A Multi-Objective Energy Storage Optimization Method for Power Systems Based on an Improved Genetic Algorithm

Abstract

A multi-objective energy storage optimization method for power systems based on an improved genetic algorithm is proposed. First, a multi-objective optimization model is constructed with the objective function of minimizing the total cost of energy storage configuration, assuming energy storage is deployed on the power generation side, grid side, and load side. Then, a non-dominated sorting genetic algorithm based on the Pareto optimality concept is used to solve the constructed multi-objective optimization model, yielding an optimized energy storage scheme, including the location and capacity of each energy storage unit. This invention simultaneously achieves smoothing of renewable energy output on the power generation side, peak shaving and valley filling on the grid side, and cost reduction on the load side through peak-valley arbitrage.

Description

Technical Field

[0001] This invention belongs to the field of power system energy storage planning, specifically involving a multi-objective energy storage optimization method for power systems based on an improved genetic algorithm. Background Technology

[0002] Because energy storage power stations can effectively improve the absorption of new energy sources such as wind power and photovoltaics, reduce curtailment rates, and enhance the overall utilization rate of new energy sources, while also providing services such as frequency regulation and voltage support, they have become an important component of new power systems such as wind power and photovoltaics in recent years. Proper grid connection of energy storage power stations can effectively reduce network losses, reduce investment in grid upgrades, and reduce load fluctuations on power lines. However, unreasonable planning schemes may lead to increased active power losses after the energy storage power station is connected to the grid, and even reduce grid voltage stability, causing voltage exceedances at some nodes. Therefore, reasonable site selection and capacity planning for energy storage devices can effectively improve the efficiency of energy storage systems, improve power system network losses and the smoothing effect on new energy sources, and optimize the economics of system operation. In conclusion, it is necessary to rationally plan the site selection and capacity of energy storage power stations.

[0003] Domestic and international scholars have conducted extensive research on the site selection and capacity determination of energy storage. Wang Ziqi's research on energy storage site selection, capacity determination, and energy management in power grids with wind and solar power (North China Electric Power University, 2021) proposed a site selection and capacity determination planning method for energy storage power stations in power grids with wind and solar power, using voltage stability margin as an indicator. It established a pre-site selection strategy, and based on this strategy, determined the installation location and capacity of energy storage power stations with the objective of minimizing the sum of daily active power losses. However, this study considers centralized energy storage access. Li Ran's research on the optimal configuration of energy storage systems based on the economic dispatch of wind and storage combined transmission (North China Electric Power University, 2018) proposed an optimal configuration method for energy storage systems based on maximizing comprehensive energy efficiency, comprehensively considering the economic dispatch cost of thermal power, wind curtailment cost, network loss cost, and energy storage investment cost. The above studies only consider the centralized access of energy storage to a single node, which has limited effectiveness in improving power system network losses and mitigating the impact of new energy sources. Summary of the Invention

[0004] The purpose of this invention is to address the aforementioned problems in the existing technology by providing a multi-objective energy storage optimization method for power systems based on an improved genetic algorithm.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows:

[0006] A multi-objective energy storage optimization method for power systems based on an improved genetic algorithm includes the following steps:

[0007] Step A: Assuming energy storage is configured on the power supply side, grid side, and load side, construct a multi-objective optimization model for energy storage with the objective function of minimizing the total cost of configuring energy storage:

[0008] min[C investment (P s E s ,T s )+C operation (P g ,P w ,P s )]

[0009] C operation (P g ,P w ,P s ) = C generation (P g ,P w ,P s )+η loss W loss

[0010] In the above formula, C investment (P s E s ,T s P represents the investment cost for energy storage configuration. s E s T s These represent the charging / discharging power, capacity, and lifespan of energy storage, respectively. operation (P g ,P w ,P s P represents operating costs. g P w The generating capacities of the synchronous generator set and the wind turbine set are respectively, C generation (P g ,P w ,P s W represents the cost of electricity generation. loss η loss These are network losses and their unit costs, respectively.

[0011] Step B: The constructed multi-objective optimization model for energy storage is solved using a genetic algorithm based on the Pareto optimality concept of non-dominated sorting to obtain an energy storage optimization scheme, which includes the configuration location and capacity of each energy storage unit.

