- Email: [email protected]
Impact on power planning due to demand-side management (DSM) in commercial and government sectors with rebound effect—A case study of central grid of Oman
ARTICLE IN PRESS
Energy 32 (2007) 2157–2166 www.elsevier.com/locate/energy
Impact on power planning due to demand-side management (DSM) in commercial and government sectors with rebound effect—A case study of central grid of Oman Arif S. Malik College of Engineering, Sultan Qaboos University, P.O. Box 33, Al-Khod 123, Oman Received 21 June 2006
Abstract This paper presents the results of a study that estimated the DSM energy saving and load management potential in commercial and government/institutional sectors in Oman (central grid area) and evaluated its impact on generation capacity and energy savings. The end-use (lighting and air-conditioning) energy consumption data have been collected in two major segments of the commercial sector for simplicity and to save time and money. Another unique aspect of the study is the inclusion of the energy savings, in transmission and distribution (T&D) losses that are estimated by using generation expansion planning approach. The study has found that DSM is financially beneficial from customers’ point of view as the discounted payback period of investment in efficient lighting and airconditioning is between 4 and 12 years of the surveyed sample. From the utility point of view the capacity saving at the horizon year is between 372 and 596 MW and the overall energy saving for the whole planning horizon is about 29–44 TWh. The total avoided cost in generation and capacity saving is somewhere between 416 and 597 million dollars. r 2007 Elsevier Ltd. All rights reserved. Keywords: Demand-side management; Generation expansion planning; Power planning
1. Introduction The utility planning process which takes into account both the supply—as well as demand-side options to meet the demand forecast is known as integrated resource planning (IRP) [1]. IRP studies are based on energy production costs simulations and reliability analyses. In many IRP processes, the resource contribution from demand-side management (DSM) programs could be very large because of untapped resources that exist. For screening the large number of DSM options available and identifying the most promising options avoided costs method is used to estimate their value to the utility system. Selecting among demand-side resources is difficult because they are by nature diverse, diffuse and decentralized [2]. Moreover, each option is much smaller than a typical supply option. To reduce the large number of DSM options in the resource planning process the individual DSM Fax: +968 24413416.
E-mail address: [email protected] 0360-5442/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.energy.2007.05.004
options are bundled together into larger resource blocks. The economy of aggregated bundles of DSM options is tested by the avoided cost method. The avoided cost has two components of savings the energy costs savings and capacity cost savings. The planning horizon is optimized twice; one with a base case set of loads and another with load shape changed by the expected load impact of DSM. Avoided cost is then calculated by the difference in energy and capacity costs of the two optimized cases [3]. In developing countries, there have been several studies conducted on DSM [4–9]. In [4] electricity load analysis of the commercial sector of Northern Cyprus is simplified by studying only one segment, i.e., hotel buildings which constitute the largest energy-consuming segment in the commercial sector. The study is based on end-use data collection in hotel buildings and subsequent statistical analysis. The study concludes that it is possible to reduce the demand of end-uses by at least 58% by the application of DSM programs, with peak reduction corresponds to 11% of the utility summer peak of
ARTICLE IN PRESS 2158
A.S. Malik / Energy 32 (2007) 2157–2166
Northern Cyprus. The study has not done any effort to find the benefit of avoided cost of utility in doing DSM. In [5] the study has analyzed the most promising DSM technologies, namely power factor correction at end-user, energy efficient lighting in residential and commercial sector, and intelligent motor controllers in Nepal. On-site interviews were undertaken with power utilities (the Nepal Electricity Authority and its subcompanies) and various electricity end-users in Nepal to collect data, as well as walk-through auditing at the end-users’ premises. The study concludes that most of the DSM technologies are financially viable. The utility benefit is estimated in terms of delays and avoidance in the investment, from the envisaged expansion plan, such as (i) cut power imports from India, (ii) reduce power production costs and hence electricity tariffs in Nepal; (iii) avoid the investment in Budhi Ganga (106 MW); and (iv) postpone the investment of Hewa Khola (67 MW). In Sri Lanka [6] the government has created an Energy Conservation Fund and a DSM unit within the Ceylon Electricity Board supported by the World Bank, Global Environmental Facility and the Asian Development Bank. The DSM unit has implemented a broad range of programs and measures, both incentive and information-based, including efficient lighting, energy audits, power factor correction, establishing a ‘best practices’ group with industry, building code and appliance labeling. A study on Greenhouse Gas Mitigation Options in the Sri Lankan Power Sector [7] has taken into account the impact on GHG mitigation using DSM and other options such as alternative energy, reduction in T&D losses, rehabilitation of plants, etc. The study has used the generation expansion planning approach to quantify the environmental benefits. A study on DSM in Thailand [8] shows that variety of DSM programs targeting a wide range of sub-sectors and end-uses have produced savings of 566 MW in peak load and 340 GWh/year as of June 2000. The savings in capacity and energy saving are estimated from the actual reduction in load forecast. The first DSM study in Oman (1998), ‘The Study on Demand Supply Management for Power Sector in Sultanate of Oman’, was conducted by Japan International Cooperation Agency (JICA). The study identified several strategies for potential load management including (i) gas cooling systems for government buildings, hospital, hotels commercial complexes and large houses; and (ii) shifting load from peak time to off-peak time in the industrial and commercial sectors by application of ice thermal storage system and introducing time-of-use tariffs [9]. The utility benefit was estimated using avoided marginal generation capacity and energy cost concept. The study has been shelved since then and there have been no implementation on the recommendations. This paper presents the results of a recent study, undertaken at Sultan Qaboos University, that estimated
the DSM energy saving and load management potential in commercial and government/institutional sectors in Oman (central grid area) and evaluated its impact on generation capacity and energy savings [10]. The commercial sector represents about 16% and government sector represents about 18.4% of the total consumption in 2003 [11]. The commercial and government sectors in Oman includes service businesses, retail and wholesale stores, hotels and restaurants, government ministries, schools and universities, and hospitals and health clinics. The diversity of the end-uses and the technologies available in these buildings requires a large stratified sample. To do a survey with such a large sample size would be difficult because of the poor availability of statistical data and obtaining financial support for such a project. In this paper a different approach is used. Instead of trying to represent the whole commercial and government sector the energy saving potential is worked out in only two end-use functions, i.e., lighting and air conditioning, for large customers of commercial and government/institutional sectors only. These two end-use functions constitute major energy consumption (79–96% of the surveyed samples) in these sectors regardless of the activity in both physical and economic terms. Another unique aspect of the study is the estimation of the energy savings, in transmission and distribution (T&D) losses that are estimated by using generation expansion planning approach. The widely available software packages for generation expansion planning such as WASP, EGEAS, etc. do not include representation of T&D and hence the avoided cost savings, due to DSM, in T&D capacity cost and T&D losses are generally neglected. The method to include the T&D losses is described in Section 3 of this paper. To carry out the study, the load forecast and generation data was taken from Ministry of Housing, Electricity and Water (MHEW) of Oman and the software used to optimize the generation plan is Wien Automatic System Planning (WASP), developed for IAEA, and used by many electric utilities in developing world [12]. The paper is arranged in eight sections. Section 1 is an introduction. Section 2 is a theoretical review of power planning concepts. Section 3 in brief provides a general methodology adopted to carry out the study. Section 4 provides an overview of electricity tariff structure in Oman. Section 5 provides the DSM results of surveyed commercial and government/institutional sectors. Section 6 provides generation and load data of interconnected central grid system. Section 7 presents the results of generation capacity and energy savings. And the last section concludes the paper. 2. Generation expansion planning The objective of electric planning study is to meet the load forecast with high reliability at a minimum cost. There are three keywords in the above statement, i.e., load forecast, reliability, and cost. A brief discussion on these three keywords is followed.
ARTICLE IN PRESS A.S. Malik / Energy 32 (2007) 2157–2166
In generation planning the total cost to be minimized include investment cost in new generation facilities, fuel cost, operation and maintenance cost, and energy not served cost. The cost is minimized depending on the financial resources, technical, environmental and political considerations. The following four questions must be answered when the study of capacity planning is done [13]: (1) (2) (3) (4)
What type of capacities will be added to the system? How much capacities will be added? When these capacities will be added? Where these capacities will be located?
