New energy station energy storage capacity configuration method based on multi-stage robust optimization

Abstract

The invention relates to a new energy station energy storage capacity configuration method based on multi-stage robust optimization. The method comprises an investment decision-making layer, a day-ahead operation scheduling strategy decision-making layer and an intra-day operation simulation layer. By establishing a multi-stage energy storage capacity planning model, new energy output uncertainty and system flexibility requirements are considered, and methods such as distributed robust optimization and cooperative game are adopted to optimize energy storage configuration. The method can effectively reduce the comprehensive cost of the system, improve the energy storage utilization rate and balance the economical efficiency and robustness of the solution. Compared with the prior art, the method has the following advantages: new energy output uncertainty and system flexibility requirements are considered more comprehensively; the optimization method is more scientific, and the accuracy and reliability of energy storage configuration can be improved; energy storage sharing and investment cost sharing are realized, and consumption of new energy and stable operation of a power system are promoted. The method has important guiding significance and application value for energy storage capacity configuration of the new energy station.

Description

Technical Field

[0001] The present invention relates to a method for configuring energy storage capacity of a new energy station based on multi-stage robust optimization. Background Art

[0002] With the global emphasis on renewable energy and its development and utilization, the proportion of new energy in the power system continues to increase. However, the inherent uncertainty and volatility of new energy have brought new challenges to the safe and stable operation of the power system. As an effective means of power regulation, energy storage systems are of great significance in alleviating the fluctuations of new energy and improving system stability.

[0003] At present, a lot of research has been carried out in China on the configuration of energy storage capacity on the power supply side, but there are still the following shortcomings:

[0004] 1. One-sided considerations

[0005] 1. Ignoring uncertainty factors: Most existing technologies do not fully consider the uncertainty of renewable energy output, and only plan based on a small number of typical scenarios. They are unable to accurately respond to complex situations in actual operation, resulting in poor practicality of planning results.

[0006] 2. Ignoring system flexibility needs: In the context of insufficient flexibility supply in the power system, the flexibility of the energy storage system is crucial to ensuring the stable operation of the power system, but existing technologies do not take system flexibility needs into consideration.

[0007] 2. Limitations of Optimization Methods

[0008] 1. Deficiencies of traditional optimization methods: The random optimization method assumes that random variables follow a deterministic distribution, but the probability distribution of wind and solar forecast errors is difficult to obtain accurately; the robust optimization method is more conservative in dealing with uncertainty parameters, resulting in overly conservative scheduling results; although the distributed robust optimization method has overcome the shortcomings of the first two methods to a certain extent, its application in power system flexibility planning is not yet mature enough.

[0009] 2. Lack of comprehensive optimization: Existing technologies often focus on a single optimization goal, such as minimizing energy storage investment costs or maximizing system operating efficiency, and fail to achieve the coordinated optimization of energy storage systems and new energy sources, as well as the rational allocation of energy storage resources among different sites.

[0010] 3. Shared energy storage issues are prominent

[0011] 1. Unreasonable profit distribution: Under the shared energy storage model, the existing technology mainly considers the marginal profit of members when distributing profits, ignoring the impact of members' joining on the overall output effect, resulting in inefficient resource allocation and affecting the enthusiasm of new energy sites to participate in shared energy storage.

[0012] 2. Single energy storage type selection: There are many types of energy storage technologies, but existing research has not fully considered the diversified needs of new power systems and the trend of lowering electrochemical energy storage costs when selecting energy storage types, which limits the performance of energy storage systems.

[0013] At present, in the combination of new energy and traditional energy, the uncertainty of new energy output is an important issue to be considered in energy storage planning. The research methods for the uncertainty of new energy output can be divided into three categories: stochastic programming (SP), robust optimization (RO), and distributionally robust optimization (DRO).

[0014] The SP method must assume that the random variables follow a certain deterministic distribution, but for the wind and solar prediction errors, it is difficult to obtain their probability distribution in practice. The RO method uses a data set to describe random variables and solves the worst scenario, and its results are often too conservative. The DRO method overcomes the shortcomings of the SP and RO methods. It does not need to assume the type and parameters of the probability distribution. Instead, it establishes a fuzzy set of probability distribution based on the data and makes decisions based on the worst probability distribution in the fuzzy set. However, the flexibility requirements of the system are not considered in the above energy storage planning. Summary of the invention

[0015] In order to solve the deficiencies of the prior art, a more complete configuration system is now provided.

