 Advances in Technology Innovation , vol. 1, no. 1, 2016, pp. 07 - 12 7 Copyright © TAETI Benefit Analysis of Emergency Standby System Promoted to Cogeneration System Shyi-Wen Wang Department of Electrical Engineering, Chien kuo Technology University, Changhua, Taiwan. Received 03 March 2016; received in revised form 25 April 2016; accept ed 30 April 2016 Abstract Benefit analysis of emergency standby system comb ined with absorption chille r promoted to cogeneration system is introduced. Economic evaluations of such upgraded projects play a majo r part in the decisions made by investors. T ime-o f-use rate structure, fue l cost and system constraints are taken into account in the evaluation. Therefore, the proble m is formulated as a mixed-integer progra mming problem. Using two-stage methodology and modified mixed-integer progra mming technique, a novel algorithm is developed and introduced here to solve the nonlinear optimization proble m. The net present value (NPV) method is used to evaluate the annual benefits and years of payback for the cogeneration system. The results indicate that upgrading standby generators to cogeneration systems is profitable and should be encouraged, especially fo r those utilities with insufficient spinning reserves, and moreover, for those having difficu lty constructing new p ower plants. Keywor ds : emergency standby system, cogeneration system, time-of-use rate structure, mixed-integer programming, nonlinear optimization 1. Introduction A marked increase in the cost of constructing a generation, transmission and distribution system have also resulted in higher de mand charges to customers in the past few years. To reduce the system's peak load and therefore reduce the system's idle stand -by capacity, seasonal and time -of-day (or peak-load pric ing) rate structures are usually applied by the utilities. This has increased the benefit of peak shaving by using dispersed-storage-and-generation (DSG) systems [1]. When made a part of an energy management system (EMS)[2] or distribution dispatch center (DDC) of a utility system, DSG may provide benefits to utilities by reducing system peak load, improving reliability, and increasing operation efficiency, which implies that the opportunity cost of generation can thus be reduced by the DSG. Cost reductions or avoided costs result fro m saving of both capital carrying charges and operation e xpenses. However, additional capita l investment will be required fo r the DSG, so the benefit obtained by the utility should be shared with partic ipating customers to promote the DSG. T wo customer incentives for DSG a re that it a llo ws a time -of-use (T OU) rate structure based on the peak-load pric ing theory and reduced fuel costs. Considering the tre mendous quantities of waste heat generated in the production of electric ity, it is apparent that there is an opportunity to save fuel by a cogeneration system (CGS). CGS can be described as the simu ltaneous generation of electrica l or mechanica l power and usable energy by a sin gle energy conversion system [3]. It has long been common here and abroad. Currently, there is renewed interest in CGS because the overall energy efficienc ies are claimed to be as high as 70 to 85%. Reducing the cost of electricity for industrial and comme rcia l users, re liev ing e xcessive demand on utilit ies, and using fuel effic iently a re certainly worthwhile goals. However, carefu l planning in design and operation is necessary to meet these goals. Therefore, in this paper, a novel optima l operation scheme for sma ll cogeneration systems that are upgraded from standby generators is introduced. * Corresponding aut hor, Email: wshin@ctu.edu.tw Advances in Technology Innovation , vol. 1, no. 1, 2016, pp. 07 - 12 8 Copyright © TAETI 2. Method Fig. 1 illustrates the structure of a small CGS investigated in this paper. The sa mple CGS system is upgraded from a 100 kW gas -engine-driven standby generator by adding an absorption refrigerator and other necessary accessories. The hot-water capacity of this absorption refrigerator is 50 RT . The electricity demands of the