Multiobjective Synergistic Scheduling Optimization Model for Wind Power and Plug-In Hybrid Electric Vehicles under Different Grid-Connected Modes
In order to promote grid’s wind power absorptive capacity and to overcome the adverse impacts of wind power on the stable operation of power system, this paper establishes benefit contrastive analysis models of wind power and plug-in hybrid electric vehicles (PHEVs) under the optimization goal of minimum coal consumption and pollutant emission considering multigrid connected modes. Then, a two-step adaptive solving algorithm is put forward to get the optimal system operation scheme with the highest membership degree based on the improved constraints method and fuzzy decision theory. Thirdly, the IEEE36 nodes 10-unit system is used as the simulation system. Finally, the sensitive analysis for PHEV’s grid connected number is made. The result shows the proposed algorithm is feasible and effective to solve the model. PHEV’s grid connection could achieve load shifting effect and promote wind power grid connection. Especially, the optimization goals reach the optimum in fully optimal charging mode. As PHEV’s number increases, both abandoned wind and thermal power generation cost would decrease and the peak and valley difference of load curve would gradually be reduced.
With the promotion of the national energy conservation policy, implementing energy-saving generation scheduling in power system has been already put in the process. This means renewable energies are facing excellent development opportunities. With wind power’s scale increasing, in 2013, wind power newly increased installed capacity was 16093.3 MW and the gross installed capacity reached 92038.49 MW. Both of the two indicator values rank first in the world. Currently, China is planning to build 8 million level wind power bases. The gross installed capacity of wind power will reach one hundred million in 2020 MW .
However, because of wind power output volatility, power sources away from the load centers, uncoordinated grid construction, and other factors, wind power absorptive problem has become the bottleneck of wind power in China. According to the statistical data released by The National Energy Board, in 2012, the total abandoned wind power reaches 20TWh in China. The average equivalent utilization hours of wind power units are 1890 h and less than that in 2011. In Mongolia East, Jilin, Qinghai, and other wind resource-rich regions, the average equivalent utilization hours of wind power units are 1400 h and the abandoned wind power problem is very serious. To achieve the target, that renewable energy proportion in primary energy reaches 15% , abandoned wind problem should be solved mainly by improving the wind power absorptive capacity and efficiency of the grid.
There are already many researches focused on improving the wind power absorptive problem. Optimization methods are mainly aimed at optimizing wind power’s backup service. They could be divided into four categories, namely, wind power-AGC collaborative grid connection [3–6], wind power and adjustable units collaborative grid connection [4–7], wind power and energy storage devices collaborative grid connection [8, 9], and wind power and demand side response collaborative grid connection [10–13]. Although all the optimization methods could bring some benefits, they also have certain deficiencies. For example, the adjustment capability of backup service is weak, and backup service cost is too high [14, 15], which promote power system to search for new backup service for wind power.
Plug-in hybrid electric vehicles (PHEVs) have chargeable and dischargeable capacity, which makes PHEV have the potential to be the backup service for wind power grid connection. PHEV’s orderly grid connection can bring multiple benefits such as achieving energy conservation by using electricity rather than oil, smoothing the load curve by controlling charging and discharging behavior, and promoting wind power grid connection. However, if PHEV’s grid connection behavior is not under control, its arbitrary charging and discharging behavior would exacerbate the volatility of system load, which would decrease wind power consumption capacity. To solve these problems, wind power-PHEV synergistic scheduling optimization method needs to be studied.
Researches on wind power-PHEV synergistic scheduling optimization are mainly focused on two aspects, namely, model building and solution algorithm. As for model building, literature [16–18] established PEHV-thermal power synergistic scheduling optimization model and taxonomic studies on 4 kinds of charging modes’ influence on the optimal combination of thermal power. In those literatures, there are only thermal units in the assumed system. Literature [19, 20], respectively, established PHEV-wind power synergistic scheduling stochastic optimization model. Literature  put forward a PHEV-wind power synergistic scheduling optimization model in multiple time scales. And using the model, the literature studied wind power-PHEV synergistic scheduling’s influence on the grid under free-charging mode. However, literature [19–21] assumes that PHEVs are in the free-charging mode, ignoring the elasticity relationships between user’s electricity consumption behavior and electricity price.
In terms of solution algorithm, minimizing pollution emission becomes another optimization objective of wind power-PHEV synergistic scheduling under energy conservation requirements. This makes the optimization problems become multiobjective problems. Multiobjective problems need suitable solution algorithms to get the optimal solution sets. Literature [22, 23] and literature , respectively, use POS, the simulated annealing, differential evolution, and other intelligent algorithms to solve unit commitment problems in wind power-PHEV synergistic scheduling. Conversional solution algorithms generally have difficulties in determining parameters and transforming constraints, which make their optimization levels not high [25, 26]. Intelligent algorithms can avoid these problems. But when individual extremes do not meet the multiobjective planning principles, intelligent algorithms are easy to fall into the local extreme point, limiting the search abilities [27–29].
