Scenario analysis of the new energy policy for Taiwan's electricity sector until 2025

Energy Policy 61 (2013) 162–171

Contents lists available at ScienceDirect

Energy Policy journal homepage: www.elsevier.com/locate/enpol

Scenario analysis of the new energy policy for Taiwan's electricity sector until 2025 Fung-Fei Chen a, Seng-Cho Chou b,n, Tai-Ken Lu c a

Taiwan Power Company Research Institute, 198, Roosevelt Road, Section 4, Taipei 100, Taiwan Department of Information Management, National Taiwan University, 85 Section 4, Roosevelt Road, Taipei 106, Taiwan c Department of Electrical Engineering, National Taiwan Ocean University, No. 2, Pei-Ning Rd., Keelung 202, Taiwan b

H I G H L I G H T S







We constructed three case scenarios based on the Taiwan government's energy policy. We employed a long-term Generation Expansion Planning optimization model. A significant gap exists between the carbon reduction target and baseline. The carbon reduction target requires a holistic resolution needed taking seriously.

art ic l e i nf o

a b s t r a c t

Article history: Received 26 June 2012 Accepted 27 May 2013 Available online 21 June 2013

For this study, we constructed the following three case scenarios based on the Taiwanese government's energy policy: a normal scenario, the 2008 “Sustainable Energy Policy Convention” scenario, and the 2011 “New Energy Policy” scenario. We then employed a long-term Generation Expansion Planning (GEP) optimization model to compare the three case scenarios' energy mix for power generation for the next å15 years to further explore their possible impact on the electricity sector. The results provide a reference for forming future energy policies and developing strategic responses. & 2013 Elsevier Ltd. All rights reserved.

Keywords: Scenario analysis Energy policy Mixed-integer linear programming

1. Introduction Under the cross impact of the greenhouse effect and the environmental and ecological protection in recent years, power utilities worldwide have been working toward developing an environmentally-friendly development framework that considers carbon reduction, efficiency improvement, and energy conservation. The International Energy Agency's “Energy Technology åPerspective” report (IEA, 2010) stated that the optimum strategies to reduce carbon dioxide emissions comprise the following åmeasures: improving end-user energy efficiency (38%), implementing carbon capture technology (19%), developing renewable energy (17%), establishing end-user alternatives (15%), developing nuclear power generation (6%), and enhancing the efficiency of power generation (5%). This indicates that the main source of reducing carbon dioxide emissions is the energy supply side, which accounts for 47%. In practice, the optimum ratio for the power generation structure must consider numerous factors, including maintaining reserve capacity requirements, ensuring a stable power supply,

n

Corresponding author. Tel.: +886 2 33661182. E-mail address: [email protected] (S.-C. Chou).

0301-4215/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.enpol.2013.05.100

determining the best ratio of various units, satisfying the power load characteristic, balancing regional supply and demand, reducing pollution emissions from power generation units, maintaining power generation methods, diversifying sources, and safely storing fuel. In addition, Taiwan's power system is independent from other nations' power grids; therefore, Taiwan cannot import electricity from other countries when power shortages occur; however, generating excess electricity is considered a serious waste of precious energy and land resources.1 Thus, when planning for the proportion of future long-term optimal power generation structure, the power sector must satisfy these requirements and actual situations, as well as consider reducing carbon dioxide emissions. Therefore, the various influencing factors must be considered. In addition, optimal planning of long-term power generation structures and properly configuring resources from the power supply side are important requirements for developing Taiwan's power industry and achieving the carbon dioxide emission reduction targets. Recently, complying with the government's energy policy has become the biggest variable for the power sector's future longterm Generation Expansion Planning (GEP). The Taiwanese

1

In 2011, 99.4% of Taiwan's energy was imported.

F.-F. Chen et al. / Energy Policy 61 (2013) 162–171

government's 2011 announcement of a new energy policy to reduce the ratio of nuclear power units and expand the ratio of gas-fired units in response to Japan's Fukushima event will significantly impact the power generation structure for the next 15 years. åIn this study, we explore the possible impact of the government's energy policy on the energy sector by using scenario analysis. åIn Section 2, we provide an overview of relevant literature. åIn Section 3, the long-term GEP optimization model is described. In Section 4, the evolution of Taiwan's energy policy and a case simulation analysis of the situation are presented. In Section 5, we analyze the simulation results. Finally, in Section 6, we summarize the results and present our conclusions.

