Assessment of Air-Pollution Control Policy ’ s Impact on China ’ s PV Power : A System Dynamics Analysis

Abstract: Recently, China has brought out several air-pollution control policies, which indicate the prominent position that PV power hold in improving atmosphere environment. Under this policy environment, the development of China’s PV power will be greatly affected. Firstly, after analyzing the influencing path of air-pollution control policies on PV power, this paper built a system dynamics model, which can be used as a platform for predicting China’s PV power development in every policy scenario during 2015–2025. Secondly, different model parameters are put into the SD model to simulate three scenarios of air-pollution control policies. Comparisons between the simulated results of different policy scenarios measure the air-pollution control policy’s impact on China’s PV power in the aspect of generation, installed capacity, power curtailment and so on. This paper points out the long-term development pattern of China’s PV power under latest incentive policies, and provides reference for the policymakers to increase the effect and efficiency of air-pollution control policies.


Introduction
China's increasingly serious air-pollution problem is caused by its unreasonable energy developing pattern and the lack of pollution-controlling consciousness [1,2].In September 2013, the Chinese government has introduced a severe policy named The Action Plan for Air Pollution Control, which shows the beginning of powerful air-pollution management.To drive the implementation of this action plan, China's energy industry introduced several supporting policies, which mention the necessity of adjusting energy structure.Among all the renewable energy generation technologies, photovoltaic (PV) power generation is regarded as the most perspective one because of its perfect match in time to the industry electrical load [3][4][5][6][7].Then, China's solar power industry issued Some Suggestions to Improving PV Industry Development to match the requirement of air-pollution controlling for PV power industry.
These air-pollution control policies fully indicate the prominent position that PV power hold in improving atmosphere environment and promoting energy substitution.The macrolevel policies will increase the relative demand of PV power generation by limiting coal consumption.In addition, microlevel policies will complete the electricity price and subsidy mechanism of PV power, give priority to the generation scheduling of PV power, and implement the full acquisition of PV power generation.This favor policy environment will bring China's PV power an unprecedented developing opportunity [8,9].However, on the other hand, the high targets for installed PV capacity set in these policies will bring PV power great challenges as well [10,11].China's PV power is growing rapidly now, with installed-capacity growing by 60% and generation growing by 200% year on year.However, in 2014, China's installed PV capacity reached only 28

Analysis of Air-Pollution Policy Contents
In September 2013, China's state council issued The Action Plan for the Control of Air Pollution.It pointed out several measures which can influence PV industry: (1) encourage electric power replacement; 2) restrict new coal-fired power generation project; (3) set non-fossil fuel energy share target in 2017 as 13%; (4) stimulate renewable energy generation; (5) limit high energy consumption industry; (6) promote green building with distributed rooftop PV installation; (7) improve price mechanism of renewable energy generation.These policies measures will undoubtedly increase the share of non-fossil fuel energy, and bring about uncertainty to electricity consumption.
In March 2014, the National Energy Administration (NEA) issued Energy Industry's Action Plan for Strengthening Air Pollution Control, which formulated several measures to support The Action Plan for the Control of Air Pollution on energy industrial level: (1) set PV power as an important development direction for energy industry; (2) set targets for installed PV capacity and distributed PV installed capacity; (3) outline the 'feed PV power to grid after self-consumption' principle for the development of distributed PV power; (4) set the target of proportion of coal consumption in total energy consumption in 2017 as 65%; (5) set target for interregional power transmission; (6) improve subsidy mechanism of distributed renewable energy generation.These measures provided PV power with good opportunity and great challenge.
In December 2014, the National Development and Reform Commission (NDRC) issued other supporting policy Interim Regulations for Reducing and Replacing Coal Consumption.Its measures, which can influence PV power including: (1) set coal consumption reduction target; (2) close down backward production facilities; (3) accelerate power grids construction.
In July 2013, the state council issued Some Suggestions to Improving PV Industry Development to make sure PV power will make great contribution to air-pollution control.This policy is first proposed at the executive meeting of the state council on February 12 as a supporting policy of The Action Plan for the Control of Air Pollution, and formally announced a little earlier than The Action Plan.It (1) set benchmark prices for different PV resource areas; (2) set whole electricity subsidy (0.42 CNY/kWh before 2020) and purchase mechanism for distributed PV power; (3) free self-consumption electricity of distributed PV system from all taxes and fees levied along with electricity price; (4) accelerate technology advancement to increase utilization hours and reduce cost.
In general, air-pollution control policy on national level will indirectly affect PV industry by promoting electric power replacement and energy structure adjustment.In addition, air-pollution control policies on energy industrial level will indirectly stimulate PV industry by restricting coal consumption and setting it as an important development direction.These policies together with PV power stimulating policies constitute the framework of air-pollution control policy in this paper, to ensure PV power will give a boost to air-pollution control.The specific effects of policy measures on PV industry are shown in Figure 2.
Because of the uncertainties China's PV power will face, it is of great significant to figure out how China's PV power will be affected by these 'air-pollution' control policies.Considering its advantages on integrity and dynamics during the complex analysis, this paper plans to establish a SD model with different parameter settings, which can be determined by the policy analyses

