Multi-target adaptive scheduling method based on optical storage integrated system

Abstract

The invention discloses a multi-target adaptive scheduling method based on a light storage integrated system, belongs to the technical field of power distribution network optimization, and aims to improve the electric energy utilization efficiency of a charging station. Firstly, a multi-objective optimization model is established by taking maximization of the electric energy utilization rate, the charging pile satisfaction rate and the photovoltaic consumption degree as objectives; secondly, in order to avoid the premature problem of the traditional NSGA-III algorithm, a self-adaptive binary mutation operator is constructed to improve the population diversity; and finally, aiming at the problem that a Pareto optimal solution set obtained by multi-objective optimization solution is difficult to screen, a unique optimal solution is selected from the Pareto optimal solution set by adopting a fuzzy analytic hierarchy process, and the strategy can better realize optimal utilization of electric energy.

Description

Technical Field

[0001] The present invention relates to the technical field of power system and distribution network optimization, and in particular to a multi-objective adaptive scheduling method based on a photovoltaic storage integrated system. Background Art

[0002] As the world pays more and more attention to environmental protection and sustainable development, the popularity of electric vehicles has become an irreversible trend as a key means to reduce carbon emissions, reduce dependence on fossil energy and improve traffic pollution. However, with the surge in the number of electric vehicles, their charging demand poses a huge challenge to the power grid. The charging demand of electric vehicles shows a high degree of uncertainty and transientity in terms of time and space distribution, especially the peak and trough phenomenon, which greatly impacts the safe and stable operation of the power grid and affects the quality of power supply.

[0003] At present, photovoltaic and energy storage charging stations face two core problems: First, the phenomenon of photovoltaic abandonment is serious, that is, during the peak period of photovoltaic output, due to insufficient charging capacity of the energy storage system or improper scheduling strategy, some photovoltaic power cannot be effectively stored or utilized, resulting in waste of resources; second, the phenomenon of charging piles reducing power supply frequently occurs, that is, during the peak period of electric vehicle charging demand, the energy storage system’s discharge capacity is insufficient to meet the full-power operation requirements of all charging piles, affecting the user’s charging experience.

[0004] A good energy scheduling strategy for a photovoltaic and energy storage charging station can greatly bring into play the role of photovoltaics and energy storage, and plays a vital role in the operation optimization of the photovoltaic and energy storage charging station. The energy scheduling strategy affects the charging and discharging capacity of the energy storage system by changing the power state of the energy storage system. If the discharge capacity of the energy storage system is insufficient, the charging pile will not be able to supply enough power when the peak period of EV charging demand arrives, so the charging pile needs to reduce the power supply, affecting the charging quality of the EV. In addition, if the charging capacity of the energy storage system is weak, it will not be able to fully absorb the photovoltaic output during the peak period of photovoltaic output, that is, photovoltaic abandonment occurs, wasting the photovoltaic output capacity. Regarding the common problems of photovoltaic abandonment and charging pile power reduction in photovoltaic charging stations, there is currently no relevant research work that proposes an energy scheduling solution that takes into account the optimization of both.

[0005] Therefore, the prior art still needs to be improved and enhanced. Summary of the invention

[0006] In view of the above-mentioned deficiencies in the prior art, the purpose of the present invention is to provide a multi-objective adaptive scheduling method based on a photovoltaic storage integrated system, aiming to solve the problems of serious photovoltaic abandonment and easy power reduction of charging piles in photovoltaic storage charging stations in the prior art.

[0007] In order to achieve the above object, the present invention adopts the following technical solutions:

[0008] A multi-objective adaptive scheduling method based on a photovoltaic storage integrated system, comprising:

[0009] Step 1: Obtain the energy demand data of each site in the photovoltaic storage charging station, and establish a multi-objective optimization model with the goal of maximizing the power utilization rate, charging pile satisfaction rate and photovoltaic absorption rate;

[0010] Step 2: Use the reference point-based non-dominated sorting adaptive genetic algorithm AD-NSGA-Ⅲ algorithm to solve the objectives in the multi-objective optimization model, obtain the optimal energy value, and perform energy allocation scheduling with the optimal energy value;

[0011] Step 3: Use the fuzzy hierarchical analysis method to determine the normalized weights of the multi-dimensional fitness values, select the solution with the smallest weighted sum from the optimal solution set, and verify the rationality of the energy allocation and scheduling method.

