building. If this listed number method is installed, since the chromosome model that should be canceled will not appear, analysis efficiency will improve. The optimal solution to the route searching problem of hot-water piping is a route with minimum heat release of the piping. Therefore, the fitness value of the genetic algorithm is evaluated highly, so that the value of the objective function shown in Equation (15) is small.

FO =

Period N

∑ ∑ ΔH hw,nn′,t t =1

n =1

(15)

30

Shin'ya Obara

Analysis Flow of the Search Program of the Piping Route The analysis procedure of the route planning is described below. Each item of data of the coordinate and the usage of each building in the urban area model shown in Figure 15, the energy need pattern shown in Figure 16, and the outside air temperature of a representative day of every month shown in Figure 18 is given to the analysis program. Next, two or more random chromosome models indicating the route order described in this Section are prepared. For each of these chromosome models, the fitness shown in Equation (15) is evaluated, and models of high fitness are proliferated, and low models are screened. Furthermore, crossover and mutation are added to the remaining chromosome models, and such fitnesses are evaluated. These calculations are repeated only the number of times (equal to a generation number) decided beforehand. Moreover, the model with fitness maximum is chosen among the last generation's chromosome model groups. The optimal solution is the piping route expressed by this chromosome model. Heat release ΔH hw, nn′,t of the connection piping in Equation (15) is calculated using the following procedure. Power demand E need , n,t at a certain sampling time is obtained from the power demand pattern of each building. Before Section “Output characteristics of the fuel cell” describes the capacity and output characteristics of the fuel cell placed in each building. Heat output E need ,n,t when operating the fuel cell according to the power demand pattern can be obtained from Equation (12). Fuel cell exhaust heat H f , m,t is calculated for every building in the distributed system. The fuel cell exhaust heat of the fuel cell installed in the machinery room in the central system is calculated. Both systems set the hot water temperature of the machinery room outlet to 353 K, and calculate piping heat release ΔH hw, nn′,t from Equation (14) using the piping specification described in before Section “Energy balance equation”. Moreover, a back-up boiler is operated so that the hot water temperature of the heat storage tank outlet may become 353 K. The temperature of the hot water that returns to the heat storage tank through the hot-water network shall be about 333 K, and the amount of hot-water circulating flow is decided. Heat input-and-output H st ,t of the heat storage tank is calculated by giving H f , m,t , ΔH hw, nn′,t , and heat demand H need , n,t of each building to Equation (13). When the amount of heat storage run short, the heat of H bo,t is outputted from the boiler. Figure 19 shows the heat balance model of the hot-water piping network. Connecting Buildings A to G with piping in order, hot water returns to Building

Route Planning of Heat Supply Piping…

31

A. The machinery room is set in Building A and the heat storage tank and the back-up boiler are installed, and each heat output is H st ,t and H bo,t . Buildings A to G have heat demands H need , A,t to H need ,G ,t , respectively. The distributed system has exhaust heat power output H f , A,t to H f ,G ,t with a fuel cell installed in each building. In the case of the central system, there is exhaust heat power H f , A,t from the fuel cell installed in Building A. In order to estimate heat release

ΔH hw, nn′,t ( n = A, B, C,....., G ) from the piping that connects each building, it is necessary to give the outside air temperature Tatm,t to Equation (14). So, in the analysis case of next Section, the meteorological data of Sapporo in summer (August), winter (February), and mid-term (May) shown in Figure 18 are used [21].

Temperature [K]

313

August

303 May

293 283

February

273 263

0

5

10 15 Time [O’ clock]

20

Figure 18. Outside temperature model in Sapporo.

Building A H bo , t H st , t H in , A , t H in , B , t FA H out , A ,t

B FB

H nee d , A ,t

H nee d , B ,t

H f , A, t

H f , B ,t ΔH hw , AB , t

C FC

H out , B ,t

H need , C ,t H f ,C ,t ΔH hw , BC , t

Figure 19. Heat energy network model.

G

A H in , A , t

FG H need , G , t H f ,G ,t

ΔH hw ,GA ,t

32

Shin'ya Obara

CASE STUDY Program Check by Shortest Route Search The analytic accuracy of the route planning program of piping developed in this section is checked. Therefore, the shortest route for the urban area model in Figure 15 is sought. This analysis result is shown in Figure 20. In the result in Figure 20, the hot-water piping of all the areas is a closed route. It was checked using the analysis method of TSP by Dewdney that these route results were the shortest in length [20]. Moreover, the variable installed in the genetic algorithm was changed variously, and numerical simulation was repeated. The chromosome model with a population of 2500, a crossover probability of 0.9, a mutation probability of 0.001, and a generation number of 200 had the best analysis efficiency of computation time and accuracy.

Area 2 Area 1 1 47 23 17 11 51 10 24 2 48 50 25 3 49 18 52 12 43 40 4 9 19 26 45 20 41 8 42 14 44 5 27 28 13 46 100 21 31 29 33 36 6 38 22 16 15 7 30 35 32 34 39 37 68 71 55 59 62 67 53 72 74 50 69 73 63 54 56 66 60 A rea 5 57 64 70 58 61 65 Area4 0 0 50 100 150 Distance [m] Area 3

Distance [m]

150

Figure 20. Results of minimum length route analysis.

200

Route Planning of Heat Supply Piping…

33

Power Generation Capacity of a Fuel Cell In the central system, a machinery room is set in Building 1 in Area 1 as shown in Figure 20, Building 17 in Area 2, Building 40 in Area 3, Building 53 in Area 4, and Building 65 in Area 5, and a fuel cell is installed in each building. The power generation capacity of the fuel cells is set at 26 kW, 26 kW, 36 kW, 42 kW, and 78 kW, respectively, according to the method described in before Section “Output characteristics of the fuel cell”. Moreover, the fuel cell capacity of the distributed system was also decided to be values of Table 7 according to same Section. Table 7. Fuel cell capacity of buildings

M ark

M aximum Fuel cell electricity capacity (kW) needs (kW)

S D F DH SO CS AP(1) AP(2) AP(3)

0.4 0.8 0.9 1.0 6.0 26.0 4.0 14.0 4.0

0.5 1.0 1.1 1.2 7.5 32 5.0 20.0 5.0

Result of Route Planning (1) Difference Between a Distributed System and a Central System Figures 21 and 22 show the route planning results of a representative day in winter, summer, and mid-term of the distributed system and the central system, respectively. The piping routes of each Area in Figures 21 and 22 differ from the shortest routes shown in Figure 20. This is because the heat release changes with the piping route as before Section “Heat release model of hot-water piping” and Figure 13 explain. In particular, the analysis result of Area 5 in Figures 21 (a) to (c) greatly differs from the shortest route shown in Figure 20. In this area, power

Shin'ya Obara

Distance [m]

34

1 23 17 47 11 10 150 51 48 24 50 49 25 18 2 52 26 19 12 934 43 40 13 45 8 5 44 42 41 27 20 14 100 46 36 33 3128 2921 15 7 6 38 35 3230 22 16 39 37 34 67 68 71 53 55 59 62 72 74 50 63 66 69 73 54 5756 60 70 58 61 64 65 0 0 50 100 150 200 Distance [m] (a) February 23 47 2417 11 10 1 51 48 2 52 50 4925 18 12 3 26 19 45 43 40 27 20 14 13 9 4 8 5 44 42 41 21 100 46 36 33 3128 15 7 6 29 38 35 32 30 22 16 39 3734 68 71 53 5559 62 67 72 74 50 54 56 6366 69 73 57 60 61 70 58 6465 0 0 50 100 150 200 Distance [m] (b) May

Distance [m]

150

23 1 17 11 10 51 47 2 50 4824 18 52 4925 3 12 26 19 4 45 43 40 27 20 13 9 5 14 44 41 8 42 3128 21 100 46 15 7 6 38 3633 32 29 22 34 30 16 39 37 35 68 71 67 53 55 62 72 74 50 54 56 59 63 66 69 60 73 57 6465 70 58 61 0 0 50 100 150 200 Distance [m] (c) August

Distance [m]

150

Figure 21. Results of hot-water pipe route analysis of distributed system.

