Shared energy storage configuration optimization method and system for wind farm cluster

Abstract

Disclosed in the present invention are a shared energy storage configuration optimization method and system for a wind farm cluster. The method comprises: constructing a shared energy storage double-layer programming model; on the basis of the output of a wind farm cluster, an electricity price in an electric energy market and a compensation price in a frequency-modulation auxiliary service market, obtaining multiple uncertainty sets, using an improved K-means clustering method to perform joint clustering on the multiple uncertainty sets so as to obtain a typical scenario, and on the basis of the annual occurrence frequency of the typical scenario, calculating a typical scenario probability; and using a distributionally robust optimization method to model the multiple uncertainty sets on a short-time scale so as to obtain an extreme uncertainty scenario, using the typical scenario probability and the extreme uncertainty scenario to update the shared energy storage double-layer programming model so as to obtain an updated shared energy storage double-layer programming model, and using a particle swarm algorithm to perform solving so as to obtain a configuration result. By means of the method, shared energy storage configuration optimization is performed by means of taking various uncertainty factors into consideration, thereby improving the accuracy of shared energy storage configuration optimization.

Description

A wind farm cluster shared energy storage optimization configuration method and system TECHNICAL FIELD

[0001] The present application relates to the technical field of grid energy storage optimization configuration, and particularly relates to a wind farm cluster shared energy storage optimization configuration method and system. BACKGROUND

[0002] The optimization scheduling problem of shared energy storage power stations has important significance in actual industrial users. By installing shared power stations on the industrial user side, the energy storage capacity demand of different industrial users can be met, and energy utilization efficiency can be improved. The shared energy storage mode breaks down the information sharing barriers and energy sharing barriers caused by the binding of ownership and use right of energy storage resources. In this operation mechanism, the energy storage devices of each new energy power plant meet their own main demand, and provide energy storage services for other main bodies or power grids through time-sharing reuse of surplus energy storage resources and collect corresponding service fees. The energy storage devices can meet the demand of their own new energy power plants and realize energy mutual aid among multiple new energy power plants.

[0003] However, the current shared energy storage optimization configuration method is mainly based on the demand of the entire power system, and is formed based on the targets of smoothing output fluctuation and participating in frequency modulation and peak regulation to prevent wind and light abandonment problems from occurring. The influence of multiple uncertainties such as wind farm output on shared energy storage capacity configuration under multiple time scales is not considered, resulting in relatively low rationality of the configuration effect. SUMMARY

[0004] In order to solve the above technical problems, the embodiment of the present application provides a wind farm cluster shared energy storage optimization configuration method and system, which considers multiple uncertain factors to optimize the configuration of shared energy storage, and improves the accuracy of shared energy storage optimization configuration.

[0005] The first aspect of the embodiment of the present application provides a wind farm cluster shared energy storage optimization configuration method, which comprises:

[0006] A shared energy storage double-layer planning model is constructed, wherein the upper model of the double-layer planning model takes the highest economic benefit of shared energy storage as the objective function, and the lower model takes the lowest operation cost of the wind farm cluster as the objective function;

[0007] A multiple uncertainty set is obtained according to the wind farm cluster output, the electricity market price and the frequency modulation auxiliary service market compensation price, the improved K-means clustering method is used for joint clustering of the multiple uncertainty set to obtain a typical scenario, and the typical scenario probability is calculated according to the annual occurrence frequency of the typical scenario;

[0008] The distributed robust optimization method is used to model multiple uncertainty sets in a short time scale to obtain uncertainty extreme scenarios, the typical scenario probability and the uncertainty extreme scenarios are used to update the shared energy storage bi-level programming model to obtain an updated shared energy storage bi-level programming model, and the particle swarm algorithm is used to solve the configuration result.

[0009] In a possible implementation of the first aspect, the shared energy storage bi-level programming model is constructed, including:

[0010] A shared energy storage economic benefit is taken as an objective function to construct an upper model, wherein the upper model is:

[0011] In the formula, I R,d is a wind farm cluster capacity leasing daily income; I DNL,d is an electricity market daily income; I TP,d is a frequency modulation auxiliary service market daily income; C in is a shared energy storage investment construction cost; C op is a shared energy storage annual operation and maintenance cost;

[0012] In the formula, is a shared energy storage wind farm cluster leasing power at the dth day and the tth time, p r,d is a shared energy storage unit power leasing price at the dth day, P DNL,d is a shared energy storage electricity market out-of-clearing power at the dth day and the tth time, p DNL,d is an electricity market out-of-clearing price at the dth day and the tth time, P TP,d is a shared energy storage frequency modulation auxiliary service market bid capacity at the dth day and the tth time, p TP,d is a frequency modulation auxiliary service market frequency modulation compensation price at the dth day and the tth time, C p is a shared energy storage unit power investment construction cost, C s is a shared energy storage unit capacity investment construction cost, N y is a shared energy storage full life cycle, P max is a shared energy storage maximum charging and discharging power, E SES is a shared energy storage optimized configuration capacity, g is a discount rate, h is a replacement number of the shared energy storage device, n is a total number of replacements of the shared energy storage device in the full life cycle, C pom is a shared energy storage unit power operation and maintenance cost, C som is a shared energy storage unit capacity operation and maintenance cost, W(t) is a shared energy storage full life cycle charging and discharging amount;

[0013] The lower-level model, with the objective function of minimizing the operating cost of the wind farm cluster, is as follows: f d =min(C p +C r -I s )

[0014] In the formula, C p For the daily deviation assessment cost of wind farm clusters, C r For the daily rental cost of shared energy storage in wind farm clusters, I s For the daily electricity sales revenue of the wind farm cluster, Let be the actual wind power output of the wind farm cluster at time t. Let p be the day-ahead forecast of wind power output of the wind farm cluster at time t. wind The price per unit of wind power deviation assessment, p DNL (t) represents the clearing price of the electrical energy market at time t.

[0015] In one possible implementation of the first aspect, an improved K-means clustering method is used to jointly cluster sets of multiple uncertainties, resulting in typical scenarios, including:

[0016] The set of multiple uncertainties is extended to a long-term set, and the data in the long-term set is subjected to a quantitative transformation to obtain the processed data.

[0017] Multiple data points are randomly selected as cluster centers. The distance difference and trend difference between all processed data points and each cluster center are calculated. The comprehensive difference is obtained based on the distance difference and trend difference.

[0018] Data with a comprehensive difference less than a preset value is assigned to the corresponding cluster centers to obtain multiple initial typical scenarios. The mean of the multiple uncertainty set in each initial typical scenario is recalculated, and the mean is used as the new cluster center. The comprehensive difference of all processed data to the new cluster center is calculated and assigned until the cluster center no longer changes, thus obtaining multiple typical scenarios.

