[TuG2-4]

8th International Conference on Power Electronics - ECCE Asia May 30-June 3, 2011, The Shilla Jeju, Korea

Optimal configuration of energy supply system in a microgrid with steam supply from a municipal waste incinerator
S. Bando1, H. Asano1, K. Sasajima2, N. Odajima3, M. Sei4, and T. Ogata4
1

Central Research Institute of Electric Power Industry, 1-6-1 Otemachi, Chiyoda-ku, Tokyo, Japan 2 Nihon Sekkei Inc., 6-5-1, Nishi-shinjuku, Shinjuku-ku, Tokyo, Japan 3 Shimizu Corporation, 1-2-3, Shibaura, Minato-ku, Tokyo, Japan 4 Tokyo Gas Co., 1-7-7, Suehiro-cho, Tsurumi-ku, Yokohama, Japan

Abstract-- We present the optimal design of energy supply system of a urban microgrid by using a Mixed Integer Non Linear Programming model. We have conducted a feasibility study of energy supply system using waste steam from incinerator plants via a steam distribution network in the urban redevelopment project area in Tokyo as a case study. We compared the optimal configuration of both of a district heat and cooling (DHC) system and a microgrid system with sufficient self generation capacity as an energy supply system in the redevelopment area. We present the economic and environmental effectiveness of the microgrid with large scale cogeneration and steam supply from incinerator plants as unharnessed energy. Large scale cogeneration with high efficiency has capability to reduce 20% of primary energy consumption and 2% of the amount of CO2 emissions compared with a new constructed DHC system. Steam supply from incinerator plants has capability to reduce about 3% of CO2 emissions. Pay-back time of the microgrid for commercial and residential area is 4 to 6 years. The microgrid with large scale gas engine generators is economically viable. Index TermsCogeneration, District Heating and Cooling, Microgrid, Mixed Integer Non-Linear Programming, Optimal configuration

gas engine (GE) generators has improved remarkably. This is frequently called scale merit of efficiency. Furthermore, initial cost (investment cost) of the engine generator declines as the capacity of the generator increases. This is scale economy of investment cost. A lot of large scale gas engine generators, say, MW class, are often installed as generators for district heat and cooling (DHC) system in Europe and Japan. Recently in Japan, large scale gas engine generators are focused as a key technology of urban redevelopment and construction of sustainable city, and the development of design method of energy supply system is necessary with consideration of the scale economy and scale merit of gas engine generator. Though penetration of renewable energy resource, such as photovoltaic, is another key to realize sustainable city, the capacity of renewable energy can be deployed in city area is far smaller than electric demand scale in the area because of buildings verticalization. Therefore, we focus on the waste heat from incinerator plant as unutilized energy. The optimal configuration of energy supply system in a DHC area near the steam distribution network from incinerator plant was calculated. II. NOMENCLATURE Cl [JPY]: daily cost of each representative day, l CD [kW]: the contract demand of electricity purchased from the utility COP [-]: COP of each heat supply system D [-]: the number of startups and shutdowns of a GE EGE [kWh]: sum of the electricity generated by a GE F [m3]: the volume of gas consumed I [-]: interest rate IC [JPY/kW]: initial cost of each equipment K [kW]: rated power of each equipment Pd [kW]: electric power demand PGE [kW]: output from GE generator PEsell [kW]: selling electric power PEbuy [kW]: buying electric power PTR [kW]: electric power consumed by turbo refrigerators Paux [kW]: electric power consumed by auxiliary equipments QdCW [kW]: cool water demand

I. INTRODUCTION The recent rolling blackout experiences have demonstrated the vulnerability of the interconnected electric power system to grid failure caused by natural disasters. A restructured electric distribution network which employs a large number of small distributed energy resources (DERs) units can improve the level of system reliability and provide service differentiations. With the advance of small distributed-generation (DG) technology and the ongoing deregulation of the electricity market, the microgrid concept is expected to be widely implemented in the coming years[1]. The concept of a microgrid involves interconnection of both small DG and loads through a local grid. These microgrid systems can be connected to the main power network or operated autonomously in an islanded mode when the main power network is seriously disrupted. Economy and energy efficiency of a microgrid depend on the scale and generation efficiency of controllable prime movers. Especially, the efficiency of large scale