Distance [m]

Distance [m]

Distance [m]

Route Planning of Heat Supply Piping…

47 2423 17 11 10 1 150 51 2 25 52 50 48 49 26 18 3 12 19 45 43 40 27 20 13 9 4 14 42 41 8 5 100 46 44 33 31 28 21 3635 15 7 6 32 29 38 22 16 30 34 39 3755 71 53 59 62 67 72 74 50 54 56 6366 68 69 73 60 57 70 58 61 6465 0 0 50 100 150 200 Distance [m] (a) F ebruary 47 2423 17 11 10 1 150 51 48 2 52 50 49 25 18 12 3 26 19 9 4 45 43 40 14 13 8 5 41 27 20 21 100 46 443642 31 28 6 29 15 7 35 3332 30 22 38 16 39 37 34 68 71 53 5559 6267 72 74 50 54 5660 63 66 69 73 57 64 61 58 65 70 0 0 50 100 150 Dis tance [m] (b) May

200

23 17 47 11 10 1 150 51 24 2 52 50 4825 18 3 494026 19 12 9 4 45 43 41 27 2014 13 5 21 8 28 100 46 44 4233 31 36 29 38 35 32 30 22 15 7 6 16 39 37 34 67 68 71 55 59 53 62 72 74 50 54 56 6366 69 73 57 60 58 61 64 65 70 0 0 50 100 150 200 Dis tance [m] (c) A ugust

Figure 22. Results of hot-water pipe route analysis of central system.

35

36

Shin'ya Obara

1.2

Area 4 Area 5

Area 1 Area 2 Area 3

0.8

0.4

Distribute system Figure 23. Heat loss of piping.

st Au gu

Ma y

Feb rua ry

st Au gu

Ma y

0.0 Feb r ua ry

Quantity of heat loss / Heat needs [-]

consumption is large compared with the surrounding buildings. However, two convenience stores (Buildings 70 and 71) with little heat consumption are included. The exhaust heat of the fuel cells installed in these two buildings becomes surplus in the distributed system. Correspondence of this analysis program when there is large heat surplus is not taken into consideration because the analytic accuracy of the route planning falls. Compared with this, in the analysis result of the piping route of the central system shown in Figure 22, large heat surplus is not generated on the route. Therefore, compared with the result of Area 5 in the distributed system, the result of the central system is close to the shortest route. Figure 23 shows the ratio of the heat release of the hot-water piping to the heat demand of the central system and the distributed system. When the distributed system and the central system are compared regarding the piping heat release of each area, the heat release of the distributed system is about 25% of the central system (Area 4 on a representative day in August) at maximum. Moreover, when the heat release of all the areas in each season is added, the distributed system is about 75% of the central system. Therefore, if optimization analysis is installed into the route planning problem of the hot-water piping network, it is possible to greatly reduce heat release. The heat release in the case of the shortest piping length shown in Figure 20 for reference was calculated. As a result, the heat demand amount was almost the same as heat release in both systems.

Central system

Route Planning of Heat Supply Piping…

1 2 3 12 13 9 4 5 14 8 6 16 15 7 Area 1 47 51 50 48 52 43 49 40 45 46 44 42 41 Area 3 11 10

37

23 24 17 25 18 26 27 19 20 28 36 33 31 29 21 38 35 22 39 37 34 32 30 Area 2 53 55 59 62 68 7172 63 67 74 54 56 69 73 57 60 66 65 70 58 61 64 Area 4 Area 5

Figure 24. Results of hot-water pipe route analysis of distributed system with load fluctuation of 15% in February.

9.0 Load fluctuation 0% 8.0 6.0

±15%

±30% Area 1 Area 2 Area 3 Area 4

4.0

Fe b ru

ary

0.0

Ma y Au gus t Fe bru ary Ma y Au gus t Fe bru ary Ma y Au gus t

2.0

Figure 25. Heat loss with load fluctuations.

(2) Route Planning of Hot-Water Piping with Load Fluctuation The result of the route analysis in the distributed system when adding less than ±15% random load fluctuation to the power and heat demand pattern on a winter representative day is shown in Figure 24. However, the route patterns shown here do not comprise all of the results. If the analysis that added load fluctuation is repeated, several different route patterns will appear in any area. Moreover, Figure 25 shows the result of the hot-water piping heat release of these analyses. When ±30% or less load fluctuation is added, the heat release is changed from -19% to 3% in the range to an analysis result without load fluctuation, and this is averaged at -6%. In this way, if load fluctuation occurs, heat release will

38

Shin'ya Obara

drop. Furthermore, when fluctuation is added to a small power load, a no-load condition will occur.

Piping Route Plan with a Solar Module The influence of a hot-water piping route plan when connecting solar module equipment to the power network of the distributed system is investigated. The solar module equipment is 24 m2 in the power generation module area, and the power output of the AC-DC converter outlet is a maximum of 3.0 kW. Figure 26 shows the power generation characteristic model in Sapporo for solar module equipment [17]. The random output change within ±25% and ±50% is added to Figure 26, and these data are used for analysis. The power obtained by the solar module is supplied to each building through the power network. Since each fuel cell connected to the network differs at each load factor, the generation efficiency of each fuel cell is different. Consequently, the fuel cell is stopped when generation efficiency is low instead of using solar power.

Fe bruary May Augast

Electric powe r output per hour ? 103 [MJ]

3.0

2.0

1.0

0.0 0

3

6

9

12

15

18

21

24

Time [O’ clock] Figure 26. solar mocdule output.

Figure 27 shows the analysis results when adding less than ±30% random load fluctuation to the power and heat demand on a representative day in winter, and adding less than ±25% random solar power. As before Section describes, if

Route Planning of Heat Supply Piping…

39

analysis is repeated, two or more route results will appear in all the areas. Since some fuel cells stop if solar module equipment is connected to the network, compared with Figure 24, the piping routes may differ greatly.

11

10

23 17 24 25 18 19 26 20 27 28 21 31 38 36 3533 29 22 39 37 34 32 30

1

2 3 12 9 4 13 14 5 8 16 7 6 15 Area 1 47 51 50 48 52 43 40 45 46 44 42 41

Area 2

53 5559 62 54 63 56 60 57 58 Area 3 67 68 66 65

61 64 Area 4

71 72 74 69 73 70 Area 5

Figure 27. Result of hot-water pipe rout analysis of distributed system with load fluctuation of +/- 30% and solar power output fluctuation of +/- 25% in February.

Figure 28 shows the result of the heat release of hot-water piping when the fluctuation rate of the power and heat load is less than ±30%, and the output change rates of the solar module are less than ±25% and less than ±50%. Compared with solar power with ±25% output change, the result of the heat release with ±50% output change is -16% to 5%. The average reduction value of heat release is 2%. Moreover, if solar module equipment is connected to the network, compared with when it is not connected, heat release fluctuation of -7% to 1% will occur. The average heat release is reduced by 2%. If power is supplied to the network from solar module equipment, some of the fuel cells connected to the network will stop. As a result, the fuel cell exhaust heat outputted to the hotwater network decreases, and heat release decreases. In this case, the capacity of solar module equipment is small and heat release reduction of the hot-water piping is about 2%. However, if the capacity of the solar module connected to the network is extended, the heat release of the hot-water piping will become smaller.

Shin'ya Obara Solar output flactuation ±25%

8.0

±50% Area 1 Area 2 Area 3 Area 4

6.0 4.0 2.0

st Au gu

Ma y

ry Fe b rua

gu st Au

Ma y

rua r

y

0.0 Fe b

Quantity of heat loss ×103 [MJ/Day]

40

Figure 28. Heat loss with load fluctuation 30% and solar device outputs.

CONCLUSION A route planning program that optimizes the hot-water piping route of a fuel cell energy network was developed. The view of TSP is installed into the analysis program, and the fuel cell central system and fuel cell distributed system can be examined. An urban area model of 74 buildings in Sapporo was analyzed, and the piping route and the hot-water piping heat release were investigated. Furthermore, the piping route and heat release of the fuel cell network that connected the buildings with load fluctuation and solar module equipment with output fluctuation were examined. As a result, the piping heat release in one year from the distributed system is about 75% compared to the central system. The piping heat release with power load fluctuation of ±30% shows an average of a 6% reduction compared with no load fluctuation. This is because a fuel cell idle state of load zero will occur if the load is varied when the power demand is small. Moreover, compared with when solar module equipment is not connected, there is an average 2% reduction of heat release at the time of ±50% output change of the solar module equipment. This is because some of the distributed fuel cells are stopped due to the power supply to the solar module network. Although hot-water piping with an inside diameter of 60 mm was assumed in this section, a reduced size is actually taken.