[0019] In one possible implementation of the first aspect, the distance difference and trend difference from all processed data to each cluster center are calculated, and a comprehensive difference is obtained based on the distance difference and trend difference, including:

[0020] The distance difference from all processed data to each cluster center is calculated using Euclidean distance. The trend difference is then calculated using the trend difference formula, where the formula for calculating the distance difference is:

[0021] In the formula, a distance difference of the dth processed data to the kth cluster center, a wind power sequence of the kth cluster center of the multiple-uncertainty-set wind farm cluster, DNL,k a clearing price sequence of the kth cluster center of the multiple-uncertainty-set electricity market, TP,k a frequency modulation compensation price sequence of the kth cluster center of the multiple-uncertainty-set frequency modulation auxiliary service market,

[0022] The trend difference calculation formula is:

[0023] In the formula, a trend difference of the dth processed data to the kth cluster center;

[0024] The distance difference and the trend difference of each processed data to the cluster center are normalized by using an exponential, to obtain a distance difference value and a trend difference value, wherein the distance difference value calculation formula is:

[0025] In the formula, γ1(d,k) is the distance difference value of the dth processed data to the kth cluster center;

[0026] The trend difference value calculation formula is:

[0027] In the formula, γ2(d,k) is the distance difference value of the dth processed data to the kth cluster center;

[0028] The distance difference value and the trend difference value are fused by using a proportional coefficient method, to obtain a comprehensive difference value of the processed data to each cluster center, wherein the comprehensive difference value expression is: γ(d,k)=αγ1(d,k)+βγ2(d,k)

[0029] In the formula, γ(d,k) is the comprehensive difference value of the processed data to each cluster center, and α and β are proportional coefficients, which take values of relative importance degrees of the distance difference value and the trend difference value, and α+β=1.

[0030] In a possible implementation manner of the first aspect, the multiple-uncertainty-set is modeled on a short time scale by using a distribution robust optimization method, to obtain uncertainty extreme scenarios, including:

[0031] The multiple-uncertainty-set on the short time scale is constructed by using a robust parameter, wherein the set is:

[0032] In the formula, a wind power real value of the wind power sequence of the wind farm cluster at the tth time on the dth day, Let be the day-ahead forecast of wind power for the wind farm cluster at time t on day d. Let be the maximum value of the wind power fluctuation range of the wind farm cluster at time t on day d. Let p be the distribution bar parameter of the wind power sequence of the wind farm cluster on day d. DNL,d (t) represents the true electricity price at time t in the clearing price sequence of the electricity market on day d. Let t be the day-ahead electricity price forecast for wind power at time t, representing the clearing price sequence of the electricity market on day d. χ represents the maximum fluctuation range of wind power at time t in the clearing price sequence of the electricity market on day d. DNL,d Let p be the parameter of the Blob bar for the clearing price sequence of the electricity market on day d. TP,d (t) represents the true value of the frequency modulation compensation price sequence of the ancillary services market at time t on day d. This refers to the day-ahead forecast of the compensation price at time t in the frequency modulation compensation price sequence for the ancillary services market on day d. χ represents the maximum range of compensation price fluctuations in the frequency modulation compensation price sequence of the ancillary service market at time t on day d. TP,d For the sub-Bluer bar parameters of the frequency modulation compensation price series of the ancillary services market on day d;

[0033] The formula for calculating the parameters of the Brussels rod is:

[0034] In the formula, μ DNL,d and μ TP,d Sequences Sequence Y DNL,d and sequence Y TP,d Expectations; and Sequences Sequence Y DNL,d and sequence Y TP,d Standard deviation; α DNL,d and α TP,d The confidence levels for wind farm cluster output, electricity market price, and frequency regulation ancillary service market compensation price in the sub-Browser optimization process are respectively. It is the probability distribution function of the standard normal distribution.

[0035] In one possible implementation of the first aspect, the updated shared energy storage two-level planning model is: f u =max(π) k [I R,d (ωk )+I DNL,d (ω k )+I TP,d (ω k )]-C in -C op )

[0036] In the formula, I R,d (ω k ), I DNL,d (ω k ) and I TP,d (ω k ) are daily income of wind farm cluster capacity leasing, daily income of electricity market and daily income of frequency modulation auxiliary service market under the kth typical scene respectively, C p is daily deviation evaluation cost of the wind farm cluster under the kth typical scene, C r is daily leasing cost of the wind farm cluster leasing shared energy storage under the kth typical scene, I s is daily electricity sale income of the wind farm cluster under the kth typical scene, and ψ is a multiple uncertainty set of wind farm cluster output and electricity market clearing price. is the worst scene of the multiple uncertainty set.

[0037] The second aspect of the embodiment of the application provides a wind farm cluster shared energy storage optimization configuration system, the system comprising a construction module, a typical scene calculation module and a configuration module,

[0038] The construction module is used for constructing a shared energy storage bi-level programming model, wherein an upper model of the bi-level programming model takes the highest economic benefit of the shared energy storage as an objective function, and a lower model of the bi-level programming model takes the lowest operation cost of the wind farm cluster as an objective function.

[0039] The typical scene calculation module is used for obtaining a multiple uncertainty set according to wind farm cluster output, electricity market price and frequency modulation auxiliary service market compensation price, performing joint clustering on the multiple uncertainty set by using an improved K-means clustering method to obtain a typical scene, and calculating a typical scene probability according to an annual appearance frequency of the typical scene.

[0040] The configuration module is used for modeling the multiple uncertainty set in a short time scale by using a distribution robust optimization method to obtain an uncertainty extreme scene, updating the shared energy storage bi-level programming model by using the typical scene probability and the uncertainty extreme scene to obtain an updated shared energy storage bi-level programming model, and solving the updated shared energy storage bi-level programming model by using a particle swarm algorithm to obtain a configuration result.

[0041] In a possible implementation manner of the second aspect, the shared energy storage bi-level programming model is constructed, comprising:

[0042] The upper model is constructed with the shared energy storage economic benefit as the objective function, wherein the upper model is:

[0043] In the formula, I R,d is the daily income of wind farm cluster capacity leasing; I DNL,d is the daily income of electricity market; I TP,d is the daily income of frequency modulation auxiliary service market; C in is the shared energy storage investment construction cost; C op is the shared energy storage annual operation and maintenance cost;

[0044] In the formula, is the shared energy storage leasing power of the wind farm cluster at the tth moment of the dth day, p r,d is the shared energy storage unit power leasing price on the dth day, P DNL,d is the shared energy storage market out-of-clearing power at the tth moment of the dth day, p DNL,d is the out-of-clearing price of the electricity market at the tth moment of the dth day, P TP,d is the shared energy storage bid capacity in the frequency modulation auxiliary service market at the tth moment of the dth day, p TP,d is the frequency modulation compensation price of the frequency modulation auxiliary service market at the tth moment of the dth day, C p is the shared energy storage unit power investment construction cost, C s is the shared energy storage unit capacity investment construction cost, N y is the shared energy storage full life cycle, P max is the maximum charge-discharge power of the shared energy storage, E SES is the shared energy storage optimal configuration capacity, g is the discount rate, h is the replacement number of the shared energy storage device, n is the total replacement number of the shared energy storage device in the full life cycle, C pom is the shared energy storage unit power operation and maintenance cost, C som is the shared energy storage unit capacity operation and maintenance cost, and W(t) is the full life cycle charge-discharge amount of the shared energy storage;

