Multi-energy storage joint dispatching method and equipment adapted to peak loads in extreme scenarios

Abstract

The present application discloses a multi-energy storage joint scheduling method and device that adapts to peak loads in extreme scenarios, and relates to the field of energy storage scheduling. The method includes: using a clustering algorithm oriented to peak-to-valley difference characteristics to generate an extreme scenario set; using a fast convex hull algorithm to generate an extreme scenario uncertainty set; using a multi-objective generalized normal distribution optimization algorithm to generate a diversified energy storage power station aggregation model; based on the diversified energy storage power station aggregation model and the extreme scenario uncertainty set, a robust optimization model of wind, solar, thermal and storage in the region is constructed, and the black-winged kite optimization algorithm is used to solve the charging and discharging power of the diversified energy storage power station; according to the charging and discharging power of the diversified energy storage power station, a microhabitat-based gradient directed evolution algorithm and a scheduling plan prediction model are used to generate an internal joint scheduling plan for the diversified energy storage power station. The present application can realize rapid decision-making of the internal joint scheduling plan of the diversified energy storage power station under extreme scenarios, so as to ensure the stable supply capacity of the power grid under extreme scenarios.

Description

Technical Field

[0001] The present application relates to the field of energy storage scheduling, and in particular to a multi-energy storage joint scheduling method and device that is adaptable to peak loads in extreme scenarios. Background Art

[0002] With the continuous large-scale grid connection of new energy, the proportion of new energy electricity has increased year by year. The output of new energy fluctuates violently under the frequent occurrence of extreme weather. At the same time, the rapid development of the tertiary industry has further widened the peak-to-valley difference of electricity load. The power grid under the influence of two-way uncertainty is facing a tense situation of power supply and demand balance. At present, new energy storage represented by electrochemistry has the characteristics of bidirectional rapid charging and discharging and energy transfer. In conjunction with the compressed air energy storage system, it provides guarantees for the reliable and stable development of new power systems in terms of charging and discharging power and charging and discharging time. Under the current situation, the mechanism of diversified energy storage power stations supporting stable power supply of the power grid in extreme scenarios is still unclear. At the same time, the different scales of energy storage systems inside the diversified energy storage power stations are relatively independent at the model level, lack of method guidance at the joint dispatch level, and lack of effective means to conduct model aggregation equivalence in terms of external output characteristics of energy storage power stations. Therefore, it is necessary to study a diversified energy storage joint dispatching method that adapts to peak loads in extreme scenarios to ensure the stable supply capacity of the power grid in extreme scenarios. Summary of the invention

[0003] The purpose of this application is to provide a multi-energy storage joint scheduling method and equipment that can adapt to peak loads in extreme scenarios, which can realize rapid decision-making of the internal joint scheduling plan of the diversified energy storage power station under extreme scenarios, so as to ensure the stable supply capacity of the power grid under extreme scenarios.

[0004] To achieve the above objectives, this application provides the following solutions.

[0005] In the first aspect, the present application provides a multi-energy storage joint scheduling method that adapts to peak loads in extreme scenarios, which is applied to diversified energy storage power stations, including: based on the wind power output, photovoltaic output and electric load data in the historical time period, a clustering algorithm oriented to the peak-to-valley difference characteristics is used to generate an extreme scenario set; based on the wind power output, photovoltaic output and electric load data under the extreme scenario set, a fast convex hull algorithm is used to generate an extreme scenario uncertainty set; based on the general energy storage model expressed in the form of state of charge, a multi-objective generalized normal distribution optimization algorithm is used to generate an aggregation model of a diversified energy storage power station; based on the aggregation model of the diversified energy storage power station and the extreme scenario uncertainty set, a robust optimization model of wind, solar, thermal and storage in the region is constructed, and the black kite optimization algorithm is used to solve the charging and discharging power of the diversified energy storage power station; based on the charging and discharging power of the diversified energy storage power station, a microhabitat-based gradient directed evolution algorithm and a scheduling plan prediction model are used to generate an internal joint scheduling plan of the diversified energy storage power station; the scheduling plan prediction model is constructed based on a convolutional neural network, an enhanced sparrow search algorithm and a bidirectional gated recurrent unit.

[0006] Optionally, based on the wind power output, photovoltaic output and electric load data of the historical time period, a clustering algorithm oriented to the peak-valley difference characteristics is used to generate an extreme scenario set, including: obtaining the wind power output, photovoltaic output and electric load data of the historical time period; constructing an indicator set of wind power characteristics, photovoltaic characteristics and electric load characteristics based on the wind power output, photovoltaic output and electric load data of the historical time period; clustering is performed using an improved k-means clustering algorithm based on the indicator set of wind power characteristics, photovoltaic characteristics and electric load characteristics to obtain a clustering division result; the improved k-means clustering algorithm is based on the k-means clustering algorithm, and uses the clustering results of the small sample set as the initial clustering center of the full sample set to perform clustering division on the full sample set to obtain a clustering division result; the full sample set is an indicator set of the wind power characteristics, photovoltaic characteristics and electric load characteristics; the small sample set is obtained by extracting a set proportion from the full sample set; and an extreme scenario set is generated based on the clustering division result.

[0007] Optionally, based on the wind power output, photovoltaic output and electric load data under the extreme scenario set, a fast convex hull algorithm is used to generate an extreme scenario uncertainty set, including: based on the wind power output, photovoltaic output and electric load data under the extreme scenario set, a kernel density estimation method is used to construct the wind power output, photovoltaic output and electric load probability density functions under the extreme scenario; based on the wind power output, photovoltaic output and electric load probability density functions under the extreme scenario, a Monte Carlo sampling method is used for random sampling to generate a random scenario set; based on the random scenario set, a fast convex hull algorithm is used to generate an extreme scenario uncertainty set.

[0008] Optionally, based on the general energy storage model in the form of state of charge expression, a multi-objective generalized normal distribution optimization algorithm is used to generate a diversified energy storage power station aggregation model, including: based on the general energy storage model in the form of state of charge expression, a multi-objective diversified energy storage power station aggregation model is constructed; using a multi-objective generalized normal distribution optimization algorithm to solve the multi-objective diversified energy storage power station aggregation model to obtain a Pareto optimal solution set; based on the Pareto optimal solution set, a compromise solution strategy considering fairness is used to transform the multi-objective diversified energy storage power station aggregation model to generate a diversified energy storage power station aggregation model.