[0016] A method for configuring energy storage capacity of new energy stations based on multi-stage robust optimization, including investment decision-making layer, day-ahead operation scheduling strategy decision-making layer and intraday operation simulation layer;

[0017] The objective function of the overall energy storage capacity configuration method is:

[0018] Formula 1: minC Total =C inv +T D (C op +C risk )

[0019] Among them, C Total It represents the annual comprehensive cost of the system;

[0020] C inv represents the equivalent annual value investment cost of energy storage;

[0021] C op represents the operation and scheduling cost of the day-ahead system;

[0022] C risk Indicating the cost of penalty for lack of flexibility;

[0023] T D Represents the number of days in a year.

[0024] Furthermore, the investment decision layer uses the energy storage capacity to be planned as a decision variable, and minimizes the annual investment cost of energy storage on the basis of satisfying the capacity constraints of the energy storage system;

[0025] Formula 2:

[0026] In the formula, Represents the set of energy storage systems to be planned;

[0027] r represents the discount rate;

[0028] y ESS Indicates the actual operating life of the energy storage system;

[0029] C ESS Represents the investment cost per unit capacity of the energy storage system;

[0030] represents the rated capacity of the kth energy storage system to be planned;

[0031] The upper and lower limits of the capacity of the energy storage system to be planned must be met, as shown in Formula 3;

[0032] Formula 3:

[0033] In the formula, are the upper and lower limits of the required energy storage configuration capacity respectively;

[0034] Day-ahead operation scheduling strategy decision layer: The day-ahead operation scheduling strategy decision layer is a two-layer optimization model, including outer optimization and inner optimization.

[0035] In order to obtain better intraday simulation data and conduct analysis. Intraday operation simulation layer: The flexibility of the power system is defined as follows: In the active power balance of the time scale of interest, the power system optimizes the allocation of various available resources to adapt to the random changes of power generation, power grid and load at a certain cost. The intraday operation simulation layer takes the minimum penalty cost of insufficient system operation flexibility as the optimization goal, as shown in Formula 19:

[0036] Formula 19:

[0037] Formula 20:

[0038] Formula 21:

[0039] Formula 22:

[0040] In the formula, c up 、c dn They are the penalty coefficients for insufficient flexibility increase and decrease;

[0041] are the load shedding, wind and solar power abandonment caused by insufficient system flexibility in time period t;

[0042] ΔP L,t represents the increase or decrease of net load within the dispatch period interval ΔT, and ΔP L,t Take positive value;

[0043] They are respectively the upward and downward climbing rate limits that the thermal power unit can provide during time period t;

[0044] are the upward and downward climbing rate limits that the energy storage system can provide during time period t; P L,t , P L,t+ΔT are the net load power at time t and time t+Δt respectively;

[0045] are the maximum charging and discharging power of energy storage system k respectively; Represents the energy stored in the energy storage system k at time t.

[0046] The outer optimization is to minimize the total cost of starting and stopping the thermal power unit, fuel cost, cost of providing ramp backup capacity and operating cost of the energy storage system, given the predicted output of renewable energy, with the pre-dispatched output of the thermal power unit and the energy storage system as the decision variable, and the optimization goal is to minimize the total cost of starting and stopping the thermal power unit, fuel cost, cost of providing ramp backup capacity and operating cost of the energy storage system;

[0047] The inner optimization is to take the real-time adjustment amount of the thermal power unit in response to the wind and solar forecast error as the decision variable under the conditions of given unit combination arrangement, ramp reserve capacity arrangement, energy storage arrangement and new energy output error probability distribution function, and to minimize the expected value of the thermal power unit output adjustment cost under the probability distribution of wind and solar forecast error as the optimization goal;

[0048] Formula 4:

[0049] Formula 5:

[0050] in,

[0051] G represents the set of all thermal power units;

[0052] T represents the set of operating time periods of thermal power units, T = {1, 2, ..., 24};

[0053] K represents the set of energy storage systems to be planned;

[0054] p g,t It represents the pre-dispatched output of thermal power unit g during period t;

[0055] Δp g,t It represents the adjusted output of thermal power unit g in time period t in response to the uncertainty of renewable energy output;

[0056] They represent the cost of upward and downward ramping provided by thermal power unit g respectively;

[0057] They represent the upward and downward ramping capacities of thermal power unit g in period t respectively;

[0058] They represent the startup cost and shutdown cost of thermal power unit g in period t respectively;

[0059] c ess Represents the unit power operating cost of the energy storage system;

[0060] Represents the charging and discharging power of energy storage system k in time period t, which is positive when discharging and negative when charging;

[0061] The probability distribution function representing the forecast error of renewable energy output;

[0062] represents the expected value of the objective function under the probability distribution;

[0063] F g The coal consumption function of unit g can be expressed as a quadratic function of its output;

[0064] a g , b g 、c g They represent the coal consumption coefficient of unit g respectively;

