Equivalent model construction method and system of power distribution network

Abstract

The invention belongs to the technical field of power distribution network equivalent model modeling, and particularly relates to a power distribution network equivalent model construction method and system, and the method comprises the steps: determining equipment contained in a power distribution network of a to-be-constructed equivalent model, and building a power distribution network model; establishing a comprehensive objective function of power loss and cost based on the power distribution network model, and defining constraint conditions of the comprehensive function; based on the comprehensive objective function and constraint conditions thereof, defining particle swarm algorithm parameters and a solution space, initializing a specified number of particles, and performing iterative updating on the initialized specified number of particles; operating a particle swarm algorithm, and carrying out iterative updating on the optimized and initialized specified number of particles based on a comprehensive objective function until an optimal solution meeting a specified convergence condition or an optimal solution reaching the maximum number of iterations is obtained; and verifying the optimal solution based on the comprehensive objective function and the constraint condition. By performing equivalent modeling on the power distribution network after the distributed new energy power generation system is accessed, stable and reliable analysis on the power distribution network system is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution network equivalent model building, and in particular relates to a distribution network equivalent model building method and system. Background Art

[0002] With the rapid development of new energy technologies, more and more distributed new energy power generation systems are connected to the traditional distribution network. The access of distributed new energy power generation systems will greatly affect the load characteristics of the distribution network. The equivalent model of the traditional distribution network is based on a constant power source. However, as more and more distributed new energy power generation systems are connected to the traditional distribution network, it is obvious that they are not constant power sources. Therefore, the equivalent model of the traditional distribution network cannot be applied to the distribution network with distributed new energy power generation systems.

[0003] Therefore, how to model the equivalent model of the distribution network after the distributed renewable energy power generation system is connected to achieve stable and reliable analysis of the distribution network system is a technical problem that needs to be solved urgently. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for constructing an equivalent model of a distribution network to solve the technical problem.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is as follows: In a first aspect, a method for constructing an equivalent model of a distribution network is provided, comprising the following steps: S1: Determine the equipment included in the distribution network for which the equivalent model is to be constructed, divide the distribution network into multiple branches based on the connection status of the equipment in the distribution network, and establish a distribution network model based on the equipment and the multiple branches; S2: establishing a comprehensive objective function based on the power loss and cost of the multiple branches of the distribution network model, and defining the constraint conditions of the comprehensive function as the capacity of the multiple branches and the voltage stability; S3: Based on the comprehensive objective function and its constraints, define the particle swarm algorithm parameters and solution space, initialize a specified number of particles, iteratively update the initialized specified number of particles, and randomly distribute the iteratively updated initialized specified number of particles in the solution space; S4: running the particle swarm algorithm, and iteratively updating the optimized initialized specified number of particles based on the comprehensive objective function until an optimal solution that meets a specified convergence condition or an optimal solution with a maximum number of iterations is obtained; S5: verifying the optimal solution based on the comprehensive objective function and constraint conditions, calculating the fitness of the optimal solution, and when the fitness of the optimal solution is greater than a specified threshold, operating the distribution network model with the optimal solution.

[0006] Preferably, the equipment included in the distribution network in step S1 includes a photovoltaic power generation system, a mains access module, a transformer, an inverter, a converter, a filter, a capacitor, an inductor and a load; The particle swarm algorithm parameters defined in step S3 include the number of particles and the number of iterations.

[0007] Preferably, the formula of the comprehensive objective function of the power loss and cost of multiple branches in step S2 is as follows: ; in, f g ( x ) is the comprehensive objective function, C1( x ) is the distribution network loss cost, C2( x ) is the electricity cost of energy storage operation, C3( x ) is the power purchase cost of the upper power grid, C4( x ) are other specified costs, i is the branch number, Ri( x ) is the i The resistance of the branch, Pi( x ) is the i Active power of the branch, Qi( x ) is the i Reactive power of the branch, Ui( x ) is the i The voltage at the end of the branch.