[0045] The lower model with the wind farm cluster operation cost as the objective function is as follows: f d = min(C p +C r -I s )

[0046] In the formula, C p is the daily deviation assessment cost of the wind farm cluster, C r is the daily leasing cost of the wind farm cluster leasing shared energy storage.s P is the daily electricity sales revenue of the wind farm cluster, P is the actual value of the wind power of the wind farm cluster at the tth moment, P is the day-ahead forecast value of the wind power of the wind farm cluster at the tth moment, wind P is the unit wind power deviation assessment price, DNL P (t) is the clearing price of the electricity market at the tth moment.

[0047] A third aspect of the embodiment of the present application provides a computer device, comprising:

[0048] a memory for storing a computer program;

[0049] a processor for implementing the wind farm cluster shared energy storage optimization configuration method according to the first aspect when executing the computer program.

[0050] A fourth aspect of the embodiment of the present application provides a storage medium, and the storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the wind farm cluster shared energy storage optimization configuration method according to the first aspect.

[0051] The wind farm cluster shared energy storage optimization configuration method provided by the embodiment of the present application, by constructing a shared energy storage double-layer planning model, the upper model of the double-layer planning model with the highest economic benefit of the shared energy storage as the objective function and the lower model with the lowest operation cost of the wind farm cluster as the objective function, obtaining a multiple uncertainty set according to the wind farm cluster output, the electricity market price and the frequency modulation auxiliary service market compensation price, using an improved K-means clustering method to jointly cluster the multiple uncertainty set to obtain a typical scenario, calculating the typical scenario probability according to the annual occurrence frequency of the typical scenario, using a distribution robust optimization method to model the multiple uncertainty set in a short time scale to obtain an uncertainty extreme scenario, using the typical scenario probability and the uncertainty extreme scenario to update the shared energy storage double-layer planning model to obtain an updated shared energy storage double-layer planning model, and using a particle swarm algorithm to obtain a configuration result. Through the above method, the shared energy storage optimization configuration is performed by considering multiple uncertainty factors, and the accuracy of the shared energy storage optimization configuration is improved. BRIEF DESCRIPTION OF DRAWINGS

[0052] FIG. 1 is a configuration flowchart of the wind farm cluster shared energy storage optimization configuration method according to an embodiment of the present application;

[0053] FIG. 2 is a shared energy storage optimization configuration method flowchart of the wind farm cluster shared energy storage optimization configuration method according to an embodiment of the present application;

[0054] Fig. 3 is a system block diagram of an embodiment of the wind farm cluster shared energy storage optimal configuration method provided by the present application. DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative labor fall within the protection scope of the present application.

[0056] Please refer to Fig. 1, which is a flowchart of an embodiment of the power grid equipment fault scene and extreme operation scene generation method provided by the present application, including steps S101-S103, and each step is specifically as follows:

[0057] S101, a shared energy storage double-layer planning model is constructed, wherein the upper model of the double-layer planning model takes the highest economic benefit of shared energy storage as the objective function, and the lower model takes the lowest operation cost of the wind farm cluster as the objective function.

[0058] In the present embodiment, as shown in Fig. 2, a shared energy storage double-layer planning model is constructed, the upper model takes the highest economic benefit of shared energy storage as the objective function to solve the optimal capacity planning of shared energy storage in a long time scale, and the lower model takes the lowest operation cost of the wind farm cluster as the objective function to solve the coordinated optimal operation problem of the wind farm cluster and shared energy storage.

[0059] In some embodiments, the step S101 of “constructing a shared energy storage double-layer planning model” includes but is not limited to the following steps:

[0060] The upper model is constructed with the highest economic benefit of shared energy storage as the objective function, wherein the upper model is:

[0061] In the formula, I R,d is the daily income of wind farm cluster capacity leasing; I DNL,d is the daily income of electricity market; I TP,d is the daily income of frequency modulation auxiliary service market; C in is the investment and construction cost of shared energy storage; C op is the annual operation and maintenance cost of shared energy storage;

[0062] In the formula, is the wind farm cluster leasing power of shared energy storage at the t th time on the d th day, p r,d is the unit power leasing price of shared energy storage on the d th day, P DNL,d(t) is the market clearing power of the electricity energy market shared by the energy storage at the tth time on the dth day, p DNL,d (t) is the market clearing price of the electricity energy market at the tth time. TP,d (t) is the bid capacity of the frequency regulation auxiliary service market shared by the energy storage at the tth time on the dth day, p TP,d (t) is the frequency regulation compensation price of the frequency regulation auxiliary service market at the tth time on the dth day, C p C is the unit power construction cost of the shared energy storage, s N is the unit capacity construction cost of the shared energy storage, y P is the full life cycle of the shared energy storage, max E is the maximum charge-discharge power of the shared energy storage, SES g is the discount rate, h is the number of replacements of the shared energy storage device, n is the total number of replacements of the shared energy storage device in the full life cycle, C pom C is the unit power operation and maintenance cost of the shared energy storage, som W(t) is the full life cycle charge-discharge amount of the shared energy storage;

[0063] The lower model takes the minimum operation cost of the wind farm cluster as the objective function, and the lower model is: d = min(C p +C r -I s )

[0064] In the formula, C p is the daily deviation assessment cost of the wind farm cluster, C r is the daily leasing cost of the wind farm cluster leasing the shared energy storage, I s is the daily electricity sales revenue of the wind farm cluster, p is the actual value of the wind power of the wind farm cluster at the tth time, p is the day-ahead forecast value of the wind power of the wind farm cluster at the tth time, p wind is the unit wind power deviation assessment price, p DNL (t) is the market clearing price of the electricity energy market at the tth time.