2. Literature overview Following the recently emerging issues regarding carbon reduction, most energy planning models emphasize the simulation of low-carbon scenarios, especially in European countries. Issues such as analyzing the energy proportion of the electricity sector under CO2 reduction objectives, increasing renewable energy proportions, improving the efficiency of power generation, adjusting energy prices, and purchasing and selling through the market aim at transforming the current system into that of lowcarbon emission energy to determine the optimal solution by conducting various simulation analyses (Nagl et al., 2011). Keles et al. (2011) explored diverse studies regarding scenario simulation analysis of the international energy market in 2010. These studies were classified into three main groups: moderate, climate protection, and resource scarcity and high fossil fuel prices. Analyzing the German energy market necessitates creating a fourth scenario group that considers the possible revision of the resolved nuclear energy phase out. Obviously, with respect to climate change, the energy conservation requirements, and carbon reduction objectives, exploring a low-carbon energy proportion structure strategy is the primary current concern in the energy market. Scenario analysis is an essential strategic analysis method; various carbon reduction scenarios among countries have resulted in diverse scenario cases for analysis (Keles et al., 2011; Finon and Pignon, 2008; Moreno et al., 2010; Nagl et al., 2011). In addition to scenario analysis, model application problems are included in the energy planning model. Certain studies have applied inter-industry analysis (i.e., input–output analysis) to analyze the degree of influence that inter-industrial production activities have on each other. The interactive relationship among energy, environment, and economy (3E) is investigated to estimate the CO2 emissions of different industrial sectors (Kim, 1998; Chang, 2008; Chung et al., 2009). Additional studies have established multiple objective programming, which is a mathematical programming model that considers multiple decision objectives in the decision procedure; these studies have explored accommodation relationships for conflicts that exist among objectives (Heinrich et al., 2007; Chang, 2008). Therefore, when different models are used, the main consideration is the size set of the simulation analysis scope. Inter-industry analysis is a preferable model for examining 3E, whereas multiple objective programming is used to determine the solutions and accommodations for different objectives. Other studies have adopted hybrid models, such as combining inter-industry analysis and multiple objective programming to determine the simulation analysis solution (Chang, 2008). Furthermore, to determine the solution of the optimization model, a linear planning theory model is commonly used. The electricity industry exhibited a liberal trend in the 1980s and 1990s, when numerous algorithms were used in the field of electrical energy planning. He and David (1995) proposed the global optimization of long-term power supply planning. Kannan et al. (2007) organized related theories and determined six long-term power

163

supply planning algorithms by conducting experiments. The six algorithms include differential evolution (DE), evolutionary programming (EP), simulated annealing (SA), evolutionary strategy (ES), hybrid approach (HA), and dynamic programming (DP), and are used to determine the optimization algorithm of solutions. Mixedinteger linear programming (MILP) is a derivative of linear programming applications that is used to determine the solution of linear programming models comprising variable types of general and integer variables. When long-term power supply planning uses a generator set as the unit of analysis, the optimization model related to generating the capacity planning of generator sets is fairly complicated. The model must consider the generating capacity and state characteristics of the sets. Commonly, binary integer variables with values of 0 and 1 are used to indicate whether the set is in operation. Therefore, using MILP in the optimization model that uses sets as the unit of analysis is appropriate (Yong and Shahidehpour, 2007; Li et al., 2007; Aminifar et al., 2009). The energy planning model includes a discussion of power generating technologies. Previous studies have explored cost structures of generating technologies. Fritsche (2006) analyzed the generating costs of the German electricity sector, where coal-fired cogeneration and photovoltaics had the lowest and highest generating cost, respectively. This showed that different generator set structures also influenced the total generating cost. In addition, developing advanced generating technologies, including IGCC, CCS, and fuel cells, can significantly influence the carbon reduction effect of electricity sectors. However, according to the “Energy Technology Perspective, Scenarios and Strategies to 2050” by IEA (2006), the majority of the advanced technologies that can effectively eliminate CO2 emissions from fossil fuel generation will not be effective until difficulties in technologies, costs, and other dimensions are solved by 2030. In summary, the established optimization energy planning model, which uses generator sets as the unit of analysis, focuses on the analysis of energy proportion variations among different case scenarios based on the Taiwanese power system. Thus, we adopted a mixed integer linear programming (MILP) model to solve the GEP problem, the details are shown in Section 3. The model compared and analyzed the energy proportion variations among various case scenarios to provide a reference for energy policy makers and electricity sectors to develop corresponding policies and strategies.

3. Optimization model The primary objective of long-term generation expansion planning (GEP) is to formulate the most efficient power supply arrangements that satisfy long-term load demands by examining factors that include the following: government laws and regulations, international energy prices, the greenhouse effect, power generation technology, and operating environments. For example, advanced generators can be used as standby generator sets for GEP, and long-term load forecast software can be employed to formulate programs to address long-term load demands by considering factors such as economic development, demand management, and power efficiency. In addition, international energy prices can be estimated using energy price forecast software. This study adopted a mixed integer linear programming (MILP) model to solve the GEP problem; its objective function and significant constraints are explained below. The brief description of the long-term GEP optimization model is shown in Appendix A. 3.1. Objective function The objective function is the fuel costs of all the units during the planning period, including changes in operation and maintenance costs, carbon taxes, and fixed operation and maintenance costs.