Based on the analysis above, to overcome the deficiencies of the study on wind power-PHEV synergistic scheduling, this paper mainly contains 7 sections. Section 2 analyzed PHEV’s charging and discharging characteristics and put forward the PHEV charging and discharging power output model. Section 3 simulates the uncertainty of wind power and constructs wind power scenarios simulation and scenario reduction strategy. Section 4 considered system demand and supply balance and units’ output and backup service constraints, established wind power-PHEV synergistic scheduling models, respectively, and set the minimum coal cost and pollution emission cost as the optimization objectives during power generation. Section 5 puts forward a two-step solution algorithm based on the improved constraints and fuzzy decision theory. Section 6 introduced the solving progress of the solution algorithm. Section 7 used the improved IEEE36 points system as the simulation system to analyze the optimization results in different charging and discharging modes and then made a sensitive analysis for the grid connection PHEV’s number.
2. PHEV Charging and Discharging Models
2.1. Charging Modes
In terms of charging model, there are some research results. PHEV’s charging modes can be divided from two aspects, namely, ordered or disordered charging, distinguish time periods, or continue charging. In this way, its charging modes can be classified into 3 kinds as follows: the no-control charging mode, the continuous charging mode, and the delayed charging mode . We can compare these three charging modes with the fully optimized charging mode. Their charging load curves are as shown in Figures 1, 2, and 3. Their characteristics can be described as follows.
2.2. Discharging Modes
PHEV’s discharging mode is influenced by battery type, capacity, related parameters, discharging cycle, discharging loop power, and other factors. Because it is hardly to decide discharging parameters, this paper uses existing research results  as data basis of discharging load distribution.
PHEV is used as a kind of backup service for wind power grid connection, in a certain extent; it is determined by controlling and estimating PHEV’s discharging capacity. So the uncertainty of PHEV’s discharging time and number brings considerable difficulties to accurately estimate the discharge capacity. Literature  has already studied PHEV’s battery type, capacity, discharge cycle, discharge power, and other parameters’ influence on PHEV’s discharging behavior. This paper uses its result to make further study. Setting the maximum discharge capacity to be 9.6 kW, PHEV’s discharging time and capacity 11 curve are shown in Figure 4. In the figure, almost all PHEVs can finish their discharging work in 6 hours.
2.3. PHEV Charging and Discharging Power Model
PHEVs need storage battery as their energy storage component to achieve charging and discharging behavior. Load change of storage battery can be tracked by energy storage controller . When it comes to the maximum capacity of the storage battery, the energy storage controller would stop storage battery from charging to ensure the battery life and operation safety. And when it comes to the minimum capacity, the energy storage controller would stop storage battery from discharging. Consider where and are the upper and lower limit of the storage battery capacity; is the storage electricity of PHEV at time .(1)When PHEV is in the charging status, (2)When PHEV is in the discharging status, where and are the storage electricity at time of PHEV and PHEV ; is the self-consumed electricity during its discharging period of PHEV ; is the self-consumed electricity during its charging period of PHEV ; and are discharging and charging power at time of PHEV and PHEV .
Generally, PHEV’s charging or discharging power should not exceed 20% of the maximum capacity of the storage battery . The constraints can be described as
3. Simulate the Uncertainty of Wind Power
Wind power output is limited by the income-wind velocity. However, if the income-wind velocity is lower than the cut-in wind velocity or is higher than the cut-out wind velocity, the wind farm will not generate power. The relationship between output power and the income-wind velocity could be expressed as where is the actual available output of wind power unit at time ; is the rated output power of wind power units; is the cut-in speed; is the cut-out speed; is the rated wind speed; is the actual speed at time .
3.1. Wind Power Scenarios Simulation
In the actual scheduling progress, wind power output can be determined through two ways, namely, forecast and simulation. In terms of wind power forecast method, there are classical forecast method [32, 33], modern forecast method [34–36], and intelligent forecast method [37–39]. The development of intelligent forecast method, especially, largely promoted wind power forecast technology. However, the practical applications show that the forecasting accuracy of these methods is still not able to reach the requirements of system scheduling. Therefore, this paper does not forecast wind farm’s output power but uses scenario analysis method to simulate each power generation scenario. Assume the random variable of wind power output is ; its composition is detailed as follows: where is the forecast power of wind power; is the forecast error of wind power forecast. The forecast error is assumed to follow the normal distribution ; then could be though to follows the normal distribution .