[0065] The first and second items on the right side of the equal sign in formula 4 are the objective functions of the outer layer optimization, and the third item is the objective function of the inner layer optimization;

[0066] The time constraints and cost constraints for starting and shutting down thermal power units are:

[0067] Formula 6:

[0068] Formula 7:

[0069] Where T S , T O They are the minimum shutdown time and the minimum startup time of thermal power units respectively;

[0070] u g,tIndicates the start and stop state variables of thermal power unit g during period t. When starting, u g,t =1, when the machine stops g,t =0;

[0071] H g , J g are the single startup and shutdown costs of thermal power unit g respectively;

[0072] For collection Remove the set of the first period, that is

[0073] The ramp constraints of thermal power units and the output constraints considering the ramp reserve of units are:

[0074] Formula 8:

[0075] Formula 9:

[0076] in, It represents the maximum ramp rate of thermal power unit g in a unit period;

[0077] P -g are the maximum and minimum outputs of thermal power unit g respectively;

[0078] The state of charge constraints are shown in equations 10 and 11, the energy state constraints at the beginning and end of the scheduling cycle are shown in equation 12, the maximum charge and discharge power constraints are shown in equation 13, and the number of charge and discharge times in a scheduling cycle is shown in equation 14;

[0079] Formula 10: S k,min ≤S k,t ≤S k,max k∈K,t∈T

[0080] Formula 11:

[0081] Formula 12: S k,T =S k,0

[0082] Formula 13:

[0083] Formula 14:

[0084] In the formula, S k,t-1 , S k,t are the charge states of energy storage system k in time periods t-1 and t respectively; S k,max , S k,min are the upper and lower limits of the state of charge of the energy storage system k; S k,0 , Sk,T are the charge states of the energy storage system k at the beginning and end of the dispatch cycle; T represents the number of time periods in the dispatch cycle; η represents the charge and discharge efficiency of the energy storage system; ΔT represents the dispatch interval; represents the maximum charge and discharge power of energy storage system k; X k,t , Y k,t They are the charging and discharging state variables of energy storage system k in time period t. When charging, there is X k,t =1, Y k,t =0, when discharging, there is X k,t =0, Y k,t =1, when neither charging nor discharging, there is X k,t =Y k,t =0; N1 and N2 are the maximum charge and discharge limit times of the energy storage system respectively.

[0085] Single startup cost H g and shutdown costs J g Determined according to the type and performance of the thermal power unit, the specific value range is as follows:

[0086] For conventional thermal power units, H g The value range is US$8,000-160,000 per time. g The value range is US$6,400-120,000 per time.

[0087] For high-efficiency thermal power units, H g The value range is US$5,600-136,000 per time. g The value range is 4200-84000 yuan / time.

[0088] The new energy sites include wind farms and photovoltaic power stations.

[0089] Constraints for inner optimization:

[0090] The power balance constraint of the system is shown in Equation 15; the opportunity constraint of the thermal power unit to adjust the output in real time in response to the wind and solar forecast error is shown in Equation 16, which describes the probability that the real-time adjustment output meets the constraint condition is not less than 1-α gen (α gen The significance level of real-time adjustment of output constraints for thermal power units);

[0091] Formula 15:

[0092] Formula 16:

[0093] The opportunity constraint of DC power flow security based on the generation transfer distribution factor is shown in Equation 17, which describes the probability that the line capacity does not exceed the limit is not less than 1-α grid (αgrid is the significance level of the DC power flow security constraint of the power grid);

[0094] Formula 17:

[0095] Formula 18:

[0096] Beneficial effects:

[0097] Compared with the traditional capacity configuration system, this system can improve the utilization rate of the energy storage system. By establishing a multi-stage energy storage capacity planning model that takes flexibility into account, including the investment decision-making layer, the day-ahead operation scheduling strategy decision-making layer, and the intraday operation simulation layer, it can comprehensively consider the investment cost, operation scheduling cost, and the penalty cost of insufficient flexibility of the energy storage system, and achieve the optimal configuration of the energy storage system on different time scales. For example, in practical applications, by reasonably arranging the charging and discharging time and power of the energy storage system, it can better adapt to the uncertainty of new energy output, reduce the idle time of the energy storage system, and improve the utilization rate of the energy storage system. Statistics based on experimental data and actual operation conditions.

[0098] At the same time, the method based on distributed robust optimization is adopted, combined with the uncertainty of the output of new energy stations, and the probability distribution fuzzy set of renewable energy power generation is established through the Wasserstein distance, which can effectively deal with the uncertainty of new energy output and improve the robustness of energy storage configuration. Compared with traditional random optimization and robust optimization methods, the present invention can better balance the economy of energy storage configuration while ensuring the robustness of energy storage configuration. Specifically, there is an optimal value for the curve of system comprehensive cost changing with the energy storage configuration capacity. Under the optimal energy storage configuration capacity, the system comprehensive cost is the minimum.