The first three questions can be answered by using any generation expansion planning software. However, to know where to locate the new facilities a detailed feasibility study has to be carried out. The objective function of the least cost-planning software is normally the following: minimize Bj for all j, n X CC i SV i þ FC i þ OM i þ UE i , Bj ¼
ð1Þ
i¼1
where CC is the capital cost, SV is the salvage value, FC is the fuel cost, OM is the operation and maintenance cost, UE is the energy not served cost, n is number of years in study period, and over bar (—) on the above costs represents present worth of all the costs. 2.2. Generation planning study period The study period normally spans from 20 to 25 years. The study period consists of three sub periods: pre-planning period, planning period and post-planning period [14]. Preplanning period is the first 3–4 years in which the planning was done earlier. This is included in the study period to check the energy production cost and reliability of the system. The planning period is between 4 and 10 years in which the decision has to be taken now about the plants that have to be added in future to meet the forecast. The post-planning period is added in the study period to avoid the end-effects by making the calculations continue for another 10 years or more so that a proper trade-off between construction costs and operating costs are found. Therefore, the long-term forecasts of 20–25 years of electricity consumption and demand are used in the planning of investment in generating capacity and the development of fuel supplies.
which simply states that the installed capacity Pt must lie between the given maximum and minimum reserve margins, at and bt, respectively, above the peak demand Dtp in year t. Another popular reliability criterion is the loss-of-load probability (LOLP) index which may be calculated for each period and a limiting value may be suggested by the user. However, it may be noted in Eq. (1) that the reliability is also embedded in terms of energy not served cost. This cost is the penalty for unreliability. If more capacity is added the capital cost would be high but because of more capacity in the system the total cost of energy not served would be low as the system would be more reliable. On the other hand, if the capacity added is less than the actual required then there would be more customer outages and hence higher cost of unserved energy. The basic concept of reliability-cost/reliabilityworth evaluation is relatively simple and can be presented by the cost/reliability curves, as shown in Fig. 1 [15]. The utility curve shows that the investment cost increases with higher reliability, i.e. more investment is required in order to improve reliability. On the other hand, the customer outages cost decreases as the reliability increases. The total cost of the two curves is the sum of the utility cost and customer cost, which exhibits the optimal cost for the society. 3. General methodology DSM economic energy saving potential is worked out from both the customer point of view as well as the electric utility point of view. The estimate of DSM potential is based on a survey of six large commercial (shopping malls) customers and six large government/institutional customers. The annual energy consumption of these 12 customers represents about 4.6% (127.6 GWh) of the total consumption (2775 GWh, year 2003) in these two sectors. The energy saving potential is worked out in only two enduse functions, i.e., lighting and air conditioning, because these two end-use functions constitute the major energy
Annual Costs ($)
2.1. Cost minimization
2159
Utility Cost Total Cost
Customers Cost
2.3. Reliability In expanding the system the reliability constraints can be exogenously added into the cost minimization process. The following reliability constraints may be added to (1): ð1 þ at ÞDtp XPt Xð1 þ bt ÞDtp ,
(2)
System Reliability Fig. 1. Relation between consumers, utility and total annual cost with respect to reliability.
ARTICLE IN PRESS A.S. Malik / Energy 32 (2007) 2157–2166
2160
consumption (79–96% of the surveyed samples) in these sectors regardless of the size of the activity in both physical and economic terms. To work out the benefit of DSM, in these two sectors, from customer point of view the following procedure is adopted: 1. A survey questionnaire was designed to get the data needed from the commercial and government/institutional sectors. The first section of the survey requests some general information about the building such as floor area, electric end-use penetration rate, and annual power consumption. The second section surveyed on lighting end-use share, percentage efficiency and the lifetime of the lighting system and devices. The last section of the survey requires information about the airconditioning system in the sector, its end-use share,
Sectoral energy Consumption data
percentage efficiency and lifetime of the air-conditioning equipment. 2. From the catalogue information available for the efficient cooling and lighting devices [16,17], the cost of efficient cooling and lighting functions were then worked out as 9.4 and 2.1 OR/m2, respectively (1 Omani Rial (OR) ¼ 1000 Baizas (Bz) ¼ 2.58 US$). 3. The estimate of total energy saved in lighting and airconditioning function for each customer was worked out using the percentage of efficiency save in lighting and air conditioning, had the existing lighting and air-conditioning systems were replaced with more efficient equipment, and multiplying the annual consumption in lighting and air-conditioning, respectively. 4. The financial benefit of energy saved component from the customer point of view is calculated by
Forecast sectoral energy consumption till 2024 (end of study period)
Forecast Peak demand in commercial and government sectorsusing (3)
Peak coincident factors data
Contribution of peak demand of both sectors toward the system peak
Load factor data
Average system losses data
Calculate peak losses using (4&5)
Forecast combined commercial and government peak demand to be supplied from the generation end using (6)
Estimated percentage reduction from the sample data
Forecast combined reduction inpeak in both sectors considering the reduction upper and lower limits with 10% and 40% rebound effects respectively
Modify system peak demand forecast
Calculate the utility benefit interms of avoided cost of generation capacity and energy by optimizing the planning horizon thrice; one with a base case set of loadsand the other two cases with loadshape changed because of DSM that is having 40% rebound effect, and 10% rebound effect Fig. 2. Flowchart of avoided cost calculation.
ARTICLE IN PRESS A.S. Malik / Energy 32 (2007) 2157–2166
multiplying the energy saved and the price of the commercial tariff. 5. The present worth of the net benefit and the discounted payback time for each customer were then worked out using discount rate of 10% and the estimated lifetime of lighting (replacement of lighting devices every 6 years) and air-conditioning systems of 20 years. The avoided cost benefit of DSM in these two sectors, from the utility point of view, is shown in a flowchart in Fig. 2 and the details are delineated as follows: 1. An aggregate energy saving potential is worked out with the assumption that the reduction in power consumption due to DSM in the sample data is representative of the actual potential in both the sectors. However, standard error analysis (95% confidence interval) is also done from the sample data by normalizing the energy saved in the 12 customers’ sample. Sector-wise analysis of variance has shown that the standard error of average reduction in power consumption was not significantly different (8.8% and 8.3% for commercial and government sectors, respectively) and the pooled standard error is 8.6%. Moreover, the rebound effect is also studied. Rebound effect suggests that the technical saving potential may not be a correct indicator as consumers may use more energy in other areas. Rebound effect varies across countries even for a particular DSM program. A study done in commercial/household sectors in Norway found that the rebound effect is 10–40% [18]. Using these two limits the range of DSM benefit from electric utility point of view is estimated. 2. The sectoral energy consumption from 1999 to 2003 taken from Ref. [11] is extrapolated and forecasted till 2024 (end of study period). 3. The energy forecast in commercial and government sectors is then converted to peak demand forecast by using the following expression: peak demand in both sectors ðMWÞ ¼
annual energy consumption in both sectors ðMW hÞ , LF 8760 h
ð3Þ where LF is the system load factor. Here, an assumption is made that the load factor of both government and commercial sectors are same as the average system load factor of 53%. The average system load factor was calculated using the past data of average demand divided by the actual system peak demand. 4. To get the contribution of peak demand of both sectors toward the system peak another assumption is made that the peak coincident factor of government sector is 100% of the system peak and commercial sector is 80% of the system peak. This assumption is based on the fact that the system peak occurs between 2:00 and
2161
4:00 p.m. in July (actual data) when all the government offices are working (100% coincidence, assuming it is basically air-conditioning load) and major commercial stores and buildings are open while some of the commercial retailers are closed (about 80% coincidence assumed). 5. The peak losses are calculated from the average system losses from the following expressions [19]: peak losses ¼
average system losses . peak loss factor
(4)
The average of system losses of 8 years from 1995 to 2002 is 18.25% [11]. The peak loss factor in Eq. (4) is the ratio of the energy loss in the system during a given time period to the energy loss that would result if the system peak loss had persisted throughout that period. The following empirical approximation is used to calculate the peak loss factor [19]: peak loss factor ¼ 0:3LF þ 0:7LF 2 .
(5)
6. The combined commercial and government peak demand to be supplied from the generation end is then calculated by the following formula: combined peak to be supplied ¼
peak ðstep 4 after taking into account the coincident factorsÞ . 100% peak losses
ð6Þ 7. The combined reduction in peak in both sectors is then worked out with the estimated percentage reduction from the sample data and considering the reduction upper and lower limits with 10% and 40% rebound effects, respectively. The system peak forecast is then modified accordingly. 8. The utility benefit in terms of avoided cost of generation capacity and energy is then worked out by optimizing the planning horizon thrice; one with a base case set of loads and the other two cases with load shape changed because of DSM that is having 40% rebound effect and 10% rebound effect. 4. Electricity tariff structure The government of Oman provides a large subsidy to the electricity sector, where the domestic sector takes the most of this subsidy because it constitutes the main service receiver among the power consumers. However, the average returns of the electricity sector from the domestic slab constitute 50% of the total subsidy of the government. In addition, the government of Oman provides generous subsidy to other sectors. There are several electricity tariff structures in different sectors. These sectors are domestic, commercial, industrial, agriculture and fisheries, and, hotels and tourist establishments. The domestic tariff increases with the quantity of consumption and starts
ARTICLE IN PRESS A.S. Malik / Energy 32 (2007) 2157–2166
2162
from 10 Bz/kWh for the first 3000 kWh of consumption and end at 30 Bz/kWh for consumption more than 10,000 kWh. The commercial tariff is 20 Bz/kWh without any maximum unit for consumption. The industrial tariff in summer is 24 and 12 Bz/kWh in winter season. The summer is taken from May to August for central grid system. The other tariff rates can be found in annual report of MHEW [11]. 5. Results of surveyed commercial and government/ institutional sectors Tables 1 and 2, respectively, show the results of DSM energy saving potential in commercial, and government/ institutional sectors. These results are from the customers’ point of view. As can be seen from the tables that with the exception of one customer the discounted payback on investment in efficient air-conditioning is less than 10 years and most of them have less than 6 years. The discounted payback on investment in efficient lighting system is about 10 years for most of the surveyed sample.