164

F.-F. Chen et al. / Energy Policy 61 (2013) 162–171

3.1.1. Fuel costs The unit fuel cost is the total fuel cost during the planning period. The annual unit fuel costs are calculated by multiplying the t annual total power generation of the unit (MWhi ), the annual fuel t unit costs (PriceFueli ), and the annual fuel unit consumption t (FuelMWhi ), as follows: t

t

t

∑ ∑ðMWhi  PriceFueli  FuelMWhi Þ t

ð1Þ

i

3.1.2. Changes in operation and maintenance costs The changes in operation and maintenance costs for each unit are the total annual changes in unit operation and maintenance costs during the planning period. The annual changes in unit operation and maintenance costs are calculated by multiplying the annual unit power generation and the annual changes in unit operation and maintenance costs (VCost ti ) as follows: t ∑ ∑ðMWhi t i



V Cost ti Þ

ð2Þ

3.1.3. Carbon tax Carbon tax is the total carbon tax incurred during the planning period. The annual carbon tax for the unit is calculated by multiplying the annual unit power generation, the annual unit carbon emissions (CO2 Emissionti ), and the annual carbon tax (CO2 Taxt ) as follows: t

∑t ∑i ðMWhi  CO2 Emissionti Þ  CO2 Taxt

ð3Þ

3.1.4. Fixed operation and maintenance costs The fixed operation and maintenance costs for the unit are the annual fixed unit operation and maintenance costs during the planning period. The annual fixed operation and maintenance costs are calculated by multiplying the unit's installed capacity and fixed operation and maintenance costs (FCost ti ) as follows: ( ) ∑ t

∑

ðMW i  FCost ti Þ þ

i∈fz ¼ 1g

∑

i∈fz ¼ 2g

ðU ti  MW i  FCost ti Þ

ð4Þ

3.2. Constraints The restrictions for long-term GEP comprises numerous factors, such as power supply and demand balance of energy load, CO2 emission restrictions, power allocation according to fuel type, reserve capacity, power allocation according to region, power allocation for base, intermediate, and peak loads, annual fuel consumption and annual capacity factor for generator sets. 3.2.1. Supply and demand balance of energy load The electricity industry must ensure that the amount of electricity produced satisfies the sum of anticipated energy loads t (DemandMWh ) and line loss (Losst ). Energy balance constraint is as follows: t

MWh þ Losst

∑ MW i ≤ð∑ ∑ MW i Þ  FuelRateUptf

ð6Þ

∑ MW i ≥ð∑ ∑ MW i Þ  FuelRateDowntf

ð7Þ

i∈ff g

ð5Þ

t

Here MWhi∈ff ¼ 7g =0:7 is the power consumption of pumped storage generator sets. 3.2.2. Power allocation according to fuel type The government's policies (regarding low-carbon energy, energy safety etc.) determine the installed capacity ratio for various fuel units. For example, the total installed capacity of gas-fired units must be ≥30% of the total installed capacity. Fuel is

f i∈ff g

i∈ff g

f i∈ff g

Here Fuel Rate Uptf and Fuel Rate Downtf are the upper and lower limits of unit fuel ratio respectively. 3.2.3. The annual carbon dioxide emissions The annual carbon dioxide emissions cap (MaxCOt2 ) is as follows: t

∑ðMWhi  CO2 Emissionti Þ ≤MaxCOt2

ð8Þ

i

3.2.4. Reserve capacity In consideration of electricity supply safety and the prevention of power brownout caused by the localized generator set failure, percent reserve margins (SRt ) are defined to stipulate that the electricity industry's overall installed capacity should exceed the peak load for power. Reserve capacity is at least 16% higher than the peak load (DemandMW t ): ∑

MW i þ

i∈fz ¼ 1g

∑

i∈fz ¼ 2g

U ti  MW i ≥ð1 þ SRt Þ  DemandMW t

ð9Þ

3.2.5. Power allocation according to region To prevent unstable system operations caused by excessive power concentration in specific regions and line loss resulting from long distance transmission, the electricity industry has defined power allocation for various regions to ensure regional balances of power supply and demand. Region is divided into 4 categories, namely, (y: 1 ¼ north, 2 ¼central, 3 ¼south, 4¼ east). ∑ MW ≤ð∑ ∑ MW i Þ  Area Rate Upty