[0008] Preferably, the specific process of initializing a specified number of particles and iteratively updating the initialized specified number of particles in step S3 is as follows: S31: Define the size of the initialized particle population as N, and the decision variable as n , then the particle population is initialized as P(0)={ X 1, X 2, X 3,···, X N}; X i =( x i1 , x i2 , x i3 ,···, x in ); in, X 1, X 2, X 3,···,X N is a population of N particles, X i For the i A particle population, x i1 , x i2 , x i3 ,···, x in For the i In the particle population n Particles; S32: Based on the objective function, the fitness of the individuals in the initialized particle population is evaluated, and a number of individuals are randomly selected from the individuals whose fitness values ​​are greater than the preset fitness threshold to compete, and the winning individuals are selected to enter the next generation, and the random selection of a number of individuals for competition is repeated to obtain multiple winning individuals to enter the next generation; S33: randomly selecting multiple crossover points from the multiple winning individuals to perform gene exchange to generate multiple new individuals. The specific formula is as follows; ; in, m is the number of intersection points, k 1, k 2, k 3. ··· , k m-1 is the intersection point; S34: The individuals in step S33 are Y i No. j Gene location Y ij Perform mutation processing. The specific formula for mutation processing is as follows: ; in, is a random disturbance, which obeys normal distribution, P m is the mutation probability; S35: Iteratively execute steps S32 to S34 to iteratively update the initialized particle population P(0).

[0009] Preferably, after iteratively executing steps S32 to S34 and iteratively updating the initialized particle population P(0), the iteratively updated particle population is optimized. The specific process is as follows: S36: For the iteratively updated particles, d In the dimensional solution space, the position of the best individual particle in the iteratively updated particles is expressed as: X 1* , X 2 * , X 3 * ,···, X d * ), individual particles X j The position is ( X 1 j , X 2 j , X 3 j ,···, X d j ); S37: Based on individual particles X j Under the influence of the best individual particle, calculate the next generation of individual particles X j+1 Location( X 1 j+1 , X 2 j+1 , X 3 j+1 ,···, X d j +1 ), the specific calculation formula is: X k j+1 = X k * -(2 a · r 1- a )·|2 r 2. X k * - X k j |; in, X k j+1 Represents the solution space coordinates X j+1 No. k Quantity, a As the number of iterations increases, it decreases linearly from 2 to 0. r 1 and r 2 is a random number between 0 and 1; S38: For the next generation of particle individuals X j+1 Perform spiral position update, the specific formula is: X k j+1 = X k * +| X k * - X k j | e bl cos (2π l ); in, b is the logarithmic spiral shape constant, l A random number between -1 and 1; S39: Establish a mathematical model of search and predation, X k j+1 = X k rand - A ·| X k * - X k j |, and iteratively execute the search-predation mathematical model to obtain the best particle swarm; in, A is the coefficient, X k rand For particles X k of random locations.

[0010] Preferably, the specific process of iteratively updating the optimized initialized specified number of particles based on the comprehensive objective function in step S4 is as follows: S41: Evaluate the current position of each particle in the best particle group in step S39, and calculate the objective function value of each particle in the best particle group based on the comprehensive objective function; S42: Update the speed and position of each particle in the optimal particle group, where the speed update formula is: v [ i ]= wv [i]+ c 1 r 1*( pBest [ i ]- x [ i ])+ c 2 r2*( gBest [ i ]- x [ i ]); The position update formula is: x [ i ]= x [ i ]+ v [ i ]; in, v [ i ] is the i The velocity vector of a particle, x [ i ] is the first i The position vector of a particle, w is the inertia weight, c 1 and c 2 are learning factors r 1 and r 2 is a random number, pBest [ i ] is the i The individual optimal position vector of each particle is gBest [ i ] is the current global optimal position vector ‌; S43: Iteratively execute step S41 and step S42 until a maximum number of iterations is met or a specified preset condition is met.

[0011] In a second aspect, a distribution network equivalent model construction system is provided, which is used to implement any one of the distribution network equivalent model construction methods, including a device acquisition module, a model building module, an objective function creation module, a particle swarm algorithm module, an algorithm optimization module and a verification module, wherein the device acquisition module is connected to the model building module, the model building module is connected to the objective function creation module, the objective function creation module is connected to the particle swarm algorithm module, and the particle swarm algorithm module is connected to the algorithm optimization module; The device acquisition module is used to acquire the devices in the distribution network to be modeled, including photovoltaic power generation systems, city power access modules, transformers, inverters, converters, filters, capacitors, inductors and loads; The model building module is used to divide the distribution network into multiple branches based on the connection relationship of the equipment, and to build a distribution network model based on the equipment and the multiple branches; The objective function creation module is used to establish a comprehensive objective function based on the power loss and cost of the multiple branches of the distribution network model, and define the constraint conditions of the comprehensive function as the capacity of the multiple branches and the voltage stability; The particle swarm algorithm module is used to define the particle swarm algorithm parameters and solution space based on the comprehensive objective function and its constraints, and initialize a specified number of particles, and iteratively update the initialized specified number of particles, and randomly distribute the iteratively updated initialized specified number of particles in the solution space; The algorithm optimization module is used to iteratively update and optimize the optimized initialized specified number of particles based on the comprehensive objective function until an optimal solution that meets a specified convergence condition or reaches an optimal solution with a maximum number of iterations is obtained; The verification module is used to verify the optimal solution based on the comprehensive objective function and constraint conditions, calculate the fitness of the optimal solution, and when the fitness of the optimal solution is greater than a specified threshold, operate the distribution network model with the optimal solution.