[0099] Furthermore, a shared energy storage power station planning model based on cooperative game theory can be designed in the future, taking into account the cooperative configuration of energy storage systems among new energy sites. By designing comprehensive evaluation indicators for shared energy storage systems and using the improved Shapley scoring method to allocate the investment costs of each new energy site, energy storage sharing and investment cost sharing among new energy sites can be achieved. This not only improves the utilization rate of energy storage devices, but also reduces the investment cost of energy storage systems for new energy sites. In practical applications, through cooperative game theory, new energy sites can reasonably configure energy storage resources according to their own needs and resource advantages to achieve mutual benefit and win-win results. BRIEF DESCRIPTION OF THE DRAWINGS

[0100] Figure 1 It is a multi-timescale coordinated planning model framework;

[0101] Figure 2 It is the model solution flow chart;

[0102] Figure 3is the curve of cost changing with energy storage configuration capacity;

[0103] Figure 4 It is the change of wind curtailment and load shedding power with the energy storage configuration capacity;

[0104] Figure 5 The impact on the model solution time;

[0105] Figure 6 is the impact of ρ and N' on system cost;

[0106] Figure 7 This is a comparison chart of the degree of anti-peak regulation of new energy stations before and after the cooperation

[0107] Figure 8 This is a comparison of the summer output of new energy stations before and after the cooperation

[0108] Fig. 9 This is a comparison chart of how energy storage systems improve the output of new energy. DETAILED DESCRIPTION

[0109] In order to deepen the understanding of the present invention, the present invention will be further described in detail below in conjunction with examples. The examples are only used to explain the present invention and do not constitute a limitation on the protection scope of the present invention.

[0110] Embodiment 1:

[0111] A case analysis is carried out based on the IEEE-RTS24-node system.

[0112] The parameters of thermal power units, lines, and example calculations are shown in Table 1, Table 2, and Table 3, respectively. Because the prediction error of wind power is greater than the prediction error of photovoltaic power generation, two wind farms are selected for the new energy station, and it is assumed that the two wind farms are connected to nodes 4 and 16 respectively. Their output data are selected from, and the energy storage device is set to be fully charged in 4 hours at the fastest, that is, the maximum charging and discharging power of the energy storage is 1 / 4 of its capacity. The rated capacity range of the energy storage to be planned is [40, 100] MW·h, and it is connected to nodes 4 and 16.

[0113] Table 1 Parameters of thermal power units

[0114]

[0115]

[0116] Table 2 RTS24 node line capacity parameters

[0117]

[0118] Table 3 Calculation parameters of the example

[0119]

[0120] The node system, thermal power units, lines and other related parameters used in the shared energy storage power station planning model based on cooperative game are the same as those in the energy storage capacity planning model based on distributed blue stick optimization mentioned above. The new energy station uses two wind farms and one photovoltaic power station. The installed capacities of wind farm 1, wind farm 2 and photovoltaic power station are 400MW, 300MW and 300MW respectively, and they are connected to nodes 4, 6 and 9. The specific data of time-of-use electricity prices are shown in Table 4.

[0121] Table 4 Time-of-use electricity price

[0122]

[0123] The method of combining multi-stage iterative optimization with linear decision rule transformation is adopted to simplify the problem solving difficulty and improve the calculation efficiency. The particle swarm optimization algorithm is used to solve the energy storage investment decision stage. The solution flow chart is as follows: Figure 2 The specific solution steps are as follows.

[0124] Step 1: Initialize data. Input the relevant data of thermal power units, node loads, system topology, and historical data of renewable energy output to generate a renewable energy output data set, which is divided into a training set and a test set.

[0125] Step 2: Initialize the particle population. Set the population size, maximum number of iterations and other parameters of the particle swarm optimization algorithm, initialize the position and velocity of each particle in the particle population, the individual optimal value, and the group optimal value, where the particle represents the energy storage capacity to be planned, and the fitness function is shown in formula (1).

[0126] Step 3: Input the energy storage capacity generated in step 2 as the initial value into stage 2 of the model, and input the training set of renewable energy output to solve for the day-ahead schedule of energy storage, the start and shutdown plan of thermal power units, and the estimated output.

[0127] Step 4: Use the day-ahead dispatch results and renewable energy output test set obtained in step 3 as the initial values ​​of model stage 3, and solve the effect of the energy storage configuration results in out-of-sample applications according to equations (20)-(22), that is, the penalty cost for insufficient intraday flexibility.