ð10Þ

∑ MW ≥ð∑ ∑ MW i Þ  Area Rate Downty

ð11Þ

i∈fyg

Here z is the generator set status (1: operational, 2: under planning), U ti is the unit status (1: planned, 0: not planned) and MW i is the installed capacity of unit i.

t t ∑ðMWhi Þ− ∑ ðMWhi =0:7Þ ¼ Demand i i∈ff ¼ 7g

divided into 7 categories, namely, (f: 1 ¼oil, 2 ¼coal, 3¼ gas, 4¼ nuclear energy, 5 ¼hydroelectricity, 6 ¼renewable energy, 7¼ pumped storage hydro). Fuel ratio constraints are as follows:

i∈fyg

i

y i∈fyg

i

y i∈fyg

Here Area Rate Upty and AreaRateDownty are the upper and lower limits of unit regional ratio respectively. 3.2.6. Power allocation for base, intermediate, and peak loads In consideration of peak and off-peak loads, the electricity industry must install generator sets of appropriate proportions according to set properties to accommodate load patterns and reduce operating costs. Load is divided into 3 categories, namely (x: 1 ¼base, 2 ¼intermediate, 3¼ peak). ∑ MW i ≤ð∑ ∑ MW i Þ  Type Rate Uptx

ð12Þ

∑ MW i ≥∑ ∑ MW i  Type Rate Downtx

ð13Þ

x i∈fxg

i∈fxg

x i∈fxg

i∈fxg

Here TypeRateUptx and Type Rate Downtx are the upper and lower limits of unit load ratio respectively. 3.2.7. Annual fuel consumption The electricity industry's fuel procurement strategies limit the annual fuel consumption (Fuel Maxtf ) by generator sets that åuse specific types of fuels. For example, the current contract procurement for natural gas limits the energy production of generator sets that consume natural gas. t

t

∑ ðMWhi Þ  FuelMWhf ≤Fuel Maxtf

i∈ff g

ð14Þ

F.-F. Chen et al. / Energy Policy 61 (2013) 162–171

165

Table 1 Case analysis of the scenario conditions. Power ratio

Renewable energy (excluding pumped storage hydro)

Current ratio (capacity％) January 2010

Until 2025 BAU case

Sustainable case


Highest capacity at 15%


Power generation accounts for

Traditional combustion Natural gas


Installed capacity accounts for

more than 8%

5.70


Nuclear plants 1 to Nuclear

New Energy case




12.80

more than 20%


Promote energy diversification
Promote nuclear energy as a

3 decommissioned according to schedule Nuclear plant 4 becomes operational according to schedule


Nuclear plants 1–3

carbon-free alternative




decommissioned according to schedule Nuclear plant 4 becomes operational according to schedule

36.50

–

Coal

29.60

–

Power generation accounts for more Installed capacity accounts for than 25% of the system approximately 40%
Introduce clean coal technology
Develop carbon capture and – storage methods

Fuel oil (light and heavy oil)

9

–

–




–

Carbon dioxide emission limits

3.2.8. Annual capacity factor for generator sets Various generator sets may be unable to perform year-round power generation because of maintenance shutdown following system failure. Therefore, the electricity industry currently defines the annual capacity factor for various generator sets based on their failure rates to establish a ceiling for annual actual power generation. t

MWhi ≤MW i  CF ti  8760; t

i∈fz ¼ 1g

MWhi ≤U ti  MW i  CF ti  8760;

i∈fz ¼ 2g


Reach 2008 emission levels

ð15Þ

–


Same with the Sustainable Energy

between 2016 and 2020 Reach 2000 emission levels in 2025

Policy Convention

which will become increasingly significant if the cost of extracting shale oil is reduced. Therefore, this study added the fourth case scenario (i.e., the ING case), where the natural gas price was 20% lower than that of the BAU case, to compare the difference between the two cases. The comparison results were used to analyze the influence of declining natural gas prices on the longterm GEP. The analysis details are provided in Section 5. 4.1. Business As Usual case (BAU case)

ð16Þ

Here CF ti is the annual capacity factor of unit i. 4. The evolution of Taiwan's energy policy and the scenarios analysis Since the Kyoto Protocol was implemented on February 16, 2005, Taiwan has held the “National Energy Conference,” “National Sustainable Development Conference,” and “Taiwan's Sustainable Economic Development Conference.” In addition, Taiwan government has announced the “Sustainable Energy Policy Convention” in 2008 to set carbon dioxide emission reduction targets. In response to the Fukushima nuclear crisis caused by the 2011 earthquake in Japan, the Taiwanese government announced the “New Energy Policy” to expand the ratio of natural gas and renewable energy and implement a nuclear decommissioning schedule. Table 1 shows the case scenario based on the government's energy policy and carbon dioxide emission control targets. The forecasted peak load and power supply information for the three cases and the required conditions and parameters for simulation analysis were obtained from government information.2 The three simulated situations are presented below. The long-term GEP will be continued for the next 15 years. The current natural gas prices have experienced a declining trend, 2 Quoted from the “Future Electricity Supply and Demand Planning” published by Taiwan's Bureau of Energy, Ministry of Economic Affairs in 2010.