[0009] Optionally, a robust optimization model of wind, solar, thermal and storage in the region is constructed based on the aggregation model of diversified energy storage power stations and the uncertainty set of extreme scenarios, and the Black Kite optimization algorithm is used to solve the charging and discharging power of the diversified energy storage power station, including: constructing a robust optimization model of wind, solar, thermal and storage in the region based on the aggregation model of diversified energy storage power stations and the uncertainty set of extreme scenarios; converting the robust optimization model of wind, solar, thermal and storage in the region into a maximization-minimum problem model; using strong duality theory to convert the maximization-minimum problem model into a single-layer maximization problem model; using the Black Kite optimization algorithm to solve the single-layer maximization problem model to obtain the charging and discharging power of the diversified energy storage power station.

[0010] Optionally, according to the charging and discharging power of the diversified energy storage power station, a microhabitat-based gradient directed evolution algorithm and a scheduling plan prediction model are used to generate an internal joint scheduling plan for the diversified energy storage power station, including: taking the maximum capacity of the diversified energy storage power station as the upper limit, using a random function to generate a number of target charging powers and target discharging powers, and generating corresponding scheduling scenarios according to the target charging power and target discharging power; constructing an internal joint scheduling model for the diversified energy storage power station with the goal of meeting the charging power and discharging power of the diversified energy storage power station; using a microhabitat-based gradient directed evolution algorithm to solve the internal joint scheduling model for the diversified energy storage power station to obtain target scheduling plans under different scheduling scenarios; using the target charging power, target discharging power and target scheduling plans under the corresponding scheduling scenarios of the diversified energy storage power station as basic data, generating a neural network training data set; constructing a scheduling plan prediction model based on a convolutional neural network, an enhanced sparrow search algorithm and a bidirectional gated recurrent unit; using the neural network training data set to train the scheduling plan prediction model to obtain a trained scheduling plan prediction model; inputting the charging and discharging power of the diversified energy storage power station into the trained scheduling plan prediction model to obtain an internal joint scheduling plan for the diversified energy storage power station.

[0011] Optionally, the diversified energy storage power station includes an electrochemical energy storage system and a compressed air energy storage system; the target scheduling plan is a joint scheduling plan of the electrochemical energy storage system and the compressed air energy storage system within the diversified energy storage power station.

[0012] Optionally, the electrochemical energy storage system includes a solenoid valve, a circulation pump and a battery reactor; the compressed air energy storage system includes a compression unit, an expansion unit, a gas storage tank and a heat storage tank; with the goal of meeting the charging power and discharging power of the diversified energy storage power station, a diversified energy storage power station internal joint scheduling model is constructed, including: based on the loss of the solenoid valve and the circulation pump and the loss of the battery reactor, a power loss model of the electrochemical energy storage system is constructed; based on the self-discharge problem of the reduction of power during the storage period after the battery is charged, a self-discharge rate coefficient is introduced to construct an equivalent model of the electrochemical energy storage system; based on the operation of the compression unit, the expansion unit, the gas storage tank and the heat storage tank, an operation model of the compressed air energy storage system is constructed; with the goal of meeting the charging power and discharging power of the diversified energy storage power station, the power loss model of the electrochemical energy storage system, the equivalent model of the electrochemical energy storage system, the operation model of the compressed air energy storage system and the relationship between the charging power and the discharging power are used as constraints to construct a diversified energy storage power station internal joint scheduling model.

[0013] Optionally, the relationship between the charging power and the discharging power is that the product of the charging power and the discharging power is equal to 0.

[0014] In a second aspect, the present application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the multi-energy storage joint scheduling method adapted to peak loads in extreme scenarios.

[0015] According to the specific embodiments provided in the present application, the present application discloses the following technical effects: the present application provides a multi-energy storage joint scheduling method and equipment that adapts to peak loads in extreme scenarios. According to the wind power output, photovoltaic output and electric load data in the historical time period, a clustering algorithm oriented to the peak-to-valley difference characteristics is used to generate an extreme scenario set, which can overcome the problem of inconsistent results of traditional algorithms and stabilize the algorithm effect; according to the wind power output, photovoltaic output and electric load data under the extreme scenario set, a fast convex hull algorithm is used to generate an extreme scenario uncertainty set, which can consider the uncertainty factors within the extreme scenario set and improve the generation efficiency of the uncertainty set; based on the general energy storage model expressed in the form of state of charge, a multi-objective generalized normal distribution optimization is adopted. The algorithm generates an aggregation model of diversified energy storage power stations, which can accelerate the speed of solving multi-objective problems; based on the aggregation model of diversified energy storage power stations and the uncertainty set of extreme scenarios, a robust optimization model of wind, solar, thermal and storage in the region is constructed, and the Black Kite optimization algorithm is used to solve the charging and discharging power of the diversified energy storage power station, which can improve the global search capability and convergence speed of the algorithm; according to the charging and discharging power of the diversified energy storage power station, the microhabitat-based gradient directed evolution algorithm and the scheduling plan prediction model are used to generate the internal joint scheduling plan of the diversified energy storage power station, which can prevent premature convergence and fall into local solutions, enhance the global search capability, and realize rapid decision-making of the internal joint scheduling plan of the diversified energy storage power station under extreme scenarios, thereby ensuring the stable supply capacity of the power grid under extreme scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 The main flow chart of the multi-energy storage joint dispatching method adapted to peak loads in extreme scenarios provided in this application.

[0018] Figure 2 A detailed schematic diagram of a multi-energy storage joint dispatching method adapted to peak loads in extreme scenarios provided in this application. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0020] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0021] In an exemplary embodiment, the present application provides a multi-energy storage joint dispatching method adapted to peak loads in extreme scenarios. Figure 1 and Figure 2 As shown, the method includes the following steps 1 to 5.

[0022] Step 1: Based on the wind power output, photovoltaic output and power load data in the historical time period, a clustering algorithm oriented to the peak-to-valley difference characteristics is used to generate an extreme scenario set, which specifically includes the following steps 1-1 to 1-4.

[0023] Step 1-1: Obtain wind power output, photovoltaic output and power load data for the historical period.

[0024] Step 1-2: Based on the wind power output, photovoltaic output and electric load data in the historical time period, construct an indicator set of wind power characteristics, photovoltaic characteristics and electric load characteristics.