[0128] Step 5: Update the particle population. Calculate the fitness value of each particle based on the solution results of steps 3 and 4, update the individual optimal value and the global optimal value, and perform evolution and boundary condition processing on the particle speed and position.

[0129] Step 6: Repeat steps 3 to 5 until the set threshold is met or the maximum number of iterations is reached.

[0130] Step 7: Output the optimal energy storage capacity.

[0131] At the same time, when the energy storage configuration capacity is not less than 10% of the wind farm capacity and not more than 25%, the energy storage configuration capacity of a single wind farm ranges from [40,100] MWh. The number of training samples for the DRO model is N′=20, the number of test samples is O=10, the radius of the Wasserstein sphere is 0.005, and the confidence level is 1-α=0.95. The curves of system comprehensive cost, operation and dispatch cost, penalty cost for insufficient flexibility and energy storage investment cost with the change of energy storage configuration capacity in wind farm 1 are shown in the figure below. Figure 3 shown.

[0132] Depend on Figure 3 (a) It can be seen that the system comprehensive cost first increases slowly with the increase of energy storage configuration capacity, then decreases significantly, and then increases sharply again. There is an optimal energy storage configuration capacity that minimizes the comprehensive cost. Therefore, the multi-stage optimization problem in this paper is convergent, and it can be seen that the optimal capacity is between [70,75]MWh. Figure 3 (b) It can be seen that: when the energy storage configuration capacity is small, the system operation and dispatching cost decreases slightly with the increase of energy storage configuration capacity; when the energy storage configuration capacity increases to a certain extent, the operation and dispatching cost decreases significantly and then gradually stabilizes. Figure 3 (c) It can be seen that the penalty cost of insufficient flexibility decreases as the storage capacity increases. Figure 3 (d) It can be seen that the energy storage investment cost increases with the increase of energy storage configuration capacity. In summary, when the energy storage configuration capacity is small, as the configuration capacity increases, the operation and dispatching costs and the penalty costs for insufficient flexibility decrease, but the reduced costs are not enough to compensate for the increase in energy storage investment costs, so the comprehensive cost increases slowly; when the energy storage configuration capacity reaches a certain value, the increase in configuration capacity has a significant impact on the operation and dispatching costs, so the comprehensive cost of the system is greatly reduced and finally reaches the minimum value; when the energy storage configuration capacity continues to increase, the operation and dispatching costs and the penalty costs for insufficient flexibility tend to be flat with the change of configuration capacity, so the comprehensive cost rises sharply again. The relationship between wind abandonment and load shedding power and the change of energy storage configuration capacity is as follows: Figure 4 As shown in the figure, the wind power abandonment and load shedding power decrease with the increase of energy storage configuration capacity, and when the energy storage configuration capacity reaches a certain value, the reduction of wind power abandonment tends to be gentle.

[0133] Compared with the DRO method, the energy storage configuration capacity obtained based on the SP method is larger, and the comprehensive cost and the penalty cost of insufficient flexibility are both higher. This is because in the pre-dispatch stage, the SP method regards wind power output as satisfying a certain probability distribution, which reduces the conservatism of the dispatch result, but the flexibility in the operation stage is obviously insufficient, so the system needs to configure a larger capacity of energy storage to provide flexibility resources; compared with the DRO method, the energy storage configuration capacity obtained based on the RO method is slightly smaller, but its comprehensive cost is higher than that of the DRO method. This is because in the pre-dispatch stage, the RO method is more conservative in handling uncertainty parameters than the DRO method, so the dispatch result is also more conservative, which significantly increases the cost of the pre-dispatch stage. Although the penalty cost of insufficient flexibility is lower than that of the DRO method, the comprehensive cost is greater, that is, the RO method pays too much dispatch cost. When the DRO method is used, although the energy storage investment cost and the penalty cost of insufficient flexibility are slightly higher than those of the RO method, the operation dispatch cost is lower than that of the RO method, and the DRO method has the smallest comprehensive cost. Compared with the energy storage planning schemes in the literature, the energy storage configuration capacity of the SP method in this paper is higher, while the energy storage configuration capacity of the RO method is smaller. This is because in the planning model of this paper, the pre-dispatch stage mainly considers the uncertainty of wind power output borne by thermal power units, and in stage 3, energy storage is used as the main flexibility resource. Therefore: when the SP method is used, the cost of thermal power units in the pre-dispatch stage is relatively small, while the system flexibility in the daily operation stage is insufficient, so more energy storage is needed to provide flexibility; when the RO method is used, the cost of thermal power units in the pre-dispatch stage is relatively large, while the flexibility in the daily operation stage is better, so the energy storage capacity is relatively small.