The basic case scenario involves the extended use of Taiwan's existing power system, where Nuclear Power Plants 1–3 are decommissioned according to schedule, and Nuclear Power Plant 4 becomes operational according to schedule. The carbon dioxide emission limits were simulated based on Taipower's power system emissions3 standard. 4.2. Sustainable energy policy convention case (Sustainable case) Based on the previous performance of Taipower's power system (〈http://www.taipower.com.tw〉), in this case, we assumed that the power sector's carbon dioxide emissions limit is 115.2 million tons by 2020 and 90.1 million tons by 2025. In 2025, the low-carbon energy generation in total energy generation will account for more than 55%, renewable energy power generation will account for more than 8%, and natural gas power generation will account for more than 25%. Thus, the operation of Nuclear Power Plants 1–3 will be extended and Nuclear Power Plant 4 will become operational according to schedule, where Unit 1 becomes operational in 2014 and Unit 2 is activated in 2016. 4.3. New Energy Policy case The important conditions of the New Energy Policy case (New Energy case) are that Nuclear Power Plants 1–3 will be 3 Quoted from the “Electric Power Development Program,” which was published in 2010.

166

F.-F. Chen et al. / Energy Policy 61 (2013) 162–171

Table 2 Case analysis results. Case

Performance

BAU case

Year

2010

2020

Installed capacity (GW)/(ratio %) Renewable energy Nuclear Traditional energy Natural gas Coal Oil Reserve capacity (%) Annual electricity generation (TW h) Renewable energy Nuclear Traditional energy Natural energy Coal Oil Carbon dioxide emissions (Tg) Emission intensity (kg/kW h)

(5.7) (12.8)

7.67 (14) 6.57 (12)

Sustainable case 2025

8.98 (15) 3.65 (6)

2020

New Energy case 2025

2020

2025

15.6 (23) 7.84 (12)

33.3 (41) 7.84 (10)

24.6 (31) 6.57 (8)

34.7 (37) 3.65 (4)

19.2 (23) 15.9 (19) 2.43 (3) 40.1

24.3 (31) 18.0 (23) 2.28 (3) 43.8

33.7 (36) 15.9 (17) 2.0 (2) 47.4

(36.5) (29.6) (9) 23.4

16.5 (30) 18.0 (33) 2.28 (4) 18.3

19.8 (35) 18.2 (33) 2.0 (4) 17.7

19.5 (29) 18.0 (27) 2.3 (4) 33.3

5.2 40.0

24.8 48.9

29.1 27.2

50.9 58.4

108.4 58.4

83.6 48.9

58 83.8 7.8 111 0.55

119.6 100.8 5.1 154 0.51

132.8 135.4 4.5 192 0.57

130.3 54.4 5.2 116 0.38

128.2 28.4 5.4 91 0.27

83.1 78.4 5.1 116 0.38

112.4 27.2 183.9 0.9 4.5 91 0.27

Notes: The “Performance” comes from the actual value of power system in Taiwan in 2010. The initial generation resource types and their capacities from the optimization model shown in Appendix B.

decommissioned according to schedule, and Nuclear Power Plant 4 will become operational according to schedule. The capacity of installed natural gas in total installed capacity will account for approximately 40% in 2025. The capacity of renewable energy will increase to approximately 20% in 2025.

approximately one quarter of that of the BAU case (Fig. 1). The significant decline in the use of coal-fired units will cause the redundancy problems for coal-fired units. However, the substantial increase in renewable energy will cause problems for new structures attempting to obtain power and the surrounding environment set to be assessed.