Analysis of Air-Pollution Policy Contents
In September 2013, China's state council issued The Action Plan for the Control of Air Pollution.It pointed out several measures which can influence PV industry: (1) encourage electric power replacement; (2) restrict new coal-fired power generation project; (3) set non-fossil fuel energy share target in 2017 as 13%; (4) stimulate renewable energy generation; (5) limit high energy consumption industry; (6) promote green building with distributed rooftop PV installation; (7) improve price mechanism of renewable energy generation.These policies measures will undoubtedly increase the share of non-fossil fuel energy, and bring about uncertainty to electricity consumption.
In March 2014, the National Energy Administration (NEA) issued Energy Industry's Action Plan for Strengthening Air Pollution Control, which formulated several measures to support The Action Plan for the Control of Air Pollution on energy industrial level: (1) set PV power as an important development direction for energy industry; (2) set targets for installed PV capacity and distributed PV installed capacity; (3) outline the 'feed PV power to grid after self-consumption' principle for the development of distributed PV power; (4) set the target of proportion of coal consumption in total energy consumption in 2017 as 65%; (5) set target for interregional power transmission; (6) improve subsidy mechanism of distributed renewable energy generation.These measures provided PV power with good opportunity and great challenge.
In December 2014, the National Development and Reform Commission (NDRC) issued other supporting policy Interim Regulations for Reducing and Replacing Coal Consumption.Its measures, which can influence PV power including: (1) set coal consumption reduction target; (2) close down backward production facilities; (3) accelerate power grids construction.
In July 2013, the state council issued Some Suggestions to Improving PV Industry Development to make sure PV power will make great contribution to air-pollution control.This policy is first proposed at the executive meeting of the state council on February 12 as a supporting policy of The Action Plan for the Control of Air Pollution, and formally announced a little earlier than The Action Plan.It (1) set benchmark prices for different PV resource areas; (2) set whole electricity subsidy (0.42 CNY/kWh before 2020) and purchase mechanism for distributed PV power; (3) free self-consumption electricity of distributed PV system from all taxes and fees levied along with electricity price; (4) accelerate technology advancement to increase utilization hours and reduce cost.
In general, air-pollution control policy on national level will indirectly affect PV industry by promoting electric power replacement and energy structure adjustment.In addition, air-pollution control policies on energy industrial level will indirectly stimulate PV industry by restricting coal consumption and setting it as an important development direction.These policies together with PV power stimulating policies constitute the framework of air-pollution control policy in this paper, to ensure PV power will give a boost to air-pollution control.The specific effects of policy measures on PV industry are shown in Figure 2.
Because of the uncertainties China's PV power will face, it is of great significant to figure out how China's PV power will be affected by these 'air-pollution' control policies.Considering its advantages on integrity and dynamics during the complex analysis, this paper plans to establish a SD model with different parameter settings, which can be determined by the policy analyses mentioned above, to simulate the development of China's PV industry in different policy scenarios.Further, comparing the differences between simulation results in different policy scenarios can help us assess the effect of air-pollution control policy on China's PV industry.

Influence on solar power Industry
Figure 2. Air-pollution control measures and its effects on PV industry.

Methodology
China's PV power development with and without air-pollution control policy can be predicted by the same SD model with different parameter settings.As a systems modeling and dynamic simulation methodology, SD can analyze dynamic complexities in socio-economic and bio-physical systems [30,31].SD model can better simulate China's PV power development than other types of quantitative modeling, because of its advantage of analyzing complex relationships among the variables involved in energy development system, and simulating the feedback effects of backward variables on forward variables.It also enables us to formulate relevant policies to control the dynamic behaviors of whole system, and understand the PV power system overall [32].

Causal Loop Diagram
Our SD model will involve main power generation and consumption sections, and the process of PV power investment, generation and transmission.As the first step of building SD model, a causal loop diagram of China's PV power development under the influence of air-pollution control policies is made to analyze the relationship of key variables and mechanism of air-pollution control policies, as shown in Figure 3. Positive links, which means the forward elements stimulate the

Methodology
China's PV power development with and without air-pollution control policy can be predicted by the same SD model with different parameter settings.As a systems modeling and dynamic simulation methodology, SD can analyze dynamic complexities in socio-economic and bio-physical systems [30,31].SD model can better simulate China's PV power development than other types of quantitative modeling, because of its advantage of analyzing complex relationships among the variables involved in energy development system, and simulating the feedback effects of backward variables on forward variables.It also enables us to formulate relevant policies to control the dynamic behaviors of whole system, and understand the PV power system overall [32].

Causal Loop Diagram
Our SD model will involve main power generation and consumption sections, and the process of PV power investment, generation and transmission.As the first step of building SD model, a causal loop diagram of China's PV power development under the influence of air-pollution control policies is made to analyze the relationship of key variables and mechanism of air-pollution control policies, as shown in Figure 3. Positive links, which means the forward elements stimulate the backward elements, are marked with plus sign.In addition, negative links, which means the forward elements restrain the backward elements, are marked with minus sign.Positive loops, which have an even number of negative links, are marked with plus sign.Negative loops, which have an odd number of negative links, are marked with minus sign.
Energies 2016, 9, 336 5 of 23 backward elements, are marked with plus sign.In addition, negative links, which means the forward elements restrain the backward elements, are marked with minus sign.Positive loops, which have an even number of negative links, are marked with plus sign.Negative loops, which have an odd number of negative links, are marked with minus sign.