In order to gain the wind power scenarios, interval method is used to simulate wind power output. Divide wind power output into several intervals. Set the value of a point in the interval as wind power output expectation. When the number of intervals is sufficient, the forecast value could be regarded as the real output. The details are shown in Figure 1.
According to Figure 5, in each interval, there are three statues, namely, high, normal, and low. The expectation of wind power output in each statue is , where are, respectively, corresponding with three statues. The probabilities of each output statue are ; then the wind power output set in each scenario is and the scenario probability of wind power output is . Figure 6 shows the wind power output forecast simulation scene scenario of three statues.
3.2. Wind Power Scenario Reduction Strategy
The basic concept of scenario reduction is comparing a scenario with other scenarios and removing the closest one. And the bigger the scenario number, the bigger the workload of scenario reduction. To overcome this problem, this paper introduces the Kantorovich distance , and set minimize the Kantorovich distance between the initial scenario and the reduced scenario as the optimize objective.
Assume is the initial wind power scenario set and is the reduced scenario set. Then the Kantorovich distance between the initial scenario set and the reduced scenario set can be defined as
Then the Kantorovich distance would be
Define to be equal to the summation of the scenario’s occurrence probability in the initial scenario and the closest deleted scenario. This is detailed as
Then build wind farm scenario reduction optimization method based on (7)–(12). The optimized scenario reduction method can be expressed as
According to (13), the scenario reduction mechanism would influence the reduction result directly. To set an appropriate scenario deleted number, this paper puts forward the maximum reduction strategy: where (14) is to ensure the similar degree of the reduced scenario set and the initial scenario set in the required range. The scenario reduction model consists of (7)–(14). To solve this model, this paper uses the multistage heuristic algorithm, referring to literature .
4. Wind Power-PHEV Synergistic Scheduling Optimization Model
4.1. Objective Functions
4.1.1. Minimize Power Generation Coal Consumption Cost Objective Function
For thermal units, coal consumption cost mainly includes power generation coal consumption cost and startup-shutdown coal consumption cost; the objective function is where is the quantity of power generation units; is the total optimization period and, in this paper, ; is a 0-1 variable and when it means that unit is in operation at time and means unit is shutdown at time ; is the startup-shutdown cost of unit at time ; is the output of unit at time ; is the generation coal consumption cost function of unit at time . Consider where are coal cost coefficients of unit determined by the regression of historical data of power generation. Consider where is the cold-start cost of unit ; is the hot-start cost of unit ; is the allowed minimum downtime; is the continuous downtime of unit at time ; is the cold-start time of unit ; is the summation of minimum downtime and cold-start time of unit .
4.1.2. Minimize Pollutant Emission Cost Objective Function
Pollutant emission cost of power generation can be calculated by the least squares method using historical pollutant emission data. The objective function is where is the number of pollutant types; in this paper and , respectively, means that the pollutant is CO2, SO2, and ; , , and are pollutant emission parameters of unit .
4.2.1. Load Balancing Constricts
Consider where is the number of all PHEVs; is the number of PHEVs in charging; is the load demand before PHEVs' grid connection at time ; is the number of wind power units; is the self-consumption rate of unit ; is the output of wind power at time ; is the power consumption rate of the wind farm.
4.2.2. Thermal Units Output Constraints
Thermal power output constraints include power generation upper and lower constraints, power climbing constraints, and the minimum startup and downtime constraints, as described in formulas (14)–(17). Consider where and are the upper and lower power generation limitation of thermal unit . Consider where and are the upper and lower power climbing limitation of thermal unit . Consider where is the minimum startup time of thermal unit ; is the continuous operation time of thermal unit at time . Consider where is the minimum downtime of thermal unit ; is the continuous downtime of thermal unit at time .
4.2.3. Wind Power Output Constraints
Consider wherein is the upper output limitation of wind power units at time .
4.2.4. PHEV Charging and Discharging Constraints
Consider where and are, respectively, the number of charging and discharging PHEVs and are, respectively, the maximum number of chargeable and dischargeable PHEVs.
4.2.5. System Spinning Reserve Capacity
Consider the following: During the charging period, system’s spinning reserve capacity is where is the spinning reserve capacity before wind power’s grid connection; and are, respectively, increased upper and lower limitation of spinning reserve capacity after wind power’s grid connection. During the discharging, system’s spinning reserve capacity is where and are, respectively, increments of upper and lower limitation of spinning reserve capacity after PHEV’s grid connection
4.3.1. Objective Function Linearization
To simplify the solving process, we need to do a linearization for the quadratic objective function. Here we divide unit ’s power limit into segments, so that quadratic functions can be expressed as piecewise functions. When , where ; .
4.3.2. Contracts Linearization
To facilitate solving, we need to do linearization of formula (10) and formula (11).