[0012] In step A, the power generation cost C generation (P g ,P w ,P s The following formula is used to calculate:

[0013] C generation (P g ,P w ,P s )=f gen (P g )+f λ (P g ,P w ,P s )

[0014]

[0015]

[0016] In the above formula, f gen (P g f represents the operating cost of the synchronous generator set. λ (P g ,P w ,P s ) is the total penalty factor. The weighting coefficients for the power supply side penalty factor f1, the grid side penalty factor f2, and the load side penalty factor f3 are η1, η2, and η3. 3t These represent the smoothing output cost coefficient, the curtailment cost coefficient, and the peak-valley arbitrage cost coefficient for time period t, respectively, where T is the total number of time periods and N is the number of wind turbine units. P represents the maximum output power of the j-th wind turbine unit during a typical day's time period t. wjt To consider the average power output of the j-th wind turbine during time period t after wind curtailment, P sj1t P sj2t P sj3t Let P be the average power of the j1st energy storage unit on the power source side, the j2nd energy storage unit on the grid side, and the j3rd energy storage unit on the user side during time period t, respectively, where Δt is the duration of each time period. loadjt Let t represent the electrical power consumed by the load during the time period t.

[0017] In step A, the investment cost C of the energy storage configuration investment (P s E s ,T s The following formula is used to calculate:

[0018]

[0019] In the above formula, η P η s These represent the unit cost of the energy storage's charging and discharging power and capacity, respectively, while M represents the fixed cost.

[0020] In step A, the constraints of the objective function include:

[0021] Network loss constraints:

[0022]

[0023] In the above formula, U it U jt G represents the voltages at nodes i and j during time period t. ij Let θ be the line susceptance between nodes i and j. ijt Let be the phase angle difference between the voltages at nodes i and j during time period t;

[0024] Current constraints:

[0025] P git +P wit +P sit -P Lit =U it ΣU jt (G ij cosθ ijt +B ij sinθ ijt )

[0026] Q git +Q wit +Q sit -Q Lit =U it ΣU jt (G ij sinθ ijt -B ij cosθ ijt )

[0027] In the above formula, P git P wit P sit P Lit Let Q represent the active power of the synchronous generator, wind turbine, energy storage, and load at node i during time period t. git Q wit Q sit Q Lit Let B represent the reactive power of the synchronous generator set, wind turbine set, energy storage, and load at node i during time period t. ij The line conductance between nodes i and j;

[0028] Power constraints:

[0029] P gmin ≤P gt ≤P gmax

[0030] Q gmin ≤Q gt ≤Q gmax

[0031] Pwmin ≤P wt ≤P wmax

[0032] Q wmin ≤Q wt ≤Q wmax

[0033] P smin ≤P st ≤P smax

[0034] In the above formula, P gt P wt P st The total active power output of the synchronous generator set, wind turbine set, and energy storage unit during time period t is P. gmin P wmin P smin The minimum total active power output of the synchronous generator set, wind turbine set, and energy storage unit, respectively, P. gmax P wmax P smax Q represents the maximum total active power output of the synchronous generator set, wind turbine, and energy storage, respectively. gt Q wt Q represents the total reactive power output of the synchronous generator set and the wind turbine set during time period t. gmin Q wmin Q represents the minimum total reactive power output of the synchronous generator set and the wind turbine set, respectively. gmax Q wmax These are the maximum total reactive power output of the synchronous generator set and the wind turbine set, respectively.

[0035] Voltage constraint:

[0036] U imin ≤U it ≤U imax

[0037] In the above formula, U imin U imax These are the minimum and maximum allowable voltages for node i, respectively;

[0038] Capacity constraints:

[0039] E smin ≤E st ≤E smax

[0040] In the above formula, E st Let E be the energy storage capacity during time period t. smin E smax These are the minimum and maximum allowable capacities for energy storage, respectively.

[0041] Line capacity constraints:

[0042] -P lijmax ≤P lijt ≤P lijmax

[0043] In the above formula, P lijt P lijmax These represent the power transmitted on the line between nodes i and j and the maximum power allowed to be transmitted, respectively.