978-1-61284-957-7/11/$26.00 2011 IEEE

QdST [kW]: steam demand QGB [kW]: steam generated from gas boilers QGE [kW]: steam recovered from GEs QGL [kW]: cool water output from steam and hot water absorption refrigerator; QGLH +QGLS QGLH [kW]: cool water generated from hot water by steam and hot water absorption refrigerator QGLHin [kW]: hot water consumed by steam and hot water absorption refrigerator QGLS [kW]: cool water generated from steam by steam and hot water absorption refrigerator QGLSin [kW]: steam consumed by steam and hot water absorption refrigerator QNT [kW]: steam purchased from incinerator plant QRS [kW]: cool water output from a steam absorption refrigerator QRSin [kW]: steam consumed by steam absorption refrigerator QSTORE [kW]: cool water output from thermal storage tank; if QSTORE>0, cool water is released from tank, and vice versa. QTR [kW]: cool water output from turbo refrigerator QWS [kW]: waste steam QWH [kW]: waste hot water R: capital recovery factor Tl: the number of the day for each representative day in a year UBUE [JPY/kW/month]: backup demand charge of electricity (= 0.367*UDC) UDC [JPY/kW/month]: demand charge of electricity UD GE [JPY/number]: cost of starting and stopping a gas engine UEbuy [JPY/kWh]: charge rate for energy purchased from the utility UEsell [JPY/kWh]: charge rate for energy sold to the utility Ugas_ec [JPY/m3]: energy charge rate of gas Umtn [JPY/kWh]: maintenance cost of a GE V [m3]: capacity of the thermal storage tank x [-]: load factor of electric power output Z [JPY/year]: annual cost [-]: on-off binary variable of engine operation : lifetime E [-]: efficiency of generation H-ST [-]: thermal recovery efficiency from GE for steam H-HW [-]: thermal recovery efficiency from GE for hot water Subscript i: unit number of gas engines j: time Acronyms GB: gas boiler GE: Gas Engine GL: steam and hot water absorption refrigerator RS: steam absorption refrigerator TR: turbo refrigerator

III. THE ECONOMIC OPTIMIZATION OF MICROGRID WITH DHC The area investigated here has its urban redevelopment project. Table I shows the predicted mix customers in the area after the redevelopment. In another large redevelopment project, a steam distribution network from incinerator factory is also under consideration. We simulated the optimal configuration of both of DHC system and microgrid system as energy supply system in this area. Electric and thermal demands are estimated from each customers floor space. A. Models of Microgrid A microgrid system investigated here as schematically shown in Fig.1 is composed of large gas-engine generators (up to 5,200kW), steam-absorption refrigerator, steam and hot water absorption refrigerator, gas boiler, turbo refrigerator, and steam network from waste incinerator plants. In this system, electricity is supplied to a part of an existing DHC area in Tokyo by the parallel operation of multiple gas engine-driven generators and by power purchased from the utility grid. Steam is supplied to the DHC area by exhausted heat from gas engines, gas boiler and steam distribution network, and cool water is supplied by turbo refrigerator, steam/hot water absorption refrigerator, and steam absorption refrigerator. To compare the energy savings and cost savings potential of steam supply from waste incinerator plants and thermal storage, we set six case studies (See Table II); the first one is DHC case without distributed generation (Case 1), the second is DHC with steam network (Case 2), the third one is a microgrid without steam network (Case 3), the forth is a microgrid with steam network (Case 4), the fifth is a microgrid with steam network and their own thermal storage (Case 5), and the last one is microgrid with steam network and large own thermal storage (Case 6). The thermal storage tank is used for cool water storage only.
TABLE I ESTIMATED CUSTOMER CONFIGURATION Floor space [m2] Office 715,000 Retail 195,000 Hotel 65,000 Apartment 260,000 Musium 65,000

(a)

Electric demand in a microgrid

(c) (b) Steam demand in a microgrid

Case 3

(d) (c) Cool water demand in a microgrid Fig.1. Demand of a microgrid

Case 4

Fig. 2. (a) Case 1 Case No.