[0012] The beneficial effects of the present invention include: The invention provides a method and system for constructing an equivalent model of a distribution network. The method and system establish a distribution network model by determining the equipment included in the distribution network of which the equivalent model is to be constructed; establish a comprehensive objective function based on the power loss and cost of the distribution network model, and define the constraints of the comprehensive function; define particle swarm algorithm parameters and solution space based on the comprehensive objective function and its constraints, initialize a specified number of particles, and iteratively update the initialized specified number of particles; run the particle swarm algorithm, and iteratively update the optimized initialized specified number of particles based on the comprehensive objective function until an optimal solution that meets the specified convergence condition or an optimal solution that reaches the maximum number of iterations is obtained; and verify the optimal solution based on the comprehensive objective function and the constraints, so as to realize dynamic equivalent modeling of the distribution network after the distributed new energy power generation system is connected, thereby realizing stable and reliable analysis of the distribution network after the distributed new energy power generation system is connected, and realizing operation of the entire distribution network with minimum comprehensive power consumption and cost through dynamic equivalent modeling.

[0013] Firstly, by establishing the formula of the comprehensive objective function of power loss and cost of multiple branches, and realizing the dynamic solution of the optimal solution of the comprehensive objective function of power loss and cost of multiple branches based on the particle swarm algorithm, it is convenient to realize the operation of the distribution network with the minimum comprehensive power consumption and cost based on the optimal solution.

[0014] Secondly, by defining the initial particle population size as N, the decision variable is n, initialize the particle population, evaluate the fitness of the individuals in the initialized particle population based on the objective function, and randomly select several individuals from the individuals whose fitness values ​​are greater than the preset fitness threshold to compete, and select the winning individuals to enter the next generation, randomly select multiple crossover points from multiple winning individuals for gene exchange, generate multiple new individuals, and perform mutation processing to achieve iterative update of the initialized particle population, realize the optimization of the particle swarm algorithm, and improve the accuracy of obtaining the optimal solution through the optimized particle swarm algorithm.

[0015] Again, through the iteratively updated particles, d In the dimensional solution space, the position of the best individual particle in the iteratively updated particles is obtained, and based on the individual particle X Under the influence of the best individual particle, the position of the next generation of individual particles is calculated, the position of the next generation of individual particles is updated in a spiral manner, a search-predation mathematical model is established, and the search-predation mathematical model is iteratively executed to obtain the best particle group. In the evolution of each generation, the evolution direction of the population is guided by fitness function evaluation and selection operations, while the diversity of the population and the coverage of the search space are maintained through crossover and mutation operations. The particle swarm algorithm is further optimized to further improve the accuracy of the optimal solution.

[0016] Finally, the current position of each particle in the best particle swarm is evaluated, and the objective function value of each particle in the best particle swarm is calculated based on the comprehensive objective function; the speed and position of each particle in the best particle swarm are updated, and the speed and position updates are iteratively executed until the maximum number of iterations is met or the specified preset conditions are met. Through the particle speed and position update mechanism, it is possible to jump out of the local optimal solution and explore different areas of the solution space, so that it has a strong global search capability. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the flow chart of the method for constructing an equivalent model of a distribution network of the present invention.

[0018] Figure 2 A schematic diagram of a process of initializing a specified number of particles and iteratively updating the initialized specified number of particles according to the present invention.