Initial state constraints are as follows: where is the number of operating thermal units at initial state and if there is no initial state, then ; is the number of thermal units in shutdown statue at initial state. Set as the period of operating thermal units at initial state; then
Startup and shutdown constraints are as follows:
5. Two-Step Adaptive Solving Algorithm
5.1. The Improved Constraints
When solving multiobjective models, if a solution can make all objective functions achieve optimum, we could define as an absolutely optimal solution. However, general contradictions exist between objectives and there is no absolutely optimal solution, but a set of optimal solutions, called Pareto optimal solution set .
This paper selects the improved constraints method to calculate multiobjective optimization problems. And then it uses the determination method in literature  to select feasible solutions. That means if meets and , then is a feasible solution. The constraints method has many advantages. But there are the following two big problems: this method can only be used in the range of feasible solutions, which makes optimization results easy to fall into partial optimum; results may not satisfy Pareto optimal solution’s feasibility and are nondominant. According to literature [44, 45], lexicographic optimization  and enhanced constraints method can overcome these problems .
However, the enhanced constraints method does not take different important degrees of each objective into consideration during its solving progress. This paper combines the enhanced constraints method, lexicographic optimization method, and weight coefficient to get an improved constraints method. And by defining an iteration parameter, , we can get the feasible solution for the multiobjective problem, as shown in formula (31). Consider where is the direction of objective ; equals −1 means that objective needs to be minimized, while equals +1 means that objective needs to be maximized. To avoid objectives’ scale expand problems we pull into objective functions. is the weight of objective ; is the residual variables of constraints; is the range of objective function, which is determined by the decision attribute table and its determination progress is detailed in literature . Different from conventional weighting method, the improved constraints method regards weights as the optimization variable. And the iteration parameter would update the weights during the solving progress.
5.2. Fuzzy Decision Theory
To meet decision-makers’ demand, this paper chooses fuzzy decision method to calculate membership degrees of the Pareto optimal solution set. Define membership function , which stands for the optimal degree of objective in the Pareto optimization scheme . Then the membership degree function of Pareto optimization scheme can be described as where is the weight of objective in (35); is the number of elements in Pareto solution set; and , respectively, stand for the value and membership degrees of objective and in Pareto optimization scheme . The membership functions of the optimization objective are shown in Figure 7.
6. Solving Progress
Based on the two-step adaptive solving algorithm, we can analyze different wind power-PHEV effectiveness in multigrid connected mode. Its solving steps are as follows.Input the original data of the model. The objectives and are, respectively, set as a single-objective optimization model and it could get objective values under different optimization objectives. The results form a decision attribute table of the objectives, as shown in Table 1.Based on step , objective functions and would adjust the maximum value of each objective function (, ) according to decision makers’ wishes and preferences as follows: Based on step and step , the value range of objective function can be determined. And the improved constraints method would help to solve the model and get the optimal solution sets.Based on step , formula (36) could get the Pareto solutions and membership degrees of the objective functions. Choosing the solution set with the highest membership degree can get the satisfactory solution for the optimization model. Figure 8 shows the details.
7.1. Basic Data
This paper uses IEEE36 nodes 10-unit system as a simulation base. Add in wind farms at nodes 10, 15, and 24, respectively. The installed capacities of those wind farms are 200 MW, 300 MW, and 150 MW. Power generation coal consumption and pollutant emission parameters are listed in Table 3. The emission parameters are referenced from paper . Based on wind power scenario simulation method, MATLAB is used to simulate 100 wind power output scenarios. Then the scenarios are reduced according to (7)–(14). Finally we got 20 basic scenarios and wind power output, use the average output value of wind power in 20 basic scenarios as its available output power. The details are listed in Table 2. Load demand of a typical load day is also listed in Table 2.
Currently, there are three main types of PHEV, namely, BEVs, V2G, Triple-VG2 (equivalent to three V2G), and PCEV. Compared with other types, V2G has the advantage of being rechargeable and dischargeable, which makes it have better prospects. Therefore, this paper chooses 50000 V2G cars and studies V2G cars grid connection influence on wind power consumption. Assume that their average charging power is 1.8 kW and the maximum charging power is 2.4 kW. Charging period lasts for 6 hours and total charged electricity is 10.8 kWh . The number of PHEVs in charging period is equal to that in the discharging period. PHEV’s power discharging consumption occupies 5.6% of its total electricity capacity, which means it can discharge 10.2 kWh. The discharging behavior is limited in 6 hours.
7.2. Simulation Results
The simulation has been implemented in GAMS optimization software using CPLEX 11.0 linear solver from ILOG solver. The CPU time required for solving the problem for different case studies with a VAIO E series laptop computer powered by core i3 processor and 2 GB of RAM was less than 10 s. Firstly, we verified the validity and applicability of the solving algorithm in Section 7.2.1 to ensure the reliability of the result. And then the optimization schemes for different charging modes are compared.