Before we provide the result of DSM economics from the utility point of view, the next section discusses the generation and load data necessary to run WASP for evaluating DSM economics. 6. Generation and load forecast data The generation expansion planning is carried out for the central grid system from 2005 to 2024. The central grid power system area comprises mainly six interconnected power stations, which are Ghubrah, Rusail, Barka, Wadi Jizzi, Alkamil and Manah stations. The interconnection between these stations is through 132 kV transmission lines. However, recently Barka power station is connected with the central grid through 220 kV lines. All the power stations are operating on natural gas as fuel. The total capacity of these stations is about 2517 MW (year 2005). The power transmission and distribution system mainly consists of 132, 33, and 11 kV and 415 V. Until recently, 220 kV line from Barka station is constructed to connect the station to the main 132 kV grid. Only few kilometers of
Table 1 Final results of the surveyed commercial sectors Hotels/ shopping malls
Al Falaj Hotel Al-Araimi Complex Al-Harthi Complex Holiday Inn Sheraton Hotel OCC
Old consumption (kWh/year)
New consumption (kWh/year)
Air-conditioned
Lights
NPV (OR)
Discounted payback (years)
NPV (OR)
Discounted payback (years)
8 288 750 3 248 000
5 923 141 2 192 725
105 936 44 822
9.4 10.9
105 936 6605
16.9 12.8
6 270 216
3 510 116
119 497
9.3
22 014
10.2
6 896 498 17 963 229
4 288 242.5 11 270 130
218 165 559 277
5.0 4.6
9861 84 382
12.0 6.1
6 182 317
2 866 127
266 211
5.3
18 324
10.3
Table 2 Final results of the surveyed Government/institutional sectors Government/ institutional buildings
Ministry of Civil Services Ministry of Health Nizwa Hospital Royal Hospital Sohar Hospital SQU (College of Arts)
Old consumption (MWh)
New consumption (MWh)
Air-conditioned
Lights
NPV (OR)
Discounted payback (years)
NPV (OR)
Discounted payback (years)
4 561 800
3 228 842
119 871
7.4
7596
10.1
4 257 229
2 613 939
203 677
5.0
19 048
8.5
18 787 806
15 242 735
350 629
6.3
65 320
7.1
28 900 640
22 197 136
569 187
6.6
91 516
8.0
19 943 821 2 313 384
16 459 635 1 545 109
331 949 60 390
6.6 8.1
76 017 8183
6.4 10.5
ARTICLE IN PRESS A.S. Malik / Energy 32 (2007) 2157–2166
66 kV line also exists. The overall system losses reported in 2003 were 20.83%. The generation data collected from MHEW was manipulated, refined, and made suitable for WASP-IV input. For some of the missing data, typical values found elsewhere were used. The generation data for fixed plants, committed plants and candidate plants are briefly presented here followed by load data and economic data.
2163
6.3. Candidate plants The Candidate power plants are plants with which we do our expansion. The candidates are selected based on their economic and technical suitability for the system. Their capital costs are included in the objective function. The data for the candidate units are shown in Table 5. 6.4. Load data
6.1. Fixed system The fixed system is the existing generation system consisting of six plants Ghubrah, Rusail, Wadi Al-Jizzi, Manah, Barka and Al-Kamil. These plants contain many different types and sizes of units. Some of these units are common to most of these plants. Since the WASP model does not include the transmission representation the fictitious power plants were made by grouping same type and size of units. The details of the net capacity, minimum loading level, fixed O&M, variable O&M, forced outage rate, net heat rate at minimum load and average incremental heat rate for each units of the fixed fictitious plants, is given in Table 3. Note the last plant (SOHR) in the list is showing zero numbers of sets. This is because it is a committed plant (year 2005 study) and its two sets are supposed to be added in 2006 and one in 2007.
The surveyed government and commercial buildings annual consumptions are 78.8 and 48.9 GW h, which respectively represents about 5.3% and 3.8% of the total consumption in these sectors in Oman (2003) [11]. An aggregate energy saving potential is worked out with the assumption that the reduction in power consumption due to DSM in the sample data is representative of the actual potential in both the sectors. Table 6 shows the percentage of energy consumption reduction by implementing DSM in commercial and governmental sectors. Also shown are the standard error and the percentage reduction in peak because of 10% and 40% rebound effect. Table 7 shows the actual energy consumption in different sectors from 1999 to 2003. Fig. 3 shows the energy consumption forecast of government and commercial sectors up to 2024 using linear regression on the 5-year data shown in Table 7. The DSM energy forecast is made,
6.2. Committed and retiring plants The committed power plants are the plants which already have been decided to be added to the system in the near future. They are also part of the fixed system. These plants are in different phases of construction and their capital costs are considered sunk. In this study, the plants that are going to be added to the system in the first 3 years of study are considered as committed plants. Similarly the plants that are going to be retired from the systems during the study horizon are also taken into account. Table 4 summarizes the additions and retirements of plants.
Table 4 Summary of additions and retirements from 2005 to 2024 Name Years 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 FR5P 5 FR6B ST13 3 ST56 FR9B SOHR 2
1
1 2
2
5 6
1
2 2
1
1
Table 3 Summary description of thermal plants in year 2005 No.
1 2 3 4 5 6 7 8 9
Plant name
FR5P FR6B FR9E ST13 STM4 ST56 FR9B BRKA SOHR
No. of sets
Minimum load (MW)
Capacity (MW)
Base load
12 15 9 3 1 2 6 1 0
13 20 67 8 35 21 60 363 121
18 27 93 8 50 30 83 427 165
3900 3529 3284 3294 2985 3529 3467 1980 2324
Fast spin (res%)
FOR (%)
Incr. load
Fuel cost (cents/106) (kcal)
O&M (FIX) ($/kWmonth)
O&M (VAR) ($/kWh)
2637 2219 2100 3294 2652 2219 2102 1520 1672
608 608 608 608 608 608 608 608 608
0 0 10 0 10 5 10 5 10
5 5 5 5 5 5 5 5 5
1.37 0.91 0.31 1.5 1 0.91 0.31 0.2 0.25
3.1 2.8 2.6 3.3 2.7 2.8 2.6 1.0 1.2
Heat rates (kcal/kWh)
ARTICLE IN PRESS A.S. Malik / Energy 32 (2007) 2157–2166
2164 Table 5 Summary description of candidate plants No.
1 2 3
Name
Min. load (MW)
FRM9 FRM6 BRKA
67 20 363
Cap acity (MW)
93 27 427
Heat rates (kcal/kWh) Base load
Incr. Load
Fuel cost (cents/106) (kcal)
3284 3529 1980
2100 2219 1520
608 608 608
Fast spin (res%)
FOR (%)
O&M (FIX) ($/kWmonth)
O&M (VAR) ($/kWh)
Capital cost ($)
Construc tion time
Plant life (years)
10 0 10
5 5 5
0.31 0.91 0.2
2.6 2.8 1
531 728 780
3.0 2.0 3.5
20 20 20
Table 6 Implementing DSM in Governmental and commercial sectors Actual Modified energy Energy energy consumption saved consumption (GWh) (GWh) Governmental 78.76 sector Commercial 48.85 sector
Estimated percentage reduction (0% rebound effect) (%)
Standard error in Percentage reduction the energy saved with 40% rebound (%) effect (%)
Percentage reduction with 10% rebound effect (%)
62.12
16.64
21.1
8.8
12.7
19.0
30.05
18.8
38.5
8.3
23.1
34.6
Table 7 Energy consumption in different sectors from 1999 to 2003 Year
Governmental (GWh)
Residential (GWh)
Commercial (GWh)
Industrial (GWh)
Other (GWh)
Total (GWh)
1999 2000 2001 2002 2003
1346 1354 1423 1511 1483
3785 3966 4226 4414 4632
971 1037 1156 1221 1292
271 330 356 438 490
129 146 165 159 161
6502 6833 7326 7743 8058
3000
700 Govt
Comm
Govt-DSM
Comm-DSM
2500 2000 1500 1000 500 0 1999
2004
2009
2014
2019
2024
Years
Fig. 3. Forecast of energy consumption in government and commercial sector with and without DSM.
also shown in Fig. 3, using estimated percentage reductions of Table 6. The energy forecast of government and commercial sectors is then converted to peak demand forecast using (3). Fig. 4 shows the peak demand forecast of both the sectors with and without DSM. The system peak reduction from generation end is worked out as explained in the steps of
Peak Demand (MW)
Energy (GWh)
3500
Govt
Govt-DSM
Comm
Comm-DSM
600 500 400 300 200 100 0 2005
2010
2015
2020
Years
Fig. 4. Forecast of peak demand in government and commercial sectors with and without DSM.