[0019] Figure 3 A schematic diagram of the architecture of the equivalent model construction system of the power distribution network of the present invention. DETAILED DESCRIPTION

[0020] The following is combined with Figure 1~Figure 3 The present invention is further described in detail: Example 1 See attached Figure 1As shown, a method for constructing an equivalent model of a distribution network includes the following steps: S1: Determine the equipment included in the distribution network for which the equivalent model is to be constructed, and divide the distribution network into multiple branches based on the connection status of the equipment in the distribution network, and establish a distribution network model based on the equipment and the multiple branches, wherein the multiple branches include multiple photovoltaic power generation system branches and city power branches, and one branch is established for each photovoltaic power generation source; S2: Establishing a comprehensive objective function of the power loss and cost of the multiple branches based on the distribution network model, and defining the constraints of the comprehensive function as the capacity of multiple branches and voltage stability. The present invention hopes to obtain an optimal solution for minimizing the value of the comprehensive objective function of the power loss and cost of multiple branches with the capacity of multiple branches and voltage stability as constraints based on the established distribution network equivalent model, and based on the optimal solution, dynamically minimize the power loss and cost of the distribution network; S3: Based on the comprehensive objective function and its constraints, the particle swarm algorithm parameters and solution space are defined, and a specified number of particles are initialized. Each particle represents a potential solution in the solution space. Initializing particles means assigning a random position and speed to each particle, and iteratively updating the positions and speeds of the initialized specified number of particles. The iteratively updated initialized specified number of particles are randomly distributed in the solution space. A particle swarm algorithm is established for the comprehensive objective function of power loss and cost of multiple branches. The core of the algorithm is to simulate the behavior of a flock of birds. Each "particle" adjusts its own moving direction and speed by memorizing its own experience and interacting with other particles. In the algorithm, multiple particles move at the same time, using their own experience and the experience of other particles in the group to find the optimal solution. This mechanism makes the particles not only affected by their own evolution, but also learn the evolution between groups, so as to achieve the best adjustment. However, when dealing with functions with multiple local extreme points, the existing particle swarm algorithm is prone to fall into local extreme points, resulting in the inability to obtain the global optimal solution, and it is easy to fall into the local optimum, and its convergence accuracy is often low. During the iteration process, the algorithm may stop searching prematurely, thereby missing the global optimal solution. Therefore, the present invention optimizes the initialization process of the particle swarm, and improves the global search capability of the optimization algorithm by introducing different optimization methods and optimization logics, avoids premature convergence, accelerates the convergence speed, and improves the accuracy of obtaining the optimal solution, providing an effective solution to the complex optimization problem of the comprehensive objective function of power loss and cost of the distribution network.

[0021] S4: running the particle swarm algorithm, and iteratively updating the optimized initialized specified number of particles based on the comprehensive objective function until an optimal solution that meets a specified convergence condition or an optimal solution with a maximum number of iterations is obtained; S5: verifying the optimal solution based on the comprehensive objective function and constraint conditions, calculating the fitness of the optimal solution, and when the fitness of the optimal solution is greater than a specified threshold, operating the distribution network model with the optimal solution.

[0022] In this embodiment, the equipment included in the distribution network in step S1 includes a photovoltaic power generation system, a mains access module, a transformer, an inverter, a converter, a filter, a capacitor, an inductor and a load; The particle swarm algorithm parameters defined in step S3 include the number of particles and the number of iterations.

[0023] The formula of the comprehensive objective function of the power loss and cost of multiple branches in step S2 is as follows: ; in, f g ( x ) is the comprehensive objective function, C1( x ) is the distribution network loss cost, C2( x ) is the electricity cost of energy storage operation, C3( x ) is the power purchase cost of the upper power grid, C4( x ) are other specified costs, i is the branch number, Ri( x ) is the i The resistance of the branch, Pi( x ) is the i Active power of the branch, Qi( x ) is the i Reactive power of the branch, Ui( x ) is the i The voltage at the end of the branch.

[0024] Example 2 Based on Example 1, the specific process of initializing a specified number of particles and iteratively updating the initialized specified number of particles in step S3 is as follows: S31: Define the size of the initialized particle population as N, and the decision variable as n , then the particle population is initialized as P(0)={ X 1, X 2, X 3,···, X N}; X i =( x i1 , x i2 , x i3 ,···,x in ); in, X 1, X 2, X 3,···, X N is a population of N particles, X i For the i A particle population, x i1 , x i2 , x i3 ,···, x in For the i In the particle population n Particles; S32: Based on the objective function, the fitness of the individuals in the initialized particle population is evaluated, and a number of individuals are randomly selected from the individuals whose fitness values ​​are greater than the preset fitness threshold to compete, and the winning individuals are selected to enter the next generation, and the random selection of a number of individuals for competition is repeated to obtain multiple winning individuals to enter the next generation; S33: randomly selecting multiple crossover points from the multiple winning individuals to perform gene exchange to generate multiple new individuals. The specific formula is as follows; ; in, m is the number of intersection points, k 1, k 2, k 3. ··· , k m-1 is the intersection point; S34: The individuals in step S33 are Y i No. j Gene location Y ij Perform mutation processing. The specific formula for mutation processing is as follows: ; in, is a random disturbance, which obeys normal distribution, P m is the mutation probability; S35: Iteratively execute steps S32 to S34 to iteratively update the initialized particle population P(0).