Flow Diagram
After drawing the causal loop diagram, we built the following flow diagram of China's PV power development under the influence of air-pollution control policies, as shown in Figure 4, using Vensim software.Flow diagram is a good tool for modeling the cause and effect relations between various variables of the SD model using directive arrows.This flow diagram contains three subsystems and over 70 variables including state variables (represent cumulative results, shown in boxes), rate variables (represent changing rate of state variables, shown with double triangles), and auxiliary variables (the rest relevant variables).Key variables and its initial values are provided in Table A1.There are over sixty controlling functions used to express the quantitative relationships among all the variables.Due to the limited length of the article, only the functional relationships with great significance and something unclear in the flow diagram are enumerated.

Flow Diagram
After drawing the causal loop diagram, we built the following flow diagram of China's PV power development under the influence of air-pollution control policies, as shown in Figure 4, using Vensim software.Flow diagram is a good tool for modeling the cause and effect relations between various variables of the SD model using directive arrows.This flow diagram contains three subsystems and over 70 variables including state variables (represent cumulative results, shown in boxes), rate variables (represent changing rate of state variables, shown with double triangles), and auxiliary variables (the rest relevant variables).Key variables and its initial values are provided in Table A1.There are over sixty controlling functions used to express the quantitative relationships among all the variables.Due to the limited length of the article, only the functional relationships with great significance and something unclear in the flow diagram are enumerated.

Flow Diagram
After drawing the causal loop diagram, we built the following flow diagram of China's PV power development under the influence of air-pollution control policies, as shown in Figure 4, using Vensim software.Flow diagram is a good tool for modeling the cause and effect relations between various variables of the SD model using directive arrows.This flow diagram contains three subsystems and over 70 variables including state variables (represent cumulative results, shown in boxes), rate variables (represent changing rate of state variables, shown with double triangles), and auxiliary variables (the rest relevant variables).Key variables and its initial values are provided in Table A1.There are over sixty controlling functions used to express the quantitative relationships among all the variables.Due to the limited length of the article, only the functional relationships with great significance and something unclear in the flow diagram are enumerated.

Power Generation and Consumption Subsystem
To stimulate the process of power generation and power consumption, we built the power generation and consumption subsystem, as shown in Figure 4. We assume that China's total power generation equals total power consumption.Total power consumption consists of the primary industrial electricity consumption, the second industrial electricity consumption, the third industrial electricity consumption, and residential electricity consumption.To simplify model, we assume total power generation only consists of PV power generation, thermal power generation, hydropower generation, wind power generation and nuclear power generation.
As the core of our SD model, PV power generation (GP) is determined by total electricity consumption (TC), thermal power generation (GT), hydropower generation (GH), wind power generation (GW), nuclear power generation (GN) and transmission limit factor (TF), as in Equation (1).TF, which is the major restricted factor for PV generation, is set to a piecewise function of the gap between interregional power transmit demand (ITD) and actual value of transmission capacity (IT) as in Equation ( 2).
GP " pTC ˆ0.995 ´GT ´GH ´GW ´GNq ˆTF TF " IF THEN ELSE p pITD ´ITq ă 0, 1, 1 ´pITD ´ITq { 10000q Further with the assumption that distributed PV power has the same annual utilization hours as PV power plant, total PV generation is distributed according to the proportion of distributed PV installed capacity (CPD) and PV plant installed capacity (CPP) in installed PV capacity.Thus, distributed PV power generation (GPD) is as in Equation ( 3) and PV plant power generation (GPP) is as in Equation ( 4).

Power Transmit Subsystem
To stimulate the restriction of interregional power transmission capacity on PV power generation, we built the power transmit subsystem, as shown in Figure 5.The transmission limit factor is determined by actual interregional power transmission capacity and transmission demand.Increasing rate of interregional transmission capacity is stimulated by PV power curtailment, which depends on the actual value and theoretical value of PV power generation.

Power Generation and Consumption Subsystem
To stimulate the process of power generation and power consumption, we built the power generation and consumption subsystem, as shown in Figure 4. We assume that China's total power generation equals total power consumption.Total power consumption consists of the primary industrial electricity consumption, the second industrial electricity consumption, the third industrial electricity consumption, and residential electricity consumption.To simplify model, we assume total power generation only consists of PV power generation, thermal power generation, hydropower generation, wind power generation and nuclear power generation.
As the core of our SD model, PV power generation (GP) is determined by total electricity consumption (TC), thermal power generation (GT), hydropower generation (GH), wind power generation (GW), nuclear power generation (GN) and transmission limit factor (TF), as in Equation (1).TF, which is the major restricted factor for PV generation, is set to a piecewise function of the gap between interregional power transmit demand (ITD) and actual value of transmission capacity (IT) as in Equation (2).
Further with the assumption that distributed PV power has the same annual utilization hours as PV power plant, total PV generation is distributed according to the proportion of distributed PV installed capacity (CPD) and PV plant installed capacity (CPP) in installed PV capacity.Thus, distributed PV power generation (GPD) is as in Equation ( 3) and PV plant power generation (GPP) is as in Equation ( 4).