[0065] In this embodiment, the upper layer takes the maximum annual economic benefit of the shared energy storage as the objective function to determine the configuration capacity and the maximum charge-discharge power of the shared energy storage. The annual economic benefit includes five parts: the annual revenue of the wind farm cluster capacity leasing, the annual revenue of the electricity energy market, the annual revenue of the frequency regulation auxiliary service market, the construction cost of the shared energy storage (averaged to years in the full life cycle), and the annual operation and maintenance cost of the shared energy storage. Specifically:

[0066] In the formula, I R,d is the wind farm cluster capacity leasing daily income; I DNL,d is the electricity market daily income; I TP,d is the frequency modulation auxiliary service market daily income; C in is the shared energy storage investment construction cost; C op is the shared energy storage annual operation and maintenance cost;

[0067] In the formula, is the shared energy storage wind farm cluster leasing power at the dth day and the tth moment, p r,d is the shared energy storage unit power leasing price at the dth day, P DNL,d (t) is the electricity market out-of-clearing power at the dth day and the tth moment, p DNL,d (t) is the out-of-clearing price of the electricity market at the dth day and the tth moment, P TP,d (t) is the shared energy storage bid capacity in the frequency modulation auxiliary service market at the dth day and the tth moment, p TP,d (t) is the frequency modulation compensation price of the frequency modulation auxiliary service market at the dth day and the tth moment, C p is the shared energy storage unit power investment construction cost, C s is the shared energy storage unit capacity investment construction cost, N y is the shared energy storage full life cycle, P max is the shared energy storage maximum charging and discharging power, E SES is the shared energy storage optimal configuration capacity, g is the discount rate, h is the replacement number of the shared energy storage device, n is the total number of replacements of the shared energy storage device in the full life cycle, C pom is the shared energy storage unit power operation and maintenance cost, C som is the shared energy storage unit capacity operation and maintenance cost, W(t) is the shared energy storage full life cycle charging and discharging amount;

[0068] The lower layer takes the minimum wind farm cluster daily operation cost as the objective function to determine the optimal operation of the wind farm cluster and the shared energy storage. Specifically, f d = min(C p +C r -I s )

[0069] In the formula, C p is the wind farm cluster daily deviation assessment cost, C r is the daily leasing cost of the wind farm cluster leasing shared energy storage, I s is the wind farm cluster daily electricity sales income, is the actual value of the wind power of the wind farm cluster at the tth moment, P is the day-ahead forecast value of the wind power of the wind farm cluster at the tth time point wind P is the unit wind power deviation assessment price DNL (t) is the clearing price of the electricity market at the tth time point.

[0070] S102, according to the wind farm cluster output, the electricity market price and the frequency modulation auxiliary service market compensation price, a multiple uncertainty set is obtained, an improved K-means clustering method is used for joint clustering of the multiple uncertainty set, typical scenarios are obtained, and the typical scenario probabilities are calculated according to the annual occurrence frequencies of the typical scenarios.

[0071] In the embodiment, by considering the influence of the wind farm cluster output, the electricity market price and the frequency modulation auxiliary service market compensation price uncertainty on the shared energy storage upper optimization configuration model, the scene reduction concept is introduced, the improved K-means clustering method is used for joint clustering of the multiple uncertainty in a long time scale, the multiple types of scenarios in the long time scale are reduced to typical scenarios, and the probabilities of different typical scenarios are calculated according to the annual occurrence frequencies of the typical scenarios. The scene quantity (i.e. annual occurrence frequency) corresponding to the multiple typical scenarios determined by the improved K-means clustering method is calculated according to the normalization method.

[0072] In the formula, π k is the probability corresponding to the kth typical scenario; N k is the number of clusters corresponding to the kth typical scenario.

[0073] In some embodiments, the step S102 of “using the improved K-means clustering method to jointly cluster the multiple uncertainty set to obtain typical scenarios” includes but is not limited to the following steps:

[0074] The multiple uncertainty set is extended to a long time scale set, the data in the long time scale set is processed by non-dimensionalization, and the processed data is obtained;

[0075] A plurality of data is randomly selected as clustering centers, the distance difference and the trend difference of all the processed data to each clustering center are calculated, and the comprehensive difference value is obtained according to the distance difference and the trend difference;

[0076] The processed data with a comprehensive difference value less than a preset value is allocated to the corresponding clustering center, a plurality of initial typical scenarios is obtained, the mean value of the multiple uncertainty set in each initial typical scenario is recalculated, the mean value is taken as a new clustering center, the comprehensive difference value of all the processed data to the new clustering center is continuously calculated and allocated, until the clustering center does not change, and a plurality of typical scenarios is obtained.

[0077] In the embodiment, the uncertainties of the wind farm cluster output, the electricity market price and the frequency regulation auxiliary service market compensation price in the shared energy storage bi-level planning model are described as a multiple uncertainty set, wherein the multiple uncertainty set is:

[0078] wherein s is a set of uncertain factors in the shared energy storage bi-level planning model, is the wind power of the wind farm cluster at time t, p DNL (t) is the clearing price of the electricity market at time t, p TP (t) is the frequency regulation compensation price of the frequency regulation auxiliary service market at time t.

[0079] Considering the influence of multiple uncertainties on the upper shared energy storage capacity planning in a long time scale, an improved K-means clustering method is used to jointly cluster the multiple uncertainties, and the multi-type configuration scenarios in the long time scale are reduced to typical scenarios. The multiple uncertainty set is extended to the long time scale to obtain a long time scale set, which is specifically expressed as follows: p DNL,d ={p DNL,d (1),p DNL,d (2),...,p DNL,d (t),...,p DNL,d (T)} p TP,d ={p TP,d (1),p TP,d (2),...,p TP,d (t),...,p TP,d (T)}

[0080] wherein s d is the long time scale basic period representation of the uncertainty set; T is the long time scale basic period of each element of the uncertainty set, T = 24; is the wind power sequence of the wind farm cluster on the dth day; p DNL,d is the clearing price sequence of the electricity market on the dth day; p TP,d is the frequency regulation compensation price sequence of the frequency regulation auxiliary service market on the dth day.

[0081] The specific steps of jointly clustering the long time scale set by using the improved K-means clustering method are as follows:

[0082] S11: Normalize the data in the long time scale set;

[0083] S12: Randomly select K uncertainty set sample data as clustering centers, k ∈ {1, 2,..., K};

[0084] S13: Calculate the difference between all sample data in the uncertainty set and different cluster centers. The difference calculation includes distance difference and trend difference. The comprehensive difference is used for characterization.

[0085] S14: According to the principle of minimum comprehensive difference, all sample data in the uncertainty set are allocated to different cluster centers. The average of the uncertainty set of different typical scenarios is recalculated, and the average of each typical scenario uncertainty set is taken as a new cluster center.

[0086] S15: Repeat S13 and S14 until each cluster center is not changed. K typical scenarios ω are obtained. k .