[0044] Step B includes the following steps in sequence:

[0045] Step B1: First, set the population size N. p Maximum number of iterations (GEN) max The crossover and mutation probabilities are then used to set six gene positions on each individual using a decimal encoding method. Three of these positions are node positions on the power supply side, grid side, and load side, and the other three positions are energy storage capacities corresponding to the power supply side, grid side, and load side.

[0046] Step B2: Select nodes on the power supply side, grid side, and load side respectively, and randomly generate an energy storage capacity of N. p The initial population is then used, and each individual is fed into the objective function, i.e. the fitness function, to calculate its fitness value.

[0047] Step B3: Based on the Pareto dominance relationship, first select all non-dominant individuals in the population according to their fitness values ​​and set them as the first layer. Assign them a shared virtual fitness value m1. Then, continue to select non-dominant individuals in the population according to the Pareto dominance relationship and set them as the second layer. Assign them a shared virtual fitness value m2 that is less than m1. Repeat this process until all individuals in the population are stratified. Then, calculate the crowding degree of the individuals in each layer. At this point, each individual in the population has virtual fitness and crowding degree attributes. The higher the virtual fitness, the better the individual in that layer. In the same layer, the higher the crowding degree of the individual, the better.

[0048] Step B4: First, select the better individuals from the above population and perform crossover and mutation operations to generate a population of size N. p The offspring population is then merged with the parent population to form a population of size 2N. p The population is then sorted using the method described in step B3, and the top N populations are selected. p Select the better individuals as the next generation of the population;

[0049] Step B5: Repeat step B4 until the maximum number of iterations (GEN) is reached. max At this point, the optimal individual is selected as the energy storage optimization scheme.

[0050] In step B4, the crossover operation uses decimal crossover, where the j-th gene of individuals k and l is crossed in a certain ratio b:

[0051] a kj =a lj b+a kj (1-b)

[0052] a lj =a kj b+a lj (1-b)

[0053] In the above formula, a kj a lj These are the j-th genes of individuals k and l, respectively;

[0054] The mutation operation is achieved by randomly selecting the j-th gene of individual i for mutation:

[0055] a′ ij =a ij +σ*randn(n)

[0056] In the above formula, a ij 、a′ ij Let be the j-th gene of individual i before and after the mutation, respectively; σ be a constant matrix; and randn(n) be a random number generated according to the normal distribution.

[0057] In step B1, each individual is also equipped with 3T gene locations, which are the energy storage charging and discharging strategies for the power supply side, grid side, and load side in T time periods.

[0058] In step B5, the energy storage optimization scheme also includes energy storage charging and discharging strategies for each time period.

[0059] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0060] 1. This invention proposes a multi-objective energy storage optimization method for power systems based on an improved genetic algorithm. This method involves configuring energy storage on the power generation side, grid side, and load side. First, a multi-objective optimization model for energy storage is constructed with the objective function of minimizing the total cost of energy storage configuration. Then, a non-dominated sorting genetic algorithm based on the Pareto optimality concept is used to solve the constructed multi-objective optimization model, yielding an optimized energy storage scheme. This method comprehensively considers factors such as power fluctuations, peak shaving and valley filling, and efficiency of renewable energy units from three different perspectives: power source, grid, and load. It simultaneously optimizes the power structure and smooths renewable energy output on the power generation side, performs peak shaving and valley filling and improves power supply reliability on the grid side, and reduces costs on the load side through peak-valley arbitrage. This effectively enhances the system's capacity to accommodate renewable energy, and it is superior in improving power system network losses, mitigating renewable energy impacts, and optimizing system operating economics. Furthermore, this method is applicable to large-scale energy storage integration and can be applied to large power grids. Therefore, this invention simultaneously achieves smoothing renewable energy output on the power generation side, peak shaving and valley filling on the grid side, and cost reduction on the load side through peak-valley arbitrage.