Power Transmit Subsystem
To stimulate the restriction of interregional power transmission capacity on PV power generation, we built the power transmit subsystem, as shown in Figure 5.The transmission limit factor is determined by actual interregional power transmission capacity and transmission demand.Increasing rate of interregional transmission capacity is stimulated by PV power curtailment, which depends on the actual value and theoretical value of PV power generation.As shown in Equation ( 5), the value of transmission capacity (IT) at time t depends on its value at time t-dt and newly-added interregional transmission capacity (ITN), which is the product of IT and its increasing rate (ITI).The increasing rate of interregional transmission capacity (ITI) is the As shown in Equation ( 5), the value of transmission capacity (IT) at time t depends on its value at time t-dt and newly-added interregional transmission capacity (ITN), which is the product of IT and its increasing rate (ITI).The increasing rate of interregional transmission capacity (ITI) is the product Energies 2016, 9, 336 7 of 23 of fundamental increasing rate (ITIF) and stimulating factor of PV power curtailment (CPF), which is a piecewise function of PV power curtailment (PC).PC is set as an IF function of theoretical wind power generation (GPT) and actual generation (GP) as in Equation (6).

Installed PV Capacity Subsystem
To stimulate the development situation of China's PV power, we built the installed PV capacity subsystem, as shown in Figure 6.At present, China has price subsidy for full generation of distributed PV power, which can be feed back to grid after self-consumption at coal-fired benchmark price.PV power plants can sale all electricity to grid at PV benchmark price.These electricity price and price subsidies determine the revenue of PV power industry, which will be partly used to make investment in new installed capacity.The investment is also influenced by the gap between installed PV capacity target and actual value.Therefore, the government can control installed PV capacity by adjusting capacity target.Newly-added installed PV power capacity is determined by its investment and construction cost per capacity.We assume that the utilization hours of distributed PV power and PV plant are equal.Thus, the composition of installed PV capacity decided the composition of PV power generation.

Model Validation
The validity of the model is tested using an authenticity test and a sensitivity test when the model design is finished.This paper used an authenticity test to justify the appropriateness of the proposed model by analyzing the error rate of eight variables between true value and simulation in 2012-2014.These variables cover all level variables and some key variables, which have overall situation and representativeness.Authenticity test results are provided in Table A2.It  We can calculate the revenue of distributed PV power (RPD) as in Equation ( 7), and calculate the revenue of PV power plant (RPP) as Equation (8).Because distributed PV power has electricity price subsidy (PS) for the whole generation, and its excessive generation can be fed back to grid after self-consumption at desulfurating electricity price (PS).PV power plants can only get revenue (RPP) from selling all their generation (GPP) to grid at PV benchmark price (PB).Investment of distributed PV power (IPD) depends on RPD, discretionary investment proportion (IPDP), stimulating factor of policy (PFD) and proportion of investment subsidy (ISP) as in Equation ( 9).In addition, investment of PV power plant (IPP) is very similar as in Equation (10).Distributed PV installed capacity (CPD) at time t depends on CPD at time t-dt and newly-added capacity (CPDN), which determined by IPD and building costs per unit (CPD).The capacity of PV power plant (CPP) has very similar function as in Equation ( 12).

Model Validation
The validity of the model is tested using an authenticity test and a sensitivity test when the model design is finished.This paper used an authenticity test to justify the appropriateness of the proposed model by analyzing the error rate of eight variables between true value and simulation in 2012-2014.These variables cover all level variables and some key variables, which have overall situation and representativeness.Authenticity test results are provided in Table A2.It is worth mentioning that the true values of interregional transmission capacity are estimated by the development plan of China's State Grid Corporation, which may underestimate the actual development situation.It may lead to an error rate of 7.80% in 2014.However, the errors of other variables are almost all controlled within ´5%~5%, which completely meet the requirement of authenticity test.In addition, the simulation period has significance on behalf of those typical periods of China's PV power development under the influence of 'air-pollution control policy'.Thus, our SD model is valid, and can be used to predict China's PV power development under different policy environments, if there are no structural changes in the future.
The sensitivity test is conducted to estimate the policy effects on PV power.PV power generation, installed PV capacity, PV power curtailment and the proportion of distributed PV installed capacity are chosen as studied variables, because they can basically reflect the development and utilization situation of PV power.In addition, ten parameters, which can simulate the effect of air-pollution control policies, are chosen as studied parameters.The sensitivities of four variables to ten parameters are shown in Table A3.The test results indicate that the proportion of coal consumption in total energy consumption, the increasing rate of interregional transmission capacity, interregional power transmit proportion and the proportion of the third industrial output value have a greater effect on PV power generation.The target of installed PV capacity, average benchmark price of PV power plant, and electricity price subsidy of distributed PV power have a greater effect on installed PV capacity.Average utilization hours, the target of distributed PV installed capacity, average benchmark price of PV power plant and increasing rate of interregional transmission capacity have a greater effect on PV power curtailment.The target of installed capacity, electricity price subsidy of distributed PV power and the increasing rate of interregional transmission capacity have a greater effect on the proportion of distributed PV installed capacity.Thus, these significant effect factors need to be further studied in policy scenario analysis.