system a re usually supplied fro m a public electric utility. However, the cogenerator may operate in a para lle l manner to optimize the energy supplies of both the electric ity and cooling de mands for some suitable time period. In order to optimize the operation procedure, the parameters of the system are obtained by fie ld testing the sample CGS under fu ll and some partia l load conditions. On the basis of the data of the fie ld tests, the fuel cost curve of this system can be obtained by the curve-fitting techniques as shown in Fig. 2 by the dotted line when the fuel cost is 7.84 NT$/m 3 . In th is system the heat recov ered by the absorption chiller is transformed to equivalent p o we r and so is de duc te d fro m th e po we r d e ma nd o f t he c en t rifug a l ch ille r. On t h e Gas engine Absorption chiller Exhaust boiler Centrifigual chiller GCity gas Generator Purchased electricity Mixing tank Pump water steam exhaust air exh aust air Space cooling demand exhaust air Electricity demand Fig. 1 Structure of a Gas-engine Cogeneration System Fig. 2 Fuel-cost Curve of the Sample CGS System contrary, for the conventional standby generator the o n ly output is electric power. Its power output curve is also shown in Fig. 2 for comparison with that of the CGS. 2.1. Problem Formulation For customers with a CGS that is upgraded fro m a standby generator, the total operating cost can be e xpressed simply as the sum of the payments for electric ity to the electric utility and for gas consumption to the gas company. The monthly electricity cost that is the cost function of the optima l operation scheduling of a CGS is proposed as follows:     ),0(++ + +),( 30 24 1 CDDMaxRCDR HPPSD TPHPOCPSf mPCD GiGiii i iGiGii      (1) Where i: time interval PSi: pseudo switch used to indicate the on/off (1/0) status of CGS during time interval i OC(PGi, HGi): operation cost of the CGS during time interval i (NT$/h) PGi: output power of the CGS during time interval i (kWh) HGi: equivalent power of recovered heat of the CGS during time interval i (kWh) TPi: TOU rate structure during time interval i (NT$/kWh) Di: demand during time interval i (kWh) RCD: rate when demand is under the contract capacity RP : rate if demand exceeds contract capacity CD: contract capacity Dm: monthly maximu m demand On the basis of the formulat ion of the cost function, the optima l operation scheme of the sample CGS and the optima l use of power fro m the utility can therefore be determined, ultimately, so as to minimize the monthly electric ity cost. Control variables of this optimization proble m include binary and continuous variables that correspond to the on/off status and the load level of CGS, respectively. Thus, this problem is formulated as a mixed-integer programming problem. 2.2. Optimal Operation Scheme Using a two-stage methodology and a modified mixed-integer progra mming technique, 0 50 100 150 200 250 300 20 40 60 80 100 120 140 c o st ( N T /h ) power output (kW/h) Sta n… Advances in Technology Innovation , vol. 1, no. 1, 2016, pp. 07 - 12 9 Copyright © TAETI a novel algorithm is developed and introduced here to solve the nonlinear optimization proble m. In the first stage, the reduction of the capacity charge for power due to the decrease of contract capacity caused by installation of the CGS is not take into consideration. Therefore, at this stage a linear progra mming algorithm can be applied. However, in the second stage the neglected factor is taken into account and the decomposition and alternative policy method is adopted [4]. The total operating cost of CGS during each time interval can be expressed as :   )P(DTPPOCf GiiiGii  1,2...24i  (2) where PGi is the sum of the electric power and heat output of the CGS. The heat should be converted to its equivalent electric power before the summat ion can be made. The fuel cost curve shown in Fig. 2 can therefore be formulated as : OC(P ) K P K P K Gi 1 Gi 2 Gi 3      2 1,2...24i  (3) The coefficients K1, K2, and K3 depend on the type, capacity, and maker of the CGS. Substituting (3) into (2), and then minimizing the obtained equation by linear