[0012] Furthermore, the step 2 includes:

[0013] Step 1.1, initialize various parameters such as population size and maximum number of iterations in the AD-NSGA-Ⅲ algorithm, create a reference point set, and generate an initial population;

[0014] Step 1.2: According to the particle update principle, the population is crossovered using a binary crossover operator, and the population is mutated using the improved adaptive polynomial mutation operator in the algorithm to generate a descendant population;

[0015] Step 1.3, merge the parent generation and the offspring generation to form a new population, and perform non-dominated sorting on the initial population according to the non-dominated sorting method;

[0016] Step 1.4: Standardize the population according to the update algorithm, find the reference points of the associated individuals, and use the reference points to construct the optimal solution set;

[0017] Step 1.5: Perform elite retention operation to retain elite individuals to the next generation;

[0018] Step 1.6: Select the optimal solution with the smallest weighted sum from the optimal solution set, form the optimal energy scheduling strategy based on the mapping from individuals to scheduling results, and the algorithm terminates.

[0019] Furthermore, the step 1.1 includes:

[0020] An individual represents an energy scheduling strategy. The chromosome X of individual i i As shown in the formula:

[0021]

[0022] In the formula, x′ i(t=1,2,…,m) represents the energy storage charging and discharging power coefficient of individual i in the tth period, m represents the number of time periods in a day, the energy storage charging and discharging power can be calculated, and then the power relationship is used to obtain the power of each period of the system;

[0023] According to the power of each unit in the photovoltaic charging station at each time period, the target value of each dimension r1, r2, r3…r i , then the fitness value of the j-th dimension of individual i can be defined as shown in the formula:

[0024]

[0025] Among them, r j i represents the target value of the jth dimension of individual i, r j min represents the minimum target value of the jth dimension of all individuals in the current population, r j max Indicates the maximum target value of the jth dimension of all individuals in the current population;

[0026] The individual update steps of the population t generation are as follows: First, the parent generation P with a population size of n t and offspring Q t Merge into a new population R t , that is, R t =P t ∪Q t , from the new population R t Select n individuals from the 2n individuals to form the next generation population; in order to achieve this selection process, it is necessary to first t Perform non-dominated sorting and divide it into multiple non-dominated layers (F1, F2, ..., F n ), and then build a new population S from the first layer F1 t , gradually add the next layer to S t until its size is greater than or equal to n, the last layer of this process is recorded as the lth layer, and all individuals above the l+1th layer are eliminated; if S t If the number of individuals in the lth layer is greater than n, it is necessary to remove some individuals from the last accepted layer to keep the number of individuals in the population at n. At this time, it is necessary to measure the diversity of the population to select the individuals to be retained from the lth layer, and the selection mechanism based on the reference point is adopted for the selection of the lth layer.

[0027] Furthermore, the step 1.2 includes:

[0028] The adaptive mutation probability of individual i can be expressed as follows:

[0029]

[0030] Among them, pi m represents the adaptive mutation probability of individual i, d i p i m It represents the distance between the fitness value of individual i and the ideal point, and dim represents the dimension of the optimization target.

[0031] Furthermore, the step 1.2 also includes:

[0032] The input of the algorithm includes the parent population Pop, the multi-dimensional fitness of the parent population F = {F1, F2, ... F m} and distribution index η; where the multidimensional fitness value F of individual i i ={f1 i ,f2 i ,…f dim i}; First, initialize the offspring population to the parent population. The purpose of initialization is to make the encoding of individuals that have not mutated consistent with the parent population; after initialization, traverse the individuals in the population, according to F i Calculate the mutation probability p of individual i i m ; If the random decimal s1 with a value in [0,1] is less than p i m The mutation operation is performed, otherwise no mutation operation is performed.