7.2.1. Verification of Algorithm’s Validity and Applicability
This paper sets the no-control charging mode as the basic simulation scenario and chooses improved multiobjective nondominated sorting genetic algorithm-II (NSGA-II)  to solve the wind power-PHEV synergistic scheduling optimization model, where the maximum iterations time of NSGA-II algorithm is 10000, population size is 100, genetic operations crossover probability is 0.95, and mutation probability is 0.05. Results comparing the two algorithms are shown in Table 4.
According to Table 4, compared with the NSGA-II algorithm, the algorithm put forward by this paper could get a better optimal solution with lower iteration number and solving time. The grid-connected electricity of wind power is also more than that in the NSGA-II algorithm.
Analyzing the solution of the algorithm put forward by this paper can verify its applicability of solving wind power-PHEV synergistic scheduling optimization model under multiobjective functions. Figure 9 shows charging and discharging load distribution in the no-control charging mode. PHEV’s charging time is concentrated in low load periods (1:00–5:00, 16:00-17:00, and 23:00-24:00). Its discharging time is concentrated in peak load periods (9:00–12:00 and 19:00–21:00). So the optimization result shows that PHEVs tend to charge in low load period and discharge in high load period, consistent with the expectation.
7.2.2. Comparison between Different Charging Modes
In this section, we will compare thermal units’ output structure, wind power grid connection statue, power generation cost, and pollutant emission after PHEV’s grid connection under 4 charging modes. Table 5 lists Pareto sets in different charging modes. According to Table 4, (32) is applied to get the optimal result with the highest membership degree.
(1) Thermal Units’ Output Structure. After PHEV’s grid connection, thermal units’ total output decreased. The output structure is significantly optimized. For example, 7# and 8# units no longer generate power and 6# unit only generates power in the unrestricted charging mode. And 3# and 4# units generate less while units 1# and 2# generate more. In terms of the output structure, consider the following.(1)Compared with the no-control charging mode, in the other 3 charging modes, 1# unit achieved full capacity operation and the large capacity generators’ output increased, namely, 2# and 3# units.(2)In the continuous charging mode, charge and discharge times are not limited. Therefore, PHEV’s grid connection is the maximum, which makes thermal units’ output decrease. But for units with small installed capacity, like 4# and 5# units, their output would be more than that in the delayed charging mode or the fully optimal charging mode.(3)In the fully optimal charging mode, thermal units’ output allocation is determined by their installed capacity. That means units with big installed capacity would output much more than the small ones. In this way, the output structure comes to the optimal. Table 5 shows output allocation of thermal power units in four modes.
According to Table 6, thermal units’ output before and after PHEV grid connection can be compared. The output allocation in the fully optimal charging mode is described as Figure 10 and that before PHEV grid connection is described as Figure 11. From these two figures, we can see the characteristics of the output allocation in the fully optimal charging mode as follows.(1)Thermal units’ output decreased obviously and the output curve is relatively smooth.(2)In section of the base load, unit 1# and unit 2# supply the basis load demand.(3)In section of the waist load, units 3# and 6# supply the waist load demand before PHEV grid connection and that changed to be unit 3# and unit 4# in the fully optimal charging mode.(4)In section of the peak load, units 7# and 8# supply the peak load demand before PHEV grid connection.
And that changed to be unit 5# in the fully optimal charging mode.
(2) Wind Power Grid Connection Situation. Before PHEV’s grid connection, abandoned wind was 1511.39 MW. But after PHEV’s grid connection, abandoned wind decreased obviously. It is because in delayed charging mode or fully optimal charging mode, users’ charging behaviors are influenced by the load curve. That makes load curve gentler and provide a stronger backup for wind power. In these two charging modes, the abandoned winds are, respectively, 233 MWh and 26 MWh. Wind power output in 4 charging modes is shown in Figure 12.
(3) Economic and Environmental Benefits. PHEV’s economic and environmental benefits in 4 modes are shown in Table 7. After its grid connection, both economic and environmental benefits increased. In the fully optimal charging mode, thermal units’ startup and shutdown costs decreased by 50%, coal consumption cost decreased by 14%, emission of CO2, SO2, and , respectively, decreased by 7%, 6%, and 6%, and abandoned wind decreased to 25.64 MWh. PHEV’s grid connection would bring, obviously, economic and environmental benefits. To achieve the maximum benefit, electric enterprise should take relative incentive mechanisms to guide PHEV users’ rational charging and discharging.