[0087] In some embodiments, the step S102 of "calculating the distance difference and trend difference between all processed data and each cluster center, and obtaining the comprehensive difference according to the distance difference and the trend difference" includes but is not limited to the following steps:

[0088] The distance difference between all processed data and each cluster center is calculated by using the Euclidean distance, and the trend difference between all processed data and each cluster center is calculated by using the trend difference calculation formula. The calculation formula of the distance difference is:

[0089] In the formula, is the distance difference between the dth processed data and the kth cluster center, is the wind power sequence of the kth cluster center of the multi-uncertainty set wind farm cluster, p DNL,k is the clearing price sequence of the kth cluster center of the multi-uncertainty set electricity market, p TP,k is the frequency modulation compensation price sequence of the kth cluster center of the multi-uncertainty set frequency modulation auxiliary service market;

[0090] The trend difference calculation formula is:

[0091] In the formula, is the trend difference between the dth processed data and the kth cluster center;

[0092] The distance difference and the trend difference between each processed data and the cluster center are normalized by using the exponential respectively, and the distance difference value and the trend difference value are obtained. The calculation formula of the distance difference value is:

[0093] In the formula, γ1(d, k) is the distance difference value between the dth processed data and the kth cluster center;

[0094] The calculation formula of the trend difference value is:

[0095] wherein γ2(d, k) is the distance difference value of the dth processed data to the kth cluster center;

[0096] The distance difference value and the trend difference value are fused by using a proportional coefficient method to obtain a comprehensive difference value of the processed data to each cluster center, and an expression of the comprehensive difference value is γ(d, k) = aγ1(d, k) + βγ2(d, k).

[0097] wherein γ(d, k) is the comprehensive difference value of the processed data to each cluster center, a and β are proportional coefficients, and the values of a and β represent the relative importance of the distance difference value and the trend difference value, and a + β = 1.

[0098] In the embodiment, the distance difference of the data in the dimensionless uncertainty set to different cluster centers is calculated by using the Euclidean distance, and the distance difference is specifically as follows:

[0099] wherein, is the distance difference of the dth processed data to the kth cluster center, is the wind power sequence of the kth cluster center of the multiple uncertainty set wind farm cluster, p DNL,k is the clearing price sequence of the kth cluster center of the multiple uncertainty set electricity market, p TP,k is the frequency modulation compensation price sequence of the kth cluster center of the multiple uncertainty set frequency modulation auxiliary service market;

[0100] The distance difference of the data in the uncertainty set to different cluster centers is normalized by using an index to obtain a distance difference value, and the distance difference value is specifically as follows:

[0101] wherein γ1(d, k) is the distance difference value of the dth uncertainty set data to the kth cluster center.

[0102] The trend difference of the uncertainty set sample data to different cluster centers is calculated by using the following formula, and the trend difference is specifically as follows:

[0103] wherein, is the trend difference of the dth processed data to the kth cluster center;

[0104] The trend difference of the data in the uncertainty set to different cluster centers is normalized by using an index to obtain a trend difference value, and the trend difference value is specifically as follows:

[0105] wherein γ2(d, k) is the trend difference value of the dth uncertainty set data to the kth cluster center.

[0106] The proportional coefficient method is used to fuse distance difference and trend difference to obtain the comprehensive difference of the processed data to each cluster center. The expression for the comprehensive difference is: γ(d,k)=αγ1(d,k)+βγ2(d,k)

[0107] In the formula, γ(d,k) is the comprehensive difference between the processed data and each cluster center, and α and β are both proportionality coefficients, which take the relative importance of the distance difference and the trend difference, and α+β=1.

[0108] S103: The multi-uncertainty set is modeled on a short time scale using the sub-Bruker optimization method to obtain the uncertainty extreme scenario. The shared energy storage two-layer planning model is updated using the typical scenario probability and the uncertainty extreme scenario to obtain the updated shared energy storage two-layer planning model. The configuration result is obtained by solving the particle swarm algorithm.

[0109] In this embodiment, by considering the impact of uncertainties in wind farm cluster output, electricity market prices, and frequency regulation ancillary service market compensation prices on the shared energy storage lower-level optimization operation model, a split-Brow bar optimization method is used to model the multiple uncertainty set on a short time scale, and the worst-case scenario is used to represent the extreme cases of multiple uncertainties. Then, based on the processing results of the multiple uncertainty set, namely the probability of typical scenarios and the uncertainty extreme scenarios, the shared energy storage bi-layer programming model is updated and corrected, and a particle swarm optimization algorithm with accompanying commercial solution software is used for joint solution.

[0110] In some embodiments, step S103, "modeling the multiple uncertainty set using the sub-Bruker optimization method on a short time scale to obtain the uncertainty extreme scenario," includes, but is not limited to, the following steps:

[0111] A robust parameter is used to construct a set of the multiple uncertainties on a short time scale, wherein the set is:

[0112] In the formula, Let be the actual wind power value of the wind farm cluster at time t on day d. Let be the day-ahead forecast of wind power for the wind farm cluster at time t on day d. Let be the maximum value of the wind power fluctuation range of the wind farm cluster at time t on day d. Let p be the distribution bar parameter of the wind power sequence of the wind farm cluster on day d. DNL,d (t) represents the true electricity price at time t in the clearing price sequence of the electricity market on day d. Let t be the day-ahead electricity price forecast for wind power at time t, representing the clearing price sequence of the electricity market on day d. χ DNL,d is the distribution robust parameter of the day-ahead market clearing price sequence in the dth day, p TP,d (t) is the real value of the frequency regulation compensation price sequence in the dth day of the ancillary service market at the tth time, is the day-ahead forecast value of the frequency regulation compensation price sequence in the dth day of the ancillary service market at the tth time, χ TP,d is the distribution robust parameter of the frequency regulation compensation price sequence in the dth day of the ancillary service market;

[0113] The calculation formula of the distribution robust parameter is:

[0114] wherein, μ DNL,d and μ TP,d are the expectations of the sequences and the sequence Y DNL,d and the sequence Y TP,d and are the standard deviations of the sequences and the sequence Y DNL,d and the sequence Y TP,d α DNL,d and α TP,d are the confidence levels of the wind farm cluster output, the electricity market price and the frequency regulation ancillary service market compensation price in the distribution robust optimization process, is the probability distribution function of the standard normal distribution.

[0115] In the embodiment, considering the smoothing effect of the output prediction and the price prediction, the robust parameters are used to construct the uncertainty set of the wind farm cluster output, the electricity market price and the frequency regulation ancillary service market compensation price in the short time scale as:

[0116] wherein, is the real value of the wind power sequence of the wind farm cluster in the dth day at the tth time, is the day-ahead forecast value of the wind power sequence of the wind farm cluster in the dth day at the tth time, is the maximum fluctuation range of the wind power sequence of the wind farm cluster in the dth day at the tth time, is the distribution robust parameter of the wind power sequence of the wind farm cluster in the dth day, p​​DNL,d (t) is the true value of the electricity price of the electricity market in the d-th day at the t-th time point of the clearing price sequence, is the day-ahead forecast value of the wind power price of the electricity market in the d-th day at the t-th time point of the clearing price sequence, is the maximum value of the fluctuation range of the wind power price of the electricity market in the d-th day at the t-th time point of the clearing price sequence, DNL,d is the distribution robustness parameter of the clearing price sequence of the electricity market in the d-th day, TP,d (t) is the true value of the compensation price of the frequency modulation compensation price sequence of the ancillary service market in the d-th day at the t-th time point, is the day-ahead forecast value of the compensation price of the frequency modulation compensation price sequence of the ancillary service market in the d-th day at the t-th time point, is the maximum value of the fluctuation range of the compensation price of the frequency modulation compensation price sequence of the ancillary service market in the d-th day at the t-th time point, TP,d is the distribution robustness parameter of the frequency modulation compensation price sequence of the ancillary service market in the d-th day.