[0033] Furthermore, the step 1.5 includes:

[0034] After initialization mutation, individual X is the parent individual Pop i , then traverse the m-bit code of individual X, and use the polynomial mutation operator to perform mutation operation on each bit, and finally change the i-th individual Child i is the mutated individual X. After the above mutation process, the offspring population Child is finally output.

[0035] Furthermore, the step 1.6 includes:

[0036] The mapping method from individuals to scheduling results is as follows. First, the update is calculated using the following formula: The formula for converting individual codes into energy storage charging and discharging power is as follows:

[0037]

[0038] According to the energy storage capacity limit, the corrected energy storage charging and discharging power is calculated as follows:

[0039]

[0040] Then, we can use the following formula to find ppv,real , p cps,get and p grid , as shown below:

[0041]

[0042] Finally, the energy storage charging and discharging power in each period can be obtained Actual photovoltaic power consumption in each period p pv,real , the actual power p obtained by the charging pile in each period cps,get , the power p of the AC distribution network connected to the DC bus grid , the power strategy of each unit in each time period is obtained, that is, the energy scheduling strategy for each time period is formed.

[0043] The technical solution adopted by the present invention has the following beneficial effects:

[0044] In this application, a multi-objective optimization model is established with the goal of maximizing the power utilization efficiency of charging stations, the satisfaction rate of charging piles, and the photovoltaic absorption degree. An improved reference point-based non-dominated sorting adaptive genetic algorithm (AD_NSGA-Ⅲ) is proposed, which can adaptively adjust the mutation probability according to the current individual multi-dimensional objective function value to achieve better exploration effect. The fuzzy hierarchical analysis method is used to determine the normalized weight coefficients of multiple target values ​​to select the only optimal solution from the Pareto optimal solution set, avoiding the subjectivity of directly using fixed weight coefficients. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A schematic diagram of the structure of a photovoltaic storage charging station microgrid based on a multi-objective adaptive scheduling method of a photovoltaic storage integrated system provided by the present invention;

[0046] Figure 2 An algorithm flow chart of a multi-objective adaptive scheduling method based on a photovoltaic storage integrated system provided by the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and effect of the present invention clearer and more specific, the present invention is further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0048] A multi-objective adaptive scheduling method based on a photovoltaic storage integrated system, comprising:

[0049] Step 1: Obtain the energy demand data of each site in the photovoltaic storage charging station, and establish a multi-objective optimization model with the goal of maximizing the power utilization rate, charging pile satisfaction rate and photovoltaic absorption rate;

[0050] Establish a multi-objective energy scheduling model, such as Figure 1As shown, the schematic diagram of the microgrid structure of a photovoltaic storage charging station provided by the present invention mainly includes the following five system units: photovoltaic power generation system, energy storage system, AC distribution network system, charging pile system and energy dispatching system. Among them, the four system units of photovoltaic power generation system, energy storage system, AC distribution network system and charging pile system complete energy transmission on the DC bus through DC / DC converter or AC / DC converter. The energy dispatching system generates and sends control instructions to the corresponding DC / DC or AC / DC unit through a multi-objective energy dispatching strategy to execute power regulation of the corresponding module and allocate the energy transmission power of each system unit on the DC bus.

[0051] Step 2: Use the reference point-based non-dominated sorting adaptive genetic algorithm AD-NSGA-Ⅲ algorithm to solve the objectives in the multi-objective optimization model, obtain the optimal energy value, and perform energy allocation scheduling with the optimal energy value;

[0052] Step 3: Use the fuzzy hierarchical analysis method to determine the normalized weights of the multi-dimensional fitness values, select the solution with the smallest weighted sum from the optimal solution set, and verify the rationality of the energy allocation and scheduling method.