7.2.3. Sensitive Analysis
As government’s support to develop PHEV is increasing, PHEV’s number grows. This means the charging and discharging behaviors of PHEV would influence power system’s load curve directly. To verify the influence of PHEV’s grid connection on power system’s load curve economic benefit and environmental benefit, this paper sets 25, 50, 75, and 1000 thousand PHEVs in succession to do system simulation. The results are shown in Figures 13 and 14 and peak and valley differences of different PHEV number are listed in Table 8.
As the number of PHEVs increases, system’s load curve becomes gentler, which means more obvious load shifting effect. When PHEV number is 25 thousand, the peak load is 2700 MW and the valley load is 1700 MW. And when PHEV number is 100 thousand, the peak and valley loads changed to 2545 MW and 1830 MW. Their difference decreased by 29%.
According to Figure 14, as the number of PHEVs increases, abandoned wind decreases. When PHEV number is 100 thousand, abandoned wind is 72.5 MW, 77% less than that of 315.4 MW when PHEV number is 25 thousand. Furthermore, part of the PHEV would use wind power to charge at the valley time and discharge at the peak time. Thermal power’s generation cost decreased when the number of PHEVs increased.
To analyze PHEV’s grid connection benefits, this paper established a wind power-PHEV synergistic scheduling optimization model and put forward a two-step adaptive solving algorithm based on improved constraints theory and fuzzy decision theory. According to the simulation results, there are some conclusions as follows.(1)This paper uses the improved constraints method to get Pareto optimal solution set and fuzzy decision theory to select a satisfactory solution with the highest membership degree. This two-step adaptive algorithm can be more suitable to solve the multiobjective wind power-PHEV synergistic scheduling optimization model.(2)PHEV’s grid connection can help in load shifting, decrease abandoned wind and power generation cost, and obviously bring economic and environmental benefits. And in delayed charging pattern and fully optimal charging pattern, system would get the optimal benefits.(3)Compared with the no-control charging mode, the delayed charging mode has the maximum grid connection electricity amount, the least thermal power generation, and little abandoned wind. However, since users’ casually charging behavior, PHEV grid connection can bring big impact on system’s safe and stable operation. So the abandoned wind in this mode is higher than that in the delayed charging mode or fully optimal charging mode.(4)In the fully optimal charging mode, abandoned wind comes to be the least. That is because in this mode users would adjust their charging behavior according to the load curve. Then the load shifting effect would be increased and PHEV grid connection can help wind power grid connection, increasing economic and environment benefits.(5)Based on the conclusions above, to make full use of PHEV grid connection, we need to commence from two aspects. On one hand, use incentive policies to promote the development of electric vehicles. And on the other hand, use reasonable demand side electricity to guide users’ charging behavior, which would be an important research direction on wind power-PHEV synergistic scheduling.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
This paper is supported by the National Science Foundation of China (Grant nos. 71071053 and 71273090).
M. B. McElroy, X. Lu, C. P. Nielsen, and Y. Wang, “Potential for wind-generated electricity in China,” Science, vol. 325, no. 5946, pp. 1378–1380, 2009.View at: Publisher Site | Google Scholar
D. Yu, J. Liang, X. Han, and J. Zhao, “Profiling the regional wind power fluctuation in China,” Energy Policy, vol. 39, no. 1, pp. 299–306, 2011.View at: Publisher Site | Google Scholar
H. Banakar, L. Changling, and B. T. Ooi, “Impacts of wind power minute-to-minute variations on power system operation,” IEEE Transactions on Power Systems, vol. 23, no. 1, pp. 150–160, 2008.View at: Google Scholar
T. Ackermann, “Wind power in power systems,” in Power System Requirements for Wind Power, chapter 8, pp. 146–150, John Wiley & Sons, Chichester, UK, 2004.View at: Google Scholar
Z. Xuanyu, L. Peng, W. Hao et al., “Novel performance assessment method and corresponding AGC control model,” Proceedings of the CSEE, vol. 30, no. 6, pp. 93–97, 2010 (Chinese).View at: Google Scholar
Q. Ying and L. Zongxiang, “Wind farms active power control considering constraints of power grids,” Automation of Electric Power Systems, vol. 33, no. 22, pp. 88–93, 2009 (Chinese).View at: Google Scholar
T. Zhu and J. Yu, “Peak load dispatch modeling considering regulation between NON-AGC units and AGC units,” Automation of Electric Power Systems, vol. 31, no. 3, pp. 9–14, 2007 (Chinese).View at: Google Scholar
F. Giraud, Analysis of a utility-interactive windphotovoltaic hybrid system with battery storage using neural network [ProQuest dissertations and theses], University of Massachusetts, Lowell, Mass, USA, 1999.