[0025] Both wind power characteristic indicators and photovoltaic characteristic indicators include maximum output peak-to-valley difference, average output and maximum output fluctuation.

[0026] (1)

[0027] (2)

[0028] (3)

[0029] Where: , and They are the peak-to-valley difference, average output and maximum output fluctuation of wind power and photovoltaic power respectively. is the output of wind power and photovoltaic power in the t period, is the output of wind power and photovoltaic power in the t+1 period, is the total scheduling period, max( ) means taking the maximum value, and min( ) means taking the minimum value.

[0030] The electrical load characteristic indicators include the maximum load peak-to-valley difference, average load and maximum load fluctuation.

[0031] (4)

[0032] (5)

[0033] (6)

[0034] Where: , and They are the maximum load peak-to-valley difference, average load and maximum load fluctuation of the electric load, is the electric load in the t period, It is the electric load situation in the t+1 period.

[0035] Step 1-3: Based on the index set of wind power characteristics, photovoltaic characteristics and electric load characteristics, the improved k-means clustering algorithm is used for clustering to obtain the clustering division result. The improved k-means clustering algorithm is based on the k-means clustering algorithm, and the clustering result of the small sample set is used as the initial clustering center of the full sample set to perform clustering division on the full sample set to obtain the clustering division result; the full sample set is the index set of the wind power characteristics, photovoltaic characteristics and electric load characteristics; the small sample set is obtained by extracting a set proportion from the full sample set. The set proportion is preferably 10%.

[0036] The data within a day is taken as a unit, and one time period is represented by one data point. One day can be divided into 24 time periods by hour. In the clustering process, a small sample set of 10% is first extracted from the full sample set; then the first cluster center is selected from the small sample set by a random method; thereafter, the samples farther away from the cluster center have the priority to be selected, which can ensure that as many points with more obvious features as possible can be selected during the selection process, and the clustering results can reflect the characteristics of the original samples as much as possible, ensuring the validity of the calculation results; the clustering results obtained from the small sample set are used as the initial cluster center of the full sample set, and then the full sample set is clustered, which overcomes the problem of inconsistent results obtained by the traditional k-means clustering algorithm.

[0037] The definition formula of the Euclidean distance based on the indicator set is as follows.

[0038] (7)

[0039] Where: is the Euclidean distance between sample No. m and sample No. n, and are the maximum peak-to-valley differences of the mth and nth samples, respectively. and are the average values ​​of samples No. m and No. n, respectively. and are the maximum fluctuation values ​​of samples No. m and No. n respectively.

[0040] Solve the Euclidean distance between the samples of each cluster and their cluster centers to ensure that the sum of the Euclidean distances between the samples and their cluster centers is minimal. After the clustering results are divided, the distance between the samples divided into a cluster and the cluster center is smaller than the distance to other cluster centers, and the index characteristics of the samples in each cluster are similar.

[0041] Step 1-4: Generate an extreme scene set based on the clustering results. Specifically, based on the clustering results, calculate the scene effects of each group of clustered samples, and use the group of clustered samples with the largest scene effect as the calculation scene set.

[0042] According to the clustering results, the scene effect of the clustered samples is calculated using formula (8), and the item with the largest result is selected as the extreme scene set.

[0043] (8)

[0044] Where: Characterize the scene effect, is the number of samples after clustering.

[0045] Step 2: Based on the wind power output, photovoltaic output and electric load data under the extreme scenario set, a fast convex hull algorithm is used to generate an extreme scenario uncertainty set, which specifically includes the following steps 2-1 to 2-3.

[0046] Step 2-1: Based on the wind power output, photovoltaic output and electric load data under the extreme scenario set, a kernel density estimation method is used to construct the probability density functions of wind power output, photovoltaic output and electric load under the extreme scenario. The kernel density estimation method is preferably a non-parametric kernel density estimation method.

[0047] According to the wind power output, photovoltaic output and electric load data under the extreme scenario set, taking the same time period as the benchmark and the Gaussian function as the kernel function, the non-parametric kernel density estimation method is used to establish the wind power output probability density function, photovoltaic output probability density function and electric load probability density function under extreme scenarios respectively.

[0048] (9)

[0049] Where: is a Gaussian function, , and They are wind power output probability density function, photovoltaic output probability density function and electric load probability density function, respectively. To contribute to wind power, is the wind power output under the nth sample, For photovoltaic power output, is the photovoltaic output under the nth sample, is the electrical load, is the electrical load under the nth sample, is the number of samples in the extreme scene set.

[0050] Step 2-2: Based on the probability density functions of wind power output, photovoltaic output and electric load under extreme scenarios, random sampling is performed using the Monte Carlo sampling method to generate a random scenario set.

[0051] Based on the above probability density function, the Monte Carlo sampling method is used for random sampling to generate random scenes and form a random scene set.

[0052] Step 2-3: Based on the random scene set, the fast convex hull algorithm is used to generate the extreme scene uncertainty set.

[0053] Based on the random scene set in the previous step, the fast convex hull algorithm is used to generate the uncertainty set of extreme scenes, that is, the three-dimensional uncertainty convex hull at each time period .

[0054] (10)

[0055] Where: For elements belonging to symbols.

[0056] The specific implementation process of the fast convex hull algorithm is as follows: In the initialization stage, the algorithm scans all data points and selects four extreme points to construct an initial tetrahedron, and removes the points inside the tetrahedron. After that, the remaining external points are iteratively processed. During the iteration, the internal points are deleted and the remaining points in each face are redistributed to its sub-faces. After the iteration is completed, a convex polyhedron can be obtained.

[0057] At the same time, in order to further simplify the convex hull of the uncertainty set, the boundary points of the convex polyhedron are taken as the selected set, and an 8-point 6-face convex space is constructed to represent the uncertainty set. With the maximum space coverage as the goal, the simplified three-dimensional uncertainty convex hull is obtained. .

[0058] (11)

[0059] Where: is the spatial coverage rate in the tth period. and They are the volumes of the three-dimensional uncertainty convex hull before and after simplification respectively.

[0060] This method takes into account the uncertainty factors within the extreme scenario set interval, combines the sampling method to more comprehensively retain the uncertainty interval range, comprehensively balances robustness and optimization benefits, and uses a fast convex hull optimization algorithm to improve the generation efficiency of the uncertainty set.