[0053] Further, see Figure 2 , the step 2 includes:

[0054] Step 1.1, initialize various parameters such as population size and maximum number of iterations in the AD-NSGA-Ⅲ algorithm, create a reference point set, and generate an initial population;

[0055] Step 1.2: According to the particle update principle, the population is crossovered using a binary crossover operator, and the population is mutated using the improved adaptive polynomial mutation operator in the algorithm to generate a descendant population;

[0056] Step 1.3, merge the parent generation and the offspring generation to form a new population, and perform non-dominated sorting on the initial population according to the non-dominated sorting method;

[0057] Step 1.4: Standardize the population according to the update algorithm, find the reference points of the associated individuals, and use the reference points to construct the optimal solution set;

[0058] Step 1.5: Perform elite retention operation to retain elite individuals to the next generation;

[0059] Step 1.6: Select the optimal solution with the smallest weighted sum from the optimal solution set, form the optimal energy scheduling strategy based on the mapping from individuals to scheduling results, and the algorithm terminates.

[0060] Specifically, in the AD-NSGA-Ⅲ algorithm, crossover operators and mutation operators are used to generate new individuals to construct the offspring population. Individuals crossover according to a fixed probability, and then mutate according to a fixed probability. The basic NSGA-III algorithm with a fixed mutation probability has an indifferent mutation mode, that is, the mutation rates of better individuals and worse individuals are the same. At the same time, the adaptive mutation probability proposed in this patent is used to increase the mutation probability of individuals with poor fitness and reduce the mutation probability of individuals with good fitness, so that the optimal solution can always be found. The specific implementation steps are as follows:

[0061] (1) An individual represents an energy scheduling strategy. The chromosome X of individual i i As shown in the formula:

[0062]

[0063] In the formula, x′ i (t=1, 2, ..., m) represents the energy storage charging and discharging power coefficient of individual i in the tth period, and m represents the number of time periods in a day. The energy storage charging and discharging power can be calculated, and then the power relationship can be used to obtain the power of the system in each period.

[0064] (2) According to the power of each unit in the system in each time period, the target values ​​r1, r2 and r3 of the three dimensions are calculated. Then the fitness value of the jth dimension of individual i can be defined as shown in the formula:

[0065]

[0066] Among them, r j i represents the target value of the jth dimension of individual i, r j min represents the minimum target value of the jth dimension of all individuals in the current population, r j max It indicates that the maximum target value of the jth dimension of all individuals in the current population is negatively correlated with the target value of each dimension. The smaller the fitness value, the larger the target value and the more ideal the individual.

[0067] (3) The steps for updating individuals in the tth generation of the population are as follows. First, the parent generation P whose population size is n t and offspring Q t Merge into a new population R t , that is, R t =P t ∪Q t From the new population R t In order to realize this selection process, we need to select n individuals from the 2n individuals to form the next generation population. t Perform non-dominated sorting and divide it into multiple non-dominated layers (F1, F2, ..., F n), and then build a new population S from the first layer F1 t , gradually add the next layer to S t Until its size is greater than or equal to n, the last layer of this process is recorded as the lth layer, and all individuals above the l+1th layer are eliminated. t If the number of individuals in the population is greater than n, it is necessary to remove some individuals from the last accepted layer (the lth layer) to keep the number of individuals in the population at n. At this time, it is necessary to measure the diversity of the population to select the individuals to be retained from the lth layer, and the selection mechanism based on the reference point is adopted for the selection of the lth layer.