(e) Case 5 and Case 6 Structure of a microgrid connected to steam network. TABLE II CASE SETTINGS Steam supply from steam network No Yes: 5.2 ton/h No Yes: 5.2 ton/h Yes: 5.2 ton/h Yes: 14.4 ton/h

Electricity supply system

Case 1 DHC system Case 2 DHC system Case 3 Microgrid* Case 4 Microgrid* Case 5 Microgrid* Case 6 Microgrid* 1 [ton/h] = 626 [kW] * interruptible load is required for islanded mode operation (b) Case 2

Thermal storage tank No No No No Yes Yes

Volume of thermal storage tank 0 [m3] 0 [m3] 0 [m3] 0 [m3] 2,800 [m3] 5,600 [m3]

B. Mathematical formulation of constraint functions Technological advances in gas engine generators have been significant. Thermal efficiency of a gas engine generator is now at least 45% (LHV) higher than the 5,000kW class gas engine currently used in power generation systems. In general, a gas engine generator has the lowest load factor, typically 50% of the rated capacity. The partial load efficiency depends on the load factor of

the electric output. The scale dependency of such nonlinear partial load efficiency has not yet been considered in optimal sizing of gas engine generator for a microgrid. Annual cost is minimized by considering the partial load efficiency of a gas engine generator and its scale economy, and the optimal number and capacity of each piece of equipment and the annual operational schedule are determined by using the optimal planning method. The power generation efficiency of an engine generator is approximated by a regression expression for the rated power and the load factor (3): E E 4.008 10 5 K GE 0.2038 xGE GE (1)
E 2.323 10 5 K GE xGE 0.1842
H ST E 5.969 10 6 K GE 0.02257 xGE GE E 9.312 10 6 K GE xGE 0.2058 H HW E 2.302 10 6 K GE 0.05011xGE GE E 2.273 10 5 K GE xGE 0.1326

(2) (3) (4)

0.5 GEi j xi j 1 GEi j

The gas engine generator can be operated at a load factor between 50% and 100% as in equation (4). Pd j PTR j Paux j PGEi j PEbuy j PEsell j (5)
ST j QS j QGB j QNT j QRSin j QGLSin j QWS j Qd

(6) QdCW j QRS j QGL j QTR j QSTORE ( j ) (7) (8) QHW j QGLH j QWH j
PEbuy j CD (9),
i
i in

PEbuy j PEsell j 0

PEsell j CD (10)

(11)

Eqs. (13-1) to (13-6) are constraints about the amount of steam purchased from incinerator. In Case 1(Eq.13-1) and Case 3 (Eq.13-3), the amounts of steam purchased from incinerator are set to zero. Eqs. (13-2), (13-4), (135) and (13-6) show the amounts of steam bought from steam distribution network is determined steady at every hour, every season in Case 2, Case 4, Case 5 and Case 6 respectively. These assumptions are based on the predicted use cases of steam network because the incinerator factories usually burn the gathered garbage at steady pace in every hour and every season in Tokyo. In Case 2, Case 4 and Case 5, the contracted volume velocity from steam network is set as the minimum value can be consumed every hour if gas engines stop at night. In Case 5, we set storage tank volume 2,800 [m3] under the balance between the cost of construction of storage tank and the benefit of thermal storage in this microgrid in the test calculation. In Case 6, we determine the amount of steam from network as 14.4 ton/h, which is maximum steady volume velocity of steam from incinerator that can be consumed up in this microgrid with using fully large storage tank in a day; it equals the minimum value of the sum of hourly steam demand and the average hourly cool water demand in a representative day. Storage tank volume is set to 5,600 [m3] which is enough capacity to consume cool water generated from network steam in every day through a year. We also set the operation pattern of storage tank; cool water is produced by steam from network and stored in the tank in night time, and is released in daytime in order to shave the peak demand of cool water. Storage tank volume and operation pattern of the tank are not variable determined by the optimization.