[0061] Step 3: Based on the general energy storage model in the form of charge state expression, a multi-objective generalized normal distribution optimization algorithm is used to generate a diversified energy storage power station aggregation model, which specifically includes the following steps 3-1 to 3-3.

[0062] Step 3-1: Based on the general energy storage model in the form of state of charge expression, a multi-objective diversified energy storage power station aggregation model is constructed.

[0063] A diversified energy storage power station aggregation model is constructed based on the energy storage general model expressed in the form of State of Charge (SoC). The energy storage general model expressed in the form of SoC is as follows.

[0064] (12)

[0065] Where: and They represent the charging power and discharging power of the energy storage in the tth period respectively, and are the charging efficiency and the discharging efficiency, and are the rated capacity and rated energy of energy storage respectively, is the state of charge of the energy storage at time period t, is the state of charge of the energy storage at time period t-1, and are the minimum state of charge and the maximum state of charge, respectively. and are the charge states at the initial and final moments, respectively. Indicates a time interval.

[0066] When aggregating diversified energy storage power station models, the goal is to minimize the expected net load fluctuations and maximize the regulation capacity of the energy storage power station; the operation of the electrochemical energy storage system and the compressed air energy storage system meets the above SoC constraints.

[0067] The objective function and constraints of the multi-objective problem (multi-objective diversified energy storage power station aggregation model) are as follows.

[0068] (13)

[0069] (14)

[0070] Formula (13) is the objective function of the multi-objective problem, including the expected net load fluctuation and energy storage power station regulation capacity ; In formula (13): , and They represent the average output of electric load, wind power and photovoltaic power in the t period respectively, which are the input data; and are the charging power and discharging power of the energy storage power station in period t, and They are the maximum charging capacity and the maximum discharging capacity of the energy storage power station in period t respectively; and are the charging powers of the electrochemical energy storage system and the compressed air energy storage system in the energy storage power station at time period t. Similarly, and are the discharge powers corresponding to the tth period respectively. In formula (14): Characterizing electrochemical energy storage systems, Characterizing compressed air energy storage systems; and are the rated capacities of the electrochemical energy storage system and the compressed air energy storage system, respectively; and are the rated energies of the electrochemical energy storage system and the compressed air energy storage system respectively; and are the charging efficiencies of the electrochemical energy storage system and the compressed air energy storage system, respectively; and are the discharge efficiencies of the electrochemical energy storage system and the compressed air energy storage system, respectively; and are the charge states of the electrochemical energy storage system and the compressed air energy storage system at time period t-1 respectively; and are the minimum charge states of the electrochemical energy storage system and the compressed air energy storage system, respectively; and are the maximum states of charge of the electrochemical energy storage system and the compressed air energy storage system, respectively; Formula (12) in Formula (14) is the general energy storage model in the form of SoC expression.

[0071] Step 3-2: Use a multi-objective generalized normal distribution optimization algorithm to solve the multi-objective diversified energy storage power station aggregation model and obtain the Pareto optimal solution set.

[0072] In order to solve the above multi-objective problems, a multi-objective generalized normal distribution optimization algorithm (MOGNDO) is used. This algorithm adds an archive mechanism to the generalized normal optimization algorithm to store non-dominated Pareto optimal solutions and record the best results in detail; at the same time, a leader selection mechanism is introduced to identify and select the best solution from the archive to guide the optimization process.

[0073] Based on the generalized normal distribution optimization (GNDO) algorithm, exploration and exploitation are used to obtain the best solution. First, all populations need to be evaluated by the objective function. Further, for each solution in the population, the generated random number is used to switch between exploration mode and exploitation mode. In the global exploration mode, the entire search space is scanned to identify potential areas. The local exploitation mode involves seeking improved solutions near the existing positions of all individuals in the search space.

[0074] The exploration equation and development equation are shown in equation (15) and equation (16) respectively.

[0075] (15)

[0076] (16)

[0077] Where: is the location to be updated. and is the middle position of the decision variable, is the current best position, To adjust the parameters, and is a random variable that follows a standard normal distribution; represents the generalized average position of the ith individual, represents the generalized standard deviation, as a penalty factor.

[0078] The archive mechanism (AM) and the leader selection mechanism (LSM) are further introduced. The archive mechanism is designed to save solutions that have not been dominated so far, and consists of an archive controller and a grid. The archive controller decides whether to include a solution in the archive. Solutions that already exist in the archive will be immediately excluded. On the contrary, undominated solutions will be added to the archive. When the archive reaches its capacity limit, the grid mechanism will be activated. The role of the grid is to maintain the diversity of solutions in the archive as much as possible and divide the space into multiple areas.

[0079] The leader selection mechanism takes the best solution obtained so far as the current best position. This strategy uses archived information to select a leader from a less dense area in the search space and provides one of the non-dominated solutions as the new best position, where the probability of each proposed new leader is detailed in formula (17): The significance of this is that the search direction always tends to the area in the search space that has not been thoroughly explored or exposed.

[0080] (17)

[0081] Where: The probability of updating the leader for the wth region, is a constant greater than 1, is the number of Pareto optimal solutions obtained.

[0082] Step 3-3: Based on the Pareto optimal solution set, a compromise solution strategy that considers fairness is adopted to transform the multi-objective diversified energy storage power station aggregation model to generate a diversified energy storage power station aggregation model.

[0083] The result obtained by the multi-objective generalized normal distribution optimization algorithm is the Pareto optimal solution set, but only one solution is actually needed. The problem of finding a suitable solution from the Pareto optimal solution set belongs to a multi-attribute decision-making problem. This application adopts a compromise solution strategy considering fairness to achieve multi-objective optimization decision-making, that is, selecting a point with an increment ratio close to 1 as a fair compromise solution. The fairness of the Pareto solution set is defined as follows.

[0084] (18)

[0085] Where: and are the net load fluctuation increment and the corresponding minimum value respectively, and They are the increment of regulation capacity of energy storage power station and the corresponding minimum value respectively.

[0086] The final equivalent SoC model of the diversified energy storage power station is shown below.

[0087] (19)

[0088] Step 4: Based on the diversified energy storage power station aggregation model and the extreme scenario uncertainty set, a robust optimization model of wind, solar, thermal and storage in the region is constructed, and the Black Kite optimization algorithm is used to solve the charging and discharging power of the diversified energy storage power station, which specifically includes the following steps 4-1 to 4-4.