[0061] 2. This invention discloses a multi-objective energy storage optimization method for power systems based on an improved genetic algorithm. In solving the multi-objective energy storage optimization model using a non-dominated sorting genetic algorithm based on the Pareto optimality concept, considering the configuration of energy storage on the power source side, grid side, and load side, six gene positions are set for each individual during encoding: the node positions on the power source side, grid side, and load side, and the corresponding energy storage capacity. An initial population is randomly generated by randomly selecting nodes on the power source side, grid side, and load side, and arbitrary energy storage capacities. Then, the population is stratified using the Pareto optimality concept, and each stratum is assigned a virtual fitness value, with the value of the next stratum being less than that of the previous stratum. Subsequently, crowding is calculated for each stratum. Furthermore, during the evolutionary process, parent and offspring populations are merged for selection to ensure that a better population participates in the evolution. This algorithm achieves precise optimization of the multi-objective energy storage problem. Therefore, this invention achieves precise optimization of the multi-objective energy storage problem. Attached Figure Description

[0062] Figure 1 This is a flowchart of the present invention.

[0063] Figure 2 This is a model diagram of the IEEE-14 node system used in Example 1.

[0064] Figure 3 The charging and discharging status of the power supply side energy storage device in Example 1 over 24 hours is shown (negative values ​​indicate discharging, and positive values ​​indicate charging).

[0065] Figure 4 The charging and discharging status of the grid-side energy storage device in Example 1 over 24 hours is shown (negative values ​​indicate discharging, and positive values ​​indicate charging).

[0066] Figure 5 The charging and discharging status of the load-side energy storage device in Example 1 over 24 hours is shown (negative values ​​indicate discharging, and positive values ​​indicate charging).

[0067] Figure 6 This is a diagram illustrating the process of selecting the next generation population in Example 1. Detailed Implementation

[0068] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings.

[0069] This invention proposes a multi-objective energy storage optimization method for power systems based on an improved genetic algorithm. This method configures energy storage on the power source side, grid side, and load side. It can not only achieve energy storage location and capacity determination under the premise of determining the energy storage charging and discharging strategy, but also cooperate with the energy management strategy of energy storage for energy storage planning and operation strategy optimization.

[0070] Example 1:

[0071] See Figure 1 A multi-objective energy storage optimization method for power systems based on an improved genetic algorithm is presented in this embodiment. Figure 2 The IEEE-14 node system shown (containing 14 nodes, 17 branches, 3 transformer branches, and 5 power sources consisting of synchronous generators and renewable energy generators, with an energy storage device installed on the power source side, grid side, and load side, and the 24-hour charge and discharge status of each energy storage device as shown below) Figure 3 , 4 As shown in Figure 5 (the specific capacity and node locations are yet to be planned), this is the optimization target, and the following steps will be followed in sequence:

[0072] 1. Construct a multi-objective optimization model for energy storage, using the configuration nodes and capacity of energy storage as decision variables, with the objective function being the minimum total cost of configuring energy storage:

[0073] min[C investment (P s E s ,T s )+C operation (P g ,P w ,P s )]

[0074]

[0075] C operation (P g ,P w ,P s ) = C generation (P g ,P w ,Ps )+η loss W loss

[0076] C generation (P g ,P w ,P s )=f gen (P g )+f λ (P g ,P w ,P s )

[0077]

[0078]

[0079] In the above formula, C investment (P s E s ,T s P represents the investment cost for energy storage configuration. s E s T s These represent the charging / discharging power, capacity, and lifespan of energy storage, respectively. operation (P g ,P w ,P s P represents operating costs. g P w The generating capacities of the synchronous generator set and the wind turbine set are respectively, η P η s These represent the unit cost of energy storage charging / discharging power and capacity, respectively; M is the fixed cost; and C is the unit cost. generation (P g ,P w ,P s The cost of power generation is W. Since the charging and discharging status of the energy storage device is determined in this embodiment, this value is constant. loss η loss These are network losses and their unit costs, respectively. The magnitude of network losses is related to the installation location of the energy storage. gen (P g f represents the operating cost of the synchronous generator set. λ (P g ,P w ,P s ) is the total penalty factor. Here, f1, f2, and f3 are the weighting coefficients for the power generation penalty factor f1, the grid penalty factor f2, and the load penalty factor f3, respectively. f1, f2, and f3 represent the costs of smoothing renewable energy output on the power generation side, reducing power curtailment on the grid side, and utilizing peak-valley price arbitrage on the load side, respectively.3t These represent the smoothing output cost coefficient, the curtailment cost coefficient, and the peak-valley arbitrage cost coefficient for time period t, respectively, where T is the total number of time periods and N is the number of wind turbine units. P represents the maximum output power of the j-th wind turbine unit during a typical day's time period t. wjt To consider the average power output of the j-th wind turbine during time period t after wind curtailment, P sj1t P sj2t P sj3t Let P be the average power of the j1st energy storage unit on the power source side, the j2nd energy storage unit on the grid side, and the j3rd energy storage unit on the user side during time period t, respectively, where Δt is the duration of each time period. loadjt The power consumption of the load during time period t;