[0025] The process of initializing the particle swarm in the particle swarm algorithm was optimized by defining the initialization particle population size as N and the decision variable as n, initialize the particle population, evaluate the fitness of the individuals in the initialized particle population based on the objective function, and randomly select several individuals from the individuals whose fitness values ​​are greater than the preset fitness threshold to compete, and select the winning individuals to enter the next generation, randomly select multiple crossover points from multiple winning individuals for gene exchange, generate multiple new individuals, and perform mutation processing to achieve iterative update of the initialized particle population, realize the optimization of the particle swarm algorithm, and improve the accuracy of obtaining the optimal solution through the optimized particle swarm algorithm.

[0026] Example 3 On the basis of Example 1 or Example 2, after iteratively executing steps S32 to S34 and iteratively updating the initialized particle population P(0), the iteratively updated particle population is optimized. The specific process is as follows: S36: For the iteratively updated particles, d In the dimensional solution space, the position of the best individual particle in the iteratively updated particles is expressed as: X 1 * , X 2 * , X 3 * ,···, X d * ), individual particles X j The position is ( X 1 j , X 2 j , X 3 j ,···, X d j ); S37: Based on individual particles X j Under the influence of the best individual particle, calculate the next generation of individual particles X j+1 Location( X 1 j+1 , X 2 j+1 , X 3 j+1 ,···, X d j +1 ), the specific calculation formula is: X k j+1 = X k * -(2 a ·r 1- a )·|2 r 2. X k * - X k j |; in, X k j+1 Represents the solution space coordinates X j+1 No. k Quantity, a As the number of iterations increases, it decreases linearly from 2 to 0. r 1 and r 2 is a random number between 0 and 1; S38: For the next generation of particle individuals X j+1 Perform spiral position update, the specific formula is: X k j+1 = X k * +| X k * - X k j | e bl cos (2π l ); in, b is the logarithmic spiral shape constant, l A random number between -1 and 1; S39: Establish a mathematical model of search and predation, X k j+1 = X k rand - A ·| X k * - X k j |, and iteratively execute the search-predation mathematical model to obtain the best particle swarm; in, A is the coefficient, X k rand For particles X k of random locations.

[0027] By iteratively updating the particles, d In the dimensional solution space, the position of the best individual particle in the iteratively updated particles is obtained, and based on the individual particle X Under the influence of the best individual particle, the position of the next generation of individual particles is calculated, the position of the next generation of individual particles is updated in a spiral manner, a search-predation mathematical model is established, and the search-predation mathematical model is iteratively executed to obtain the best particle swarm, further optimizing the particle swarm algorithm and further improving the accuracy of the optimal solution.

[0028] In this embodiment, the specific process of iteratively updating the optimized initialized specified number of particles based on the comprehensive objective function in step S4 is as follows: S41: Evaluate the current position of each particle in the best particle group in step S39, and calculate the objective function value of each particle in the best particle group based on the comprehensive objective function; S42: Update the speed and position of each particle in the optimal particle group, where the speed update formula is: v [ i ]= wv [i]+ c 1 r 1*( pBest [ i ]- x [ i ])+ c 2 r 2*( gBest [ i ]- x [ i ]); The position update formula is: x [ i ]= x [ i ]+ v [ i ]; in, v [ i ] is the i The velocity vector of a particle, x [ i ] is the first i The position vector of a particle, w is the inertia weight, c 1 and c 2 are learning factors r 1 and r 2 is a random number, pBest [ i] is the i The individual optimal position vector of each particle is gBest [ i ] is the current global optimal position vector‌.

[0029] S43: Iteratively execute step S41 and step S42 until a maximum number of iterations is met or a specified preset condition is met.