Section 3. Fig. 5 shows the system peak demand forecast, taken from MHEW, and the modified system peak demand with 40% rebound effect and 10% rebound effect. The annual chronological load data of year 2003 taken from MHEW was used for making normalized load duration curve (LDC) shapes. The year was divided into two seasons and two LDC shapes were formed [10].
ARTICLE IN PRESS A.S. Malik / Energy 32 (2007) 2157–2166
6.5. Economic data The discount rate used for the study is 10.0% and the cost of unserved energy as 1.6 $/kWh. The cost of unserved energy was worked out in [20] and is consistent with the optimal reserve margins found using the cost of unserved energy and the average reserve margins existing in Oman.
2165
over the planning horizon (avoided cost ¼ Bj,baseBj,DSM) with 40% and 10% rebound effect is about 416 and 597 million dollars, respectively. At the horizon year the capacity and energy difference from the base case is about 372 MW and 29.2 TWh for 40% rebound effect, and 596 MW and 43.8 TWh for 10% rebound effect. 8. Conclusions
7. Results and discussion
System Peak Demand (MW)
It has been shown in Section 5 that DSM implementation in commercial and government sectors is beneficial from the customer point of view. To investigate the DSM benefit from the utility point of view the avoided cost method is used. The avoided cost of generation capacity and energy is worked out by optimizing the planning horizon thrice; first with a base case set of loads (Case 1), second with load shape changed by the expected load impact of DSM with 40% rebound effect (Case 2) and the third with 10% rebound effect (Case 3). Table 8 provides the summary of the optimization results. The table shows the present worth of total cost (cumulative objective function of (1) in thousands of dollars), the average LOLP index, and the cumulative candidate plants selected over the planning horizon. The table also provides the cumulative capacity selected in each case and the energy requirement for all the three scenarios. The difference in cost, capacity and energy of Cases 1 and 2, and Cases 1 and 3 are also shown. The overall cumulative cost difference 6500 6000 5500 5000 4500 4000 3500 3000 2500 2000 2005
System Peak Load Forecast System Peak Load Forecast- DSM with 40% rebound effect System Peak Load Forecast- DSM with 10% rebound effect
2010
2015
2020
Years Fig. 5. System peak demand forecast with and without DSM.
This paper presented the results of a study that estimated the DSM energy saving and load management potential in commercial and government/institutional sectors in Oman (central grid area) and evaluated its impact on generation capacity and energy savings. The unique aspect of the study is the estimation of the energy savings, in transmission and distribution (T&D) losses that are estimated by using generation expansion planning approach. The study results have shown that it is beneficial to implement DSM in commercial and government/institutional sectors from customers’ point of view. According to the energy calculations of the surveyed sample, we found that applying DSM programs in the government and commercial sectors the energy consumption can be reduced up to 21% and 38%, respectively, with 8.6% standard pooled error. For the surveyed models the net present value of the investment is positive with discounted pay back period (between 4 and 12 years) which encourage sectors to apply DSM. From the supplier (MHEW) point of view the capacity saving at the horizon year ranges between 372 and 596 MW considering 40% and 10% DSM rebound effect, respectively. Likewise, the overall energy saving for the whole planning horizon ranges between 29 and 44 TWh and the total avoided cost saving from 416 to 597 million dollars. One of the problems which faced the countries that applied DSM is motivating costumers to pay for replacing their inefficient devices. Financing DSM remains a significant issue in most places. Many countries have imposed DSM taxes within power tariff schemes. Other governments have opted to fund DSM programs for lowincome customers as a social program, and effectively pay the utility to implement the program [8]. Finally, we can conclude with confidence that demandside management has positive outcomes for consumers,
Table 8 Results of optimal generation expansion plan from 2005 to 2024 Cases
Base (Case 1) DSM with 40% rebound DSM with 10% rebound Difference (Case 1Case Difference (Case 1Case
effect (Case 2) effect (Case 3) 2) 3)
Present worth cost (thousand $) objective function Bj (cumulative)
Average LOLP (%)
3 801 320 3 385 609 3 204 479 415 711 596 841
0.797 0.496 0.527 0.301 0.270
Cumulative candidate plants (no. of units) FRM9
FRM6
BRKA
16 12 5 4 11
0 0 0 0 0
6 6 7 0 (1)
Cumulative capacity (MW)
Cumulative energy (GWh)
6217 5845 5621 372 596
413 752 384 588 369 961 29 164 43 791
ARTICLE IN PRESS 2166
A.S. Malik / Energy 32 (2007) 2157–2166
MHEW, environment and the national income which makes it strongly recommended to start thinking about implementing it in Oman. Acknowledgments The authors would like to acknowledge Internal Research Grant for this project from Sultan Qaboos University and Power Planning Department of Ministry of Housing, Electricity and Water for providing necessary data for carrying out this study. Disclaimer: The views expressed in the paper are entirely ours and do not reflect the position of MHEW. All errors are ours. References [1] Kreith F. Integrated resource planning. J Energy Resour Technol 1993;115:80–5. [2] Busch J, Eto J. Estimation of avoided costs for electric utility demand-side planning. Energy Sour 1996;18:473–99. [3] Malik A. Modeling and economic analysis of DSM programs in generation planning. Int J Electric Power Energy Syst 2001;23:413–9. [4] Atikol U. A demand-side planning approach for the commercial sector of developing countries. Energy 2004;29:257–66. [5] Yang M. Demand-side management in Nepal. Energy 2006;31: 2677–98. [6] World Bank. Project Appraisal Document for Energy Services Delivery Project Report No. 16063-CE. Energy and Project Finance Division Country Department, 1 South Asia Region, Washington, DC, February 1997. [7] World Bank. Greenhouse gas mitigation options in the Sri Lanka power sector. Joint UNDP/World Bank Energy Sector Management Assistance Programme (ESMAP). Washington, DC, February 2003.
[8] World Bank. DSM in Thailand: a case study. A working paper written by Jas Singh and Carol Mulholland, October 2000. [9] Japan International Cooperation Agency (JICA). The study on demand supply management for power sector in Sultanate of Oman. A final report of JICA to Ministry of Housing, Electricity and Water (MHEW). MHEW Central Office: Al-Khuwair, Muscat, Sultanate of Oman, November 1998. [10] Malik A, Al-Khrousi A, Al-Hindi M, Al-Mahrouqi Y. Demandside management energy saving potential in commercial and governmental/institutional sectors and its impact on power planning—a case study of central grid of Oman. In: Proceedings of the second international green energy conference, Oshawa, 25–29, June 2006. [11] Annual statistical book of year 2003 of electrical sectors, Muscat, Oman. Al-Khuwair, Muscat, Sultanate of Oman: Ministry of Housing, Electricity, and Water, MHEW Central Office; 2003. [12] Wien automatic system planning package (WASP-IV). Vienna: IAEA; 2003. [13] Buehring W, Huber C, Marques de Souza J. Expansion planning for electrical generating systems—a guidebook. Technical report series no. 241. Vienna: IAEA; 1984. [14] Wang J, McDonald J. Modern power system planning. London: McGraw-Hill; 1994. [15] Billinton R, Allan R. Reliability evaluation of power systems. New York: Plenum Press; 1996. [16] Bahwan Engineering Company (BEC). Catalogue of cooling devices. BEC Head Office: Ruwi-112, Sultanate of Oman; 2004. [17] Khimji Ramdas Company (KRC). Catalogue of lighting devices. KRC Corporate Office: Muscat-113, Sultanate of Oman; 2003. [18] Roland K, Haugland T. Joint implementation: difficult to implement. In: Jepma CJ, editor. The feasibility of joint implementation. Dordrecht: Kluwer Academic Publishers; 1994. p. 359–66. [19] Pabla A. Electric power distribution. New York: McGraw-Hill; 2005. [20] Al-Farsi F, Al-Shihi M, Al-Shokaili Y. Optimal generation reserve margin for Muscat interconnected system. Final year project report. Muscat, Oman: SQU, Electrical Engineering Department; 2000.