[0089] Step 4-1: Construct a robust optimization model of wind, solar, thermal and energy storage in the region based on the diversified energy storage power station aggregation model and extreme scenario uncertainty set.

[0090] A robust optimization model considering wind, solar, thermal and storage resources in the region is proposed. The robust optimization model considering wind, solar, thermal and storage resources is as follows.

[0091] (20)

[0092] (twenty one)

[0093] Formula (20) is the objective function of the robust optimization model of wind, solar, thermal and storage in the region, where: , , and They are the total cost, coal consumption cost, new energy power abandonment cost and energy storage power station dispatch cost within the optimization dispatch cycle; , and are the actual outputs of thermal power, wind power, and photovoltaic power in period t, is the actual output of thermal power in period t-1, and are the predicted outputs of wind power and photovoltaic power in period t, and are the actual charging power and discharging power of the energy storage power station in period t respectively; and They are the coal consumption cost per unit of electricity and the penalty cost per unit of renewable energy abandonment. is the time-of-use electricity price in the tth period. Formula (21) is the constraint condition of the robust optimization model of wind, solar, thermal and storage in the region, where: , , They are the installed capacity of thermal power, wind power and photovoltaic power respectively. is the grid-connected capacity of thermal power in period t, , , , They are the minimum technical output coefficient, maximum technical output coefficient, climbing rate coefficient and descending rate coefficient of thermal power respectively; For the Active power flow of the branches, and Respectively The minimum active power flow limit and the maximum active power flow limit of each branch; is the actual output of the electric load in the tth period; Formula (19) in Formula (21) is the final equivalent SoC formal model of the diversified energy storage power station, that is, the aggregation model of the diversified energy storage power station.

[0094] Step 4-2: Convert the robust optimization model of wind, solar, thermal and storage in the region into a maximization-minimum (max-min) problem model.

[0095] The robust optimization model can be expressed as the following mathematical problem, namely the maximization-minimization problem model.

[0096] (twenty two)

[0097] (twenty three)

[0098] Where: is the uncertainty set of wind and solar load, For The relevant variable space, is a set of variables; is the inequality coefficient matrix, is the equation coefficient matrix, is the coefficient matrix of the uncertainty correlation equation of wind and solar load; is the objective function coefficient vector, for The transpose of is the inequality coefficient vector, is the coefficient vector of the equation, is the uncertainty equation vector of wind and solar load.

[0099] Step 4-3: Use strong duality theory to transform the maximization-minimization problem model into a single-layer maximization (max) problem model.

[0100] Furthermore, the strong duality theory is used to transform the robust optimization max-min problem into a single-layer max problem.

[0101] (twenty four)

[0102] Where: , and is the introduced dual variable, , , , , and They are , , , , , The transpose of .

[0103] Step 4-4: Use the Black Kite optimization algorithm to solve the single-layer maximization problem model and obtain the charging and discharging power of the diversified energy storage power station.

[0104] The Black-winged kite algorithm (BKA) uses two models of attack behavior and migration behavior for iterative optimization. The algorithm's combination mode strikes a good balance between exploring global solutions and using current information, improving the algorithm's global search capability and convergence speed. Both attack behavior and migration behavior are divided into two strategies, as shown below.

[0105] (25)

[0106] (26)

[0107] Where: is the j-th dimension position information of the e-th black-winged kite at the b-th iteration, is the j-th dimension position information of the e-th black-winged kite at the b+1-th iteration, is a random number between 0 and 1, is a constant, is the iteration coefficient, which is related to the current number of iterations and the total number of iterations; represents the j-th dimension position information of the current best black-winged kite at the b-th iteration, and They represent the fitness value of the j-th dimension current position information and the j-th dimension random position obtained by any black-winged kite at the b-th iteration, represents a Cauchy mutation.

[0108] The specific solution process of using the black kite algorithm to solve the converted robust optimization model (single-layer maximization problem model) is as follows: 1) Initialize the population, that is, give each black kite initial position information. The position of a black kite represents a solution to the model; 2) Use attack behavior and migration behavior strategies to continuously update the position of each black kite; 3) Update and record the current best black kite position information and target value; 4) Continue to iterate until the set total number of iterations is reached, end the iteration, and complete the solution.

[0109] Step 5: According to the charge and discharge power of the diversified energy storage power station, a microhabitat-based gradient directed evolution algorithm and a scheduling plan prediction model are used to generate a joint scheduling plan within the diversified energy storage power station, specifically including the following steps 5-1 to 5-7. The scheduling plan prediction model is constructed based on a convolutional neural network, an enhanced sparrow search algorithm, and a bidirectional gated recurrent unit.

[0110] Step 5-1: Taking the maximum capacity of the diversified energy storage power station as the upper limit, a random function is used to generate a number of target charging powers and target discharging powers, and corresponding scheduling scenarios are generated according to the target charging powers and target discharging powers.

[0111] Step 5-2: To meet the charging power and discharging power of the diversified energy storage power station, a joint scheduling model is constructed within the diversified energy storage power station, specifically including the following steps 5-2-1 to 5-2-4. The diversified energy storage power station includes an electrochemical energy storage system and a compressed air energy storage system. The electrochemical energy storage system includes a solenoid valve, a circulation pump and a battery reactor. The compressed air energy storage system includes a compression unit, an expansion unit, a gas storage tank and a heat storage tank.

[0112] Taking the charging power and discharging power of the diversified energy storage power station in the previous step as the boundary, an internal joint scheduling model of the diversified energy storage power station is established, which is composed of an electrochemical energy storage system and a compressed air energy storage system.

[0113] Step 5-2-1: Construct a power loss model for the electrochemical energy storage system based on the losses of the solenoid valve and the circulation pump and the losses of the battery reactor.

[0114] The electrochemical energy storage system in this application is composed of batteries, and its power loss includes the operation of the solenoid valve and the circulation pump and the battery reactor loss, as shown below.

[0115] (27)

[0116] Where: is the power loss of the solenoid valve and circulation pump in the tth period; and are the rated powers of the solenoid valve and circulation pump of the oth battery, is the startup state of the oth battery in the tth period, is the number of batteries; and are the battery reactor charging loss and discharging loss in the tth period, and are the charging efficiency and discharging efficiency of the battery reactor, and are respectively the charging rated power and the discharging rated power of the battery reactor of the oth battery.