B. Borowy and Z. M. Salameh, “Methodology for optimally sizing the combination of a battery bank and PV array in a wind/PV hybrid system,” IEEE Transactions on Energy Conversion, vol. 11, no. 2, pp. 367–375, 1996.View at: Google Scholar
H. Daneshi and A. K. Srivastava, “Security-constrained unit commitment with wind generation and compressed air energy storage,” IET Generation, Transmission and Distribution, vol. 6, no. 2, pp. 167–175, 2012.View at: Publisher Site | Google Scholar
S. A. Gabriel, A. J. Conejo, M. A. Plazas, and S. Balakrishnan, “Optimal price and quantity determination for retail electric power contracts,” IEEE Transactions on Power Systems, vol. 21, no. 1, pp. 180–187, 2006.View at: Publisher Site | Google Scholar
W. Qingran, X. Guohui, and Z. Lizi, “An integrated generation-consumption dispatch model with wind power,” Automation of Electric Power Systems, vol. 35, no. 5, pp. 15–18, 2011 (Chinese).View at: Google Scholar
P. S. Moura and A. T. de Almeida, “The role of demand-side management in the grid integration of wind power,” Applied Energy, vol. 87, no. 8, pp. 2581–2588, 2010.View at: Google Scholar
R. Wenjun, W. Beibei, L. Yang et al., “Customer response behavior in time-of-use priced,” Power System Technology, vol. 36, no. 7, pp. 86–93, 2012 (Chinese).View at: Google Scholar
L. Wenyuan, The Safety of the Electricity System Economyoperation-Model and Method, Chongqing University Press, Chongqing, China, 1988, (Chinese).
X. Fei, L. Xianzhang, Z. Yebiao, and et al, “A brand-new approach of connecting electrical vehicles with smart grid from vehicle-to-grid mode to battery-to-grid mode,” Power System Technology, vol. 36, no. 2, pp. 29–34, 2012 (Chinese).View at: Google Scholar
L. Lingrong, W. Fushuan, X. Yusheng et al., “Unit commitment in power system with plug-in electric vehicles,” Automation of Electric Power Systems, vol. 35, no. 21, pp. 16–20, 2011 (Chinese).View at: Google Scholar
K. Parks, P. Denholm, and T. Markel, “Cost and emissions associated with plug in hybrid electric vehicle charging in the Xcel energy Colorado service territory[R/OL],” 2007, http://www.nrel.gov/vehiclesandfuels/pdfs/41410.pdf.View at: Google Scholar
H. Zechun, S. Yonghua, X. Zhiwei et al., “Impacts and utilization of electric vehicles integration into power systems,” Proceedings of the CSEE, vol. 32, no. 4, pp. 1–11, 2012 (Chinese).View at: Google Scholar
C. Defu, Q. Bin, and C. Jinfu, “Analysis on dynamic probabilistic characteristic of power grid connected with electric vehicle charging load and wind power,” Power System Technology, vol. 37, no. 3, pp. 590–596, 2013 (Chinese).View at: Google Scholar
L. Huiling, B. Xiaomin, T. Wen et al., “Coordination control for grid-connection of plug-in hybrid electric vehicles and distributed generation,” Power System Technology, vol. 37, no. 8, pp. 2108–2115, 2013 (Chinese).View at: Google Scholar
Y. Dayang, S. Shuguang, Z. Bo et al., “Synergistic dispatch of PEVs charging and wind power in Chinese regional power grids,” Automation of Electric Power Systems, vol. 25, no. 14, pp. 24–29, 2011 (Chinese).View at: Google Scholar
H. Haiying, H. Jinghan, W. Xiaojun et al., “An improved particle swarm optimization-based load response strategy with participation of vehicle to grid,” Power System Technology, vol. 35, no. 10, pp. 165–169, 2011 (Chinese).View at: Google Scholar
H. Wang, Q. Huang, C. Zhang, and A. Xia, “A novel approach for the layout of electric vehicle charging station,” in Proceedings of the IEEE International Conference on Apperceiving Computing and Intelligence Analysis (ICACIA '10), pp. 64–70, Chengdu, China, December 2010.View at: Publisher Site | Google Scholar
S. Yue, X. Changyun, and C. Haozhong, “Reconstruction sequence algorithm for replacing overhead lines by underground cables based on priority weights method,” Power System Technology, vol. 28, no. 12, pp. 5–9, 2004 (Chinese).View at: Google Scholar
L. Suhua, Y. Xinmei, X. Xinyin et al., “Unit commitment using improved discrete particle swarm optimization algorithm,” Proceedings of the CSEE, vol. 25, no. 8, pp. 30–35, 2005 (Chinese).View at: Google Scholar
S. Bo, L. Qiangqiang, and X. Pinjie, “A cost-benefit analysis model of vehicle-to-grid for peak shaving,” Power System Technology, vol. 36, no. 10, pp. 30–34, 2012 (Chinese).View at: Google Scholar
S. Wang, The Optimization Model and Method for Comprehensive Utilization of Power Resources Under the Constraint of Emission Reduction, North China Electric Power University, Beijing, China, 2013.