Energy 32 (2007) 2157–2166 www.elsevier.com/locate/energy
Impact on power planning due to demand-side management (DSM) in commercial and government sectors with rebound effect—A case study of central grid of Oman Arif S. Malik College of Engineering, Sultan Qaboos University, P.O. Box 33, Al-Khod 123, Oman Received 21 June 2006
Abstract This paper presents the results of a study that estimated the DSM energy saving and load management potential in commercial and government/institutional sectors in Oman (central grid area) and evaluated its impact on generation capacity and energy savings. The end-use (lighting and air-conditioning) energy consumption data have been collected in two major segments of the commercial sector for simplicity and to save time and money. Another unique aspect of the study is the inclusion of the energy savings, in transmission and distribution (T&D) losses that are estimated by using generation expansion planning approach. The study has found that DSM is financially beneficial from customers’ point of view as the discounted payback period of investment in efficient lighting and airconditioning is between 4 and 12 years of the surveyed sample. From the utility point of view the capacity saving at the horizon year is between 372 and 596 MW and the overall energy saving for the whole planning horizon is about 29–44 TWh. The total avoided cost in generation and capacity saving is somewhere between 416 and 597 million dollars. r 2007 Elsevier Ltd. All rights reserved. Keywords: Demand-side management; Generation expansion planning; Power planning
1. Introduction The utility planning process which takes into account both the supply—as well as demand-side options to meet the demand forecast is known as integrated resource planning (IRP) [1]. IRP studies are based on energy production costs simulations and reliability analyses. In many IRP processes, the resource contribution from demand-side management (DSM) programs could be very large because of untapped resources that exist. For screening the large number of DSM options available and identifying the most promising options avoided costs method is used to estimate their value to the utility system. Selecting among demand-side resources is difficult because they are by nature diverse, diffuse and decentralized [2]. Moreover, each option is much smaller than a typical supply option. To reduce the large number of DSM options in the resource planning process the individual DSM Fax: +968 24413416.
E-mail address: [email protected] 0360-5442/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.energy.2007.05.004
options are bundled together into larger resource blocks. The economy of aggregated bundles of DSM options is tested by the avoided cost method. The avoided cost has two components of savings the energy costs savings and capacity cost savings. The planning horizon is optimized twice; one with a base case set of loads and another with load shape changed by the expected load impact of DSM. Avoided cost is then calculated by the difference in energy and capacity costs of the two optimized cases [3]. In developing countries, there have been several studies conducted on DSM [4–9]. In [4] electricity load analysis of the commercial sector of Northern Cyprus is simplified by studying only one segment, i.e., hotel buildings which constitute the largest energy-consuming segment in the commercial sector. The study is based on end-use data collection in hotel buildings and subsequent statistical analysis. The study concludes that it is possible to reduce the demand of end-uses by at least 58% by the application of DSM programs, with peak reduction corresponds to 11% of the utility summer peak of
ARTICLE IN PRESS 2158
A.S. Malik / Energy 32 (2007) 2157–2166
Northern Cyprus. The study has not done any effort to find the benefit of avoided cost of utility in doing DSM. In [5] the study has analyzed the most promising DSM technologies, namely power factor correction at end-user, energy efficient lighting in residential and commercial sector, and intelligent motor controllers in Nepal. On-site interviews were undertaken with power utilities (the Nepal Electricity Authority and its subcompanies) and various electricity end-users in Nepal to collect data, as well as walk-through auditing at the end-users’ premises. The study concludes that most of the DSM technologies are financially viable. The utility benefit is estimated in terms of delays and avoidance in the investment, from the envisaged expansion plan, such as (i) cut power imports from India, (ii) reduce power production costs and hence electricity tariffs in Nepal; (iii) avoid the investment in Budhi Ganga (106 MW); and (iv) postpone the investment of Hewa Khola (67 MW). In Sri Lanka [6] the government has created an Energy Conservation Fund and a DSM unit within the Ceylon Electricity Board supported by the World Bank, Global Environmental Facility and the Asian Development Bank. The DSM unit has implemented a broad range of programs and measures, both incentive and information-based, including efficient lighting, energy audits, power factor correction, establishing a ‘best practices’ group with industry, building code and appliance labeling. A study on Greenhouse Gas Mitigation Options in the Sri Lankan Power Sector [7] has taken into account the impact on GHG mitigation using DSM and other options such as alternative energy, reduction in T&D losses, rehabilitation of plants, etc. The study has used the generation expansion planning approach to quantify the environmental benefits. A study on DSM in Thailand [8] shows that variety of DSM programs targeting a wide range of sub-sectors and end-uses have produced savings of 566 MW in peak load and 340 GWh/year as of June 2000. The savings in capacity and energy saving are estimated from the actual reduction in load forecast. The first DSM study in Oman (1998), ‘The Study on Demand Supply Management for Power Sector in Sultanate of Oman’, was conducted by Japan International Cooperation Agency (JICA). The study identified several strategies for potential load management including (i) gas cooling systems for government buildings, hospital, hotels commercial complexes and large houses; and (ii) shifting load from peak time to off-peak time in the industrial and commercial sectors by application of ice thermal storage system and introducing time-of-use tariffs [9]. The utility benefit was estimated using avoided marginal generation capacity and energy cost concept. The study has been shelved since then and there have been no implementation on the recommendations. This paper presents the results of a recent study, undertaken at Sultan Qaboos University, that estimated
the DSM energy saving and load management potential in commercial and government/institutional sectors in Oman (central grid area) and evaluated its impact on generation capacity and energy savings [10]. The commercial sector represents about 16% and government sector represents about 18.4% of the total consumption in 2003 [11]. The commercial and government sectors in Oman includes service businesses, retail and wholesale stores, hotels and restaurants, government ministries, schools and universities, and hospitals and health clinics. The diversity of the end-uses and the technologies available in these buildings requires a large stratified sample. To do a survey with such a large sample size would be difficult because of the poor availability of statistical data and obtaining financial support for such a project. In this paper a different approach is used. Instead of trying to represent the whole commercial and government sector the energy saving potential is worked out in only two end-use functions, i.e., lighting and air conditioning, for large customers of commercial and government/institutional sectors only. These two end-use functions constitute major energy consumption (79–96% of the surveyed samples) in these sectors regardless of the activity in both physical and economic terms. Another unique aspect of the study is the estimation of the energy savings, in transmission and distribution (T&D) losses that are estimated by using generation expansion planning approach. The widely available software packages for generation expansion planning such as WASP, EGEAS, etc. do not include representation of T&D and hence the avoided cost savings, due to DSM, in T&D capacity cost and T&D losses are generally neglected. The method to include the T&D losses is described in Section 3 of this paper. To carry out the study, the load forecast and generation data was taken from Ministry of Housing, Electricity and Water (MHEW) of Oman and the software used to optimize the generation plan is Wien Automatic System Planning (WASP), developed for IAEA, and used by many electric utilities in developing world [12]. The paper is arranged in eight sections. Section 1 is an introduction. Section 2 is a theoretical review of power planning concepts. Section 3 in brief provides a general methodology adopted to carry out the study. Section 4 provides an overview of electricity tariff structure in Oman. Section 5 provides the DSM results of surveyed commercial and government/institutional sectors. Section 6 provides generation and load data of interconnected central grid system. Section 7 presents the results of generation capacity and energy savings. And the last section concludes the paper. 2. Generation expansion planning The objective of electric planning study is to meet the load forecast with high reliability at a minimum cost. There are three keywords in the above statement, i.e., load forecast, reliability, and cost. A brief discussion on these three keywords is followed.
ARTICLE IN PRESS A.S. Malik / Energy 32 (2007) 2157–2166
In generation planning the total cost to be minimized include investment cost in new generation facilities, fuel cost, operation and maintenance cost, and energy not served cost. The cost is minimized depending on the financial resources, technical, environmental and political considerations. The following four questions must be answered when the study of capacity planning is done [13]: (1) (2) (3) (4)
What type of capacities will be added to the system? How much capacities will be added? When these capacities will be added? Where these capacities will be located?