[0117] Step 5-2-2: Based on the self-discharge problem of the battery during storage after charging, the self-discharge rate coefficient is introduced to construct an equivalent model of the electrochemical energy storage system.

[0118] Considering the self-discharge problem of the battery during storage after charging, the self-discharge rate coefficient is introduced , the equivalent model of the electrochemical energy storage system is established as follows.

[0119] (28)

[0120] (29)

[0121] In formula (28): and are the equivalent charging efficiency and discharging efficiency of the electrochemical energy storage system in period t, is the rated energy of the battery reactor; in formula (29): and are the equivalent charging power and discharging power of the electrochemical energy storage system in period t, is the energy status of the electrochemical energy storage system at time period t, is the energy status of the electrochemical energy storage system at time t-1, and They are the minimum state of charge and maximum state of charge of the electrochemical energy storage system, respectively.

[0122] Step 5-2-3: Based on the operation of the compression unit, expansion unit, air storage tank and heat storage tank, construct an operation model of the compressed air energy storage system.

[0123] The compressed air energy storage system in this application is mainly composed of a compression unit, an expansion unit, an air storage tank and a heat storage tank. Its operation model is as follows.

[0124] (30)

[0125] (31)

[0126] In formula (30): and are the charging power and discharging power of the compressed air system in period t, and Respectively The charging power and discharging power of the compressed air system during the period, and are the compression charging coefficient and the expansion releasing coefficient, and are the gas mass flow rates flowing into and out of the gas storage tank at time period t, respectively; and are the initial pressure value of the gas tank and the pressure value of the tth period respectively; and are the internal heat of the heat storage tank at the initial moment and the internal heat of the tth period, and Respectively Compression heat storage power and expansion heat power in each period; For the The heat energy power output from the heat storage tank to the outside during the period, is the minimum scheduling period. In formula (31): is the heat loss coefficient, is the thermal power output from the heat storage tank to the outside in period t, is the actual thermal power output after considering the loss in the t period, is the upper limit of the output thermal power, and are the lower and upper limits of the heat storage tank, and They are the lower and upper limits of the gas pressure in the gas tank respectively.

[0127] Step 5-2-4: To meet the charging power and discharging power of the diversified energy storage power station as the goal, the power loss model of the electrochemical energy storage system, the equivalent model of the electrochemical energy storage system, the operation model of the compressed air energy storage system, and the relationship between the charging power and the discharging power are constrained to construct a joint dispatch model within the diversified energy storage power station. Among them, the relationship between the charging power and the discharging power is that the product of the charging power and the discharging power is equal to 0.

[0128] The internal joint dispatching model of the diversified energy storage power station aims to meet the charging power and discharging power of the diversified energy storage power station. The problem can be expressed as follows.

[0129] (32)

[0130] Where: is the objective function of the internal joint dispatch model of the diversified energy storage power station, and are the target charging power and target discharging power of the diversified energy storage power station in the tth period respectively; Formula (27-31) in Formula (32) refers to the power loss model of the electrochemical energy storage system, the equivalent model of the electrochemical energy storage system and the operation model of the compressed air energy storage system.

[0131] Step 5-3: Use the niche-based gradient directed evolution algorithm to solve the internal joint scheduling model of the diversified energy storage power station to obtain the target scheduling plan under different scheduling scenarios. The target scheduling plan is a joint scheduling plan for the electrochemical energy storage system and the compressed air energy storage system in the diversified energy storage power station.

[0132] For the above minimization optimization problem, a gradient directed evolution algorithm based on microhabitats is used to solve it. The algorithm uses gradient information to guide the mutation direction and adopts a microhabitat method to protect the population diversity, prevent premature convergence and fall into a local solution, and enhance the global search capability.

[0133] The specific solution process is as follows: first, use random numbers to initialize the population, and the population size is M; second, use the microhabitat method to divide the population into multiple groups according to the fitness value of the individual, and the population size in each group is as shown in formula (33), which maintains the diversity of the population; third, based on formulas (34)-(35), use the mutation operator to iterate each group of microhabitats to obtain candidate individuals, and update the individuals according to formula (36); finally, continue to iterate until the upper limit of the number of iterations is reached to obtain the global optimal solution. That is, a refined joint scheduling plan for the electrochemical energy storage system and the compressed air energy storage system in the diversified energy storage power station is obtained to meet the charging and discharging goals of the diversified energy storage power station.

[0134] (33)

[0135] (34)

[0136] (35)

[0137] (36)

[0138] Where: is the total population size, For the The number of individuals in the niche, is a constant; and Respectively The maximum and minimum values ​​of individuals in the group, and Respectively Group The objective function value and standardized function value of the individual, is a constant between 0 and 1; is the candidate individual for the k+1th iteration, and are the kth iterations respectively. Group The position and gradient of the individual is the kth iteration The position of the best individual in the group; is a mutation factor, is the step length, and are two individuals randomly selected from the total population; is the kth iteration Closest in group of individuals.

[0139] Step 5-4: Generate a neural network training data set based on the target charging power, target discharging power of the diversified energy storage power station and the target scheduling plan under the corresponding scheduling scenario.

[0140] Step 5-5: Construct a scheduling plan prediction model based on convolutional neural network, enhanced sparrow search algorithm and bidirectional gated recurrent unit.

[0141] Step 5-6: Use the neural network training data set to train the scheduling plan prediction model to obtain a trained scheduling plan prediction model.

[0142] Step 5-7: Input the charging and discharging power of the diversified energy storage power station into the trained scheduling plan prediction model to obtain the internal joint scheduling plan of the diversified energy storage power station.

[0143] A CNN-ISSA-BiGRU neural network model suitable for rapid decision-making in the dispatch of diversified energy storage power stations is established. The specific process is as follows.

[0144] In the first step, with the maximum capacity of the diversified energy storage power station as the upper limit, a large number of target charging powers and target discharging powers are generated by random functions. Further, based on the product of the target charging power and the target discharging power being 0, one side of the power is randomly set to zero, thereby generating a large number of scheduling scenarios about the target charging power and the target discharging power of the diversified energy storage power station, and using the microhabitat-based gradient directed evolution algorithm to solve the scheduling plan of the electrochemical energy storage system and the compressed air energy storage system inside the diversified energy storage power station under the scheduling scenario; further, based on the target charging and discharging power of the diversified energy storage power station and the refined joint scheduling plan of the electrochemical energy storage system and the compressed air energy storage system inside the diversified energy storage power station as basic data, a neural network training data set is generated.