[0030] In the above process of updating the speed and position of each particle in the optimal particle group, it is also necessary to check the fitness of each particle's current position and the individual historical optimal position. If the current position is better, the individual optimal position is updated. If the current position is not better, the individual historical optimal position is retained. The speed and position of the particle are updated by considering the individual historical optimal position and the group optimal position.

[0031] The particle speed update consists of two parts: individual experience terms and social experience terms. The individual experience term represents the ability of the particle to search based on its own optimal solution, that is, to move in the direction of the individual optimal solution; the social experience term represents the ability of the particle to search based on the global optimal solution, that is, to move in the direction of the global optimal solution. By adjusting the weights of the individual and social experience terms, the ability of local search and global search can be balanced. By evaluating the current position of each particle in the best particle swarm and calculating the objective function value of each particle in the best particle swarm based on the comprehensive objective function, the speed and position of each particle in the best particle swarm are updated, and the speed and position updates are iteratively executed until the maximum number of iterations or the specified preset conditions are met. Through the particle speed and position update mechanism, it is possible to jump out of the local optimal solution and explore different areas of the solution space, so that it has a strong global search capability. A distribution network equivalent model construction system, used to implement any one of the distribution network equivalent model construction methods, including a device acquisition module, a model building module, an objective function creation module, a particle swarm algorithm module, an algorithm optimization module and a verification module, wherein the device acquisition module is connected to the model building module, the model building module is connected to the objective function creation module, the objective function creation module is connected to the particle swarm algorithm module, and the particle swarm algorithm module is connected to the algorithm optimization module. The device acquisition module is used to acquire the devices in the distribution network to be modeled, including photovoltaic power generation systems, city power access modules, transformers, inverters, converters, filters, capacitors, inductors and loads; the model building module is used to divide the distribution network into multiple branches based on the connection relationship of the devices, and establish a distribution network model based on the devices and multiple branches.

[0032] The objective function creation module is used to establish a comprehensive objective function based on the power loss and cost of the multiple branches of the distribution network model, and define the constraints of the comprehensive function as the capacity of multiple branches and voltage stability; the particle swarm algorithm module is used to define the particle swarm algorithm parameters and solution space based on the comprehensive objective function and its constraints, and initialize a specified number of particles, and iteratively update the initialized specified number of particles, and randomly distribute the iteratively updated initialized specified number of particles in the solution space.

[0033] The algorithm optimization module is used to iteratively update and optimize the optimized and initialized specified number of particles based on the comprehensive objective function until an optimal solution that meets the specified convergence conditions or an optimal solution with a maximum number of iterations is obtained. The verification module is used to verify the optimal solution based on the comprehensive objective function and constraint conditions, calculate the fitness of the optimal solution, and when the fitness of the optimal solution is greater than a specified threshold, the distribution network model is operated with the optimal solution.

[0034] In summary, the method and system for constructing an equivalent model of a distribution network provided by the present invention establish a distribution network model by determining the equipment included in the distribution network of the equivalent model to be constructed; establish a comprehensive objective function based on the power loss and cost of the distribution network model, and define the constraints of the comprehensive function; based on the comprehensive objective function and its constraints, define the particle swarm algorithm parameters and solution space, initialize a specified number of particles, and iteratively update the initialized specified number of particles; run the particle swarm algorithm, and iteratively update the optimized initialized specified number of particles based on the comprehensive objective function until an optimal solution that meets the specified convergence conditions or an optimal solution with a maximum number of iterations is obtained; based on the process of verifying the optimal solution based on the comprehensive objective function and the constraints, dynamic equivalent modeling of the distribution network after the distributed renewable energy power generation system is connected is realized, and then stable and reliable analysis of the distribution network after the distributed renewable energy power generation system is connected is realized, and through dynamic equivalent modeling, the operation of the entire distribution network with the lowest comprehensive power consumption and cost is realized.