The first three questions can be answered by using any generation expansion planning software. However, to know where to locate the new facilities a detailed feasibility study has to be carried out. The objective function of the least cost-planning software is normally the following: minimize Bj for all j, n X CC i SV i þ FC i þ OM i þ UE i , Bj ¼
ð1Þ
i¼1
where CC is the capital cost, SV is the salvage value, FC is the fuel cost, OM is the operation and maintenance cost, UE is the energy not served cost, n is number of years in study period, and over bar (—) on the above costs represents present worth of all the costs. 2.2. Generation planning study period The study period normally spans from 20 to 25 years. The study period consists of three sub periods: pre-planning period, planning period and post-planning period [14]. Preplanning period is the first 3–4 years in which the planning was done earlier. This is included in the study period to check the energy production cost and reliability of the system. The planning period is between 4 and 10 years in which the decision has to be taken now about the plants that have to be added in future to meet the forecast. The post-planning period is added in the study period to avoid the end-effects by making the calculations continue for another 10 years or more so that a proper trade-off between construction costs and operating costs are found. Therefore, the long-term forecasts of 20–25 years of electricity consumption and demand are used in the planning of investment in generating capacity and the development of fuel supplies.
which simply states that the installed capacity Pt must lie between the given maximum and minimum reserve margins, at and bt, respectively, above the peak demand Dtp in year t. Another popular reliability criterion is the loss-of-load probability (LOLP) index which may be calculated for each period and a limiting value may be suggested by the user. However, it may be noted in Eq. (1) that the reliability is also embedded in terms of energy not served cost. This cost is the penalty for unreliability. If more capacity is added the capital cost would be high but because of more capacity in the system the total cost of energy not served would be low as the system would be more reliable. On the other hand, if the capacity added is less than the actual required then there would be more customer outages and hence higher cost of unserved energy. The basic concept of reliability-cost/reliabilityworth evaluation is relatively simple and can be presented by the cost/reliability curves, as shown in Fig. 1 [15]. The utility curve shows that the investment cost increases with higher reliability, i.e. more investment is required in order to improve reliability. On the other hand, the customer outages cost decreases as the reliability increases. The total cost of the two curves is the sum of the utility cost and customer cost, which exhibits the optimal cost for the society. 3. General methodology DSM economic energy saving potential is worked out from both the customer point of view as well as the electric utility point of view. The estimate of DSM potential is based on a survey of six large commercial (shopping malls) customers and six large government/institutional customers. The annual energy consumption of these 12 customers represents about 4.6% (127.6 GWh) of the total consumption (2775 GWh, year 2003) in these two sectors. The energy saving potential is worked out in only two enduse functions, i.e., lighting and air conditioning, because these two end-use functions constitute the major energy
Annual Costs ($)
2.1. Cost minimization
2159
Utility Cost Total Cost
Customers Cost
2.3. Reliability In expanding the system the reliability constraints can be exogenously added into the cost minimization process. The following reliability constraints may be added to (1): ð1 þ at ÞDtp XPt Xð1 þ bt ÞDtp ,
(2)
System Reliability Fig. 1. Relation between consumers, utility and total annual cost with respect to reliability.
ARTICLE IN PRESS A.S. Malik / Energy 32 (2007) 2157–2166
2160
consumption (79–96% of the surveyed samples) in these sectors regardless of the size of the activity in both physical and economic terms. To work out the benefit of DSM, in these two sectors, from customer point of view the following procedure is adopted: 1. A survey questionnaire was designed to get the data needed from the commercial and government/institutional sectors. The first section of the survey requests some general information about the building such as floor area, electric end-use penetration rate, and annual power consumption. The second section surveyed on lighting end-use share, percentage efficiency and the lifetime of the lighting system and devices. The last section of the survey requires information about the airconditioning system in the sector, its end-use share,
Sectoral energy Consumption data
percentage efficiency and lifetime of the air-conditioning equipment. 2. From the catalogue information available for the efficient cooling and lighting devices [16,17], the cost of efficient cooling and lighting functions were then worked out as 9.4 and 2.1 OR/m2, respectively (1 Omani Rial (OR) ¼ 1000 Baizas (Bz) ¼ 2.58 US$). 3. The estimate of total energy saved in lighting and airconditioning function for each customer was worked out using the percentage of efficiency save in lighting and air conditioning, had the existing lighting and air-conditioning systems were replaced with more efficient equipment, and multiplying the annual consumption in lighting and air-conditioning, respectively. 4. The financial benefit of energy saved component from the customer point of view is calculated by
Forecast sectoral energy consumption till 2024 (end of study period)
Forecast Peak demand in commercial and government sectorsusing (3)
Peak coincident factors data
Contribution of peak demand of both sectors toward the system peak
Load factor data
Average system losses data
Calculate peak losses using (4&5)
Forecast combined commercial and government peak demand to be supplied from the generation end using (6)
Estimated percentage reduction from the sample data
Forecast combined reduction inpeak in both sectors considering the reduction upper and lower limits with 10% and 40% rebound effects respectively
Modify system peak demand forecast
Calculate the utility benefit interms of avoided cost of generation capacity and energy by optimizing the planning horizon thrice; one with a base case set of loadsand the other two cases with loadshape changed because of DSM that is having 40% rebound effect, and 10% rebound effect Fig. 2. Flowchart of avoided cost calculation.
ARTICLE IN PRESS A.S. Malik / Energy 32 (2007) 2157–2166
multiplying the energy saved and the price of the commercial tariff. 5. The present worth of the net benefit and the discounted payback time for each customer were then worked out using discount rate of 10% and the estimated lifetime of lighting (replacement of lighting devices every 6 years) and air-conditioning systems of 20 years. The avoided cost benefit of DSM in these two sectors, from the utility point of view, is shown in a flowchart in Fig. 2 and the details are delineated as follows: 1. An aggregate energy saving potential is worked out with the assumption that the reduction in power consumption due to DSM in the sample data is representative of the actual potential in both the sectors. However, standard error analysis (95% confidence interval) is also done from the sample data by normalizing the energy saved in the 12 customers’ sample. Sector-wise analysis of variance has shown that the standard error of average reduction in power consumption was not significantly different (8.8% and 8.3% for commercial and government sectors, respectively) and the pooled standard error is 8.6%. Moreover, the rebound effect is also studied. Rebound effect suggests that the technical saving potential may not be a correct indicator as consumers may use more energy in other areas. Rebound effect varies across countries even for a particular DSM program. A study done in commercial/household sectors in Norway found that the rebound effect is 10–40% [18]. Using these two limits the range of DSM benefit from electric utility point of view is estimated. 2. The sectoral energy consumption from 1999 to 2003 taken from Ref. [11] is extrapolated and forecasted till 2024 (end of study period). 3. The energy forecast in commercial and government sectors is then converted to peak demand forecast by using the following expression: peak demand in both sectors ðMWÞ ¼
annual energy consumption in both sectors ðMW hÞ , LF 8760 h
ð3Þ where LF is the system load factor. Here, an assumption is made that the load factor of both government and commercial sectors are same as the average system load factor of 53%. The average system load factor was calculated using the past data of average demand divided by the actual system peak demand. 4. To get the contribution of peak demand of both sectors toward the system peak another assumption is made that the peak coincident factor of government sector is 100% of the system peak and commercial sector is 80% of the system peak. This assumption is based on the fact that the system peak occurs between 2:00 and
2161
4:00 p.m. in July (actual data) when all the government offices are working (100% coincidence, assuming it is basically air-conditioning load) and major commercial stores and buildings are open while some of the commercial retailers are closed (about 80% coincidence assumed). 5. The peak losses are calculated from the average system losses from the following expressions [19]: peak losses ¼
average system losses . peak loss factor
(4)
The average of system losses of 8 years from 1995 to 2002 is 18.25% [11]. The peak loss factor in Eq. (4) is the ratio of the energy loss in the system during a given time period to the energy loss that would result if the system peak loss had persisted throughout that period. The following empirical approximation is used to calculate the peak loss factor [19]: peak loss factor ¼ 0:3LF þ 0:7LF 2 .
(5)
6. The combined commercial and government peak demand to be supplied from the generation end is then calculated by the following formula: combined peak to be supplied ¼
peak ðstep 4 after taking into account the coincident factorsÞ . 100% peak losses
ð6Þ 7. The combined reduction in peak in both sectors is then worked out with the estimated percentage reduction from the sample data and considering the reduction upper and lower limits with 10% and 40% rebound effects, respectively. The system peak forecast is then modified accordingly. 8. The utility benefit in terms of avoided cost of generation capacity and energy is then worked out by optimizing the planning horizon thrice; one with a base case set of loads and the other two cases with load shape changed because of DSM that is having 40% rebound effect and 10% rebound effect. 4. Electricity tariff structure The government of Oman provides a large subsidy to the electricity sector, where the domestic sector takes the most of this subsidy because it constitutes the main service receiver among the power consumers. However, the average returns of the electricity sector from the domestic slab constitute 50% of the total subsidy of the government. In addition, the government of Oman provides generous subsidy to other sectors. There are several electricity tariff structures in different sectors. These sectors are domestic, commercial, industrial, agriculture and fisheries, and, hotels and tourist establishments. The domestic tariff increases with the quantity of consumption and starts
ARTICLE IN PRESS A.S. Malik / Energy 32 (2007) 2157–2166
2162
from 10 Bz/kWh for the first 3000 kWh of consumption and end at 30 Bz/kWh for consumption more than 10,000 kWh. The commercial tariff is 20 Bz/kWh without any maximum unit for consumption. The industrial tariff in summer is 24 and 12 Bz/kWh in winter season. The summer is taken from May to August for central grid system. The other tariff rates can be found in annual report of MHEW [11]. 5. Results of surveyed commercial and government/ institutional sectors Tables 1 and 2, respectively, show the results of DSM energy saving potential in commercial, and government/ institutional sectors. These results are from the customers’ point of view. As can be seen from the tables that with the exception of one customer the discounted payback on investment in efficient air-conditioning is less than 10 years and most of them have less than 6 years. The discounted payback on investment in efficient lighting system is about 10 years for most of the surveyed sample.