[0145] The second step is to establish a CNN-ISSA-BiGRU neural network model as a scheduling plan prediction model. In this model, the Convolutional Neural Network (CNN) layer is used as a data processing layer to preprocess the input data, the Improved Sparrow Search Algorithm (ISSA) layer is used to search for the optimal parameters of the BiGRU layer, and the Bidirectional Gated Recurrent Unit (BiGRU) layer deeply analyzes the correlation between data. ISSA adopts a reverse learning strategy to mitigate the impact of unpredictable trends in complex functions, solving the shortcomings of low convergence accuracy and possible tendency to local optimality when dealing with complex function optimization problems.

[0146] The third step is to input the neural network training data set as input data to the CNN-ISSA-BiGRU neural network model for continuous training. After the training is completed, the finished neural network model is obtained, that is, the trained scheduling plan prediction model. The effect is that the target charge and discharge power of the diversified energy storage power station is input into the finished neural network model, and the refined joint scheduling plan of the electrochemical energy storage system and the compressed air energy storage system inside the diversified energy storage power station can be quickly obtained, that is, the internal joint scheduling plan of the diversified energy storage power station.

[0147] This application proposes a diversified energy storage joint scheduling method that adapts to peak loads in extreme scenarios. First, a k-means clustering algorithm for peak-to-valley difference characteristics is used to generate a set of extreme scenarios; second, a kernel density estimation method is used to generate a probability density function under extreme scenarios, and the sampling data is used to generate an uncertainty convex space based on the fast convex hull algorithm, which is reduced to form an extreme scenario uncertainty set; third, based on the general energy storage model expressed in the form of SoC, a multi-objective generalized normal distribution optimization algorithm is used to generate a diversified energy storage power station aggregation model; fourth, a robust optimization model of wind, solar, thermal and storage (diversified energy storage power station) in the region is proposed, and the black kite optimization algorithm is used to solve and obtain a robust optimization scheduling strategy (charging and discharging power of the diversified energy storage power station); finally, according to the output boundary of the diversified energy storage power station, a diversified energy storage power station internal joint scheduling method considering the microhabitat-based gradient directed evolution algorithm is proposed, and the CNN-ISSA-BiGRU neural network model is used to achieve rapid decision-making of the internal joint scheduling plan of the diversified energy storage power station.

[0148] The technical effects of this application are as follows: (1) The clustering algorithm for peak-valley difference characteristics proposed in this application is suitable for the rapid screening of extreme scene sets. At the same time, the clustering algorithm overcomes the problem of inconsistent results of traditional algorithms and stabilizes the algorithm effect. (2) This application uses a fast convex hull algorithm to quickly convert a random scene set into a three-dimensional uncertainty convex hull, and then reduces it to form an extreme scene uncertainty set, which more comprehensively retains the uncertainty interval range, balances robustness and optimization benefits, improves the generation efficiency of uncertainty sets, and simplifies the complexity of uncertainty sets. (3) This application uses a multi-objective generalized normal distribution optimization algorithm to quickly solve the multi-objective diversified energy storage power station aggregation model, accelerates the speed of solving multi-objective problems, and quickly generates a diversified energy storage power station aggregation model. (4) This application uses strong duality theory to transform the robust optimization model of wind, solar, thermal and storage (diversified energy storage power station) in the region, solves the problem that is difficult to solve under the uncertainty set, and uses the black kite optimization algorithm to achieve rapid solution. The combination mode of the algorithm achieves a good balance between exploring global solutions and using current information, improving the global search ability and convergence speed of the algorithm. (5) The internal joint scheduling model of the diversified energy storage power station proposed in this application can support the charging and discharging power of the energy storage power station. It adopts a microhabitat-based gradient directed evolution algorithm for solution, uses gradient information to guide the mutation direction, and adopts a microhabitat method to protect population diversity, prevent premature convergence and fall into a local solution, and enhance the global search capability; further, the CNN-ISSA-BiGRU neural network model is used to realize the rapid decision-making of the internal joint scheduling plan of the diversified energy storage power station.

[0149] In an exemplary embodiment, the present application further provides a computer device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0150] In an exemplary embodiment, the present application further provides a computer-readable storage medium storing a computer program, which implements the steps in the above-mentioned method embodiments when executed by a processor.

[0151] In an exemplary embodiment, the present application further provides a computer program product, including a computer program, which implements the steps in the above-mentioned method embodiments when executed by a processor.

[0152] In this application, all actions to obtain signals, information or data are carried out in compliance with the relevant data protection laws and policies of the country where the application is located and with the authorization of the corresponding device owner. It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant legal provisions.

[0153] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0154] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0155] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0156] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A multi-energy storage joint dispatching method adapted to extreme scene peak loads, applied to a diversified energy storage power station, characterized in that: include: Based on the wind power output, photovoltaic output and power load data in the historical period, a clustering algorithm oriented to the peak-to-valley difference characteristics is used to generate a set of extreme scenarios; According to the wind power output, photovoltaic output and electric load data under the extreme scenario set, the fast convex hull algorithm is used to generate the extreme scenario uncertainty set; Based on the general energy storage model in the form of state of charge expression, a multi-objective generalized normal distribution optimization algorithm is used to generate a diversified energy storage power station aggregation model; Based on the diversified energy storage power station aggregation model and the uncertainty set of extreme scenarios, a robust optimization model of wind, solar, thermal and storage in the region is constructed, and the charging and discharging power of the diversified energy storage power station is solved using the Black Kite optimization algorithm. According to the charging and discharging power of the diversified energy storage power station, a microhabitat-based gradient directed evolution algorithm and a scheduling plan prediction model are used to generate an internal joint scheduling plan of the diversified energy storage power station; the scheduling plan prediction model is constructed based on a convolutional neural network, an enhanced sparrow search algorithm and a bidirectional gated recurrent unit.