Chapter 3

LOAD LEVELING OF FUEL CELL SYSTEM BY OXYGEN CONCENTRATION CONTROL OF CATHODE GAS INTRODUCTION Wide usage of proton exchange membrane fuel cells (PEFC) in houses, as well as in small to middle-scale buildings requires a fuel cell stack, a reformer, a heat storage tank, and cost reduction of cost of a back-up heat source. In particular, the fuel cell stack is expensive, and is therefore used together with commercial power and requires a device for reducing the power generation capacity. Furthermore, if a fuel cell system is installed with a large load fluctuation with a town gas reformer in a house and performs load-following operation, partial load operation will occur frequently [22-24]. The dynamic characteristics of a fuel cell stack and a reformer differ greatly, and the load following rate of the reformer is very slow compared to the load following rate of the fuel cell stack [25-25]. If small partial loads occur frequently in short time intervals, a drop in efficiency will be expected for the reformer with a slow speed of response. The time shift of the electric-power supply and demand using a battery as a possible method to solve the reduction in fuel cell capacity and the dynamic characteristics of the reformer can be considered. However, when the cost of a battery, its charge-and-discharge efficiency, and its life are taken into consideration, the introduction of the battery system to a house with long operation time has many considerations. Improvement in the electric-powergeneration capacity reduction of a fuel cell and partial load operation of a

42

Shin'ya Obara

reformer has been attempted by electric power generation in time shifts using hydrogen and oxygen produced by water electrolysis using surplus electric power. Although reformed gas and air are supplied and power generation from the fuel cell is performed in time zones with little electric power demand, the power also has to be supplied to a water electrolyzer. The hydrogen and oxygen produced by water electrolysis and they are stored under compression in cylinders. On the other hand, electric power demand supplies and generates the gases in the gas cylinders to the fuel cell during a time zone with larger demand. The electric-power-generation characteristics of PEFC improve by supplying a higher concentration of oxygen to the cathode [29-32]. Therefore, hydrogen and oxygen are produced during time zones of smaller electric power demand, and the power load peak is leveled by using these gases for time zones of greater electric power demand. Although the capacity of the fuel cell stack has been decided until now using the output characteristics of the fuel cell at the time of electricity generation with reformed gas and air, in the operational method that will be proposed in this report, the capacity of the fuel cell stack can be decided from the output characteristics at the time of electricity generation with hydrogen and oxygen. In this case, since the electric-power-generation characteristics of the fuel cell improve by supplying gas and pure oxygen with a high oxygen concentration to the cathode compared with the case where just air is supplied, the cell stack capacity can be reduced. The proton exchange membrane water electrolysis system (SPE) for water electrolysis has an efficiency of 84% at 393K and 0.4MPa, and has already been established as a basic standard technology. Moreover, the SPE system cost may also reduce substantially in the future [9, 33]. In the present study, the capacity of the fuel cell facility, the town gas consumption and the operational time of auxiliary machines is investigated by assuming the introduction of a fuel cell system with a water electrolyzer into an energy-demand model of a house, a hospital, a factory, a hotel, and a small store for 24-hour operation. These models are used as analysis examples.

SYSTEM DESCRIPTION System Structure Figure 29 shows the structural drawing of a fuel cell system assuming water electrolysis. Town gas fuel of a fuel cell system is supplied to the heat source burner (BN) and reforming gas system (RM) of a reformer. Town gas is reformed

Load Leveling of Fuel Cell System…

43

by a reformer, and a dryer (DY) recovers the surplus water of reforming gas after that. After removing the carbon monoxide in reforming gas with a carbon monoxide oxidization device (MC), the anode of the fuel cell (FC) is supplied. If power is supplied to a water electrolyzer (EL) when the power load of the system is small, hydrogen and oxygen are generated. The hydrogen generated by water electrolysis other than reforming gas can be supplied to the anode of the fuel cell. The air of a blower (B3) or pure oxygen generated with EL can be supplied to the cathode of a fuel cell. Power is supplied from the system to the demand-side using a fuel cell system E IT ,t or a commercial power system Ecm,t using an interconnection device. Part of the electric power generated using the fuel cell is supplied to a water electrolyzer, and the electric power ΔEEL ,t is consumed for the production of hydrogen and oxygen. The hydrogen and oxygen thus produced are compressed into gas cylinders CDH and CDO , and are stored. The power consumption of the compressors are ΔEPH ,t and ΔE PO ,t , respectively. The stored gases can be supplied to the fuel cell and electric power generation, which shifted the time period, can be carried out. The exhaust heat of a fuel cell is stored in a heat storage tank (ST) by the heat transfer medium conveyed with a pump (PP). When there is exhaust heat input exceeding the capacity of the heat storage tank, a part for surplus is released from a radiator (RA). In addition, heat is supplied to the demand side by feeding tap water to the heat exchanger installed within the ST. Commercial power Town gas

Q BE, t Air Town gas

Q BN ,t

Town gas

Q RM ,t VA1

B3 B1 ΔE B1 ,t

B2

ΔE B2, t

BN RM

η RM ,t

CDH η Δ E PH ,t PH PH CDO η PO PO Δ EP O,t VA4

ΔE B3,t

DY

η MC ,t MC

FC VA2

η F C,t

E FC ,t H FC ,t

Ecm, t ΔE EL,t

BE EL

DC

η EL η DC

WP

Interconnection device

E s,t

Electric power load AC electric power output

IT

η IT

EIT ,t

RT PP ΔE PP ,t

Drain

VA 3

Tap water

H RM ,t ST

RA

H ST ,t

Exhaust VA 5

Figure 29. Structure of fuel cell cogeneration with electrolyzer.

Heat exhaust

Δ H RA,t

H S ,t

Heat power load Heat power output

44

Shin'ya Obara O2 from cylinder (CDO) Air from blower (B3)

Mixing chamber

Reforming gas from CO oxidation (MC)

Mixing chamber

FC

E F C ,t

H 2 from cylinder (CDH) (a) System A O2 from cylinder (CDO) VA4 Air from blower (B3) Reforming gas from VA2 CO oxidation (MC)

FC

E FC ,t

H 2 from cylinder (CDH) (b) System B

Figure 30. Structure of fuel cell cogeneration with battery.

Figure 30 shows the two methods of supplying hydrogen and oxygen to a fuel cell stack. In Figure 30 (a), hydrogen and reformed gas from a gas cylinder are mixed in one chamber, oxygen from a gas cylinder and air from the blower B3 are mixed in another chamber, and the system supplying these gases to the fuel cell stack is described as System A. Since the cathode gas in System A uses mixed air and oxygen, the oxygen concentration supplied to the fuel cell varies. On the other hand, in Figure 30 (b), valves VA 2 and VA 4 are operated, and the anode gas and cathode gas which are supplied to the fuel cell stack are changed. This system is termed as system B. Since the air flow path of the blower B3 and the flow path of the gas cylinder oxygen are changed, and cathode gas is supplied to the fuel cell, air or pure oxygen will be supplied. In System A of Figure 30 (a), the hydrogen supplied to FC is a mixed gas of MC and CDH. Moreover, the oxygen supplied to FC is a mixed gas of B3 and CDO. In this case, since the air of B3 and the oxygen of CDO are mixed and FC is supplied, the oxygen concentration changes. On the other hand, VA2 and VA4 of System B of Figure 30 (b) are switching valves. Therefore, VA2 supplies the gas of MC or CDH to FC. Moreover, VA4 supplies the gas of B3 or CDO to FC. The electrochemical

Load Leveling of Fuel Cell System…

45

reaction of the cathode in the case of supplying the gas of CDO to FC uses pure oxygen, and the electrochemical reaction of the cathode in the case of supplying the gas of B3 to FC uses air. Although the oxygen concentration in the cathode gas of System A is controllable, the oxygen supplied to cathode from CDO is pure gas in the case of System B. However, even if it mixes the air of B3 and pure oxygen of CDO of System A, since the volume of gas other than oxygen is large, the oxygen concentration cannot be increases. Since the power-generation characteristics of the fuel cell especially using pure oxygen improves compared with the fuel cell using air, it is the target to reduce the fuel cell capacity greatly by supplying at a peak period. System A and System B attempt a reduction in power-generation capacity by increasing the oxygen concentration in the gas supplied to the cathode, shifting time and supplying this gas to a fuel cell. However, System A and System B are independent structures, and combined operation is not considered. Although the operating methods of the cathode gas of each system differ by power load, the details are described later. Town gas is supplied to the burner BN with quantity of flow Q BN ,t , and the heat source of the reformer RM is acquired. Electricity consumption is ΔEB1 ,t though air is supplied to a burner by the blower B1 . Moreover, air is supplied to the dryer DY, which is used for removing water from the reformed gas, by the blower B2 , and electricity consumption is ΔEB2 ,t . The blower, which supplies air to the fuel cell stack, is B 3 and the electricity consumption is ΔE B3 ,t . For heat recovery from the fuel cell stack, electric power ΔE PP,t is consumed using the heat-transfer-medium circulated by pump PP. Although hydrogen and oxygen produced by water electrolysis are stored in each gas cylinder with operation of compressors PH and PO. The electric power consumed by the compressors are ΔEPH ,t and ΔE PO,t respectively. The electricity consumption ΔEB1 ,t , ΔEB2 ,t , and