PGEi j K GEi , QRS j K RS , QGB j K GB , QGL j K GL , QTR j K TR

K
i

GEi

MAX CD E d K TR / COPTR MAX aux

(12)

E

Eqs. (5) to (8) represent the balance between electric demand and supply, the balance between steam demand and supply, the balance between cool water demand and supply, the balance between hot water demand and supply, respectively. Eqs. (9) and (10) show the power flow between the utility and microgrid does not exceed the contract demand. And Eq. (11) expresses selling electricity is not allowed when electricity is purchased from the utility. Eqs. (12) show the outputs of each equipment do not exceed its scale.
QNT j 0
CW ST j Qd j COPRS QNT j min Qd constant

K GEi , K RS , K GL , K TR , K GB i

(14)

MAX MAX MAX QST QS K GEi K GB Q NT i

(15)

Q K K
MAX Si GEi i

GB

ST (16) K RS / COP RS KGL / COP GL

QDCMAX K RS K GL KTR Qstore

(17)

ST MAX MAX K RS / COPRS K GL / COPGL QS K GEi K GB Q NT i i

(13-2, 13-4, 13-5)

24 CW j Qd ST 1 constant QNT j min Qd j 24 COPRS

(13-6)

(18) In this research, the energy supply(13-1), system is designed (13-3) in order that it can supply the momentary maximum electric, steam and cool water load. So we set the constraint of Eqs. (14)-(18). Table III and Table IV list parameters for simulations. The initial costs of the equipment were determined based on current market price. 80 JPY equals to 1 USD.

RS GL TR GB

TABLE III PARAMETERS. Scale constraint 2,000 5,200 [kW] Load Rate Constraint 50 100 [%] Unit cost of maintenance 2.5 [JPY/kWh] Start-up and stop cost 0.286 [JPY/kW] -6.113[JPY/kW2] KGE2 Initial Cost + 99,791 [JPY/kW] KGE IC(KGE) + 11,321,987[JPY] 51,136[JPY/kW]KRS Initial cost, IC(KRS) Initial Cost, IC(KGL) 54,000 [JPY/kW] KGL COP(driven by steam) 1.51 COP(driven by hot water) 0.8 57,054 [JPY/kW] KTR Initial Cost, IC(KTR) +35,360,000[JPY] 28,409[JPY/kW]KGB Initial cost, IC(KGB) TABLE IV

where Z is annual cost, T is the number of day for each representative day in a year, Cl is the average daily cost of the representative day, UBUE is the backup demand unit charge of electricity (Eq. 21) for the penalty of unavailability of the generators, and R is capital recovery factor (Eq.22). The backup contract demand was determined according to the maximum capacity of a single gas engine (Eq.20). (21) U BUE 0.367 U CD The assumed lifetime of each equipment is 15 years. The interest rate for capital cost is 0.03.

Gas Engine

Gas Steam Selling Energy charge Electricity

UNIT COST OF ENERGY Energy charge 45 Energy charge 0.80 Demand charge 1,533 Peak (Summer 13:00-16:00) 12.02 Shoulder (Summer) 11.44 Shoulder (Spring & Autumn, 10.34 Winter) Off-peak (22:00-8:00, holyday 7.08 & Sunday) Peak (Summer 8:00-22:00) 5.5 Shoulder 4.95 Off-peak (22:00-8:00, holyday 2.45 & Sunday)

Electric Power

JPY/Nm3 JPY/MJ JPY/kW/month JPY/kWh JPY/kWh JPY/kWh JPY/kWh JPY/kWh JPY/kWh JPY/kWh

I (1 I ) (22) 1 I 1 Cl consists of energy charge of gas, charge for energy purchased from the steam network, charge for energy purchased from the utility, charge for energy sold to the utility, contract demand from the utility, maintenance cost, the cost of starting and stopping a GE, and the initial cost of all equipments. It is supposed that the scale of economy of the generator make larger size more economical while a utility backup supply contract would avoid the larger single unit of generator due to the penalty of unavailability of the generator. R

C. Mathematical formulation of objective function Annual cost is defined as the sum of the operating cost and the annualized initial cost (capital cost). The annual operational cost is calculated for a total of six representative days: three seasons (summer, winter, and spring & fall) and two day of week, weekdays and holiday. Decision variables are composed of design (capacity) and operational variables. The problem formulated here is a Mixed Integer Non Linear Programming because it includes nonlinear consumption of fuel and includes the number of gas engines in operation and the number of times the gas engine generators start-up and shut-down. The formulated problem is then solved by using GAMS (a modeling system for mathematical programming and optimization). The objective function to be minimized is defined as follows:

IV. RESULTS AND DISCUSSIONS Table V shows the configuration of each case. results of the optimal

TABLE V RESULTS OF OPTIMAL CONFIGURATION Case No. Gas engine generator [kW] The number of GE RS [USRT] GL [USRT] TR [USRT] GB[ton/h] Contracted demand [kW] Contracted volume velocity of steam from network[kW] Volume of thermal tank [m3] Payback time [year] 1 [USRT] = 3.53 [kW] 1 0 0 2 3 4 5 6 0 56,200 56,200 56,200 36,400 0 11 6,332 11 9,120 11 7

0 21,967 0 31,586 86.9

6,881 15,989 7,035 4,717 57.3

0 12,897 10,330 12,108 9,619 12,357 12,136 11,201 73.3 52.6 47.4 47.4

83,170 69,380 13,269 24,256 12,650 44,356

Z C l Tdl
l 1

(19)

3,277

3,277

3,277

9,025

U st Q NT j C l U gas FGEi j FGB j j j i, j U Ebuy j PEbuy j U E sell j PEsell j U CD CD / 30

j j

2,800

5,600

U BUE max K GE / 30 U mtn PGEi j U DGE K GEi DGEi

i, j i

4.6

5.5

3.9

3.1

IC GEi K GEi IC RS K RS / 365 R i IC K IC K IC K GL GL TR TR GB GB

(20)

In Cases 3, 4, and 5, the sum of capacities of gas engines are 56,200 kW and the number of gas engines is

11 units. In Case 6, the sum of capacities of gas engines are less than that in Cases 3, 4 and 5. This is because 30 % of thermal demand is supplied by steam from network and that they cannot consume the steam generated from more than 8 gas engine units. Thus, the microgrid in Case 6 must have about more 20,000 kW interruptible electric demand than in other cases when it is operated in islanded mode during its peak demand time. Figures 3, 4, 5 and 6 show the result of energy demand and supply in a microgrid about electric power, steam, and cool water on a weekday in spring/autumn in each microgrid case.

demand and cool water demand, however, two gas engine units continue to be operated for thermal supply and the cost is decreased by reduction of contracted demand. The electricity is purchased from the utility during off-peak time and peak time because off-peak electricity is inexpensive and the output from gas engines cannot meet the electricity demand during the peak time.

(a)

Electricity demand and supply

(a)

Electricity demand and supply

(b)

Steam demand and supply

Fig. 4 (b) Steam demand and supply

(c) Cool water demand and supply Energy demand and supply on a weekday in spring/autumn weekday in Case 4

Fig. 3

(c) Cool water demand and supply Energy demand and supply on a weekday in spring/autumn weekday in Case 3

In Case 3, gas engines are operated in all day, and the exhausted heat are utilized for steam demand and input to steam/hot water absorption refrigerator. In night time, partial exhausted steam is wasted due to low steam

In Case 4, operation of gas engines is stopped in night time on a weekday in spring/autumn. There are two reasons; the first one is that the purchased steam from incinerator is enough for steam demand and cool water demand in night time, and second one is that unit price of grid electricity is less expensive than generation cost of gas engine in case that all the exhausted heat from gas engine is wasted. In Case 5, gas engines are operated in all day, and the exhausted heat from gas engines and purchased steam from incinerator are utilized for steam demand and input to steam/hot water absorption refrigerator. In night time, partial exhausted steam from gas engine cannot be utilized because of the limited capacity of cool water storage tank, however, two gas engine units continue to be operated through night time for thermal supply and the cost reduction of contracted demand.