[0068] (4) In the NSGA-III algorithm, the process of generating new individuals relies on the crossover operator and the mutation operator to construct the offspring population. These individuals first crossover according to a fixed probability, and then mutate according to a fixed probability. However, when the basic NSGA-III algorithm uses a fixed mutation probability, the mutation rate of individuals is the same for individuals with better fitness and individuals with worse fitness. This indiscriminate mutation method causes the mutation direction to become random and blind, so it is not always possible to effectively find the optimal solution. In order to solve this problem, this patent proposes a mutation operator with adaptive mutation probability. The operator can dynamically adjust the mutation probability according to the fitness value of the individual, so that individuals with worse fitness have a higher mutation probability, while individuals with better fitness have a lower mutation probability. This method aims to optimize the mutation process and improve the efficiency of the algorithm's optimization. The adaptive mutation probability of individual i can be expressed as follows:

[0069]

[0070] Among them, p i m represents the adaptive mutation probability of individual i, d i p i m It represents the distance between the fitness value of individual i and the ideal point, and dim represents the dimension of the optimization target.

[0071] (5) The input of the algorithm includes the parent population Pop, the multi-dimensional fitness of the parent population F = {F1, F2, ... F m} and distribution index η. The multi-dimensional fitness value F of individual i i ={f1 i ,f2 i ,…f dim i First, initialize the offspring population to the parent population. The purpose of initialization is to make the encoding of individuals that have not mutated consistent with the parent population. After initialization, traverse the individuals in the population and i Calculate the mutation probability p of individual i im If the random decimal s1 with a value in [0,1] is less than p i m The mutation operation is performed, otherwise no mutation operation is performed.

[0072] (6) After initialization, individual X is the parent individual Pop i , then traverse the m-bit code of individual X, and use the polynomial mutation operator to perform mutation operation on each bit, and finally change the i-th individual Child i is the mutated individual X. After the above mutation process, the offspring population Child is finally output.

[0073] (7) The mapping method from individuals to scheduling results is as follows. First, the update is calculated using the following formula: The formula for converting individual codes into energy storage charging and discharging power is as follows:

[0074]

[0075] According to the energy storage capacity limit, the corrected energy storage charging and discharging power is calculated as follows:

[0076]

[0077] Then, we can use the following formula to find p pv,real , p cps,get and p grid , as shown below:

[0078]

[0079] Finally, the energy storage charging and discharging power in each period can be obtained Actual photovoltaic power consumption in each period p pv,real , the actual power p obtained by the charging pile in each period cps,get , the power p of the AC distribution network connected to the DC bus grid , the power strategy of each unit in each time period is obtained, that is, the energy scheduling strategy for each time period is formed.

[0080] like Figure 2 As shown, Figure 2 The algorithm flow of the present invention is given, and the advantages of this patent are explained with specific examples:

[0081] (1) After the start, the resource pool will be initialized and the reference point set will be created. The initial population will be generated according to the above steps. Then the population size, maximum number of iterations, crossover parameter, mutation parameter and crossover probability will be set.

[0082] (2) The model input is the time-of-use electricity price at different times of the day, the time-of-use electricity selling price of the charging pile, the predicted demand power of the charging pile and the predicted photovoltaic power generation power. This patent uses different energy scheduling schemes as an individual of the algorithm to generate an initial population, and calculates the energy storage charging and discharging power and the power of each unit of the system in each time period in turn. According to the power of each time period, the target values ​​of three dimensions can be calculated, and the fitness of different individuals in different dimensions can be defined. The fitness of each dimension is related to its target value. The smaller the fitness value, the larger the target value and the more ideal the individual.

[0083] (3) Compared with the later generated offspring population, its previous generation population is equivalent to its parent population. The initial population is subjected to crossover and mutation operations. Through the crossover operation, the excellent solutions of the parent population are retained and a better solution is sought through different sorting. The offspring population obtained by the simple crossover operation does not introduce new solutions, and it is difficult to obtain a more optimized solution. Therefore, through the mutation operation, the introduced new solutions are used to expand the data resource pool of the offspring population to obtain a better solution. In the process of crossover and mutation operations, crossover operators and mutation operators are introduced. The creation of offspring populations is standardized by crossover operators and mutation operators. Individuals in the offspring population will crossover according to the fixed probability of the crossover operator, and then mutate indiscriminately according to the fixed probability of the mutation operator. However, in the traditional mutation algorithm, the mutation rates of the better individuals and the worse individuals are the same, resulting in the randomness and blindness of the mutation direction, making it difficult to find the optimal solution. This patent proposes a new mutation operator with adaptive mutation probability. The operator continuously updates parameters in the iterative process and dynamically changes the mutation probability, thereby achieving a more optimized solution and reducing the system's computing time.