ΔEB3 ,t of the blowers B1 , B 2 , and B 3 , and the electricity consumption ΔE PP ,t of the pump PP are obtained from the electric power supplied from a commercial power system. The electricity consumption ΔEPH ,t and ΔE PO ,t of the compressors PH and PO are supplied by the electric power generated by the fuel cell. The system shown in Figure 29 is electric power generation system, and the thermal output system is not taken into consideration.

46

Shin'ya Obara

Fuel Cell Performance Figure 31 shows the cell performance curve for an operating temperature of 333 K, with the gas pressure at the anode and cathode being 0.1MPa [29, 30]. Although the cell efficiency changes with oxygen concentration of the cathode gas, there is no difference in the cell efficiency whether reformed gas or hydrogen is supplied to the anode. 1.0

Potential [V]

0.8 0.6 0.4 Oxygen concentration of cathode gas 0.2 Air 80 % 40% 100% 60% 0 0 100 200 300

400

500

600

700

2

Current density [mA/cm ]

Fue l cell e fficiency [%]

Figure 31. Cell performance generated with oxygen. Operating temperature 333K, and reactant flow stoichiometries 2 both hydrogen and oxygen.

80 60

69% 48%

53%

74%

60%

40 20 0

Air 40% 60% 80% 100% Oxygen volume concentration in cathode gas .

Figure 32. Efficiency of the fuel cell electric power output. Operating temperature 333K, and pressure 0.1 MPa.

Figure 32 shows the model of the generation efficiency η FC of the fuel cell stack [29, 30]. When pure oxygen is supplied to the cathode, the efficiency is 1.54

Load Leveling of Fuel Cell System…

47

times that of the case when air is supplied. When the electrode surface of an anode and cathode is 1 m2, Figure 33 shows the model of the oxygen concentration supplied to the cathode, the town gas consumption, and the power output in an inverter outlet [23, 24, 29, 30]. However, these characteristics are the models when setting the reformer efficiency η RM to be constant at 73%. Details of η RM are given below.

[kW]

8

Electric power output

6

100%

80% 60% 40% Air

4

2

0 0.0

Oxygen volume concentration in cathode gas . 0.1 0.2 0.3 0.4 -3 ′ , t [m3/s] QCON Quantity of flow ×10

0.5

Figure 33. The power output of inverter by difference in the oxygen volume concentration in cathode gas. The area of the electrode of anode and cathode of the fuel cell stack is 1 m2, respectively. Reformer efficiency is 0.73.

System Operation The operational model of the fuel cell system with water electrolysis is shown in Figure 34. In the fuel cell system, reformed gas and air are supplied and the threshold value E set of the electric power shown in the figure is generated. However, during the time period when the power load Ed ,t is less than E set , the electric power of ( E set - Ed ,t ) is supplied to the water electrolyzer. The hydrogen and the oxygen, which are produced in this time period, are compressed and stored in the gas cylinders. As shown in Figures 32 and 33, the output characteristics of the cell stack improve by supplying gas with a high oxygen concentration to the cathode rather than supplying air. Then, the threshold value E set of the electric power is compared with the power load Ed ,t , and the hydrogen and oxygen stored in the gas cylinders using Ed ,t in the large time

48

Shin'ya Obara

period are supplied to the cell stack. The hydrogen and the oxygen stored in the gas cylinders are supplied to the fuel cell by either of the two systems shown in Figure 30. With the operating method described in the Figure 34, the power generation characteristics under a high load serves as a high output compared with the power generation method with conventional reformed gas and air, and can perform peak-cut of the electric power load. The installed capacity of the conventional fuel cell stack was decided by the electrode surface product which can output the maximum value of electric power demand at the time of generating electricity with reformed gas and air. In this study, at the time of low load, hydrogen and oxygen are produced by water electrolysis using surplus electric power, and a reduction in capacity of the cell stack is attempted by generating electricity with this hydrogen and oxygen at the time of high load including at the maximum value of the power load.

Electric power output

Ed , t

Eset

Figure 34. System operation.

The operating range of the reformer is limited by always operating the fuel cell system at the electric-power threshold value Eset . In this case, consideration of partial load operation for which the efficiency of the reformer falls, and consideration of the delay time of the speed of response of the load are avoidable. Although steam reforming of the town gas quantity of flow Q RM ,t is carried out and reformed gas is produced in the reformer, the heat source for reforming reaction burns and obtains using the town gas with a flow quantity of Q BN1 ,t . The following equation defines the reformer efficiency and the maximum of this value is 73% [9].

Load Leveling of Fuel Cell System…

η RM =

WH ⋅ 100 (WC + W R )

49

(16)

With CO oxidation equipment, when burning and changing CO in the reformed gas into CO2, a part of the hydrogen also burns. Nearly 2% of hydrogen in the reformed gas will burn with CO oxidation equipment, with an associated efficiency η MC of 98%. Therefore, the hydrogen flow quantity QMC , H 2 ,t of the CO oxidization equipment outlet and the town gas flow quantity ( QCON ,t = QRM ,t + QBN1 ,t ) supplied to the reformer are connected by the following equation. Q MC , H 2 ,t = η MC ,t ⋅ η RM ⋅ QCON ,t

(17)

Water Electrolysis Water electrolysis produces hydrogen and oxygen using the proton exchange membrane electrolysis system [9]. The pure water used for water electrolysis supplies and produces tap water which carries out heat exchange to a pure-water production device inside the heat storage tank. A charcoal filter and an ion exchange filter are installed by the pure-water production device, and electric power is not consumed for its operation. Moreover, the cost of pure-water production, which uses a charcoal filter and an ion exchange filter, is inexpensive. Water electrolysis is performed under conditions of 393K and 0.4MPa, heat source burns and obtains town gas by the flow quantity of Q BN 2 ,t , and makes town gas a heat source. The efficiency η EL of water electrolysis is 84% [9], and the direct-current power generated with the fuel cell is supplied to the water electrolyzer through a DC-DC converter (efficiency η DC =95%). The hydrogen and oxygen produced by water electrolysis are compressed using the compressor, and are stored in gas cylinders. The electricity consumption in the hydrogen compressor is calculated by the following equation.

′ ,t ) ΔE PH ,t = 1 η CDH ⋅ Pin ⋅ Qin′ ,t ⋅ ln(Qin′ ,t Qout

(18)

50

Shin'ya Obara

′ ,t = 1.0 MPa . The hydrogen compressor-efficiency Here, Pout = Pin ⋅ Qin′ ,t Qout

η CDH is 50%. The electricity consumption ΔEPO,t in the compressor which compresses oxygen is similarly calculated.

Energy Balance Equation (19) is the electric-power balance in sampling time t . E FC ,t = E IT ,t η IT + ΔE EL,t η DC + ΔE PH ,t + ΔE PO,t

(19)

The left side is the power output of a fuel cell stack, and the 1st term on the right hand side is the electric power output to the interconnection device from the inverter and E IT ,t is equal to the electric power demand in sampling time t . The 2nd term on the right hand side is the electric power used for the water electrolysis operation, and the 3rd term and 4th terms are the electricity consumption in each compressor used for hydrogen and oxygen. Commercial power is used as the power supply to other blowers and the pump. Equation (20) shows the heat balance of the system. The left side of Equation (20) expresses the heat output from the fuel cell, the reformer, and the heat storage tank, and the right hand side expresses the thermal output from the system. H S ,t is equal to the heat demand amount in sampling time t .