[0134] Depend on Figure 5 It can be seen that the model solution time increases with the increase of N'. When N' exceeds 15, the solution time will increase significantly. Figure 6 (a) and (c) show that the comprehensive cost and the penalty cost of insufficient flexibility decrease as N' increases; Figure 6 (b) It can be seen that the operation scheduling cost increases with the increase of N', which indicates that as N' increases, the results of the day-ahead scheduling phase become more conservative, while the penalty cost of insufficient flexibility in the intraday scheduling phase decreases, thereby reducing the total cost. For the penalty cost of insufficient flexibility, the impact of N' increasing from 10 to 20 is much greater than the impact of N' increasing from 20 to 30, but Figure 5 From the influence of N' on the model solution time, it can be seen that the increase of N' causes the model solution time to increase exponentially. Therefore, a suitable N' should be selected in the energy storage configuration based on the DRO method. The radius of the Wasserstein ball ρ determines the size of the uncertain fuzzy set of the real distribution, which is given by Figure 6It can be seen that when N' is 20 and 30, as ρ increases, the cost of the operation scheduling stage gradually increases, the penalty cost of insufficient flexibility gradually decreases, and the comprehensive cost of the system first decreases and then increases, which shows that the appropriate amount of insufficient flexibility penalty can make the total cost of the system optimal. Based on the above analysis, it can be seen that the model proposed in this paper can control the size of the uncertain fuzzy set by changing the number of training samples of the model and the radius of the Wasserstein ball, and effectively balance the economy and robustness of the solution.

[0135] Due to the resource characteristics of wind and light, the output of renewable energy is random and volatile. The maximum daily fluctuation of wind power can reach 80% of the installed capacity, and it exhibits certain anti-peak characteristics. The anti-peak characteristics of renewable energy output are also the reason that restricts its development. According to the difference between the median of the grid-connected power of renewable energy stations during the peak load period and the valley load period, the output curve is divided into positive peak regulation and anti-peak regulation. If the median of the peak period is less than the median of the valley period, it is anti-peak regulation, otherwise it is positive peak regulation. Wind Farm 1 had a total of 194 days of anti-peak regulation, and after cooperating with the photovoltaic power station, the number of anti-peak regulation days was 150. Figure 7 The degree of anti-peak regulation before and after the cooperation between wind farm 1 and the photovoltaic power station was compared.

[0136] from Figure 7 (a) and (b) show that the cooperation between wind farms and photovoltaic power stations can not only reduce the number of days with reverse peak regulation, but also alleviate the degree of reverse peak regulation of wind farms to a certain extent. In addition, wind power output is severely seasonal, and even has a low output for several consecutive days in summer. In this case, the utilization rate of the energy storage system will be very low. Figure 8 The output of wind farm 1 in one summer month before and after the cooperation with the photovoltaic power station was compared. Figure 8 (a) and (b) show that after the cooperation between the wind farm and the photovoltaic power station, the light intensity in summer is stronger and lasts longer, which can alleviate the problem of low output of wind power in summer to a certain extent.

[0137] Wind power and photovoltaic power generation have different output characteristics, and energy storage cannot generate energy, but can only play a role in "time shifting" energy. Therefore, the configuration effect of energy storage before and after the cooperation between wind farms and photovoltaic power stations is also different. Fig. 9 The effects of using energy storage systems to improve renewable energy output were compared before and after the cooperation.

[0138] from Fig. 9From (a) and (b), it can be seen that the output characteristic of wind power is that the output is large at night, and the peak load periods are 8:00-11:00 and 18:00-22:00. Therefore, if the wind farm is equipped with energy storage alone, the energy storage will be charged at night and discharged during the two peak periods during the day, and the utilization rate of the energy storage is very low. After cooperating with the photovoltaic power station, the wind power stored in the energy storage at night can be released at 8:00-11:00, and the photovoltaic output stored in the energy storage at noon can be released at 18:00-22:00, which not only improves the peak regulation effect, but also improves the utilization rate of energy storage.

[0139] In order to make the energy storage system adapt to the load characteristics of the distribution network, this technical solution first selects the appropriate energy storage type for different application scenarios and technical requirements. In view of the economic efficiency of energy storage devices, system operation flexibility, uncertainty of new energy output, etc., a multi-stage energy storage capacity planning model is established. For application scenarios that require rapid adjustment, electrochemical energy storage with fast response speed and flexible adjustment is selected to ensure that the energy storage system can better adapt to the load changes of the distribution network during the entire planning period.