[0084] (4) The parent population and the offspring population are merged to form a new population, and each individual is encoded to achieve the effect of differentiation.

[0085] (5) The synthesized population is standardized, each individual is associated with a point to form an associated reference point, and the obtained reference point is used to construct the optimal solution set.

[0086] (6) Perform elite retention operation, sort all reference points according to the energy utilization efficiency of the photovoltaic power station and the amount of photovoltaic abandoned light, divide them into multiple non-dominated layers, retain the first n reference points, and enter the next ring solution as the parent population.

[0087] (7) If the maximum number of cycles is not reached, return to the crossover and mutation operations to form a new merged population for further cycles; if the maximum number of cycles is reached, output the optimal solution set.

[0088] (8) Calculate all strategies corresponding to the optimal solution set, obtain the optimal solution that best suits the current state, and output the optimal solution.

[0089] The multi-objective adaptive scheduling method based on the integrated photovoltaic and storage system proposed in this patent uses an adaptive algorithm to perform optimization and improve the adaptive mutation probability, which can significantly improve the energy utilization efficiency of the storage power station and reduce the amount of photovoltaic abandoned light. At the same time, it reduces the system calculation amount and improves the response speed of the overall system, which can meet the needs of maximizing the energy utilization of the photovoltaic storage charging station.

[0090] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the schemes disclosed herein. The present invention is intended to cover any variations, uses or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art that are not disclosed in this disclosure. The description and examples are to be considered exemplary only, and the true scope and spirit of the present invention are indicated by the claims.

Claims

1. A multi-objective adaptive scheduling method based on a photovoltaic storage integrated system, characterized in that: include: Step 1: Obtain the energy demand data of each site in the photovoltaic storage charging station, and establish a multi-objective optimization model with the goal of maximizing the power utilization rate, charging pile satisfaction rate and photovoltaic absorption rate; Step 2: Use the reference point-based non-dominated sorting adaptive genetic algorithm AD-NSGA-Ⅲ algorithm to solve the objectives in the multi-objective optimization model, obtain the optimal energy value, and perform energy allocation scheduling with the optimal energy value; Step 3: Use the fuzzy hierarchical analysis method to determine the normalized weights of the multi-dimensional fitness values, select the solution with the smallest weighted sum from the optimal solution set, and verify the rationality of the energy allocation and scheduling method.

2. The multi-objective adaptive scheduling method based on the photovoltaic storage integrated system according to claim 1 is characterized in that: The step 2 includes: Step 1.1, initialize various parameters such as population size and maximum number of iterations in the AD-NSGA-Ⅲ algorithm, create a reference point set, and generate an initial population; Step 1.2: According to the particle update principle, the population is crossovered using a binary crossover operator, and the population is mutated using the improved adaptive polynomial mutation operator in the algorithm to generate a descendant population; Step 1.3, merge the parent generation and the offspring generation to form a new population, and perform non-dominated sorting on the initial population according to the non-dominated sorting method; Step 1.4: Standardize the population according to the update algorithm, find the reference points of the associated individuals, and use the reference points to construct the optimal solution set; Step 1.5: Perform elite retention operation to retain elite individuals to the next generation; Step 1.6: Select the optimal solution with the smallest weighted sum from the optimal solution set, form the optimal energy scheduling strategy based on the mapping from individuals to scheduling results, and the algorithm terminates.