programming techniques, the optima l operating capacity of the CGS for each time interval can be obtained as follows: P (K TP ) / (2K ) GiOM 2 i 1   i 1,2...24 (4) Fro m (4), it is clear that fuel cost (K1,K2) and TOU rate structure (TPi) are two important factors in determining the optimal operation capacity. The constraint of the generator output of the CGS is P P P Gmin GiOM Gmax   i 1,2...24 (5) Where P Gmin ： minimum generator output of the CGS P Gmax ： maximum generator output of the CGS The minimu m and ma ximu m outputs of the sample CGS are 30 kW/h and 150 kW/h, respectively. The time interval is set to one hour in this paper. The optima l operation benefit during each time interval can therefore be expressed as : Be TP P OC(P ) iOM i GiOM GiOM    i 1,2...24 (6) Equation (6) shows that the optima l operating benefit is gained from the difference between the savings gained by reducing the electric ity charge and the cost of operating the CGS. The following steps describe the optima l operation scheme in which the benefit gained by the reduction of contracted electricity capacity is considered: Step 1: Eva luate the optima l operating capacity using (4). In this step, the reduction of the electric ity capacity charge due to the decrease of contract capacity made possible by installation of the CGS is neglected. Step 2: Sort the hourly loads so that they are in descending order, that is D D j j+1  j 1,2,...23 (7) and record the corresponding hour of each hourly load. The demand diffe rences between ma ximu m hourly load and every hourly load can be calculated as : ED D D j 1 j 1    j 1,2,...23 (8) in which D1 is the maximu m hourly load. Therefore, if the reduction of the electricity demand charge due to the decrease of contract capacity made possible by installation o f the CGS is neglected, the optimal operating capacity of the CGS during each pseudo time period (j) is : P Max (P , ED ) Gj GjOM j  j 1,2,...23 (9) Step 3: The optima l operating benefits for various operating hours are Advances in Technology Innovation , vol. 1, no. 1, 2016, pp. 07 - 12 10 Copyright © TAETI Be TP P OC( P ) + R ED / 30 k j Gj Gj j 1 k CD k       k 1,2,...24 (10) Step 4: Select the ma ximu m operating benefit from step 3. Be Max (Be ) max k  k 1,2,...24 (11) Therefore, the optimal operation scheme of the CGS for each pseudo time is : 1,...24+kk,=j 1,2,...k=j 0 P P Gj Gj     (12) Step 5: Reorder (12) by the actual t ime order using the record built at step 2. The daily maximu m demand can be decreased as : )PMax(DD Giimd  i 1,2,...24 (13) Step 6: Print out the optima l operation schedule, operation benefit, daily ma ximu m de mand, and then stop the process. 3. Results and Discussion On the basis of the proposed algorithm, a program has been developed. Three samp le customers: a hotel, a hospital and an office building are used to demonstrate the proposed algorith m. The sample CGS is assigned to operate six months per year, fro m June through September in a ll cases. TOU rates of Taipowe r are applied for a ll samp le customers. Furthermore, the CGSs of the hotel and hospital are assumed to operate 30 days a month and 24 days for that of the office building. 3.1 The demand pattern of sample customers Fig. 3 shows the daily demand patterns for air conditioning, electricity, and total demand of the three sample customers in the summer. Relatively larger hourly demands for electricity occur in the sample hotel in the period between 11:30 and 23:00, and the total demand holds almost constant between 11:00 and 21:00. On the other hand, the total demand of the sample hospital is almost constant between 9:00 and 21:00. There are lunch hours between 12:00 and 14:00. The d emands of the sample office building are more concentrated, and result in a lower load factor. (a) Hotal (b) Hospital (c) Office building Fig. 3 Daily Demand Patterns on the Summer 3.2 Optimal Operation Capacity of the Sample CGS Fig. 4 shows the optima l operation capacity and operation benefit for various fuel costs. For e xa mple , Fig. 4-(a ) indicates that if the fue l cost is 6.272 NT$/ m 3 the optima l operation