Policy Scenarios
Air-pollution control policies can be divided into fundamental policies (The Action Plan for the Control of Air Pollution), which indirectly affect China's PV industry; and supporting policies (Energy Industry's Action Plan for Strengthening Air Pollution Control, Interim Regulations for Reducing and Replacing Coal Consumption, Some Suggestions to Improving PV Industry Development), which affect China's PV industry more directly.The policy scenario with no 'air-pollution control policy' is set as Scenario 1, the scenario with only fundamental policy is set as Scenario 2, and the scenario with both fundamental and supporting policies is set as Scenario 3. The specific policy measures and policy objectives in three policy scenarios are shown in Table 1.These policy measures in different scenarios are used to set model parameters, which can simulate the control of air-pollution control policies on system behavior, as provided in Table A4.In addition, the policy objectives are set as important evaluating indexes of China's PV power development.Thus, the impact of air-pollution control policies can be assessed by comparing the predictions of evaluating indexes in different scenarios.To explain, most of the parameter settings vary every year.Our paper does not list the parameter settings of each year but the settings for time-invariant parameters and the settings for time-varying parameters in 2017 (some also in 2020) because of space cause.Year 2017 and 2020 are important years, because they are always highlighted in relevant policies.The parameter values of other years are evenly distributed on the trend-line determined by initial values and values in 2017 (some also in 2020).

Simulation Results and Analysis
First in this section, China's PV power development during 2015-2025 in different policy scenarios are simulated by the SD model we built above using Vensim software.Later, air-pollution control policy's effect on China's PV power is analyzed by comparing the differences between simulated results.

Scenario 1: With no Policy Influence
In Scenario 1, China's PV power industry will develop without any influence of air-pollution control policies.In this scenario, China will have excessive energy demand because of unreasonable industrial structure and high energy intensity.Clean power generation could not develop as expected due to the over-reliance on thermal power.In 2020, China's PV power generation will be only 153.22 billion kWh accounting for 2.09% of total power generation.PV power curtailment stays at a high level during the whole simulation period.It will peak in 2025 at 25.60% because of the lack of transmission channel.At that time, China's PV power generation will reach 220.19 billion kWh accounting for 2.59% of the total power generation.Installed PV capacity will have an accelerated growth during 2015-2020, and reach 109,855 MW in 2020.After that, the increasing rate will become steady, and installed PV capacity will reach 227,658 MW in 2025.
Without enough finance support, distributed PV power, which supposed to be the main developing trend of PV power internationally, will develop slowly in China due to high costs.In 2020 distributed PV installed capacity will reach 40,872 MW, which just meet the policy target, accounting for only 37.21% of total installed PV capacity.In addition, the generation of distributed PV power will be 96.21 billion kWh at that time.Although distributed PV power will develop steadily during the 13th five year plan period, its installed capacity will be only 88,836 MW, which lag behind the policy target (90,000 MW) in 2025.In addition, because of the relatively fast growth of PV power plant, the proportion of distributed PV power in PV power will generally remain at about 39% since 2020.The simulated results of all state variables and some important auxiliary variables in Scenario 1 are shown in Table A5.
In general, without air-pollution control policies, China can hardly get a reasonable energy structure, and PV power can hardly receive good development.Over-reliance on thermal power, high costs and outmoded transmission grid make China's PV power (especially distributed PV power) could not meet the development targets, which is quite low.

Scenario 2: With Influence of Fundamental Policy
In Scenario 2, China's PV power industry will develop with the influence of fundamental air-pollution control policy.Stimulated by this policy, China's industrial structure will be more reasonable, and energy intensity will be lower as well.Further, this policy will reduce the proportion of coal consumption in total energy consumption to 65%.In consequence, thermal power generation will be about 10% less than that in Scenario 1. Non-fossil energy including PV power will get steady growth.In 2020 China's PV power generation will be 260.89 billion kWh accounting for 3.72% of total power generation.After reaching its peak at 294.11 billion kWh in 2023, it will start to decline slowly because of the more rapid rise of nuclear and other clean power.In addition, without supporting policies, the proportion of PV power generation will start to drop from the peak in 2017 (4.06%), and will probably drop to the same level in Scenario 1.
Installed PV capacity will reach 131,229 MW in 2020, and reach 266,141 MW in 2025, which is not much higher than the simulations in Scenario 1.In order to relief the pressure of PV power curtailment, fundamental air-pollution control policy plans to enhance the construction of interregional transmission lines.However, without the supporting polices, which make plans for transmission lines to realize large-scale interregional electricity transmission, PV power curtailment will still stay at a high level in Scenario 2. At the end of our simulation, PV power curtailment will reach up to 29.28% with a rising trend.
In Scenario 2, the increase of generation will stimulate the investments of both distributed PV power and PV power plant.Although benchmark price and price subsidy are still not enough, the investment enthusiasm of PV power will be much higher than that in Scenario 1.In addition, the increasing rate of distributed PV installed capacity will be accelerated since 2015.In 2020, distributed PV installed capacity will be 67,262 MW, which exceeds the policy target (40,000 MW), accounting for 51.26% of total installed PV capacity.However, its increasing rate will slow down since 2021 as the end of electricity price subsidy of distributed PV power.Different from distributed PV power, the increasing rate of PV plant installed capacity will maintain at about 20%, because the benchmark price of PV power plant will remain stable.Thus, the proportion of distributed PV power will begin to drop since 2022.By 2025, distributed PV installed capacity will reach 121,293 MW, accounting for 45.57% of total installed PV capacity.At that time, distributed PV power generation will reach 128.31 billion kWh, and PV plant power generation will reach 147.85 billion kWh.The simulated results of all state variables and some important auxiliary variables in Scenario 2 are shown in Table A6.
In general, with fundamental air-pollution control policy, China can have a more reasonable energy structure.However, without relevant supporting policies, PV power will have lower competitiveness compared with other clean powers, which have obvious cost advantage.In this policy scenario, both distributed PV power and PV power plant will overly dependent on subsidies.Thus, the development of distributed PV power will slow down in 2021 as the end of electricity price subsidy.China's PV power still needs long-term subsidy unless technical and power transmission problem can be overcome.However, the subsidy will undoubtedly bring huge financial pressure to Chinese government.