[0035] By establishing the formula of the comprehensive objective function of power loss and cost of multiple branches, and dynamically solving the optimal solution of the comprehensive objective function of power loss and cost of multiple branches based on the particle swarm algorithm, it is convenient to realize the operation of the distribution network with the minimum comprehensive power consumption and cost based on the optimal solution. By defining the initialization particle population size as N, the decision variable is n, initialize the particle population, evaluate the fitness of the individuals in the initialized particle population based on the objective function, and randomly select several individuals from the individuals whose fitness values ​​are greater than the preset fitness threshold to compete, and select the winning individuals to enter the next generation, randomly select multiple crossover points from multiple winning individuals for gene exchange, generate multiple new individuals, and perform mutation processing to iteratively update the initialized particle population, realize the optimization of the particle swarm algorithm, and improve the accuracy of obtaining the optimal solution through the optimized particle swarm algorithm. By iteratively updating the particles, d In the dimensional solution space, the position of the best individual particle in the iteratively updated particles is obtained, and based on the individual particle X Under the influence of the best individual particle, the position of the next generation of individual particles is calculated, the position of the next generation of individual particles is updated in a spiral manner, a search-predation mathematical model is established, and the search-predation mathematical model is iteratively executed to obtain the best particle swarm, further optimizing the particle swarm algorithm and further improving the accuracy of the optimal solution.

Claims

1. A method for constructing an equivalent model of a distribution network, characterized in that: The following steps are involved: S1: Determine the equipment included in the distribution network for which the equivalent model is to be constructed, divide the distribution network into multiple branches based on the connection status of the equipment in the distribution network, and establish a distribution network model based on the equipment and the multiple branches; S2: establishing a comprehensive objective function based on the power loss and cost of the multiple branches of the distribution network model, and defining the constraint conditions of the comprehensive function as the capacity of the multiple branches and the voltage stability; S3: Based on the comprehensive objective function and its constraints, define the particle swarm algorithm parameters and solution space, initialize a specified number of particles, iteratively update the initialized specified number of particles, and randomly distribute the iteratively updated initialized specified number of particles in the solution space; S4: running the particle swarm algorithm, and iteratively updating the optimized initialized specified number of particles based on the comprehensive objective function until an optimal solution that meets a specified convergence condition or an optimal solution with a maximum number of iterations is obtained; S5: verifying the optimal solution based on the comprehensive objective function and constraint conditions, calculating the fitness of the optimal solution, and when the fitness of the optimal solution is greater than a specified threshold, operating the distribution network model with the optimal solution.

2. The method for constructing an equivalent model of a distribution network according to claim 1, characterized in that: The equipment included in the distribution network in step S1 includes a photovoltaic power generation system, a mains access module, a transformer, an inverter, a converter, a filter, a capacitor, an inductor and a load; The particle swarm algorithm parameters defined in step S3 include the number of particles and the number of iterations.

3. The method for constructing an equivalent model of a distribution network according to claim 1, characterized in that: The formula of the comprehensive objective function of the power loss and cost of multiple branches in step S2 is as follows: ； in, f g ( x ) is the comprehensive objective function, C1( x ) is the distribution network loss cost, C2( x ) is the electricity cost of energy storage operation, C3( x ) is the power purchase cost of the upper power grid, C4( x ) are other specified costs, i is the branch number, Ri( x ) is the i The resistance of the branch, Pi( x ) is the i Active power of the branch, Qi( x ) is the i Reactive power of the branch, Ui( x ) is the terminal voltage of the ith branch.

4. The method for constructing an equivalent model of a distribution network according to claim 1, characterized in that: The specific process of initializing a specified number of particles and iteratively updating the initialized specified number of particles in step S3 is as follows: S31: Define the size of the initialized particle population as N, and the decision variable as n , then the initial particle population is: P(0)={ X 1， X 2， X 3，···， X N}； X i =( x i1 ， x i2 ， x i3 ，···， x in ); in, X 1, X 2, X 3,···, X N is a population of N particles, X i For the i A particle population, x i1 , x i2 , x i3 ,···, x in For the i In the particle population n Particles; S32: Based on the objective function, the fitness of the individuals in the initialized particle population is evaluated, and a number of individuals are randomly selected from the individuals whose fitness values ​​are greater than the preset fitness threshold to compete, and the winning individuals are selected to enter the next generation, and the random selection of a number of individuals for competition is repeated to obtain multiple winning individuals to enter the next generation; S33: randomly selecting multiple crossover points from the multiple winning individuals to perform gene exchange to generate multiple new individuals. The specific formula is as follows; ； in, m is the number of intersection points, k 1, k 2, k 3. ··· , k m-1 is the intersection point; S34: The individuals in step S33 are Y i No. j Gene location Y ij Perform mutation processing. The specific formula for mutation processing is as follows: ； in, is a random disturbance, which obeys normal distribution, P m is the mutation probability; S35: Iteratively execute steps S32 to S34 to iteratively update the initialized particle population P(0).