[0140] Determining the installed capacity and power of energy storage is the key to achieving the coordination of energy storage and renewable energy. Based on a comprehensive consideration of investment costs, system dispatching costs, and penalty costs for insufficient flexibility after operation, this technical solution establishes a multi-stage energy storage capacity planning model that takes flexibility into account, which enables the installed capacity and power of energy storage to be optimally configured on long and short time scales. By reasonably setting the energy storage capacity and power, the energy storage system can not only provide sufficient regulation capabilities to meet the load requirements of the distribution network, but also ensure optimal economy and avoid over-investment. This configuration strategy ensures that the energy storage system can flexibly adapt to load fluctuations during the planning period, and effectively coordinate with the renewable energy DG to improve the overall efficiency of system operation.

[0141] The reliability and economy of the system can be effectively improved through the coordinated optimization of energy storage and new energy. This technical solution uses the distributed robust optimization method to model the energy storage capacity planning, so that the energy storage system can provide better robustness and balance under the background of uncertainty in the output of new energy. In addition, by using the shared energy storage power station planning model of cooperative game and the improved Shapley value allocation strategy, energy storage sharing and investment cost sharing between new energy sites are realized, which improves the utilization rate of energy storage and reduces the investment cost. This collaborative optimization process not only ensures the safe and reliable operation of new energy in the distribution network, but also gives full play to the regulatory role of the energy storage system to ensure the optimal economic performance of the system.

[0142] As a further improvement, the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for configuring energy storage capacity of a new energy station based on multi-stage robust optimization, characterized in that: It includes investment decision-making layer, day-ahead operation scheduling strategy decision-making layer and intraday operation simulation layer; The objective function of the overall energy storage capacity configuration method is: Formula 1: minC Total =C inv +T D (C op +C risk ) Among them, C Total It represents the annual comprehensive cost of the system; C inv represents the equivalent annual value investment cost of energy storage; C op represents the operation and scheduling cost of the day-ahead system; C risk Indicating the cost of penalty for lack of flexibility; T D Represents the number of days in a year.

2. According to claim 1, a method for configuring energy storage capacity of a new energy station based on multi-stage robust optimization is characterized in that: The investment decision layer uses the energy storage capacity to be planned as the decision variable and minimizes the annual investment cost of energy storage on the basis of satisfying the capacity constraints of the energy storage system; Formula 2: In the formula, Represents the set of energy storage systems to be planned; r represents the discount rate; y ESS Indicates the actual operating life of the energy storage system; C ESS Represents the investment cost per unit capacity of the energy storage system; represents the rated capacity of the kth energy storage system to be planned; The upper and lower limits of the capacity of the energy storage system to be planned must be met, as shown in Formula 3; Formula 3: In the formula, are the upper and lower limits of the required energy storage configuration capacity respectively; Day-ahead operation scheduling strategy decision layer: The day-ahead operation scheduling strategy decision layer is a two-layer optimization model, including outer optimization and inner optimization.

3. According to a method for configuring energy storage capacity of a new energy station based on multi-stage robust optimization according to claim 1, it is characterized in that: Intraday operation simulation layer: The flexibility of the power system is defined as follows: In the active power balance of the time scale of interest, the power system optimizes the allocation of various available resources to adapt to the random changes in power generation, power grid and load at a certain cost. The intraday operation simulation layer takes the minimum penalty cost of insufficient system operation flexibility as the optimization goal, as shown in Equation 19: Formula 19: Formula 20: Formula 21: Formula 22: In the formula, c up 、c dn They are the penalty coefficients for insufficient flexibility increase and decrease; are the load shedding, wind and solar power abandonment caused by insufficient system flexibility in time period t; ΔP L,t represents the increase or decrease of net load within the dispatch period interval ΔT, and ΔP L,t Take positive value; They are respectively the upward and downward climbing rate limits that the thermal power unit can provide during time period t; are the upward and downward climbing rate limits that the energy storage system can provide during time period t; P L,t , P L,t+ΔT are the net load power at time t and time t+Δt respectively; are the maximum charging and discharging power of energy storage system k respectively; Represents the energy stored in the energy storage system k at time t.