S. Bisanovi, M. Hajro, and M. Dlakic, “Hydrothermal self-scheduling problem in a day-ahead electricity market,” Electric Power Systems Research, vol. 78, no. 9, pp. 1579–1596, 2008.View at: Google Scholar
K. Parks, P. Denholm, and T. Markel, “Costs and emissions associated with plug in hybrid electric vehicle charging in the Xcel energy Colorado service territory[R/OL],” http://www.nrel.gov/vehiclesandfuels/pdfs/41410.pdf.View at: Google Scholar
P. Denholm and W. Short, “An evaluation of utility system impacts ad benefits of optimally dispatched plug in hybrid electric vehicles,” http://www.nrel.gov/docs/fy07osti/40293.pdf.View at: Google Scholar
B. Ge, W. Wang, D. Bi et al., “Energy storage system-based power control for grid-connected wind power farm,” International Journal of Electrical Power & Energy Systems, vol. 44, no. 1, pp. 115–122, 2013.View at: Publisher Site | Google Scholar
J. Wu, Z. Buhan, and H. Li, “Statistical distribution for wind power forecast error and its application to determine optimal size of energy storage system,” International Journal of Electrical Power & Energy Systems, vol. 55, pp. 100–107, 2014.View at: Google Scholar
N. Dongxiao, C. Shuhua, and L. Jianchang, Power Load Forecasting Technology and Its Application, China Electric Power Press, Beijing, China, 2009.
X. Chen, C. Kang, and M. Chen, “Short term probabilistic forecasting of the magnitude and timing of extreme load,” Proceedings of the Chinese Society of Electrical Engineering, vol. 31, no. 22, pp. 64–72, 2011.View at: Google Scholar
K. Jisong and L. Minqiang, “Clusterimulti-small area forecasting and spatial resolution identification in spatial load forecast,” Journal of Systems Engineering, vol. 42, no. 2, pp. 42–53, 1989.View at: Google Scholar
J. Li, J. Liu, and J. Wang, “Mid-long term load forecasting based on simulated annealing and SVM algorithm,” Proceedings of the Chinese Society of Electrical Engineering, vol. 31, no. 16, pp. 63–66, 2011.View at: Google Scholar
Y. Feng, R. Li, A. Sudjianto, and Y. Zhang, “Robust neural network with applications to credit portfolio data analysis,” Statistics and its Interface, vol. 3, no. 4, pp. 437–444, 2010.View at: Publisher Site | Google Scholar | Zentralblatt MATH | MathSciNet
N. Lu, J. Zhou, and Y. He, “Particle swarm optimization-based neural network model for short-term load forecasting,” Power System Protection and Control, vol. 38, no. 12, pp. 65–68, 2010.View at: Google Scholar
X. Zhang, G. He, K. Liu, J. Li, H. Zhai, and H. Lü, “A coordinated economic dispatch based on lower semi-absolute deviation risk,” Automation of Electric Power Systems, vol. 36, no. 19, pp. 53–59, 2012.View at: Publisher Site | Google Scholar
N. Gröwe-Kuska, H. Heitsch, and W. Römisch, “Scenario reduction and scenario tree construction for power management problems,” in Proceedings of the IEEE Power Tech Conference, pp. 152–158, Bologna, Italy, June 2003.View at: Publisher Site | Google Scholar
Z. Jin, Two-stage Power Generation Dispatch Model and Algorithm for Power Systems Including Wind Power and Pumped Storage Station, North China Electric Power University, 2012.
D. P. Kothari and J. S. Dhillon, Power System Optimization, Prentice-hall, 2004.
G. Mavrotas and D. Diakoulaki, “Multi-criteria branch and bound: a vector maximization algorithm for mixed 0-1 multiple objective linear programming,” Applied Mathematics and Computation, vol. 171, no. 1, pp. 53–71, 2005.View at: Publisher Site | Google Scholar | MathSciNet
S. J. Wang, S. M. Shahidehpour, D. S. Kirschen et al., “Short-term generation scheduling with transmission and environmental constraints using an augmented Lagangian relaxation,” IEEE Transactions on Power Systems, vol. 10, no. 3, pp. 1294–1301, 1995.View at: Google Scholar
N. Amjady, J. Aghaei, and H. A. Shayanfar, “Stochastic multiobjective market clearing of synergistic energy and reserves auctions ensuring power system security,” IEEE Transactions on Power System, vol. 24, no. 4, pp. 1841–1854, 2009.View at: Google Scholar
GWEC, Global wind resport 2012, Global Wind Energy Council, Brussels, Belgium, 2011.