capacity is 53 kW/h and operation benefit is negative (-72). There fore, it is not worthwhile to operate the CGS in off-peak periods. Fig. 4-(b) shows that the operation benefit is positive for peak periods. Hence it is worthwhile to operate the sample CGS when the fuel cost is lower than 8 NT$/m 3 . For example, if fuel cost is 6.272 NT$/m 3 , the optimal operation capacity is 150 kW/h, that is, the ma ximu m capacity of the sample CGS. In this case the operation benefit is 62 NT$/h. Advances in Technology Innovation , vol. 1, no. 1, 2016, pp. 07 - 12 11 Copyright © TAETI (a) Off-peak Period (T OU Rate Structure is 0.77 NT$/kWh) (b) Peak Period (TOU rate structure is 1.89 NT$/kWh) Fig. 4 Optimal Operation Capacity and Operation Benefit for TOU Rate Structure of Taiwan 3.3 Optimal Operation Scheme of Sample Customers Figs. 5 sho w th e opt ima l op erat ion sche me fo r fue l costs o f 9.408 NT $/ m 3 . Fo r each hou rly lo ad, op t ima l gen erat ing c apac ity of CGS, and th e a mount o f po we r be ing purchased are shown. Fig . 5-(a) sho ws th at the ma xi mu m d a ily op e ra t ion be ne fit o f th e sa mp le hot e l is p ro duc ed by ope rat in g th e CGS e le ven hou rs (fro m 12:00 t o 22:00) pe r d ay . Th e g ene rat in g c apa c it ies at t ime inte rva ls 14, 15, 17, an d 18 a re lo we r th an o pt ima l cap ac ity . He nc e , the ge ne rat ing cap ac it ies o f the CGS a re set a t the o pt ima l ca pa c ity , th at is , 117.99 kW / h . 3.4 Economic Analysis The net present value (NPV) method is used to evaluate the annual benefits and years of payback [5]. Fig. 6 shows the annual benefits of using the CGS for the samp le customers at various fuel costs. In general, the more the load coincides in t ime is, the greater the benefit obtained. The numerica l results of the payback analysis for the sample customers are sho wn in Fig. 7. (a) Hotel (b) Hospital (c) Office Building Fig. 5 Optimal Operation Scheme at Fuel Cost 9.408 NT$/m 3 Fig. 6 Annual Benefit Advances in Technology Innovation , vol. 1, no. 1, 2016, pp. 07 - 12 12 Copyright © TAETI Fig. 7 Return of Investment for Various Fuel Cost 4. Conclusions A novel optima l operation scheme for sma ll cogeneration systems upgraded fro m standby generators has been introduced in this paper. On the basis of the proposed algorithm, a p rogra m has been developed. Then three sample customers are used to demonstrate the proposed algorith m. The benefits and cos ts of these sample cases are exa mined. The fuel cost, time-of-use rate structure and system constraints are taken into account in the evaluation. The results indicate that upgrading standby generators to cogeneration systems is profitable and should be encouraged, especially for those utilit ies with insuffic ient spinning reserves, and moreover, for utilit ies having difficulty constructing new power plants. Acknowledgement The support of the Ministry of Science and Technology (Taiwan), under Grant Most 104-2221-E-270-001 is gratefully acknowledged. References [1] T. Gonen, Electric Powe r Distribution System Engineering, Ne w Yo rk: McGraw-Hill, 2008. [2] R. C. Dugan, M. F. Mc Granaghan, S. Santoso, and H. W. Beaty, Electrical Po wer Systems Quality, Ne w York: Mc Graw-Hill, 2004. [3] Public Utility Regulatory Po lic ies Act, Report No. 95-1750, Confe rence Report, Oct. 10, 2001. [4] S. Ashok and R. Banerjee, “ Optima l operation of industrial cogeneration for load manage ment,” IEEE Trans. Powe r Syst., vol. 18, no. 2, pp. 931-937, May 2003. [5] K. Methaprayoon, W. J. Lee, S. Rasmiddatta, J. R. Liao, and R. J. Rose, “Multistage artific ial neural network short-term load fo recasting engine with front-end weather forecast,” IEEE T rans. Ind. Appl., vol. 43, no. 6, pp. 1410– 1416, Nov./Dec. 2007. [6] C. T. Hsu, C. S. Chen, and C. H. Lin, “Electric powe r system analysis and design of an expanding steel cogeneration plant,” IEEE Trans. Ind. Appl. , vol. 47, no. 4, pp. 1527-1535, Jul./Aug. 2011. [7] C. H. Hsu, “A graph representation for the structural synthesis of geared kine matic chains ,” Journal of the Franklin Institute, vol. 330, pp. 131-143, Jan. 1992.