Before we provide the result of DSM economics from the utility point of view, the next section discusses the generation and load data necessary to run WASP for evaluating DSM economics. 6. Generation and load forecast data The generation expansion planning is carried out for the central grid system from 2005 to 2024. The central grid power system area comprises mainly six interconnected power stations, which are Ghubrah, Rusail, Barka, Wadi Jizzi, Alkamil and Manah stations. The interconnection between these stations is through 132 kV transmission lines. However, recently Barka power station is connected with the central grid through 220 kV lines. All the power stations are operating on natural gas as fuel. The total capacity of these stations is about 2517 MW (year 2005). The power transmission and distribution system mainly consists of 132, 33, and 11 kV and 415 V. Until recently, 220 kV line from Barka station is constructed to connect the station to the main 132 kV grid. Only few kilometers of
Table 1 Final results of the surveyed commercial sectors Hotels/ shopping malls
Al Falaj Hotel Al-Araimi Complex Al-Harthi Complex Holiday Inn Sheraton Hotel OCC
Old consumption (kWh/year)
New consumption (kWh/year)
Air-conditioned
Lights
NPV (OR)
Discounted payback (years)
NPV (OR)
Discounted payback (years)
8 288 750 3 248 000
5 923 141 2 192 725
105 936 44 822
9.4 10.9
105 936 6605
16.9 12.8
6 270 216
3 510 116
119 497
9.3
22 014
10.2
6 896 498 17 963 229
4 288 242.5 11 270 130
218 165 559 277
5.0 4.6
9861 84 382
12.0 6.1
6 182 317
2 866 127
266 211
5.3
18 324
10.3
Table 2 Final results of the surveyed Government/institutional sectors Government/ institutional buildings
Ministry of Civil Services Ministry of Health Nizwa Hospital Royal Hospital Sohar Hospital SQU (College of Arts)
Old consumption (MWh)
New consumption (MWh)
Air-conditioned
Lights
NPV (OR)
Discounted payback (years)
NPV (OR)
Discounted payback (years)
4 561 800
3 228 842
119 871
7.4
7596
10.1
4 257 229
2 613 939
203 677
5.0
19 048
8.5
18 787 806
15 242 735
350 629
6.3
65 320
7.1
28 900 640
22 197 136
569 187
6.6
91 516
8.0
19 943 821 2 313 384
16 459 635 1 545 109
331 949 60 390
6.6 8.1
76 017 8183
6.4 10.5
ARTICLE IN PRESS A.S. Malik / Energy 32 (2007) 2157–2166
66 kV line also exists. The overall system losses reported in 2003 were 20.83%. The generation data collected from MHEW was manipulated, refined, and made suitable for WASP-IV input. For some of the missing data, typical values found elsewhere were used. The generation data for fixed plants, committed plants and candidate plants are briefly presented here followed by load data and economic data.
2163
6.3. Candidate plants The Candidate power plants are plants with which we do our expansion. The candidates are selected based on their economic and technical suitability for the system. Their capital costs are included in the objective function. The data for the candidate units are shown in Table 5. 6.4. Load data
6.1. Fixed system The fixed system is the existing generation system consisting of six plants Ghubrah, Rusail, Wadi Al-Jizzi, Manah, Barka and Al-Kamil. These plants contain many different types and sizes of units. Some of these units are common to most of these plants. Since the WASP model does not include the transmission representation the fictitious power plants were made by grouping same type and size of units. The details of the net capacity, minimum loading level, fixed O&M, variable O&M, forced outage rate, net heat rate at minimum load and average incremental heat rate for each units of the fixed fictitious plants, is given in Table 3. Note the last plant (SOHR) in the list is showing zero numbers of sets. This is because it is a committed plant (year 2005 study) and its two sets are supposed to be added in 2006 and one in 2007.
The surveyed government and commercial buildings annual consumptions are 78.8 and 48.9 GW h, which respectively represents about 5.3% and 3.8% of the total consumption in these sectors in Oman (2003) [11]. An aggregate energy saving potential is worked out with the assumption that the reduction in power consumption due to DSM in the sample data is representative of the actual potential in both the sectors. Table 6 shows the percentage of energy consumption reduction by implementing DSM in commercial and governmental sectors. Also shown are the standard error and the percentage reduction in peak because of 10% and 40% rebound effect. Table 7 shows the actual energy consumption in different sectors from 1999 to 2003. Fig. 3 shows the energy consumption forecast of government and commercial sectors up to 2024 using linear regression on the 5-year data shown in Table 7. The DSM energy forecast is made,
6.2. Committed and retiring plants The committed power plants are the plants which already have been decided to be added to the system in the near future. They are also part of the fixed system. These plants are in different phases of construction and their capital costs are considered sunk. In this study, the plants that are going to be added to the system in the first 3 years of study are considered as committed plants. Similarly the plants that are going to be retired from the systems during the study horizon are also taken into account. Table 4 summarizes the additions and retirements of plants.
Table 4 Summary of additions and retirements from 2005 to 2024 Name Years 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 FR5P 5 FR6B ST13 3 ST56 FR9B SOHR 2
1
1 2
2
5 6
1
2 2
1
1
Table 3 Summary description of thermal plants in year 2005 No.
1 2 3 4 5 6 7 8 9
Plant name
FR5P FR6B FR9E ST13 STM4 ST56 FR9B BRKA SOHR
No. of sets
Minimum load (MW)
Capacity (MW)
Base load
12 15 9 3 1 2 6 1 0
13 20 67 8 35 21 60 363 121
18 27 93 8 50 30 83 427 165
3900 3529 3284 3294 2985 3529 3467 1980 2324
Fast spin (res%)
FOR (%)
Incr. load
Fuel cost (cents/106) (kcal)
O&M (FIX) ($/kWmonth)
O&M (VAR) ($/kWh)
2637 2219 2100 3294 2652 2219 2102 1520 1672
608 608 608 608 608 608 608 608 608
0 0 10 0 10 5 10 5 10
5 5 5 5 5 5 5 5 5
1.37 0.91 0.31 1.5 1 0.91 0.31 0.2 0.25
3.1 2.8 2.6 3.3 2.7 2.8 2.6 1.0 1.2
Heat rates (kcal/kWh)
ARTICLE IN PRESS A.S. Malik / Energy 32 (2007) 2157–2166
2164 Table 5 Summary description of candidate plants No.
1 2 3
Name
Min. load (MW)
FRM9 FRM6 BRKA
67 20 363
Cap acity (MW)
93 27 427
Heat rates (kcal/kWh) Base load
Incr. Load
Fuel cost (cents/106) (kcal)
3284 3529 1980
2100 2219 1520
608 608 608
Fast spin (res%)
FOR (%)
O&M (FIX) ($/kWmonth)
O&M (VAR) ($/kWh)
Capital cost ($)
Construc tion time
Plant life (years)
10 0 10
5 5 5
0.31 0.91 0.2
2.6 2.8 1
531 728 780
3.0 2.0 3.5
20 20 20
Table 6 Implementing DSM in Governmental and commercial sectors Actual Modified energy Energy energy consumption saved consumption (GWh) (GWh) Governmental 78.76 sector Commercial 48.85 sector
Estimated percentage reduction (0% rebound effect) (%)
Standard error in Percentage reduction the energy saved with 40% rebound (%) effect (%)
Percentage reduction with 10% rebound effect (%)
62.12
16.64
21.1
8.8
12.7
19.0
30.05
18.8
38.5
8.3
23.1
34.6
Table 7 Energy consumption in different sectors from 1999 to 2003 Year
Governmental (GWh)
Residential (GWh)
Commercial (GWh)
Industrial (GWh)
Other (GWh)
Total (GWh)
1999 2000 2001 2002 2003
1346 1354 1423 1511 1483
3785 3966 4226 4414 4632
971 1037 1156 1221 1292
271 330 356 438 490
129 146 165 159 161
6502 6833 7326 7743 8058
3000
700 Govt
Comm
Govt-DSM
Comm-DSM
2500 2000 1500 1000 500 0 1999
2004
2009
2014
2019
2024
Years
Fig. 3. Forecast of energy consumption in government and commercial sector with and without DSM.
also shown in Fig. 3, using estimated percentage reductions of Table 6. The energy forecast of government and commercial sectors is then converted to peak demand forecast using (3). Fig. 4 shows the peak demand forecast of both the sectors with and without DSM. The system peak reduction from generation end is worked out as explained in the steps of
Peak Demand (MW)
Energy (GWh)
3500
Govt
Govt-DSM
Comm
Comm-DSM
600 500 400 300 200 100 0 2005
2010
2015
2020
Years
Fig. 4. Forecast of peak demand in government and commercial sectors with and without DSM.