2. The multi-energy storage joint dispatching method adapted to peak loads in extreme scenarios according to claim 1 is characterized in that: Based on the wind power output, photovoltaic output and power load data in the historical period, a clustering algorithm oriented to the peak-to-valley difference characteristics is used to generate a set of extreme scenarios, including: Obtain wind power output, photovoltaic output and power load data for historical time periods; Based on the wind power output, photovoltaic output and electric load data in the historical time period, a set of indicators of wind power characteristics, photovoltaic characteristics and electric load characteristics is constructed; According to the index set of wind power characteristics, photovoltaic characteristics and electric load characteristics, an improved k-means clustering algorithm is used to perform clustering to obtain a clustering division result; the improved k-means clustering algorithm is based on the k-means clustering algorithm, and the clustering result of the small sample set is used as the initial clustering center of the full sample set to perform clustering division of the full sample set to obtain a clustering division result; the full sample set is the index set of the wind power characteristics, photovoltaic characteristics and electric load characteristics; the small sample set is obtained by extracting a set proportion from the full sample set; Generate an extreme scenario set based on the clustering results.

3. The multi-energy storage joint dispatching method adapted to peak loads in extreme scenarios according to claim 1 is characterized in that: According to the wind power output, photovoltaic output and electric load data under the extreme scenario set, the fast convex hull algorithm is used to generate the extreme scenario uncertainty set, including: Based on the wind power output, photovoltaic output and electric load data under extreme scenarios, the kernel density estimation method is used to construct the probability density functions of wind power output, photovoltaic output and electric load under extreme scenarios. According to the probability density functions of wind power output, photovoltaic output and electric load in extreme scenarios, random sampling is performed using the Monte Carlo sampling method to generate a random scenario set; According to the random scene set, the fast convex hull algorithm is used to generate the extreme scene uncertainty set.

4. The multi-energy storage joint dispatching method adapted to peak loads in extreme scenarios according to claim 1 is characterized in that: Based on the general energy storage model expressed in the form of state of charge, a multi-objective generalized normal distribution optimization algorithm is used to generate a diversified energy storage power station aggregation model, including: Based on the general energy storage model in the form of state of charge expression, a multi-objective diversified energy storage power station aggregation model is constructed; The multi-objective generalized normal distribution optimization algorithm is used to solve the multi-objective diversified energy storage power station aggregation model and obtain the Pareto optimal solution set; According to the Pareto optimal solution set, a compromise solution strategy considering fairness is adopted to transform the multi-objective diversified energy storage power station aggregation model to generate a diversified energy storage power station aggregation model.

5. The multi-energy storage joint dispatching method adapted to peak loads in extreme scenarios according to claim 1 is characterized in that: Based on the diversified energy storage power station aggregation model and the uncertainty set of extreme scenarios, a robust optimization model of wind, solar, thermal and storage in the region is constructed, and the Black Kite optimization algorithm is used to solve the charging and discharging power of the diversified energy storage power station, including: Based on the diversified energy storage power station aggregation model and the uncertainty set of extreme scenarios, a robust optimization model of wind, solar, thermal and energy storage in the region is constructed; Convert the robust optimization model of wind, solar, thermal and storage in the region into a maximization-minimization problem model; The strong duality theory is used to transform the maximization-minimization problem model into a single-layer maximization problem model; The Black Kite optimization algorithm is used to solve the single-layer maximization problem model and obtain the charging and discharging power of the diversified energy storage power station.

6. The multi-energy storage joint dispatching method adapted to peak loads in extreme scenarios according to claim 1 is characterized in that: According to the charging and discharging power of the diversified energy storage power station, the gradient directed evolution algorithm based on the niche and the scheduling plan prediction model are used to generate the internal joint scheduling plan of the diversified energy storage power station, including: Taking the maximum capacity of the diversified energy storage power station as the upper limit, a random function is used to generate several target charging powers and target discharging powers, and corresponding scheduling scenarios are generated according to the target charging powers and target discharging powers; To meet the charging and discharging power requirements of diversified energy storage power stations, a joint dispatching model within diversified energy storage power stations is constructed; The niche-based gradient directed evolution algorithm is used to solve the internal joint dispatch model of the diversified energy storage power station and obtain the target dispatch plan under different dispatch scenarios. Generate a neural network training data set based on the target charging power, target discharging power and target dispatch plan of the diversified energy storage power station under the corresponding dispatch scenario; Based on convolutional neural network, enhanced sparrow search algorithm and bidirectional gated recurrent unit, a scheduling plan prediction model is constructed; The scheduling plan prediction model is trained using a neural network training data set to obtain a trained scheduling plan prediction model; The charging and discharging power of the diversified energy storage power station is input into the trained scheduling plan prediction model to obtain the internal joint scheduling plan of the diversified energy storage power station.

7. The multi-energy storage joint dispatching method adapted to peak loads in extreme scenarios according to claim 6 is characterized in that: The diversified energy storage power station includes an electrochemical energy storage system and a compressed air energy storage system; the target scheduling plan is a joint scheduling plan for the electrochemical energy storage system and the compressed air energy storage system within the diversified energy storage power station.

8. The multi-energy storage joint dispatching method adapted to peak loads in extreme scenarios according to claim 7 is characterized in that: The electrochemical energy storage system includes a solenoid valve, a circulation pump and a battery reactor; the compressed air energy storage system includes a compression unit, an expansion unit, a gas storage tank and a heat storage tank; in order to meet the charging power and discharging power of the diversified energy storage power station, a joint scheduling model inside the diversified energy storage power station is constructed, including: Based on the losses of the solenoid valve and circulation pump as well as the losses of the battery reactor, a power loss model of the electrochemical energy storage system is constructed; Based on the self-discharge problem of the battery during storage after charging, the self-discharge rate coefficient is introduced to construct an equivalent model of the electrochemical energy storage system; Based on the operation of the compression unit, expansion unit, gas storage tank and heat storage tank, an operation model of the compressed air energy storage system is constructed; With the goal of meeting the charging power and discharging power of the diversified energy storage power station, a joint scheduling model inside the diversified energy storage power station is constructed based on the power loss model of the electrochemical energy storage system, the equivalent model of the electrochemical energy storage system, the operation model of the compressed air energy storage system and the relationship between the charging power and the discharging power as constraints.

9. The multi-energy storage joint dispatching method adapted to peak loads in extreme scenarios according to claim 8 is characterized in that: The relationship between the charging power and the discharging power is that the product of the charging power and the discharging power is equal to 0.

10. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the multi-energy storage joint scheduling method adapted to peak loads in extreme scenarios as described in any one of claims 1 to 9.

Images