5. The method for constructing an equivalent model of a distribution network according to claim 4, characterized in that: After iteratively executing steps S32 to S34 and iteratively updating the initialized particle population P(0), the iteratively updated particle population is optimized. The specific process is as follows: S36: For the iteratively updated particles, d In the dimensional solution space, the position of the best individual particle in the iteratively updated particles is expressed as: X 1 * , X 2 * , X 3 * ,···, X d * ), individual particles X j The position is ( X 1 j , X 2 j , X 3 j ,···, X d j ); S37: Based on individual particles X j Under the influence of the best individual particle, calculate the next generation of individual particles X j+1 Location( X 1 j+1 , X 2 j+1 , X 3 j+1 ,···, X d j +1 ), the specific calculation formula is: X k j+1 = X k * -(2 a · r 1- a )·|2 r 2· X k * - X k j |； in, X k j+1 Represents the solution space coordinates X j+1 No. k Quantity, a As the number of iterations increases, it decreases linearly from 2 to 0. r 1 and r 2 is a random number between 0 and 1; S38: For the next generation of particle individuals X j+1 Perform spiral position update, the specific formula is: X k j+1 = X k * +| X k * - X k j |· e bl cos (2π l )； in, b is the logarithmic spiral shape constant, l A random number between -1 and 1; S39: Establish a mathematical model of search and predation, X k j+1 = X k rand - A ·| X k * - X k j |, and iteratively execute the search-predation mathematical model to obtain the best particle swarm; in, A is the coefficient, X k rand For particles X k of random locations.

6. The method for constructing an equivalent model of a distribution network according to claim 5, characterized in that: The specific process of iteratively updating the optimized initialized specified number of particles based on the comprehensive objective function in step S4 is as follows: S41: Evaluate the current position of each particle in the best particle group in step S39, and calculate the objective function value of each particle in the best particle group based on the comprehensive objective function; S42: Update the speed and position of each particle in the optimal particle group, where the speed update formula is: v [ i ]= wv [i]+ c 1 r 1*( pBest [ i ]- x [ i ])+ c 2 r 2*( gBest [ i ]- x [ i ]); The position update formula is: x [ i ]= x [ i ]+ v [ i ]； in, v [ i ] is the i The velocity vector of a particle, x [ i ] is the i The position vector of a particle, w is the inertia weight, c 1 and c 2 are learning factors r 1 and r 2 is a random number, pBest [ i ] is the i The individual optimal position vector of each particle is gBest [ i ] is the current global optimal position vector ‌; S43: Iteratively execute step S41 and step S42 until a maximum number of iterations is met or a specified preset condition is met.

7. A distribution network equivalent model construction system, used to implement a distribution network equivalent model construction method according to any one of claims 1 to 6, characterized in that: It includes a device acquisition module, a model building module, an objective function creation module, a particle swarm algorithm module, an algorithm optimization module and a verification module, wherein the device acquisition module is connected to the model building module, the model building module is connected to the objective function creation module, the objective function creation module is connected to the particle swarm algorithm module, and the particle swarm algorithm module is connected to the algorithm optimization module; The device acquisition module is used to acquire the devices in the distribution network to be modeled, including photovoltaic power generation systems, city power access modules, transformers, inverters, converters, filters, capacitors, inductors and loads; The model building module is used to divide the distribution network into multiple branches based on the connection relationship of the equipment, and to build a distribution network model based on the equipment and the multiple branches; The objective function creation module is used to establish a comprehensive objective function based on the power loss and cost of the multiple branches of the distribution network model, and define the constraint conditions of the comprehensive function as the capacity of the multiple branches and the voltage stability; The particle swarm algorithm module is used to define the particle swarm algorithm parameters and solution space based on the comprehensive objective function and its constraints, and initialize a specified number of particles, and iteratively update the initialized specified number of particles, and randomly distribute the iteratively updated initialized specified number of particles in the solution space; The algorithm optimization module is used to iteratively update and optimize the optimized initialized specified number of particles based on the comprehensive objective function until an optimal solution that meets a specified convergence condition or reaches an optimal solution with a maximum number of iterations is obtained; The verification module is used to verify the optimal solution based on the comprehensive objective function and constraint conditions, calculate the fitness of the optimal solution, and when the fitness of the optimal solution is greater than a specified threshold, operate the distribution network model with the optimal solution.

Images