[0117] The set robustness parameters of the wind farm cluster output, the electricity market price and the compensation price of the frequency modulation ancillary service market in the short time scale can be determined by the following formula:

[0118] In the formula, μ DNL,d and μ TP,d are the expectations of the sequences and the sequence Y DNL,d and the sequence Y TP,d and are the standard deviations of the sequences and the sequence Y DNL,d and the sequence Y TP,d α DNL,d and α TP,d are the confidence levels of the wind farm cluster output, the electricity market price and the compensation price of the frequency modulation ancillary service market in the distribution robustness optimization process, is the probability distribution function of the standard normal distribution.

[0119] In some embodiments, the "updated shared energy storage bi-level planning model" in step S103 includes but is not limited to the following steps: f u = max ( π k [I R,d ( ω k ) + I DNL,d ( ω k ) + I TP,d ( ω k ) ] - C​​in -C op )

[0120] In the formula, I R,d (ω k ), I DNL,d (ω k ) and I TP,d (ω k ) are daily income of capacity leasing, daily income of electricity energy market and daily income of frequency modulation auxiliary service market of the wind farm cluster under the kth typical scene respectively, C p is daily deviation evaluation cost of the wind farm cluster under the kth typical scene, C r is daily leasing cost of leasing shared energy storage of the wind farm cluster under the kth typical scene, I s is daily electricity sale income of the wind farm cluster under the kth typical scene, ψ is a multiple uncertainty set of wind farm cluster output and electricity energy market clearing price, is the worst scene of the multiple uncertainty set.

[0121] In the embodiment, the updating and correcting result of the shared energy storage double-layer planning model is: f u = max (π k [I R,d (ω k )+I DNL,d (ω k )+I TP,d (ω k )]-C in -C op )

[0122] In the formula, I R,d (ω k ), I DNL,d (ω k ) and I TP,d (ω k ) are daily income of capacity leasing, daily income of electricity energy market and daily income of frequency modulation auxiliary service market of the wind farm cluster under the kth typical scene respectively, C p is daily deviation evaluation cost of the wind farm cluster under the kth typical scene, C r is daily leasing cost of leasing shared energy storage of the wind farm cluster under the kth typical scene, I s is daily electricity sale income of the wind farm cluster under the kth typical scene, ψ is an uncertainty set of wind farm cluster output and electricity energy market clearing price, is the worst scene of the uncertainty set.

[0123] A bi-level programming model is constructed by building an upper layer optimal shared energy storage capacity configuration and a lower layer wind farm cluster and shared energy cooperative optimization operation model. In a long time scale, the scene reduction concept is introduced, and an improved K-means clustering method is used to jointly cluster the wind farm cluster output, the electricity market price and the frequency regulation auxiliary service market compensation price, so as to reduce the influence of multiple uncertainties on the shared energy storage configuration level. In a short time scale, a distribution robust optimization method is used to model multiple uncertainties, the worst case scenario is used to represent the extreme case of multiple uncertainties, and the shared energy optimization operation strategy is formulated to reduce the influence of multiple uncertainties on the shared energy storage operation level. The multi-time scale uncertainty processing result is returned to the shared energy bi-level programming model for updating and correction, and a particle swarm algorithm is used to solve the problem jointly with a commercial solver to realize the optimal configuration capacity of the wind farm cluster leasing shared energy storage.

[0124] It should be understood that although each step in the flowchart of FIG. 1 is shown in sequence according to the direction of the arrow, these steps are not necessarily executed in sequence according to the direction of the arrow. Unless otherwise explicitly stated herein, there is no strict order limitation for the execution of these steps, and these steps can be executed in other orders. Moreover, at least part of the steps in FIG. 1 can include multiple steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least part of other steps or steps or stages in other steps.

[0125] In some embodiments, as shown in FIG. 3, a block diagram of a wind farm cluster shared energy storage optimization configuration system 300 provided by the embodiments of the present application is shown, which includes a construction module 301, a typical scene calculation module 302 and a configuration module 303, wherein:

[0126] The construction module is configured to construct a shared energy storage bi-level programming model, wherein the upper layer model of the bi-level programming model takes the highest economic benefit of the shared energy storage as the objective function, and the lower layer model takes the lowest operation cost of the wind farm cluster as the objective function.

[0127] The typical scene calculation module is configured to obtain a multiple uncertainty set according to the wind farm cluster output, the electricity market price and the frequency regulation auxiliary service market compensation price, perform joint clustering on the multiple uncertainty set by using an improved K-means clustering method to obtain a typical scene, and calculate a typical scene probability according to the annual occurrence frequency of the typical scene.

[0128] The configuration module is configured to model a plurality of uncertainty sets on a short time scale by using a distributionally robust optimization method to obtain uncertainty extreme scenarios, update a shared energy storage bi-level programming model by using typical scenario probabilities and the uncertainty extreme scenarios to obtain an updated shared energy storage bi-level programming model, and solve the updated shared energy storage bi-level programming model by using a particle swarm algorithm to obtain a configuration result.

[0129] The specific implementation of the wind farm cluster shared energy storage optimization configuration system is basically the same as that of the above-described specific embodiment of the wind farm cluster shared energy storage optimization configuration method, and thus will not be described herein.

[0130] In an embodiment of the present application, a computer device is provided, which includes a memory and a processor, the memory stores a computer program, and the processor implements the above steps when executing the computer program; the computer device provided in the embodiment has similar implementation principles and technical effects to the above-described method embodiments, and thus will not be described herein.

[0131] In an embodiment of the present application, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to implement the above steps; the computer readable storage medium provided in the embodiment has similar implementation principles and technical effects to the above-described method embodiments, and thus will not be described herein.

[0132] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing related hardware, and the computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, the computer program can include the processes of the above-mentioned embodiments. Any reference to a memory, storage, database or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0133] Any combination of the technical features in the above embodiments can be made, and for the sake of brevity, not all possible combinations are described above. However, it should be understood that any combination of the technical features is within the scope of the present disclosure.