3. The multi-objective adaptive scheduling method based on the photovoltaic storage integrated system according to claim 2 is characterized in that: The step 1.1 includes: An individual represents an energy scheduling strategy. The chromosome X of individual i i As shown in the formula: In the formula, x′ i (t=1,2,…,m) represents the energy storage charging and discharging power coefficient of individual i in the tth period, m represents the number of time periods in a day, the energy storage charging and discharging power can be calculated, and then the power relationship is used to obtain the power of each period of the system; According to the power of each unit in the photovoltaic charging station at each time period, the target value of each dimension r1, r2, r3…r i , then the fitness value of the j-th dimension of individual i can be defined as shown in the formula: Among them, r j i represents the target value of the jth dimension of individual i, r j min represents the minimum target value of the jth dimension of all individuals in the current population, r j max Indicates the maximum target value of the jth dimension of all individuals in the current population; The individual update steps of the population t generation are as follows: First, the parent generation P with a population size of n t and offspring Q t Merge into a new population R t , that is, R t =P t ∪Q t , from the new population R t Select n individuals from the 2n individuals to form the next generation population; in order to achieve this selection process, it is necessary to first t Perform non-dominated sorting and divide it into multiple non-dominated layers (F1, F2, ..., F n ), and then build a new population S from the first layer F1 t , gradually add the next layer to S t until its size is greater than or equal to n, the last layer of this process is recorded as the lth layer, and all individuals above the l+1th layer are eliminated; if S t If the number of individuals in the lth layer is greater than n, it is necessary to remove some individuals from the last accepted layer to keep the number of individuals in the population at n. At this time, it is necessary to measure the diversity of the population to select the individuals to be retained from the lth layer, and the selection mechanism based on the reference point is adopted for the selection of the lth layer.

4. The multi-objective adaptive scheduling method based on the photovoltaic storage integrated system according to claim 2 is characterized in that: The step 1.2 includes: The adaptive mutation probability of individual i can be expressed as follows: Among them, p i m represents the adaptive mutation probability of individual i, d i p i m It represents the distance between the fitness value of individual i and the ideal point, and dim represents the dimension of the optimization target.

5. The multi-objective adaptive scheduling method based on the photovoltaic storage integrated system according to claim 2 is characterized in that: The step 1.2 also includes: The input of the algorithm includes the parent population Pop, the multi-dimensional fitness of the parent population F = {F1, F2, ... F m } and distribution index η; where the multidimensional fitness value F of individual i i ={f1 i ,f2 i ,…f dim i }; First, initialize the offspring population to the parent population. The purpose of initialization is to make the encoding of individuals that have not mutated consistent with the parent population; after initialization, traverse the individuals in the population, according to F i Calculate the mutation probability p of individual i i m ; If the random decimal s1 with a value in [0,1] is less than p i m The mutation operation is performed, otherwise no mutation operation is performed.

6. The multi-objective adaptive scheduling method based on the photovoltaic storage integrated system according to claim 2 is characterized in that: The step 1.5 includes: After initialization mutation, individual X is the parent individual Pop i , then traverse the m-bit code of individual X, and use the polynomial mutation operator to perform mutation operation on each bit, and finally change the i-th individual Child i is the mutated individual X. After the above mutation process, the offspring population Child is finally output.

7. The multi-objective adaptive scheduling method based on the photovoltaic storage integrated system according to claim 2 is characterized in that: The step 1.6 includes: The mapping method from individuals to scheduling results is as follows. First, the update is calculated using the following formula: The formula for converting individual codes into energy storage charging and discharging power is as follows: According to the energy storage capacity limit, the corrected energy storage charging and discharging power is calculated as follows: Then, we can use the following formula to find p pv,real , p cps,get and p grid , as shown below: Finally, the energy storage charging and discharging power in each period can be obtained Actual photovoltaic power consumption in each period p pv,real , the actual power p obtained by the charging pile in each period cps,get , the power p of the AC distribution network connected to the DC bus grid , the power strategy of each unit in each time period is obtained, that is, the energy scheduling strategy for each time period is formed.

Images