5. Scenario results until 2025 Table 2 shows the analysis results of the three cases, which demonstrate the impact of the government's energy policy, described as follows. 5.1. BAU case For the BAU case, the nuclear power unit's installed-capacity ratio decreased by 6% in 2025. The gap was bridged by the natural gas units with minimal changes. The gradual transformation to being nuclear-free and reducing the high cost of fuel units conforms to Taiwan's energy policy. However, CO2 emissions will continue to increase to 192 Tg by 2025 and not meet the latest CO2 reduction policy requirements. The BAU case shows that because the load demand continues to grow and the nuclear units are decommissioned, the optimum ratio of the future power generation structure still relies on gas and coalfired units. 5.2. Sustainable case The Sustainable case is subject to stringent CO2 reduction goals (i.e., return to the emissions level in 2000, which is approximately 90.1 million tons, by 2025). The renewable energy supply structure will grow significantly from 3.7 GW in 2010 to 33.3 GW in 2025 (accounting for 41%). Because nuclear energy is used as a carbon-free option (i.e., installed capacity in 2025 is 7.84 GW, accounting for 10%), the installed capacity for the low-carbon gas-fired units changed minimally (i.e., 19.2 GW in 2025), which indicated that the nuclear power unit plays a decisive role in the allocation and reduction of carbon dioxide emissions. However, the installed capacity of coalfired units will be 15.9 GW in 2025, which is the existing capacity (including established and constructed standard units), and no new units will be added. In addition, the generating capacity of coal-fired units will decline substantially to 28.4 TW h by 2025, which is

5.3. New Energy case The New Energy case has the same carbon dioxide emissions reduction objective as that of the Sustainable case. However, to satisfy Taiwan's nuclear-free policy, nuclear power units will be decommissioned gradually, renewable energy units will be expanded significantly (34.7 GW by 2025, accounting for 37%), and the installed capacity of low-carbon gas units will be increased substantially (33.7 GW by 2025, accounting for 36%). Thus, the New Energy case will have the highest ranking of the three cases. The proportions of the gas and the renewable energy unit are higher than those in the optimal power generation structure. Besides facing the same challenges as those of the Sustainable case, a substantial increase in the electricity generation of natural gas units (183.9 TWh by 2025, which is significantly greater than that of the BAU case at 132.8 TWh), as shown in Fig. 2, will cause the purchase, storage, and transport of natural gas to be a significant challenge. Overall, the results of the comparative case analysis indicate that for the Sustainable case and New Energy case to achieve the policy objectives of CO2 reduction, a rapid increase in renewable energy units is required by 2014. Specifically, because the New Energy case must bear the load of the decommissioned nuclear units, the installed capacity for renewable energy is significantly higher than that of the Sustainable case, as shown in Fig. 3. Because of the low renewable energy output capacity factor (i.e., oil 0.85, coal 0.85, natural gas 0.88, nuclear 0.85, and renewable energy 0.37), the required reserve capacity is also influenced. The New Energy case is the highest in the required reserve capacity among the three cases (the BAU case was between 17% and 21%, the Sustainable case was between 18% and 40%, and the New Energy case was as high as 28–47%, as shown in Fig. 4), reaching a reserve capacity of almost 50%. This requirement is another test of the power company's operational costs and performance.

F.-F. Chen et al. / Energy Policy 61 (2013) 162–171

167

Fig. 1. Coal-fired unit generating capacity comparison.

Fig. 2. Gas-fired unit generating capacity comparison.

Fig. 3. Renewable energy unit installed capacity comparison.

5.4. ING case In this study, almost all parameters and conditions of the ING and BAU cases were identical. The only difference between the two cases was that the natural gas price of the ING case was 20% lower than that of the BAU case. The scenario analysis results show

that regarding the long-term GEP, the installed capacity of the traditional combustional types of generator sets is primarily influenced by the parameter of the unit fuel ratio instead of fuel prices, as shown in Figs. 5–7. However, the generating capacity of generator sets that use different fuels is influenced by fuel prices. By using the analysis containing 20% lower natural gas price as

168

F.-F. Chen et al. / Energy Policy 61 (2013) 162–171

Fig. 4. System reserve capacity comparison.

Fig. 5. Oil unit installed capacity comparison.

Fig. 6. Coal-combustion unit installed capacity comparison.

an example, the price of natural gas remains higher than that of coal but lower than that of oil, thereby retraining the generating capacity of the oil unit. Thus, the generating capacity of the

oil units in the ING case declines as a whole compared to that of the BAU case. The generating capacity of the oil unit is released and allocated to the coal-combustion and natural gas units.

F.-F. Chen et al. / Energy Policy 61 (2013) 162–171

169

Fig. 7. Natural gas unit installed capacity comparison.