H FC,t + H RM ,t + H ST ,t = H S ,t

(20)

Equation (21) shows the mass balance of hydrogen. The left side expresses the hydrogen quantity produced by the water electrolyzer, the hydrogen quantity of flow supplied to the fuel cell from the gas cylinder, and hydrogen quantity of flow from the CO oxidization equipment outlet. The right hand side expresses the hydrogen quantity consumed by the fuel cell. Equation (22) shows the mass balance of oxygen. The left side includes oxygen flow rate produced by the water electrolyzer, oxygen flow rate supplied to the fuel cell from the gas cylinder, and air supplied to the fuel cell by the blowers. The right hand side expresses the amount of oxygen consumed by the fuel cell at sampling time t .

Q EL , H 2 ,t + QCDH , H 2 ,t + Q MC , H 2 ,t = Q FC , H 2 ,t

(21)

Load Leveling of Fuel Cell System…

Q EL ,O2 ,t + QCDO ,O2 ,t + Q B3 ,O2 ,t = Q FC ,O2 ,t

51

(22)

Analysis Method The threshold value E set of the electric power for low load and high load is decided so that the balance of the individual amounts of hydrogen and oxygen produced by water electrolysis at the time period of low loading and the amounts consumed by the fuel cell at the time period of high load can balance. Then, E set is decided below which the maximum electric power output of the system, and the balance of the amounts of hydrogen and oxygen produced during the low loading time period and the high load period is calculated. When the power load Ed ,t exceeds the threshold value E set of electric power, the anode gas supplied to the fuel cell in System A is the mixed gas of reformed gas hydrogen and the cylinder, and the cathode gas is the mixed gas of oxygen in the air supplied by the blower and the cylinder. However, in System A, suspending the gas supply from the reformed gas and blower, and supplying only the hydrogen and the oxygen from each gas cylinder to the fuel cell can also be chosen. The choice of the method of supplying gas, which mixes gas cylinder hydrogen with reformed gas, and the gas which mixes gas cylinder oxygen with blower air to the fuel cell, and the method of supplying only gas cylinder hydrogen and oxygen is decided based upon minimal consumption of gas cylinder oxygen. On the other hand, when the power load Ed ,t exceeds the threshold value E set of electric power, electric power generation from the fuel cell in System B is performed with the hydrogen and oxygen supplied from the cylinders. In the numerical calculation which determines Eset , the balance of hydrogen and oxygen is calculated by E set , the amount of production of hydrogen and oxygen is excessive, and the value of E set whose difference of the balance is the minimum, is decided as a solution. That is, when E set > Ed ,t , the reformed gas produced by the reformer and the blower air are supplied to the fuel cell. In this case, the power output E IT ,t at the inverter outlet in Equation (19) is set to Ed ,t . The electric power consumed by the water electrolyzer is ΔEEL,t = ( Eset − Ed ,t ) η EL and sets the water electrolysis thermal efficiency to η EL . However, in order to supply the direct-current power generated by the fuel cell to the water, a DC-DC converter is required to adjust the voltage.

52

Shin'ya Obara

The efficiency of this DC-DC converter is set to η DC , and the electric power, ΔEEL,t η DC , is supplied to the water electrolyzer from the fuel cell. If the electric

power ΔEEL,t is decided, the amount Q EL, H 2 ,t of hydrogen production and the amount QEL,O2 ,t of oxygen production can be calculated. The amounts of hydrogen and oxygen for a sampling time interval of Δt which are stored in cylinders are QEL, H 2 ,t ⋅ Δt and QEL,O2 ,t ⋅ Δt , respectively. The electricity consumption ΔE PH ,t in a compressor is calculated by converting QEL, H 2 ,t ⋅ Δt into

′ , H 2 ,t , and introducing this into Qin,t in Equation (18). the volume rate of flow Q EL The electricity consumption ΔE PO ,t

in the compressor, wherein oxygen

compression occurs, is calculated similarly. The electric power E FC ,t generated

1. 0 0. 8 0. 6 0. 4 0. 2 0

0

4

8 12 16 Ti me (ho ur)

20

30

Electric po wer load [kW]

1. 2

Electri c p ower load [kW ]

Electric po wer load [kW]

by the fuel cell stack is calculated by introducing appropriate values into each term on the right hand side of Equation (19).

20

10

0

0

4

60 40

20 0

0

4

8 12 16 Time (ho ur) (d ) Hotel model

20

(b) Hos pi tal mod el

20

El ectr ic power lo ad [k W]

Electric po wer load [kW]

(a) Ind iv id ual h ou se mod el

8 12 16 Ti me (ho ur)

80 60 40 20 0

0

4

8 12 16 Ti me (ho ur)

20

(c) Facto ry mod el

30 20

Janu ary M ay Au gu st

10 0

0

4

8 12 16 2 0 Ti me (ho ur) (e) Mo del o f s mall sto re op en arou nd t he clo ck

Figure 35. Models of electric power load.

The hydrogen flow quantity Q FC , H 2 ,t supplied to a fuel cell stack is obtained using Equation (23). The η FC is calculated by creating the approximate expression for the relationship between the oxygen concentration shown in Figure 32 and the efficiency of the fuel cell stack, and introducing the oxygen concentration of the gas supplied to cathode into this approximate expression.

Load Leveling of Fuel Cell System…

53

Furthermore, QMC , H 2 ,t is calculated from Equation (21), and the town gas flow

′′ ,t + Q BN ′′ ,t supplied to a system is calculated by introducing this quantity Q RM value into Equation (24).

′ , H 2 ,t = E FC ,t (α ⋅ η FC ) Q FC

(23)

′ ,t + Q BN ′ ,t = Q MC ′ , H 2 ,t (η MC ⋅ η RM ) Q RM

(24)

For E set ＜ Ed ,t , the flow quantities QCDH , H 2 ,t and QCDO ,O2 ,t are supplied to the fuel cell from the hydrogen and oxygen stored in each gas cylinder. In System A, the oxygen concentration of the cathode gas required to reduce the power load Ed ,t at sampling time t to E set , is calculated. In order for the oxygen concentration of the gas to be the value described above, the supply oxygen flow rate QCDO ,O2 ,t in the gas cylinder added to the oxygen flow rate Q B3 ,O2 ,t in the air supplied by the blower is calculated. The hydrogen flow quantity QCDH , H 2 ,t supplied from a gas cylinder is adjusted to be twice the molar flow rate of the oxygen flow rate QCDO ,O2 ,t . The hydrogen used for the hydrogen flow quantity

Q MC , H 2 ,t in the reformed gas for QCDH , H 2 ,t is supplied to the anode of a fuel cell. On the other hand, in System B, the fuel cell is operated only by the cylinder gases, and pure hydrogen and pure oxygen are supplied and generated. The calculation results of the threshold value E set of the electric power should change with the difference in the oxygen concentration of the cathode gas in System A and System B, and differences in the power load pattern introduced into the system. When E set = Ed ,t in System A, the reformed gas produced by the reformer and the air from the blower are supplied, and the fuel cell is operated. Water electrolysis is not performed at this time. One of the two methods of operating a fuel cell, (1) with reformed gas and blower air, and (2) with cylinder hydrogen and cylinder oxygen where the value of E set is smaller, is carried out by System B.

54

Shin'ya Obara

CASE STUDY Weather Conditions in Tokyo In Tokyo, the annual average temperature for the past five years is 289K. The average temperature in January is 279K, and the highest and the lowest temperatures on a representative day for January are 283K and 275K, respectively. The average temperature in May is 292K, and the highest and the lowest temperatures on a May representative day are 296K and 288K, respectively. The highest and the lowest temperature on a July representative day for the past five years are 302K and 296K, respectively, and the average temperature is 298K [34].

Energy Demand Models The energy-demand models for an average individual house, a hospital, a factory, a hotel, and a small store with day-long business in Tokyo is shown in Figure 35. Each energy-demand pattern is for a monthly representative day of winter (January), mid-term (May), and summer (August). The space cooling load of all buildings shown in Figure 35 is included in the electric power demand. Although the space heating load of a hospital and a hotel is supplied using the fuel cell exhaust heat, the space heating load of other buildings is included in the electric power demand. The floor space of each building, the energy demand, and the reference capacities of the fuel cell installed in the buildings are shown in Table 8. The power-generation capacity in Table 8 is decided by an increase of nearly 10% in the maximum electric power load of the buildings. In a numerical analysis, as for the electric power generation characteristics of the fuel cell, the relationship shown in Figure 33 needs to be maintained relatively, and the efficiency of the fuel cell stack is obtained by giving oxygen concentration to the approximate expression which determines Figure 32. The values indicated in Table 9 are used for the efficiencies of the auxiliary machines used in the analysis. The values of such efficiency are products on the market and the products under engineering development.