(a)

Electricity demand and supply

(c) Cool water demand and supply Fig. 6 Energy demand and supply on a weekday in spring/autumn weekday in Case 6

In Case 6, operation of gas engines is stopped from 0 to 6 oclock on a weekday in spring/autumn. Purchased steam from incinerator in night time is utilized for steam demand and input to steam/hot water absorption refrigerator. Stored cool water in tank is utilized for cool water supply in day time.
(b) Steam demand and supply TABLE VI UNIT PARAMETER OF ENERGY 2.362 CO2 emission coefficient[3] Gas Heating value (LHV), G[3] 41.6 Steam 0 CO2 emission coefficient 0.425 CO2 emission coefficient[3] Utility grid Average energy efficiency 36.1 electricity coefficient

kg-CO2/Nm3 MJ/Nm3 kg-CO2/MJ kg-CO2/kWh %

Fig. 5

(c) Cool water demand and supply Energy demand and supply on a weekday in spring/autumn weekday in Case 5

(a)

Electricity demand and supply

(b)

Steam demand and supply

Figs. 7 and 8 show annual primary energy consumption and annual amount of CO2 emissions. Primary energy consumption is regarded as the sum of that of gas and electricity, doesnt include that of steam from distribution network because it is regarded as unutilized energy. Table VI shows unit parameters of energy used in the calculation of primary energy consumption and CO2 emission. CO2 emission coefficient of utility grid electricity is so low that increase of utilization percentage of exhausted heat from gas engine is a key for the reduction of CO2 emission in microgrid cases (Case 3 -6). Primary energy consumption in Case 3 is not different from that in Case 4. This is because effect of primary energy consumption reduction issued by introduction of steam from incinerator is canceled out due to increase of primary energy consumption issued by stop of cogeneration operation in night time. Amount of purchased electricity from grid utility in Case 4 is larger than that in Case 3, thus, CO2 emission in Case 4 is 3.2% smaller than that in Case 3. In Case 5, primary energy consumption is 2.2% smaller than that in Case 4 due to the long operation of cogeneration. Though, CO2 emission in Case 4 is not much different from that in Case 3. Large scale cogeneration with high efficiency has capability to reduce 20% of primary energy consumption and 2% of the amount of CO2 emissions compared with new DHC system (Case 1). And steam from incinerator has capability to reduce about 3% of CO2 emissions.

In Case 6, initial cost is about 20 % more expensive than that in Case 1. Initial cost of CGS in Case 6 is about 60% of that of Case 3. The initial cost of thermal supply system are reduced due to utilization of cool water storage tank as well as Case 5. Annual running cost is 12% lower than that in Case 1. Thus, the total cost in Case 6 get lower than that in Case 1 after 3 years operation. V. CONCLUSION
Fig. 7 Annual consumption of primary energy.

Fig. 8

Annual amount of CO2 emission.

A methodology was developed to design the number and capacity for each piece of equipment in a microgrid and to analyze the profitability of microgrids. We present the economic and environmental effectiveness of the microgrid with large scale cogeneration and steam supply from incinerator plants as unharnessed energy. Large scale cogeneration with high efficiency has capability to reduce 20% of primary energy consumption and 2% of the amount of CO2 emissions compared with new DHC system. And steam from incinerator has capability to reduce about 3% of CO2 emissions. Pay-back time of microgrid is from 4 to 6 years. This method is useful for designing sustainable city in growing energy demand area such as emerging economy. REFERENCES
[1] R.H.Lasseter, and P.Piagi, Microgrid: a conceptual solution, PESC 2004, Aachen, Germany, June 2004 [2] S. Bando, and H. Asano, Optimal Capacity Sizing of an Energy Supply System in a Microgrid by Considering the Effect of Backup Contract and Scale Economy, Intelligent Automation and Soft Computing, vol. 16 (2), pp. 269-283, February 2010. [3] The Japan Institute of Energy; Planning and Design Manual of Gas Cogeneration Systems 2005, Japan Industrial Publishing CO., LTD , Tokyo, 2005 (in Japanese)

(a)

Breakdown of initial cost ratio

(b)

Breakdown of annual running cost ratio Fig. 9 Cost analysis in each case

Initial costs in Cases 3, 4 and 5 are about 40% to 50% more expensive than that in Case 1 due to initial costs of CGS. In Case 5, utilization of thermal tank reduces the scale of refrigerators, thus, reduces initial cost of thermal supply system inexpensive. Annual running costs in Cases 3, 4 and 5 are about 20% lower than that in Case 1. After 4 to 6 years operation of microgrid, their total costs in Cases 3, 4 and 5, which is sum of running cost and initial cost, get lower than that in case that new DHC system is constructed (Case 1). (See Table V)