Scenario 3: With Influence of Fundamental and Supporting Policies
In Scenario 3, China's PV power industry will develop with influence of both fundamental and supporting air-pollution control policies.In this scenario, China's industrial structure and energy intensity are all control in reasonable range, just as Scenario 2. Stimulated by supporting air-pollution control policies, energy consumption and the proportion of coal consumption will be further reduced.With this background, the increasing rate of thermal power will be controlled within 1%.PV power generation will keep increasing at an annual rate of over 10% during 2015-2020 in Scenario 3. In 2020, China's PV power generation will reach 301.00 billion kWh accounting for 4.30% of total power generation.After that, the increasing rate of PV generation will begin to decline, and the proportion of PV generation will maintain at about 4.5%.On the other hand, supporting air-pollution control policies will keep the increasing rate of interregional transmission capacity at about 10%.Thus, China's PV power generation can gradually break the limitation of power transmission.According to the simulations, the generation will keep increasing at a low speed during 2021-2025, and reach 347.91 billion kWh accounting for 4.41% of total power generation in 2025.PV power curtailment will continue increasing during 2015-2020, because of the lag of electricity grid construction behind the increasing demand of interregional transmission.China's PV power curtailment will reach its peak at 25.67% in 2020.As electricity grid becomes more mature, China's PV power curtailment will start to drop since 2021.The curtailment will be only 8.21% in 2025, and it will probably be eliminated before 2030.
In Scenario 3, supporting policies increase the benchmark price of PV power plant and electricity price subsidy of distributed PV power, which makes installed PV capacity higher than that in Scenario 1 and 2. China's Installed PV capacity will start to grow rapidly in 2015, and will reach 247,223 MW in 2020.However, along with the changes of policy environment, the increasing rate of installed PV capacity will fail suddenly and maintain at about 8% since 2021.At the end of our simulation, China's installed PV capacity will reach 414,925 MW.
China's distributed PV power will experience a revenue surge during 2016-2020, after that its revenue will return to pre-2015 levels and maintain at around 25 billion CNY as the stop of electricity price subsidy.Its investment has the same trend as revenue.Distributed PV installed capacity will experience accelerated growth during 2015-2020 because of investment surge.It will reach 143,343 MW accounting for 57.98% of total installed PV capacity in 2020.In this stage, distributed PV power will mostly benefit from subsidy not power generation, and become the mainstream of distributed PV.Compared with distributed PV power, PV power plant will develop more steadily.During 2015-2020, its generation will be restricted by insufficient transmission grids.In addition, its capacity will reach 103,880 MW just exceeding the policy target (90,000 MW) in 2020.Since 2021, as transmission grid become more mature, PV power plant will have greater development potential.Thus, the proportion of distributed PV installed capacity will begin to fall slightly.The simulated results of all state variables and some important auxiliary variables in Scenario 3 are shown in Table A7.
In general, with both fundamental and supporting air-pollution control policies, China's PV generation will keep increasing during the whole simulation.Its proportion in total power generation will rise to 4.5% in 2021, and maintain at that level.PV power curtailment will first increase, then decrease, and probably be eliminated before 2030.In this scenario, quite a high subsidy will bring investment and capacity surge to distributed PV power, which makes it the mainstream of China's PV power.In addition, developing distributed PV power is a good solution to deal with the problem of lagging transmission grids before 2021.In Scenario 3, the transmission problem will be significantly relieved since 2021.Therefore, that it is the proper time to end up the electricity price subsidy of distributed PV power and make the benchmark price of PV power plant on the same level as thermal power.