[0134] The above-described specific embodiments further illustrate the objects, technical solutions, and advantages of the present application. It should be understood that the above-described specific embodiments are merely illustrative of the present application and are not intended to limit the scope of protection of the present application. It is specifically pointed out that any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for optimizing configuration of energy storage shared by wind farm clusters, characterized in that, The application relates to a shared energy storage double-layer planning model, wherein an upper-layer model of the double-layer planning model takes the highest economic benefit of shared energy storage as an objective function, and a lower-layer model of the double-layer planning model takes the lowest operation cost of a wind power plant cluster as an objective function. A plurality of uncertain sets are obtained according to wind power plant cluster output, electricity market price and frequency modulation auxiliary service market compensation price, the plurality of uncertain sets are jointly clustered by using an improved K-means clustering method to obtain typical scenes, and the typical scene probability is calculated according to the annual occurrence frequency of the typical scenes. The plurality of uncertain sets are modeled in a short time scale by using a distribution robust optimization method to obtain uncertain extreme scenes, the typical scene probability and the uncertain extreme scenes are used to update the shared energy storage double-layer planning model to obtain an updated shared energy storage double-layer planning model, and a particle swarm algorithm is used to solve the updated shared energy storage double-layer planning model to obtain a configuration result. The shared energy storage double-layer planning model comprises: 2.The wind farm cluster shared energy storage optimization configuration method of claim 1, wherein, The lower-layer model takes the lowest operation cost of the wind power plant cluster as an objective function, and the lower-layer model is: An upper model is constructed with a shared energy storage economic benefit as an objective function, wherein the upper model is: In the formula, I R,d is the wind farm cluster capacity leasing daily income; I DNL,d is the electricity market daily income; I TP,d is the frequency modulation auxiliary service market daily income; C in is the shared energy storage investment construction cost; C op is the shared energy storage annual operation and maintenance cost; In the formulae, Pd,t is the wind farm cluster rental power of the shared energy storage at the tth time of the dth day r,d Pd is the unit power rental price of the shared energy storage at the dth day DNL,d Pd,t is the electricity market dispatch power of the shared energy storage at the tth time of the dth day DNL,d Pd,t is the dispatch price of the electricity market at the tth time of the dth day TP,d Pd,t is the bid capacity of the shared energy storage in the frequency modulation auxiliary service market at the tth time of the dth day TP,d C d,t is the frequency modulation compensation price of the frequency modulation auxiliary service market at the tth time of the dth day p C is the unit power construction cost of the shared energy storage s N is the unit capacity construction cost of the shared energy storage y P is the full life cycle of the shared energy storage max E is the maximum charging and discharging power of the shared energy storage SES g is the discount rate, h is the replacement number of the shared energy storage device, n is the total number of replacements of the shared energy storage device in the full life cycle, C is the unit power operation and maintenance cost of the shared energy storage pom C is the unit capacity operation and maintenance cost of the shared energy storage som W(t) is the full life cycle charging and discharging quantity of the shared energy storage The plurality of uncertain sets are jointly clustered by using an improved K-means clustering method to obtain typical scenes, and the typical scene probability is calculated according to the annual occurrence frequency of the typical scenes. f d = min(C p + C r - I s ) In the formula, C p is the wind farm cluster daily deviation assessment cost, C r is the daily leasing cost of the wind farm cluster leasing shared energy storage, I s is the daily electricity sales revenue of the wind farm cluster, Pwind(t) is the actual value of the wind power of the wind farm cluster at the tth moment, is the day-ahead forecast value of the wind power of the wind farm cluster at the tth time, p wind is the unit wind power deviation assessment price, p DNL (t) is the clearing price of the electricity energy market at the tth time. 3.The method of claim 1, wherein, The plurality of uncertain sets are extended to a long time scale set, and the data in the long time scale set are processed by non-dimensionalization to obtain processed data. A plurality of data are randomly selected as clustering centers, the distance difference and the trend difference of all the processed data to each clustering center are calculated, and the comprehensive difference value is obtained according to the distance difference and the trend difference. The processed data with a comprehensive difference value less than a preset value are distributed to corresponding clustering centers to obtain a plurality of initial typical scenes, the mean value of the plurality of uncertain sets in each initial typical scene is recalculated, the mean value is taken as a new clustering center, the comprehensive difference value of all the processed data to the new clustering center is continuously calculated and distributed until the clustering center does not change, and a plurality of typical scenes are obtained. The distance difference value and the trend difference value of all the processed data to each clustering center are calculated, and the comprehensive difference value is obtained according to the distance difference and the trend difference. 4.The method of claim 3, wherein, The trend difference of the dth processed data to the kth clustering center is The distance difference of all processed data to each cluster center is calculated by using the Euclidean distance, and the trend difference of all processed data to each cluster center is calculated by using a trend difference calculation formula, wherein the calculation formula of the distance difference is: In the formulae, distance difference of data after dth day processing to kth cluster center, p is the wind power sequence of the kth cluster center of the multiple-uncertainty set wind farm cluster DNL,k p is the clearing price sequence of the kth cluster center of the multiple-uncertainty set electricity energy market TP,k p is the frequency regulation compensation price sequence of the kth cluster center of the multiple-uncertainty set frequency regulation auxiliary service market The trend difference calculation formula is: In the formulae, In the formula, gamma1(d, k) is the distance difference value of the dth processed data to the kth clustering center. The distance difference value and the trend difference value are obtained by normalizing the distance difference and the trend difference of each processed data to the cluster center by using an exponential, wherein a calculation formula of the distance difference value is: In the formula, gamma2(d, k) is the trend difference value of the dth processed data to the kth clustering center. The trend difference is calculated by the following formula: The distance difference value and the trend difference value are fused by using a proportional coefficient method to obtain the comprehensive difference value of the processed data to each clustering center, and the expression of the comprehensive difference value is: gamma(d, k) = alpha * gamma1(d, k) + beta * gamma2(d, k) In the formula, gamma(d, k) is the comprehensive difference value of the processed data to each clustering center, alpha and beta are proportional coefficients, the relative importance degrees of the distance difference value and the trend difference value are taken as the values of alpha and beta, and alpha + beta = 1. The plurality of uncertain sets are modeled in a short time scale by using a distribution robust optimization method to obtain uncertain extreme scenes, and the typical scene probability and the uncertain extreme scenes are used to update the shared energy storage double-layer planning model to obtain an updated shared energy storage double-layer planning model, and a particle swarm algorithm is used to solve the updated shared energy storage double-layer planning model to obtain a configuration result. 5.The method of claim 1, wherein, The probability distribution function is a standard normal distribution. constructing the set of multiple uncertainties using robust parameters at a short timescale, wherein the set is: In the formulae, Yd(t) is the real value of the wind power of the wind farm cluster at the tth time of the wind power sequence on the dth day, a day-ahead forecast value of the wind power at the tth time of the wind power sequence of the dth day of the wind farm cluster, the maximum value of the wind power fluctuation range of the wind power sequence of the wind farm cluster at the tth moment of the dth day, p is the distribution robust parameter of the wind power sequence of the wind farm cluster on day d DNL,d (t) is the true value of the electricity price of the tth time of the dispatch price sequence of the electricity market on day d, a wind power day-ahead electricity price forecast value for the out-cleared price sequence of the electricity market at the tth time instant of the dth day, is the maximum value of the wind power fluctuation range at the tth time of the day-d clearing price sequence of the electricity market, DNL,d is the distribution robustness parameter of the day-d clearing price sequence of the electricity market, TP,d (t) is the real value of the compensation price at the tth time of the day-d frequency modulation compensation price sequence of the ancillary service market, a day-ahead forecast value of the frequency compensation price of the auxiliary service market at the tth time of the frequency compensation price sequence of the dth day, χd,t is the maximum fluctuation range of the compensation price of the frequency regulation compensation price sequence of the auxiliary service market at the tth time on the dth day TP,d χd is the distribution robustness parameter of the frequency regulation compensation price sequence of the auxiliary service market on the dth day The calculation formula of the distribution robust parameter is: In the formulae, μ DNL,d and μ TP,d are the sequences Sequence Y DNL,d and Sequence Y TP,d desirably; and SEQ ID NO: 1, SEQ ID NO: 2, respectively the sequence Y DNL,d and the sequence Y TP,d the standard deviation of the sequence Y; α DNL,d and α TP,d are the confidence levels of wind farm cluster output, electricity market price and frequency regulation auxiliary service market compensation price in the distribution robust optimization process, respectively, ​ 6.The method of claim 1, wherein, The updated shared energy storage bi-level programming model is: u = max (π k [ I R,d (ω k )+I DNL,d (ω k )+I TP,d (ω k )]-C in -C op ) In the formula, I R,d (ω k ), I DNL,d (ω k ) and I TP,d (ω k ) are the daily income of wind farm cluster capacity leasing, daily income of electricity market and daily income of frequency modulation auxiliary service market under the kth typical scenario respectively, C p is the daily deviation evaluation cost of wind farm cluster under the kth typical scenario, C r is the daily leasing cost of wind farm cluster leasing shared energy storage under the kth typical scenario, I s is the daily electricity sale income of wind farm cluster under the kth typical scenario, and ψ is the multiple uncertainty set of wind farm cluster output and electricity market clearing price. The worst scenario of the multiple uncertainty sets.