Load Leveling of Fuel Cell System…

55

Table 8. Total floor space, power consumption and fuel cell capacity of each model Total floor 2 space (m )

Model (a) Individual house (b) Hospital (c) Factory (d) Hotel (e) Small store

128 600 800 1000 145

Power consumption of January, M ay and August (M J/day)

Fuel cell capacity (kW)

(Electric power / Heat power) 36, 32.4, 39.6 / 77.4, 48.6, 27.0 551, 745, 1130 / 3640, 1670, 1280 1620, 1490, 1800 / 1440, 1760, 2560 / 5220, 3380, 2380 1660, 1510, 1810 /

1.0 22.0 55.0 44.0 28.0

Table 9. The values of the efficiency used for analysis η RM ,t

η M C ,t

η D C ,t

η PH

η PO

ηEL

ηIT ,t

73 %

98 %

95 %

50 %

50 %

84 %

95 %

RESULTS AND DISCUSSION Operation Planning and Fuel Cell Capacity Reduction Effect Figure 36 shows the calculation results for the operational plan after the introduction of System A into each building, and Figure 37 shows the calculation results of the operational plan after the introduction of System B. In the system which supplies all the electric power loads using the fuel cell, the capacity of the fuel cell was conventionally fixed to be the value which covers the annual peak electricity demand. However, if both System A and System B are introduced, the capacity of the fuel cell can be reduced to a value which covers the electric power load shown overall in Figures 36 and 37. The production of electricity from the fuel cell stack is the value which divides the electric power needs in the figures using the inverter efficiency η IT =0.95. Moreover, for the analysis of the small store in System A, and the hotel and small store in System B, the solution by calculation of a representative day in August was not obtained. This is because the amount of oxygen required for the peak-cut of electric power load cannot be produced by water electrolysis in the time zones when power load is small.

8

12 16 Time [Hour] January

20

20 16 8 4 0

0

4

8

12 16 Time [Hour] January

20

50 40 30 20 10 00

Electric power output of the fuel cell stack [kW]

Electric power output of the fuel cell stack [kW]

12

0

4

8

12 16 Time [Hour] January

8 12 16 Time [Hour] January

Electric power output of the fuel cell stack [kW]

30 25 20 15 10 5 0

4

30 25 20 15 10 5 0

0

20

20

4

Electric power output of the fuel cell stack [kW]

4

0 0

4

8

12 16 20 Time [Hour] May (a) Individual house

20 16 12 8 4 0

0

4

8

12 16 Time [Hour] May (b) Hospital

20

50

0.4 0.2 0

0

4

8

12 16 Time [Hour] August

20

0

4

8

12 16 Time [Hour] August

20

00

4

8

12 16 Time [Hour] August

20

4

8

12 16 Time [Hour] August

20

20 16 12 8 4 0

40 30

30

20

20

10

10

30 25 20 15 10 5 0

0.6

50

40

0

0.8

Electric power output of the fuel cell stack [kW]

0

0.2

1.0

Electric power output of the fuel cell stack [kW]

Electric power output of the fuel cell stack [kW]

0

0.4

0

0

8 12 16 Time [Hour] January

4

8

4

9

20

12 16 Time [Hour] May (c) Factory

20

12 16 Time [Hour] May (d) Hotel 30 25 20 15 10 5 0

Electric power output of the fuel cell stack [kW]

0.2

0.6

Electric power output of the fuel cell stack [kW]

0.4

0.8

Electric power output of the fuel cell stack [kW]

0.6

1.0

Electric power output of the fuel cell stack [kW]

Electric power needs System A

0.8

Electric power output of the fuel cell stack [kW]

1.0

Electric power output of the fuel cell stack [kW]

Shin'ya Obara

Electric power output of the fuel cell stack [kW]

56

20

0

4

8

40 30 20 10 00

12 16 Time [Hour] May

20

(e) Small store

Figure 36. Operation planning of system A.

The analysis results for the conventional System A and the System B are shown in Table 10 for the electric power output supplying the biggest load for each sampling time of every month for representative days in winter (January),

Load Leveling of Fuel Cell System…

57

mid-term (May), and summer (August), as the power generation capacity of the fuel cell. In System A, improvements in the fuel cell efficiency by increasing the oxygen concentration as shown in Figures 32 and 33, can be performed at the time of the high load of Ed ,t ＞ E set . However, in order to boost the oxygen concentration in the air supplied by the blower, the oxygen flow quantity may increase greatly. As a result, the rate of utilization of oxygen at the cathode will fall. As shown in Table 10, in System A, the capacity reduction effect of the fuel cell is fairly small by reducing the oxygen utilization rate when compared with System B. The reduction effect of the fuel cell capacity is so large that the difference between the load for the high load sampling time of Ed ,t ＞ E set of the load pattern of a building and the low load sampling time of Ed ,t ＜ E set is large. This typical case is a load pattern of the factory, and is common in System A and System B. The capacity reduction effect of the fuel cell is influenced by frequency of appearance of the high load and the low load. There is little appearance frequency of low loads, and when the values of the loads are large, the capacity reduction effect of the fuel cell is very small. The load pattern of the hotel is a typical example of this. Moreover, when changes in the load are small through one day, the reduction effect of the fuel cell capacity is small in both systems, and the load pattern of the small store is an example of this. When there are sufficient occasions to produce hydrogen and oxygen for reducing a load peak, the reduction method of the fuel cell capacity of using the water electrolysis system is effective. The time period of low loading and high load is divided clearly, and since the difference between low loading and the high load is large, the load pattern for the hospital and the factory is sufficient for the production of hydrogen and oxygen by water electrolysis. As a result, the power-generation capacity of the fuel cell is greatly reduced. In the case of the individual house, the capacity of the electric power is decided by the cooling load in summer. In an individual house, although the load fluctuation is large, before and after night or daytime, the load is small and a large amount of hydrogen and oxygen can be produced during these periods. By introducing System B into an individual house, the effect of capacity reduction of the fuel cell stack is large.

Town Gas Consumption Figures 38 and 39 show the calculation results of the town gas consumption as a result of the introduction of System A and System B to each building,

58

Shin'ya Obara

respectively. The calculated town gas consumption on representative days each month from the analysis results are shown in Figures 40 and 41. Figure 40 shows the town gas consumption for monthly representative days in the individual house, and Figure 41 shows the results of town gas consumption in other buildings. The conventional method in Figures 40 and 41 is an operational method which follows the power load, except for the water electrolysis equipment of the system shown in Figure 29. The efficiency η RM of the reformer in the conventional method changes with the value of the load. The model showing the relationship between the power load of the fuel cell system with a reformer and the generation efficiency of this system is shown in Figure 42 [23, 24, 29, 30]. The town gas consumptions of the conventional method shown in Figures 40 and 41 are calculation results when carrying out load-following operation using the characteristic of Figure 42. In the calculation results for each monthly representative day of each load pattern for the individual house, the hospital, and the factory, the reduction effect of the town gas consumption in System B is larger than the conventional method. However, for May and August representative days for the hotel, and every monthly representative day for the small store, System A and System B cannot reduce the town gas consumption in the conventional method. This is because partial loads with low efficiency occur frequently which increases the town gas consumption, if the load following operation by the conventional method is introduced into each load pattern for the individual house, the hospital, and the factory. On the other hand, when introducing the conventional method into the load pattern of a hotel or a small store, a drop in the efficiency of the system has a small sampling time of a partial load with low efficiency few.