Analysis of Policy Effect
Based on the simulations, we analyze the impact of air-pollution control policy on China's PV power by drawing curves of the same variable in three scenarios in one graph (see Figures 7-16), and comparing the differences between them.We compared all variables presented in Tables A5-7 except for the investment of distributed PV power and PV plant, because they are not set as policy objectives, and can be reflected in installed capacity.The comparison results of PV power generation and its proportion, distributed PV power generation and PV plant power generation reflect the fulfillment of generation objectives mentioned in air-pollution control policy.The comparison results of PV power curtailment and interregional transmission capacity reflect the fulfillment of power transmission objectives.The comparison results of installed PV capacity, distributed PV installed capacity and PV plant installed capacity reflects the fulfillment of capacity objectives.In addition, the proportion of distributed PV capacity reflects the composition of PV power.In general, these ten variables cover almost all of the evaluate indexes of China's PV power development.
Scenario 3 are shown in Table A7.
In general, with both fundamental and supporting air-pollution control policies, China's PV generation will keep increasing during the whole simulation.Its proportion in total power generation will rise to 4.5% in 2021, and maintain at that level.PV power curtailment will first increase, then decrease, and probably be eliminated before 2030.In this scenario, quite a high subsidy will bring investment and capacity surge to distributed PV power, which makes it the mainstream of China's PV power.In addition, developing distributed PV power is a good solution to deal with the problem of lagging transmission grids before 2021.In Scenario 3, the transmission problem will be significantly relieved since 2021.Therefore, that it is the proper time to end up the electricity price subsidy of distributed PV power and make the benchmark price of PV power plant on the same level as thermal power.

Analysis of Policy Effect
Based on the simulations, we analyze the impact of air-pollution control policy on China's PV power by drawing curves of the same variable in three scenarios in one graph (see Figures 7-16), and comparing the differences between them.We compared all variables presented in Table A5-A7, except for the investment of distributed PV power and PV plant, because they are not set as policy objectives, and can be reflected in installed capacity.The comparison results of PV power generation and its proportion, distributed PV power generation and PV plant power generation reflect the fulfillment of generation objectives mentioned in air-pollution control policy.The comparison results of PV power curtailment and interregional transmission capacity reflect the fulfillment of power transmission objectives.The comparison results of installed PV capacity, distributed PV installed capacity and PV plant installed capacity reflects the fulfillment of capacity objectives.In addition, the proportion of distributed PV capacity reflects the composition of PV power.In general, these ten variables cover almost all of the evaluate indexes of China's PV power development.PV power generation : Scenario 1 PV power generation : Scenario 3  proportion of PV power generation : Scenario 1  proportion of PV power generation : Scenario 1 proportion of PV power generation : Scenario 2 proportion of PV power generation : Scenario 3 3 PV power curtailment : Scenario 1 PV power curtailment : Scenario 2 PV power curtailment : Scenario 3 3 interregional transmission capacity : Scenario 1 interregional transmission capacity : Scenario 2 interregional transmission capacity : Scenario 3 3  destributed PV power generation : Scenario 1 destributed PV power generation : Scenario 2 destributed PV power generation : Scenario 3 3   destributed PV power generation : Scenario 1 destributed PV power generation : Scenario 2 destributed PV power generation : Scenario 3 3 PV plant power generation : Scenario 1 PV plant power generation : Scenario 3 3 proportion of distributed PV capacity : Scenario 2 proportion of distributed PV capacity : Scenario 3    destributed PV power generation : Scenario 1 destributed PV power generation : Scenario 2 destributed PV power generation : Scenario 3 3 PV plant power generation : Scenario 1 PV plant power generation : Scenario 3 3 proportion of distributed PV capacity : Scenario 2 proportion of distributed PV capacity : Scenario 3  installed PV capacity : Scenario 1 installed PV capacity : Scenario 3 installed distributed PV capacity : Scenario 1    The differences between the curves of Scenario 1 and 2 indicate the impact of fundamental air-pollution control policy ('fundamental policy' for short) on China's PV industry.The differences between the curves of Scenario 1 and 3 indicate the impact of both fundamental and supporting air-pollution control policies ('supporting policies' for short).The differences between the curves of Scenario 2 and 3 indicate the impact of only supporting policies.We can tell from Figures 7-16 that: (1) Fundamental policy will obviously increase the generation of all clean energies including PV power by limiting emissions.However, the proportion of PV generation in Scenario 2 will eventually drop to nearly the same level in Scenario 1, because of the lack of special supports for PV power.installed PV capacity : Scenario 1 installed PV capacity : Scenario 3 installed distributed PV capacity : Scenario 1 installed distributed PV capacity : Scenario 3 installed PV plant capacity : Scenario 1 installed PV plant capacity : Scenario 3    The differences between the curves of Scenario 1 and 2 indicate the impact of fundamental air-pollution control policy ('fundamental policy' for short) on China's PV industry.The differences between the curves of Scenario 1 and 3 indicate the impact of both fundamental and supporting air-pollution control policies ('supporting policies' for short).The differences between the curves of Scenario 2 and 3 indicate the impact of only supporting policies.We can tell from Figures 7-16 that: (1) Fundamental policy will obviously increase the generation of all clean energies including PV power by limiting emissions.However, the proportion of PV generation in Scenario 2 will eventually drop to nearly the same level in Scenario 1, because of the lack of special supports for PV power.installed PV capacity : Scenario 1 installed PV capacity : Scenario 3 installed distributed PV capacity : Scenario 1 installed distributed PV capacity : Scenario 3 installed PV plant capacity : Scenario 1 installed PV plant capacity : Scenario 3 The differences between the curves of Scenario 1 and 2 indicate the impact of fundamental air-pollution control policy ('fundamental policy' for short) on China's PV industry.The differences between the curves of Scenario 1 and 3 indicate the impact of both fundamental and supporting air-pollution control policies ('supporting policies' for short).The differences between the curves of Scenario 2 and 3 indicate the impact of only supporting policies.We can tell from Figures 7-16 that: (1) Fundamental policy will obviously increase the generation of all clean energies including PV power by limiting emissions.However, the proportion of PV generation in Scenario 2 will eventually drop to nearly the same level in Scenario 1, because of the lack of special supports for PV power.Supporting policies can make PV power generation and its proportion remain stable.However these policies cannot make generation continue increasing, because wind power will still have more mature technology and more competitiveness unless new technological revolutions occur in PV power.
(2) Fundamental policy will make PV power curtailment more severe because of the lack of control measures.Supporting policies will firstly aggravate PV power curtailment during 2015-2020 because of the surge of installed PV capacity.While after that supporting policies can reduce the curtailment sharply and eventually eliminate it by accelerating the construction of transmission grid.
(3) Fundamental policy can promote the development of China's PV power temporarily, but cannot maintain its growth trend.When China's transmission grid stays immature, the supporting policies will keep PV power generation increasing by promoting distributed PV power, which does not need to be transmitted over distance.However, since 2021, transmission problem will no longer be an important factor to limit power generation of PV plant, and the saturation of architecture resources Energies 2016, 9, 336 16 of 23 will become the restriction of distributed PV power.Thus, from that time PV power plant will be the new development priority, and the proportion of distributed PV power will start to drop slightly.
(4) Fundamental policy of China's PV power only involves subsidy mechanism and capacity target, but no suitable development planning.Therefore, the development of PV power will run out of steam since 2021.Under the influence of supporting policies, the installed capacity of both distributed PV power and PV power plant will be significantly raised by high subsidy and benchmark price.As the end of subsidy, the generation and capacity increasing rate of distributed PV power will begin to stabilize since 2021.If the subsidy won't end at this appropriate time, distributed PV power will face the problem of over-capacity.Since 2020, long-distance transmission and high costs won't be the limiting factors for PV power plant.At that time, protective benchmark price of PV power plant cannot lead to increase in generation but only over-capacity.