7. A wind farm cluster shared energy storage optimization configuration system, characterized in that, The configuration module is configured to model the multiple uncertainty sets in a short time scale by using a distributionally robust optimization method to obtain an uncertainty extreme scenario, update the shared energy storage bi-level planning model by using the typical scenario probability and the uncertainty extreme scenario to obtain an updated shared energy storage bi-level planning model, and solve the updated shared energy storage bi-level planning model by using a particle swarm algorithm to obtain a configuration result. The configuration module is configured to model the multiple uncertainty sets in a short time scale by using a distributionally robust optimization method to obtain an uncertainty extreme scenario, update the shared energy storage bi-level planning model by using the typical scenario probability and the uncertainty extreme scenario to obtain an updated shared energy storage bi-level planning model, and solve the updated shared energy storage bi-level planning model by using a particle swarm algorithm to obtain a configuration result. The configuration module is configured to model the multiple uncertainty sets in a short time scale by using a distributionally robust optimization method to obtain an uncertainty extreme scenario, update the shared energy storage bi-level planning model by using the typical scenario probability and the uncertainty extreme scenario to obtain an updated shared energy storage bi-level planning model, and solve the updated shared energy storage bi-level planning model by using a particle swarm algorithm to obtain a configuration result. The configuration module is configured to model the multiple uncertainty sets in a short time scale by using a distributionally robust optimization method to obtain an uncertainty extreme scenario, update the shared energy storage bi-level planning model by using the typical scenario probability and the uncertainty extreme scenario to obtain an updated shared energy storage bi-level planning model, and solve the updated shared energy storage bi-level planning model by using a particle swarm algorithm to obtain a configuration result. 8.The wind farm cluster shared energy storage optimization configuration system of claim 7, wherein, The configuration module is configured to model the multiple uncertainty sets in a short time scale by using a distributionally robust optimization method to obtain an uncertainty extreme scenario, update the shared energy storage bi-level planning model by using the typical scenario probability and the uncertainty extreme scenario to obtain an updated shared energy storage bi-level planning model, and solve the updated shared energy storage bi-level planning model by using a particle swarm algorithm to obtain a configuration result. An upper model is constructed with a shared energy storage economic benefit as an objective function, wherein the upper model is: In the formula, I R,d is the wind farm cluster capacity leasing daily income; I DNL,d is the electricity market daily income; I TP,d is the frequency modulation auxiliary service market daily income; C in is the shared energy storage investment construction cost; C op is the shared energy storage annual operation and maintenance cost; In the formulae, For the shared energy storage of the wind farm cluster leased at time t on day d, p r,d P is the unit power rental price for shared energy storage on day d. DNL,d (t) represents the power of shared energy storage clearing out of the energy market at time t on day d, p DNL,d (t) represents the clearing price of the electricity market at time t on day d, P TP,d (t) represents the winning bid capacity of shared energy storage in the frequency regulation ancillary service market at time t on day d, p TP,d (t) represents the frequency modulation compensation price in the frequency modulation ancillary service market at time t on day d, and C p To share the investment and construction cost per unit power of energy storage, C s To share the investment and construction costs per unit capacity of energy storage, N y To share the entire lifecycle of energy storage, P max To share the maximum charge and discharge power of energy storage, E SES To optimize the configuration capacity of shared energy storage, g is the discount rate, h is the number of times the shared energy storage device can be replaced, n is the total number of times the shared energy storage device can be replaced throughout its entire life cycle, and C is the value of C. pom To share the unit power operation and maintenance cost of energy storage, C som W(t) represents the unit capacity operation and maintenance cost of shared energy storage, and W(t) represents the total charge and discharge capacity of shared energy storage over its entire life cycle. The configuration module is configured to model the multiple uncertainty sets in a short time scale by using a distributionally robust optimization method to obtain an uncertainty extreme scenario, update the shared energy storage bi-level planning model by using the typical scenario probability and the uncertainty extreme scenario to obtain an updated shared energy storage bi-level planning model, and solve the updated shared energy storage bi-level planning model by using a particle swarm algorithm to obtain a configuration result. f d = min(C p + C r - I s ) In the formula, C p is the wind farm cluster daily deviation assessment cost, C r is the daily leasing cost of the wind farm cluster leasing shared energy storage, I s is the daily electricity sales revenue of the wind farm cluster, Pwind(t) is the actual value of the wind power of the wind farm cluster at the tth moment, is the day-ahead forecast value of the wind power of the wind farm cluster at the tth time, p wind is the unit wind power deviation assessment price, p DNL (t) is the clearing price of the electricity energy market at the tth time.

9. A computer device, comprising: The memory is configured to store a computer program. The processor is configured to execute the computer program to implement the wind farm cluster shared energy storage optimization configuration method according to any one of claims 1 to 6. The storage medium stores a computer program, and the computer program is executed by the processor to implement the steps of the wind farm cluster shared energy storage optimization configuration method according to any one of claims 1 to 6.

10. A storage medium, characterized by ​

Images