Table 3 Analysis results of the generating capacity for various generator units in the BAU and ING cases. Year

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025

Oil (TW h)

Coal (TW h)

Gas (TW h)

BAU

ING

BAU

ING

BAU

ING

7.44 6.77 6.77 15.62 6.77 6.77 6.77 6.77 5.09 5.09 5.09 4.47 4.47 4.47 4.47 4.47

6.20 5.64 5.64 15.62 5.64 5.64 5.64 5.64 4.24 4.24 4.24 3.72 3.72 3.72 3.72 3.72

68.02 69.14 63.94 50.00 67.59 69.05 80.48 82.78 97.33 97.76 100.76 109.70 113.70 125.64 128.18 135.35

68.71 69.78 64.57 50.00 68.23 69.69 81.12 83.41 97.81 98.23 101.23 110.12 114.12 126.06 128.59 135.76

89.43 96.71 104.85 114.41 104.00 111.71 97.65 103.71 98.79 110.65 119.55 115.60 124.65 118.05 127.87 132.79

89.97 97.21 105.34 114.41 104.49 112.20 98.14 104.20 99.16 111.02 119.93 115.93 124.97 118.38 128.20 133.12

The allocated generating capacity is influenced by CO2 emissions restrictions, resulting in a slight increase in the generating capacity of coal-combustion and natural gas units, as shown in Table 3.

6. Conclusions Under the impact of global climate change and the pressure to save energy and reduce carbon emissions, a low-carbon power generation method has become a common development goal for electricity sectors worldwide. For Taiwan's electricity sector, carbon emissions are estimated to reach 159.5 million tons by 2020 åand 194.5 million tons by 2025. The carbon reduction target is 78.5 million tons by 2020 and 76.5 million tons by 2025. The Fukushima incident in Japan (2011) had a significant impact on Taiwan's energy policy. Taiwan announced a new energy policy to reduce the nuclear power generation ratio and expand the gas-fired and renewable energy ratio. Determining how this latest energy policy was implemented in response to public concern, and how the country's carbon reduction objectives will affect the electricity sector, is worth investigating.

For this study, we constructed three scenario cases to analyze the optimum power generation structure for the electricity sector based on the government's current energy policy requirements for carbon dioxide emissions. From the BAU case simulation, the results indicate that under the existing power generation structure, a significant gap exists between the carbon reduction target and the actual carbon emissions baseline. When nuclear energy is not an intermediate option for meeting the carbon reduction target, the electricity sector must actively seek a full range of effective carbonreduction methods. Possible directions for these efforts include continuing to promote energy conservation, strengthening current load management measures and demand-side management efficiency, improving energy-use efficiency, importing new demand response measures, implementing carbon capture technology, promoting biological carbon fixation, expanding electrical end-use alternatives, and enhancing power generation efficiency. To achieve the carbon dioxide reduction targets, the Sustainable case and the New Energy case will lead to high costs, such as increasing the ratio of renewable energy, the reserve capacity, and idle gas-fired units, and reducing the use of coal-fired units. These problems can lead to the wasting of resources and pose significant business operation challenges for the power utility. The authorities and the power sector should preemptively address these potential problems. Relevant measures to ensure power supply quality, such as configuring sufficient storage devices and ancillary services, should be prepared in advance, especially as the proportion of renewable energy increases. Furthermore, we also found that the carbon reduction target requires a holistic resolution and cannot be achieved by the power generation structure of the electricity sector alone. The Sustainable case and the New Energy case in this study simulated and analyzed scenarios based on carbon dioxide reduction target planning. Although nuclear energy could be used as an intermediate energy source in the Sustainable case, the optimum power generation structure had areas that were difficult to achieve in practice. For example, the proportion of renewable energy will reach 41% in 2025 (installed capacity 3325.2 MW) with a reserve capacity ratio of 40.1%, and the existing coal-fire units could be idled. The New Energy case forced the proportion of the installed capacity of natural gas energy sources to reach approximately 40%, and that of renewable energy to increase to 20% by 2025. The optimal power generation structure is even less likely to be realized in practice under these stringent standards. For example, in 2025, renewable energy will compose 37% (installed capacity 3475 MW) with a reserve capacity ratio of 47.4%, installed capacity

170

F.-F. Chen et al. / Energy Policy 61 (2013) 162–171

Fig. B1.

from coal-fired units will reach an excessive 3369.3 MW, gas resources could be in short supply, and the existing coalfired and gas-fired units could be idled. This study requests that policy-makers consider the practicality of carbon dioxide reduction when formulating energy policy, because (as the results of the

simulation analysis in this study demonstrate) future carbon dioxide reduction targets that fail to consider the viable framework and capacity of the electricity sector only create another type of resource waste and the beginning of more rigorous tests on the management of the electricity industry.