Conclusions
Under the stimulations of air-pollution control policies, China's PV power will develop better.It is of great significance to predict the developing trend of China's PV power, and identify whether these policies can work as expected.Hence, our paper offers a new platform for simulating China's PV power in different policy scenarios during 2015-2025.The simulations indicate that: (1) In the scenario of no air-pollution control policies, China's PV power could not develop satisfactorily due to the lack of essential incentives.PV power generation will only account for less than 3% of total power generation in 2025; (2) In the scenario of only fundamental air-pollution control policies, thermal power generation will be restricted.Although PV power will enjoy rapid grow during 2015-2020, its developing room will be gradually crowded out by other clean powers since 2021; (3) In the scenario of fundamental and supporting policies, PV power has a clear development plan, which ensures its continued growth.Distributed PV power will be energetically supported when transmission network are still immature.Since 2021, big PV plant will be given priority over distributed PV power when distributed architecture resources become saturated.The proportion of PV power generation will maintain at about 4.5%.The differences between simulations of three policy scenarios indicate air-pollution control policy's impact on China's PV power.Our paper provides a SD model, which can simulate the comprehensive relationships among various factors, for predicting the development of China's PV power.In addition, the results of sensitivity test and scenario analysis indicate the influence of single air-pollution control policy measure and a set of policies.Further, the assessments of policy efficiency provide reference for policy-making institutions.

Figure 2 .
Figure 2. Air-pollution control measures and its effects on PV industry.

Figure 4 .
Figure 4. Flow diagram for power generation and consumption subsystem.

Figure 4 .
Figure 4. Flow diagram for power generation and consumption subsystem.

Figure 4 .
Figure 4. Flow diagram for power generation and consumption subsystem.

Figure 5 .
Figure 5. Flow diagram for power transmit subsystem.

Figure 5 .
Figure 5. Flow diagram for power transmit subsystem.

Figure 6 .
Figure 6.Flow diagram for installed PV capacity subsystem.
,050 MW accounting for less than 1% of the total Industrial Level The Action Plan for Air Pollution Control Figure 1.Policy Framework.
Energies 2016, 9, 336 4 of 23 mentioned above, to simulate the development of China's PV industry in different policy scenarios.Further, comparing the differences between simulation results in different policy scenarios can help us assess the effect of air-pollution control policy on China's PV industry.

Table 1 .
Policy measures and objectives in different scenarios.

Table A4 .
Parameter settings for different policy scenarios.

Table A5 .
Simulated results of key variables in Scenario 1.

Table A6 .
Simulated results of key variables in Scenario 2.

Table A7 .
Simulated results of key variables in Scenario 3.