Reformer and Auxiliary Machines Operation The reformer is stopped in the sampling time when there is no consumption of town gas in Figures 38 and 39. As shown in each figure, since the floating loads are small, the reformer can operate at a maximum efficiency point. Since starting and stopping of the reformer takes nearly 20 minutes, it is necessary to reduce the number of times this occurs which will also reduce consumption of energy ［23, 24］. As shown in Figures 38 and 39 (a), (b), (c), in the load patterns of the individual house, the hospital, and the factory, the reformer can be stopped on many sampling time zones, respectively. Therefore, if the operational

Load Leveling of Fuel Cell System…

59

20

16 12

0

5 0 4 0 3 0 2 0 1 0 00

Electric power output of the fuel cell stack [ kW]

30 25 20 15 10 5 00

30 25 20 15 10 5 0 0

4

8 12 16 Time [Hour] January

4

8

4

4

20

12 16 20 Time [Hour] January

8 12 16 Time [Hour] January

8 12 16 Time [Hour] January

20

0.4

0.2

Electric power output of the fuel cell stack [ kW]

8 4 0

0.6

0.4

Electric power output of the fuel cell stack [kW]

8 12 16 Time [Hour] January

0.8

0.6

0.2

00

20 16 12 8 4 00

4

8 12 16 20 Time [Hour] May (a) Individual house

4

8 12 16 Time [Hour] May (b) Hospital

10 00

30 25 20 15 10 5 00

30 25 20 15 10 5 00

00

20 16 12 8 4 0

0

4

8

12 16 Time [Hour] August

20

4

8

12 16 Time [ Hour] August

20

4

8

12 16 Time [ Hour] August

20

50

40 30 20

40 30 20 10

4

5

12 16 Time [Hour] May (c) Factory

20

4

8 12 16 Time [Hour] May

20

4

8 12 16 Time [Hour] May

20

(d) Hotel

20

20

50

Electric power output of the fuel cell stack [kW]

4

20

Electr ic power output of the fuel cell stack [kW]

Electric power output of the fuel cell stack [kW]

00

Electric power output of the fuel cell stack [kW]

0.2

1.0

0.8

Electric power output of the fuel cell stack [kW]

0.6 0.4

1.0

Electric power output of the fuel cell stack [kW]

Electric power needs System B

0.8

Electric power output of the fuel cell stack [ kW]

Electric power output of the fuel cell stack [ kW]

1.0

Electric power output of the fuel cell stack [kW]

Electric power output of the fuel cell stack [kW]

methods using System A or System B are introduced, starting and stopping will only occur once or twice each day.

(e) Small store

Figure 37. Operation planning of system B.

00

60

Shin'ya Obara Table 10. Fuel cell capacity reduction effect

0.7 kW 18.2 kW 36.0 kW 40.0 kW (21.8 kW)

8

24

8

24

0. 6

16

0. 2 0

0

4

8 12 16 2 0 T ime [Hou r] (a) In di vi du al h ou se

0.49 kW 13.6 kW 23.0 kW (27.2 kW) (21.8 kW)

0

4

8 12 1 6 Time [H ou r] (b) H os pi tal

20

24

8 4 0 0

4

8 12 16 2 0 Ti me [Ho ur] (c) Fact ory

24

16

30

12

20 10 0

4 0

System B

12

12

0. 4

Con sumption of town gas [ m3/h our]

0.94 kW 22.0 kW 50.0 kW 41.0 kW, 28.0kW(May) 24.0 kW(January)

Co nsump tio n of town 3 gas [m /hou r]

System A

Co ns umption of town gas [m 3/hou r]

Co ns umption of town 3 gas [m /hou r]

Co nsump tio n of town 3 gas [m /hou r]

(a) Individual house (b) Hospital (c) Factory (d) Hotel (e) Small store

Conventional method

0

4

8 12 1 6 T im e [ Ho ur] (d ) Ho tel

20

J an uary M ay Au gu st

4 0

0

4

8 12 16 Time [ Ho ur] (e) Smal l store

20

24

8

24

0.1

0

0

4

8 12 16 20 Time [Hou r] (a) In divid ual h ouse

6 4 2 0

0

4

8 12 16 Time [Hou r] (d) Ho tel

20

0

4

8 12 16 Time [Hou r] (b) Hosp it al

20

24

6

Con sump tion of town gas [m3/h our]

Consumpti on of town gas [m3/hou r]

10 8

24

5 4 3 2 1 0

Con sump tion of town gas [m3/h our]

Consumpti on of town gas [m3/hou r]

0. 2

Con sump tion of town gas [m 3/ hour]

Figure 38. Town gas consumption using system A. 8 6 4 2 0 0

4

8 12 16 Ti me [Hou r] (c) Factory

20

24

Janu ary May A ugu st

4 2 0

0

4

8 12 16 20 Ti me [Ho ur] (e) Small sto reMay

24

Figure 39. Town gas consumption of the sytem B.

On each monthly representative day for each building, the calculation results of the number of times required for the operation of blowers B1 , B 2 and B 3 are

Load Leveling of Fuel Cell System…

61

Consumption of town gas 3 [m /day]

shown in Figure 43. There is more operational time required of the blowers for System A, except for the results of the small store, when compared to System B. The reasons for this include operation of the reformer at times of high load, mixing of the hydrogen from the gas cylinder with the reformed gas, and supplying the fuel cell.

8

Conventional system

6

System A

4

System B

2 0

January

May

August

Figure 40. Town gas consumption of the representation day of the system A and the system B introduced into the individual house.

400

300

200

Conventional system System A Syste m B

Hospital

Factory

Hotel

A ug u st

M ay

J an u ary

A ug u st

M ay

ary J anu

Aug ust

J anu

0

Aug ust Janu ary M ay

100

ary Ma y

C onsumption of town gas [ m3/d ay]

500

Sma ll stor e

Figure 41. Town gas consumption of the representation day of the system A and the system B introduced into the individual house, factory, hotel, small store and office.

Shin'ya Obara Efficiency of system electric power output [%]

62 40 30 20 10 0

0

20 40 60 80 100 Ratio of electric power load [%]

Operation time [Hour/day]

Figure 42. The related curve of the ratio of power load of a fuel cell system with reformer, and efficiency of electric power output. 24

System A System B

20 16 12 8 4

Janu ary May Aug ust Janu ary May Aug ust Janu ary May Aug ust Janu ary May Janu ary May

0

d

House

Hospital

Factory

Hotel

Small store

Figure 43. Operation time of the blower B1, B2, B3.

CONCLUSION If gas with a high oxygen concentration is supplied, rather than air, to the cathode of a fuel cell stack, the generation efficiency of the fuel cell improves. A reduction in the installed capacity of a fuel cell stack was attempted in this study using this power-generation characteristic. During periods of small electricity demand, the fuel cell is operated with reformed gas and air, along with simultaneous water electrolysis operation. Hydrogen and oxygen are produced by water electrolysis, and compression

Load Leveling of Fuel Cell System…

63

storage of these gases is carried out in cylinders. On the other hand, when the electric power load is large, the cylinder gases are supplied to the fuel cell. In this case, the town gas consumption is smaller using the method which supplies hydrogen and oxygen of cylinders to the fuel cell directly for specific load patterns, such as the individual house, the hospital, and the factory, rather than mixing hydrogen and oxygen of the cylinders with the reformer and blower air. In order to introduce a water electrolysis system and to obtain a larger reduction effect of the capacity of the fuel cell stack, sufficient hydrogen and oxygen to reduce the load peak must be produced. Therefore, the reduction effect of the fuel cell capacity changes with the load pattern. The periods of low and high loads are divided clearly, and the effect is large when the difference between the low load and high load values is large. If the load-following operation of a building with load fluctuation is performed by a fuel cell system with a reformer, partial load operation with low efficiency will occur frequently. According to the operational method of the system proposed in this section, the load fluctuation range of a reformer is narrow and drops in efficiency can be prevented. Moreover, the number of times of starting and stopping of a reformer can be limited to once or twice each day.