4. According to claim 2, a method for configuring energy storage capacity of a new energy station based on multi-stage robust optimization is characterized in that: The outer optimization is to minimize the total cost of starting and stopping the thermal power unit, fuel cost, cost of providing ramp backup capacity and operating cost of the energy storage system, given the predicted output of renewable energy, with the pre-dispatched output of the thermal power unit and the energy storage system as the decision variable, and the optimization goal is to minimize the total cost of starting and stopping the thermal power unit, fuel cost, cost of providing ramp backup capacity and operating cost of the energy storage system; The inner optimization is to take the real-time adjustment amount of the thermal power unit in response to the wind and solar forecast error as the decision variable under the conditions of given unit combination arrangement, ramp reserve capacity arrangement, energy storage arrangement and new energy output error probability distribution function, and to minimize the expected value of the thermal power unit output adjustment cost under the probability distribution of wind and solar forecast error as the optimization goal; Formula 4: Formula 5: in, G represents the set of all thermal power units; T represents the set of operating time periods of thermal power units, T = {1, 2, ..., 24}; K represents the set of energy storage systems to be planned; p g,t It represents the pre-dispatched output of thermal power unit g during period t; Δp g,t It represents the adjusted output of thermal power unit g in time period t in response to the uncertainty of renewable energy output; They represent the cost of upward and downward ramping provided by thermal power unit g respectively; They represent the upward and downward ramping capacities of thermal power unit g in period t respectively; They represent the startup cost and shutdown cost of thermal power unit g in period t respectively; c ess Represents the unit power operating cost of the energy storage system; Represents the charging and discharging power of energy storage system k in time period t, which is positive when discharging and negative when charging; The probability distribution function representing the forecast error of renewable energy output; represents the expected value of the objective function under the probability distribution; F g The coal consumption function of unit g can be expressed as a quadratic function of its output; a g 、b g 、c g They represent the coal consumption coefficient of unit g respectively; The first and second items on the right side of the equal sign in formula 4 are the objective functions of the outer layer optimization, and the third item is the objective function of the inner layer optimization; The time constraints and cost constraints for starting and shutting down thermal power units are: Formula 6: Formula 7: Where T S , T O They are the minimum shutdown time and the minimum startup time of thermal power units respectively; u g,t Indicates the start and stop state variables of thermal power unit g during period t. When starting, u g,t =1, when the machine stops g,t =0; H g , J g are the single startup and shutdown costs of thermal power unit g respectively; For collection Remove the set of the first period, that is The ramp constraints of thermal power units and the output constraints considering the ramp reserve of units are: Formula 9: in, It represents the maximum ramp rate of thermal power unit g in a unit period; P -g are the maximum and minimum outputs of thermal power unit g respectively; The state of charge constraints are shown in equations 10 and 11, the energy state constraints at the beginning and end of the scheduling cycle are shown in equation 12, the maximum charge and discharge power constraints are shown in equation 13, and the number of charge and discharge times in a scheduling cycle is shown in equation 14; Formula 10: S k,min ≤S k,t ≤S k,max k∈K,t∈T Formula 11: Formula 12: S k,T =S k,0 Formula 13: Formula 14: In the formula, S k,t-1 , S k,t are the charge states of energy storage system k in time periods t-1 and t respectively; S k,max , S k,min are the upper and lower limits of the state of charge of the energy storage system k; S k,0 , S k,T are the charge states of the energy storage system k at the beginning and end of the dispatch cycle; T represents the number of time periods in the dispatch cycle; η represents the charge and discharge efficiency of the energy storage system; ΔT represents the dispatch interval; represents the maximum charge and discharge power of energy storage system k; X k,t , Y k,t They are the charging and discharging state variables of energy storage system k in time period t. When charging, there is X k,t =1, Y k,t =0, when discharging, there is X k,t =0, Y k,t =1, when neither charging nor discharging, there is X k,t =Y k,t =0; N1 and N2 are the maximum charge and discharge limit times of the energy storage system respectively.

5. A method for configuring energy storage capacity of a new energy station based on multi-stage robust optimization according to any one of claims 1 to 4, characterized in that: Single startup cost H g and shutdown costs J g Determined according to the type and performance of the thermal power unit, the specific value range is as follows: For conventional thermal power units, H g The value range is US$8,000-160,000 per time. g The value range is US$6,400-120,000 per time. For high-efficiency thermal power units, H g The value range is US$5,600-136,000 per time. g The value range is 4200-84000 yuan / time.

6. A method for configuring energy storage capacity of a new energy station based on multi-stage robust optimization according to claim 2, characterized in that: The new energy sites include wind farms and photovoltaic power stations.

7. The method for configuring energy storage capacity of a new energy station based on multi-stage robust optimization according to claim 6 is characterized in that: Constraints for inner optimization: The power balance constraint of the system is shown in Equation 15; the opportunity constraint of the thermal power unit to adjust the output in real time in response to the wind and solar forecast error is shown in Equation 16, which describes the probability that the real-time adjustment output meets the constraint condition is not less than 1-α gen (α gen The significance level of real-time adjustment of output constraints for thermal power units); Formula 15: Formula 16:

8. The method for configuring energy storage capacity of a new energy station based on multi-stage robust optimization according to claim 2 is characterized in that: The opportunity constraint of DC power flow security based on the generation transfer distribution factor is shown in Equation 17, which describes the probability that the line capacity does not exceed the limit is not less than 1-α grid (α grid is the significance level of the DC power flow security constraint of the power grid); Formula 17: Formula 18:

Images