[0080] The constraints of the above objective function include:

[0081] Network loss constraints:

[0082]

[0083] In the above formula, U it U jt Let G be the voltages of nodes i and j in time period t. To prevent repeated calculation of line losses, j∈i is set. ij Let θ be the line susceptance between nodes i and j. ijt Let be the phase angle difference between the voltages at nodes i and j during time period t;

[0084] Current constraints:

[0085] P git +P wit +P sit -P Lit =U it ΣU jt (G ij cosθ ijt +B ij sinθ ijt )

[0086] Q git +Q wit +Q sit -Q Lit =U it ΣU jt (G ij sinθ ijt -B ij cosθ ijt )

[0087] In the above formula, P git P wit P sit P LitLet Q represent the active power of the synchronous generator, wind turbine, energy storage, and load at node i during time period t. git Q wit Q sit Q Lit Let B represent the reactive power of the synchronous generator set, wind turbine set, energy storage, and load at node i during time period t. ij The line conductance between nodes i and j;

[0088] Power constraints:

[0089] P gmin ≤P gt ≤P gmax

[0090] Q gmin ≤Q gt ≤Q gmax

[0091] P wmin ≤P wt ≤P wmax

[0092] Q wmin ≤Q wt ≤Q wmax

[0093] P smin ≤P st ≤P smax

[0094] In the above formula, P gt P wt P st The total active power output of the synchronous generator set, wind turbine set, and energy storage unit during time period t is P. gmin P wmin P smin The minimum total active power output of the synchronous generator set, wind turbine set, and energy storage unit, respectively, P. gmax P wmax P smax Q represents the maximum total active power output of the synchronous generator set, wind turbine, and energy storage, respectively. gt Q wt Q represents the total reactive power output of the synchronous generator set and the wind turbine set during time period t. gmin Q wmin Q represents the minimum total reactive power output of the synchronous generator set and the wind turbine set, respectively. gmax Q wmax These are the maximum total reactive power output of the synchronous generator set and the wind turbine set, respectively.

[0095] Voltage constraint:

[0096] U imin ≤U it ≤U imax

[0097] In the above formula, U imin U imax These are the minimum and maximum allowable voltages for node i, respectively;

[0098] Capacity constraints:

[0099] E smin ≤E st ≤E smax

[0100] In the above formula, E st Let E be the energy storage capacity during time period t. smin E smax These are the minimum and maximum allowable capacities for energy storage, respectively.

[0101] Line capacity constraints:

[0102] -P lijmax ≤P lijt ≤P lijmax

[0103] In the above formula, P lijt P lijmax These represent the power transmitted on the line between nodes i and j and the maximum power allowed to be transmitted, respectively.

[0104] 2. First, set the population size N. p Maximum number of iterations (GEN) max The crossover and mutation probabilities are then used to set six gene positions on each individual using a decimal encoding method. Three of these positions are node positions on the power supply side, grid side, and load side, and the other three positions are energy storage capacities corresponding to the power supply side, grid side, and load side.

[0105] 3. Select nodes on the power supply side, grid side, and load side respectively, and randomly generate an energy storage capacity of N. p The initial population is then used, and each individual is fed into the objective function, i.e. the fitness function, to calculate its fitness value.