INDEX A accuracy, 2, 32, 36 air, 7, 8, 24, 27, 30, 31, 42, 43, 44, 45, 47, 50, 51, 53, 57, 62 algorithm, 22, 29, 32 anode, 43, 44, 46, 47, 51, 53 application, 19, 27

B battery, 41, 44 buildings, 21, 26, 27, 29, 33, 36, 40, 41, 54, 58 burn, 49 burning, 49 burns, 48, 49 business, 1, 54 butane, 6

C calcium, 7 capacity, 2, 3, 5, 7, 13, 15, 19, 24, 30, 33, 39, 41, 42, 43, 45, 48, 54, 55, 57, 60, 62, 63 carbon, 2, 10, 43 carbon dioxide, 2, 10 carbon monoxide, 43 case study, 2, 7

catalytic, 4 cathode, 42, 43, 44, 46, 47, 51, 52, 53, 57, 62 cell, 1, 2, 4, 6, 10, 15, 18, 19, 21, 22, 23, 24, 25, 27, 30, 31, 33, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 57, 58, 60, 61, 62, 63 charcoal, 49 chemical, 6 chloride, 7 chromosome, 29, 30, 32 CO2, 49 coefficient of performance, 6 combustion, 4 commercial, 1, 2, 10, 13, 18, 19, 41, 43, 45 composition, 19 compositions, 11 compression, 7, 42, 52, 62 computation, 32 concentration, 12, 42, 44, 46, 47, 52, 53, 54, 57, 62 condensation, 6 configuration, 14 Congress, iv consumption, 2, 7, 8, 10, 11, 27, 36, 42, 43, 45, 47, 49, 50, 51, 55, 57, 58, 60, 61, 63 control, 21 cooling, 7, 54, 57 cost minimization, 13

66

Index

D demand, 2, 3, 4, 10, 13, 14, 18, 19, 21, 24, 27, 28, 30, 36, 37, 38, 40, 41, 42, 43, 48, 50, 54, 55, 62 density, 7 deregulation, 1 distribution, 1, 21

E electric energy, 10 electric power, 2, 3, 6, 7, 9, 10, 14, 15, 18, 19, 22, 42, 43, 45, 46, 47, 49, 50, 51, 52, 53, 54, 55, 56, 62, 63 electrical, 2 electrical power, 3 electricity, 9, 26, 42, 45, 48, 49, 50, 52, 55, 62 electrochemical, 44 electrochemical reaction, 45 electrolysis, 1, 2, 3, 6, 7, 10, 13, 42, 45, 47, 49, 50, 51, 53, 55, 57, 58, 62, 63 electronic, iv electrostatic, iv emission, 2, 10, 11, 13 energy, 1, 2, 4, 8, 9, 10, 13, 14, 19, 21, 22, 27, 30, 31, 40, 42, 54, 58 engineering, 54 engines, 1 environmental, 1 equipment, 1, 2, 3, 4, 9, 10, 11, 12, 13, 14, 15, 19, 22, 38, 39, 40, 49, 50, 58 exhaust heat, 3, 21, 23, 24, 27, 30, 31, 36, 39, 43, 54 expert, iv

F February, 2, 4, 9, 12, 13, 14, 15, 16, 19, 31, 37, 39 feeding, 43 fitness, 29, 30 floating, 58

flow, 9, 10, 23, 27, 30, 44, 45, 46, 48, 49, 50, 52, 53, 57 flow rate, 11, 27, 50, 53 fluctuations, 37 fuel, 1, 2, 4, 6, 8, 10, 11, 12, 15, 18, 19, 21, 22, 23, 24, 25, 26, 27, 30, 31, 33, 36, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 57, 58, 61, 62, 63 fuel cell, 1, 2, 4, 6, 10, 15, 18, 19, 21, 22, 23, 24, 25, 26, 27, 30, 31, 33, 36, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 57, 58, 61, 62, 63 fuel flow rate, 11

G gas, 2, 4, 7, 10, 11, 13, 19, 21, 22, 27, 41, 42, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 57, 58, 60, 61, 62 gases, 11, 42, 43, 44, 53, 63 gene, 2, 29 generation, 2, 21, 24, 30, 32, 33, 38, 41, 42, 43, 45, 46, 48, 51, 54, 57, 58, 62 genetic, 22, 29, 32 geo-thermal, 1, 2, 3, 6, 19 green energy, 1 greenhouse, 11, 13, 19 greenhouse gas, 11, 13, 19 greenhouse gases, 11 groups, 30

H heat, 1, 2, 3, 6, 7, 8, 9, 10, 13, 14, 15, 18, 19, 21, 22, 24, 26, 27, 29, 30, 33, 36, 37, 38, 39, 40, 41, 42, 45, 48, 49, 50, 54 heat loss, 22 heat release, 22, 24, 27, 29, 30, 31, 33, 36, 37, 39, 40 heat storage, 1, 2, 3, 7, 8, 10, 13, 14, 19, 23, 24, 27, 30, 31, 41, 43, 49, 50 heat transfer, 3, 22, 24, 27, 43 heating, 23, 54 hospital, 42, 54, 57, 58, 63

Index hospitals, 22 hot water, 3, 23, 24, 27, 30 household, 27 households, 27 hydrogen, 3, 6, 7, 10, 13, 42, 43, 44, 45, 46, 47, 49, 50, 51, 52, 53, 57, 61, 63

67

mixing, 61, 63 models, 2, 24, 29, 30, 42, 47, 54 modules, 1 mold, 27 mutation, 29, 30, 32

N I network, 21, 22, 23, 24, 26, 27, 30, 31, 36, 38, 39, 40 New York, iii, iv nonlinear, 2 numerical analysis, 54

id, 54, 57 independence, 2 injury, iv interval, 52

J January, 54, 56 Japan, 2, 4 Japanese, 11

O optimization, 1, 14, 18, 19, 36 oxidation, 49 oxygen, v, 3, 6, 7, 10, 13, 41, 42, 43, 44, 45, 46, 47, 49, 50, 51, 52, 53, 54, 55, 57, 62, 63

K P kerosene, 1, 13

L linear, 1 linear programming, 1 losses, 7

M machinery, 21, 22, 24, 30, 31, 33 machines, 27, 42, 54 magnetic, iv manipulation, 29 market, 54 measurement, 9 mechanical, iv meteorological, 31 methanol, 1, 3, 4, 5, 6 MIP, 2

PEMFC, 2, 3, 6, 19 performance, 6, 25, 46 planning, 1, 13, 16, 17, 19, 22, 30, 32, 33, 36, 40, 56, 59 polystyrene, 27 population, 32 power, 1, 3, 6, 7, 10, 13, 14, 15, 18, 19, 21, 22, 24, 26, 27, 30, 31, 33, 36, 37, 38, 39, 40, 41, 42, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 57, 58, 62, 63 power generation, 21, 24, 33, 38, 41, 42, 43, 45, 48, 51, 54, 57 preparation, iv pressure, 46 prices, 22 probability, 32 production, 7, 9, 43, 49, 51, 55, 57 program, 6, 22, 30, 32, 36, 40 programming, 1 propane, 6 property, iv

68

Index

proton exchange membrane, 1, 2, 7, 22, 41, 42, 49 pure water, 49

R radiation, 21 random, 30, 37, 38 range, 2, 6, 37, 48, 63 reactant, 46 recovery, 22, 45 reduction, 13, 39, 40, 41, 45, 48, 57, 58, 60, 62, 63 refrigerant, 6 relationship, 1, 8, 52, 54, 58 renewable energy, 1, 19 returns, 24, 30

S sampling, 5, 7, 26, 30, 50, 52, 53, 56, 58 searches, 13 searching, 29 services, iv short period, 15 simulation, 32 soil, 6 solar, 1, 2, 3, 8, 9, 13, 18, 19, 22, 38, 39, 40 solutions, 29 specific heat, 7 speed, 41, 48 speed of response, 41, 48 storage, 1, 2, 3, 7, 8, 10, 13, 14, 19, 23, 24, 27, 30, 31, 41, 43, 49, 50, 63 summer, 2, 4, 22, 27, 31, 33, 54, 57 supply, 2, 6, 8, 9, 10, 12, 19, 21, 22, 40, 41, 50, 51, 53 surplus, 3, 36, 42, 43, 48 switching, 44 symbols, 27

systems, 1, 3, 22, 24, 30, 36, 48, 57

T tanks, 3 technology, 21, 42 temperature, 3, 5, 6, 7, 8, 24, 30, 31, 46, 54 thermal, 1, 2, 3, 6, 19, 45, 50, 51 thermal efficiency, 51 threshold, 47, 48, 51, 53 time, 3, 5, 6, 7, 10, 11, 13, 21, 24, 26, 30, 32, 40, 41, 42, 43, 45, 47, 48, 50, 51, 53, 55, 56, 58, 61, 62 Tokyo, 54 transfer, 3, 7, 22, 24, 27, 43, 45 transmission, 21, 22 transport, 21

U urban, 22, 27, 30, 32, 40

V values, 5, 12, 13, 33, 52, 54, 55, 57, 63 variable, 2, 32 variables, 2

W waste, 22 water, 1, 2, 3, 6, 7, 8, 10, 13, 21, 22, 24, 27, 29, 30, 32, 33, 34, 35, 36, 37, 38, 39, 40, 42, 45, 47, 49, 50, 51, 55, 57, 58, 62, 63 wind, 1 winter, 2, 4, 9, 22, 27, 31, 33, 37, 38, 54, 56