F.-F. Chen et al. / Energy Policy 61 (2013) 162–171

171

Appendix A. The description of the long-term GEP optimization model

The model type Objective function The decision variables The input parameters

This study adopted a mixed integer linear programming (MILP) model to solve the GEP problem The objective function is the fuel costs of all the units during the planning period, including changes in operation and maintenance costs, carbon taxes, and fixed operation and maintenance costs Units installed capacity, units generating capacity and the integer variable of units on/off Load demands, energy prices, CO2 emission restrictions, power allocation according to fuel type, reserve capacity, power allocation according to region, power allocation for base, intermediate, and peak loads, annual fuel consumption and annual capacity factor for generator sets, standby generators By AMPL

The modeling language The solver

By CPLEX

Appendix B. The initial generation resource types and their capacities Please see Fig. B1. Case

BAU

Sustainable energy

Fuel type

Annual electricity Annual electricity Installed Annual electricity Installed Installed capacity, ĆGW generation, ĆTWh (%) Ćcapacity, GW generation, ĆTWh (%) Ćcapacity, GW generation, ĆTWh (%) (%) (%) (%)

Oil Coal Natural gas Nuclear Renewable energy

3.33 11.9 14.85 5.14 3.7

(8.6) (30.6) (38.1) (13.2) (9.5)

7.44 68.01 89.43 38.3 11.88

(3.5) (31.6) (41.6) (17.8) (5.5)

3.33 11.9 14.85 5.14 3.7

(8.6) (30.6) (38.1) (13.2) (9.5)

7.44 68.01 89.43 38.3 11.88

New energy

(3.5) (31.6) (41.6) (17.8) (5.5)

3.33 11.9 14.85 5.14 9.79

(7.4) (26.4) (33.0) (11.4) (21.8)

7.44 88.52 49.01 38.30 31.79

(3.5) (41.2) (22.8) (17.8) (14.8)

Note: There are stringent requirements for New Energy case shown in Table 1. For example, the installed capacity of the renewable energy in total energy should account for more than 20%. That causes the initial generation resource types and their capacities of New Energy case obviously different from BAU case and Sustainable Energy case.

References Aminifar, F., Fotuhi-Firuzabad, M., Shahidehpour, M., 2009. Unit commitment with probabilistic spinning reserve and interruptible load considerations. IEEE Transactions on Power Systems 24, 388–397. Chang, Ssu-li, 2008. Analysis of Taiwan's long-term power supply and demand planning strategy using MultEEE model. In: 2008 Academic Conference on Environmental Resource Economic, Managerial, and Systematic Analysis. Chung, Whan-Sam, Tohno, Susumu, Yul Shim, Sang, 2009. An estimation of energy and GHG emission intensity caused by energy consumption in Korea: an energy IO approach. Applied Energy. Finon, Dominique, Pignon, Virginie, 2008. Electricity and long-term capacity adequacy: the quest for regulatory mechanism compatible with electricity market. Utilities Policy 16, 143–158. Fritsche, Uwe R. 2006. Comparison of greenhouse-gas emissions and abatement cost of nuclear and alternative energy options from a life-cycle perspective. OEKO. He, Y.Q., David, A.K., 1995. Advances in global optimization for generation expansion planning. IEE Proceedings—Generation, Transmission and Distribution 142, 423–428. Heinrich, G., Howells, M., Basson, L., Petrie, J., 2007. Electricity supply industry modeling for multiple objectives under demand growth uncertainty. Energy 32, 2210–2229. IEA, 2006. 2006 Energy Technology Perspective. International Energy Agency.

IEA, 2010. 2010 Energy Technology Perspective. International Energy Agency. Keles, Dogan, Most, Dominik, Fichtner, Wolf, 2011. The development of the German energy market until 2030—a critical survey of selected scenarios. Energy Policy 39, 812–825. Kannan, S., Slochanal, S.M., Baskar, S., Murugan, P., 2007. Application and comparison of metaheuristic techniques to generation expansion planning in partially deregulated environment. IET Generation, Transmission & Distribution 1, 111–118. Kim, S.W., 1998. A study on the industry restructuring to cope with the Kyoto protocol. Korea Environment Institute. Li, T., Shahidehpour, M., Li, Z., 2007. Risk-constrained bidding strategy with stochastic unit commitment. IEEE Transactions on Power Systems 22, 449–458. Moreno, R., Barroso, L.A., Rudnick, H., Mocarquer, S., Bezerra, B., 2010. Auction approaches of long-term contracts to ensure generation investment in electricity markets: lessons from the Brazilian and Chilean experiences. Energy Policy 38, 5758–5769. Nagl, Stephan, Fürsch, Michaela, Paulus, Moritz, Richter, Jan, Johannes, Trüby, Lindenberger, Dietmar, 2011. Energy policy scenarios to reach challenging climate protection targets in the German electricity sector until 2050. Utilities Policy 19, 185–192. Yong, Fu, Shahidehpour, M., 2007. Fast SCUC for large-scale power systems. IEEE Transactions on Power Systems 22, 2144–2151.