[0106] 4. Based on the Pareto dominance relationship, first select all non-dominant individuals in the population according to their fitness values ​​and set them as the first layer. Assign them a shared virtual fitness value m1. Then, continue to select non-dominant individuals in the population according to the Pareto dominance relationship and set them as the second layer. Assign them a shared virtual fitness value m2 that is less than m1. Repeat this process until all individuals in the population are stratified. Then, calculate the crowding degree of the individuals in each layer. At this point, each individual in the population has virtual fitness and crowding degree attributes. The higher the virtual fitness, the better the individual in that layer. In the same layer, the higher the crowding degree of the individual, the better.

[0107] 5. See Figure 6 First, select several superior individuals from the above population and perform crossover and mutation operations to generate a population of size N. p The offspring population is then merged with the parent population to form a population of size 2N. p The population is selected, and non-dominated sorting and crowding calculation are performed using the method described in step 4. N is then chosen. p The superior individuals are selected as the next generation population. The crossover operation uses a decimal crossover method, where the j-th gene of individuals k and l is crossed in a certain proportion b.

[0108] a kj =a lj b+a kj (1-b)

[0109] a lj =a kj b+a lj (1-b)

[0110] In the above formula, a kj a lj These are the j-th genes of individuals k and l, respectively;

[0111] The mutation operation is achieved by randomly selecting the j-th gene of individual i for mutation:

[0112] a′ ij＝ a ij +σ*randn(n)

[0113] In the above formula, a ij 、a′ ij Let be the j-th gene of individual i before and after the mutation, respectively; σ be a constant matrix; and randn(n) be a random number generated according to the normal distribution.

[0114] 6. Repeat step 5 repeatedly until the maximum number of iterations (GEN) is reached. maxAt this point, the optimal individual is selected as the energy storage optimization scheme, including the configuration location and capacity of each energy storage, which are as follows: power supply side, node 9, capacity 300MWh; grid side, node 6, capacity 300MWh; load side, node 4, capacity 300MWh.

[0115] Example 2:

[0116] The difference from Example 1 is as follows:

[0117] In step 2, in addition to the six gene positions related to node location and energy storage capacity, 3T gene positions are added to each individual, which are the charging and discharging strategies of the energy storage devices on the power supply side, grid side, and load side in T time periods.

[0118] In step 6, the energy storage optimization scheme also includes the charging and discharging strategies of each energy storage device at each time period.

Claims

1. A power system multi-objective energy storage optimization method based on an improved genetic algorithm, characterized in that: the method comprises the following steps in turn: Step A, under the premise of configuring energy storage on the power supply side, the grid side and the load side, a multi-objective energy storage optimization model is constructed with the minimum total cost of configuring energy storage as the objective function: Step B, a non-dominated sorting genetic algorithm based on the Pareto optimal concept is used to solve the constructed multi-objective energy storage optimization model, and an energy storage optimization scheme is obtained, which includes the configuration location and capacity of each energy storage. ； ； ； ； ； In the above formula, is the investment cost of energy storage configuration, , , are the charging and discharging power, capacity and life of energy storage respectively, is the operation cost, , are the power generation of synchronous generator unit and wind turbine respectively, is the power generation cost, , are the network loss and its unit cost respectively, is the operation cost of synchronous generator unit, is the total penalty factor, , , are the weight coefficients of power supply side penalty factor , grid side penalty factor , and load side penalty factor , , , are the smoothing output cost coefficient, curtailment cost coefficient and peak-valley arbitrage cost coefficient of t period respectively, T is the total number of periods, and N is the number of wind turbines, is the maximum power output of the jth wind turbine in the t period of the wind power typical day, is the average power output of the jth wind turbine in the t period considering wind curtailment, , , are the average power of the th energy storage on the power supply side, the th energy storage on the grid side, and the th energy storage on the user side in the t period respectively, is the length of each period, is the load power in the t period; 2. The power system multi-objective energy storage optimization method based on the improved genetic algorithm according to claim 1, characterized in that:

3. The power system multi-objective energy storage optimization method based on the improved genetic algorithm according to claim 1, characterized in that: In step A, the energy storage configuration investment cost This is calculated using the following equation: ； In the above formula, , are the unit cost of the charging and discharging power and capacity of the energy storage, respectively, and M is the fixed cost. In step A, the constraint conditions of the objective function include: Network loss constraint: Power flow constraint: ； In the above formulae, , are the voltages of nodes i, j at time period t, respectively, is the line susceptance between nodes i, j, is the phase angle difference of the voltages of nodes i, j at time period t. Power constraint: ； In the above formulae, , , , are the active power of the synchronous generator, wind turbine, energy storage and load of node i at time period t, respectively, , , , are the reactive power of the synchronous generator, wind turbine, energy storage and load of node i at time period t, respectively, is the line conductance between nodes i, j. Voltage constraint: ； In the above formulae, , , are the total active power of the synchronous generator set, the wind turbine and the energy storage output respectively in the t period, , , are the minimum total active power of the synchronous generator set, the wind turbine and the energy storage output respectively, , , are the maximum total active power of the synchronous generator set, the wind turbine and the energy storage output respectively, , are the total reactive power of the synchronous generator set, the wind turbine output respectively in the t period, , are the minimum total reactive power of the synchronous generator set, the wind turbine output respectively, , are the maximum total reactive power of the synchronous generator set, the wind turbine output respectively; Capacity constraint: ； In the above formulae, , are the minimum, maximum voltage allowed for node i, respectively. Line capacity constraint: ； In the above formula, the capacity of the energy storage for the period t, , are the minimum and maximum capacity allowed for the energy storage, respectively.

4. The power system multi-objective energy storage optimization method based on the improved genetic algorithm according to claim 1, characterized in that: ； In the above formulae, , Pi,j and Pmaxi,j are the power transmitted on the line between nodes i, j and the maximum power allowed for transmission, respectively. The step B comprises the following steps in turn: Step B3, based on the pareto dominance relationship, first filter out all non-dominated individuals in the population according to the fitness value, and take them as the first layer, and assign them a shared virtual fitness value m1, then continue to filter out non-dominated individuals from the other individuals in the population as the second layer according to the pareto dominance relationship, and assign them a shared virtual fitness value m2 smaller than m1, and repeat until all individuals in the population are layered, then calculate the crowding degree of individuals in each layer, at this time, each individual in the population has virtual fitness and crowding degree attributes, and the larger the virtual fitness, the better the individual in the layer, and the larger the crowding degree of the individual in the same layer, the better; Step B1, set the population size , the maximum number of iterations , crossover and mutation probability, and then use decimal encoding method to set six gene positions on each individual, three of which are the node positions of the power supply side, the power grid side and the load side, and the other three are the energy storage capacities corresponding to the power supply side, the power grid side and the load side; Step B2, randomly generate the initial population with the scale of for the power supply side, the grid side, the load side and the arbitrary energy storage capacity respectively, and then bring each individual into the objective function, i.e. the fitness function, to calculate its fitness value; 5. The power system multi-objective energy storage optimization method based on the improved genetic algorithm according to claim 4, characterized in that: Step B4, first select the better individuals in the above population to perform crossover and mutation operation, generate a child population with a size of , then merge the child population with the parent population to form a population with a size of , and use the method described in step B3 to perform non-dominated sorting and crowding calculation, and select the top better individuals as the next generation population; Step B5, cycle repeating step B4 until the maximum number of iterations is reached At this time, the optimal individual is selected as the energy storage optimization scheme. The mutation operation is realized by randomly selecting the jth gene of individual i for mutation: In step B4, the crossover operation adopts the decimal crossover method, and the jth gene of the individual k and the individual l is crossed according to a certain proportion Hybridization: ； In the above formulae, , are the jth gene of individuals k and l, respectively; 6. The power system multi-objective energy storage optimization method based on the improved genetic algorithm according to claim 4, characterized in that: ； In the above formula, , are the jth gene of the individual i before and after variation, respectively, is a constant matrix, is a random number generated according to a normal distribution. In step B1, 3T gene positions are also set on each individual, which are the charging and discharging strategies of the energy storage on the power supply side, the grid side and the load side in T time periods; In step B5, the energy storage optimization scheme further includes the charging and discharging strategies of the energy storage in each time period. ​

Images