Section 3. Fig. 5 shows the system peak demand forecast, taken from MHEW, and the modified system peak demand with 40% rebound effect and 10% rebound effect. The annual chronological load data of year 2003 taken from MHEW was used for making normalized load duration curve (LDC) shapes. The year was divided into two seasons and two LDC shapes were formed [10].
ARTICLE IN PRESS A.S. Malik / Energy 32 (2007) 2157–2166
6.5. Economic data The discount rate used for the study is 10.0% and the cost of unserved energy as 1.6 $/kWh. The cost of unserved energy was worked out in [20] and is consistent with the optimal reserve margins found using the cost of unserved energy and the average reserve margins existing in Oman.
2165
over the planning horizon (avoided cost ¼ Bj,baseBj,DSM) with 40% and 10% rebound effect is about 416 and 597 million dollars, respectively. At the horizon year the capacity and energy difference from the base case is about 372 MW and 29.2 TWh for 40% rebound effect, and 596 MW and 43.8 TWh for 10% rebound effect. 8. Conclusions
7. Results and discussion
System Peak Demand (MW)
It has been shown in Section 5 that DSM implementation in commercial and government sectors is beneficial from the customer point of view. To investigate the DSM benefit from the utility point of view the avoided cost method is used. The avoided cost of generation capacity and energy is worked out by optimizing the planning horizon thrice; first with a base case set of loads (Case 1), second with load shape changed by the expected load impact of DSM with 40% rebound effect (Case 2) and the third with 10% rebound effect (Case 3). Table 8 provides the summary of the optimization results. The table shows the present worth of total cost (cumulative objective function of (1) in thousands of dollars), the average LOLP index, and the cumulative candidate plants selected over the planning horizon. The table also provides the cumulative capacity selected in each case and the energy requirement for all the three scenarios. The difference in cost, capacity and energy of Cases 1 and 2, and Cases 1 and 3 are also shown. The overall cumulative cost difference 6500 6000 5500 5000 4500 4000 3500 3000 2500 2000 2005
System Peak Load Forecast System Peak Load Forecast- DSM with 40% rebound effect System Peak Load Forecast- DSM with 10% rebound effect
2010
2015
2020
Years Fig. 5. System peak demand forecast with and without DSM.
This paper presented the results of a study that estimated the DSM energy saving and load management potential in commercial and government/institutional sectors in Oman (central grid area) and evaluated its impact on generation capacity and energy savings. The unique aspect of the study is the estimation of the energy savings, in transmission and distribution (T&D) losses that are estimated by using generation expansion planning approach. The study results have shown that it is beneficial to implement DSM in commercial and government/institutional sectors from customers’ point of view. According to the energy calculations of the surveyed sample, we found that applying DSM programs in the government and commercial sectors the energy consumption can be reduced up to 21% and 38%, respectively, with 8.6% standard pooled error. For the surveyed models the net present value of the investment is positive with discounted pay back period (between 4 and 12 years) which encourage sectors to apply DSM. From the supplier (MHEW) point of view the capacity saving at the horizon year ranges between 372 and 596 MW considering 40% and 10% DSM rebound effect, respectively. Likewise, the overall energy saving for the whole planning horizon ranges between 29 and 44 TWh and the total avoided cost saving from 416 to 597 million dollars. One of the problems which faced the countries that applied DSM is motivating costumers to pay for replacing their inefficient devices. Financing DSM remains a significant issue in most places. Many countries have imposed DSM taxes within power tariff schemes. Other governments have opted to fund DSM programs for lowincome customers as a social program, and effectively pay the utility to implement the program [8]. Finally, we can conclude with confidence that demandside management has positive outcomes for consumers,
Table 8 Results of optimal generation expansion plan from 2005 to 2024 Cases
Base (Case 1) DSM with 40% rebound DSM with 10% rebound Difference (Case 1Case Difference (Case 1Case
effect (Case 2) effect (Case 3) 2) 3)
Present worth cost (thousand $) objective function Bj (cumulative)
Average LOLP (%)
3 801 320 3 385 609 3 204 479 415 711 596 841
0.797 0.496 0.527 0.301 0.270
Cumulative candidate plants (no. of units) FRM9
FRM6
BRKA
16 12 5 4 11
0 0 0 0 0
6 6 7 0 (1)
Cumulative capacity (MW)
Cumulative energy (GWh)
6217 5845 5621 372 596
413 752 384 588 369 961 29 164 43 791
ARTICLE IN PRESS 2166
A.S. Malik / Energy 32 (2007) 2157–2166
MHEW, environment and the national income which makes it strongly recommended to start thinking about implementing it in Oman. Acknowledgments The authors would like to acknowledge Internal Research Grant for this project from Sultan Qaboos University and Power Planning Department of Ministry of Housing, Electricity and Water for providing necessary data for carrying out this study. Disclaimer: The views expressed in the paper are entirely ours and do not reflect the position of MHEW. All errors are ours. References [1] Kreith F. Integrated resource planning. J Energy Resour Technol 1993;115:80–5. [2] Busch J, Eto J. Estimation of avoided costs for electric utility demand-side planning. Energy Sour 1996;18:473–99. [3] Malik A. Modeling and economic analysis of DSM programs in generation planning. Int J Electric Power Energy Syst 2001;23:413–9. [4] Atikol U. A demand-side planning approach for the commercial sector of developing countries. Energy 2004;29:257–66. [5] Yang M. Demand-side management in Nepal. Energy 2006;31: 2677–98. [6] World Bank. Project Appraisal Document for Energy Services Delivery Project Report No. 16063-CE. Energy and Project Finance Division Country Department, 1 South Asia Region, Washington, DC, February 1997. [7] World Bank. Greenhouse gas mitigation options in the Sri Lanka power sector. Joint UNDP/World Bank Energy Sector Management Assistance Programme (ESMAP). Washington, DC, February 2003.
[8] World Bank. DSM in Thailand: a case study. A working paper written by Jas Singh and Carol Mulholland, October 2000. [9] Japan International Cooperation Agency (JICA). The study on demand supply management for power sector in Sultanate of Oman. A final report of JICA to Ministry of Housing, Electricity and Water (MHEW). MHEW Central Office: Al-Khuwair, Muscat, Sultanate of Oman, November 1998. [10] Malik A, Al-Khrousi A, Al-Hindi M, Al-Mahrouqi Y. Demandside management energy saving potential in commercial and governmental/institutional sectors and its impact on power planning—a case study of central grid of Oman. In: Proceedings of the second international green energy conference, Oshawa, 25–29, June 2006. [11] Annual statistical book of year 2003 of electrical sectors, Muscat, Oman. Al-Khuwair, Muscat, Sultanate of Oman: Ministry of Housing, Electricity, and Water, MHEW Central Office; 2003. [12] Wien automatic system planning package (WASP-IV). Vienna: IAEA; 2003. [13] Buehring W, Huber C, Marques de Souza J. Expansion planning for electrical generating systems—a guidebook. Technical report series no. 241. Vienna: IAEA; 1984. [14] Wang J, McDonald J. Modern power system planning. London: McGraw-Hill; 1994. [15] Billinton R, Allan R. Reliability evaluation of power systems. New York: Plenum Press; 1996. [16] Bahwan Engineering Company (BEC). Catalogue of cooling devices. BEC Head Office: Ruwi-112, Sultanate of Oman; 2004. [17] Khimji Ramdas Company (KRC). Catalogue of lighting devices. KRC Corporate Office: Muscat-113, Sultanate of Oman; 2003. [18] Roland K, Haugland T. Joint implementation: difficult to implement. In: Jepma CJ, editor. The feasibility of joint implementation. Dordrecht: Kluwer Academic Publishers; 1994. p. 359–66. [19] Pabla A. Electric power distribution. New York: McGraw-Hill; 2005. [20] Al-Farsi F, Al-Shihi M, Al-Shokaili Y. Optimal generation reserve margin for Muscat interconnected system. Final year project report. Muscat, Oman: SQU, Electrical Engineering Department; 2000.