Energy storage vehicle charging pile layout method and system based on improved adaptive particle swarm optimization

Abstract

An energy storage vehicle charging pile layout method based on an improved adaptive particle swarm algorithm comprises the steps that firstly, charging pile nodes to be distributed are subjected to site selection reduction through a Folyd shortest path algorithm to obtain a distance shortest matrix Dn, and an energy storage vehicle charging pile layout model is constructed with the minimum comprehensive cost as an objective function; then, an improved self-adaptive multi-target particle swarm algorithm is adopted to solve the model, the algorithm accelerates the convergence speed by adopting global optimal probability selection inversely proportional to the number of control particles, improves the diversity of solutions by adopting global optimal probability selection inversely proportional to the crowding distance, and searches for solutions located at the boundary by adopting boundary trimming processing; solutions located near the boundary are searched by employing exponential distribution boundary processing. According to the method, the solving difficulty can be effectively reduced through the improved self-adaptive multi-target particle swarm algorithm, the global search capability of the algorithm can be improved through crossover and mutation operators, and the stability of the layout method is effectively improved.

Description

Technical Field

[0001] The present invention relates to a layout method for charging piles of energy storage vehicles, and particularly to a layout method and system for charging piles of energy storage vehicles based on an improved adaptive particle swarm optimization algorithm. Background Art

[0002] As a clean energy, electric energy has become an important part of China's energy. Mobile energy storage vehicles have functions such as emergency power supply for users, emergency power supply for distribution network substations, and power quality governance, and are mainly applied to various power accident repair, civil disaster relief, major political power protection and other sites, and can also be used as emergency backup power for communication base station construction and in fields such as government, hospitals, and schools.

[0003] Since energy storage vehicles require charging technologies with higher voltages and larger currents, the technical requirements and safety standards of charging piles will be different. At the same time, the charging piles of energy storage vehicles tend to be laid out in areas with important grid structures or large load fluctuations to play their role in peak shaving and valley filling. Therefore, energy storage vehicles cannot directly use ordinary electric vehicle charging piles, and the layout planning is also different from that of ordinary electric vehicle charging piles.

[0004] In the initial stage of the popularization of mobile energy storage vehicles, due to the lack of a scientific and perfect layout planning guide for charging piles, a scientific charging pile layout system has not been formed, resulting in unreasonable site selection planning for charging stations, poor user experience, cost spillover of mobile energy storage systems, and low grid energy utilization rate. With the gradual development and maturity of the intelligent scheduling system for mobile energy storage devices, reasonable charging locations are of great significance for reasonably arranging charging times, improving grid utilization rate, reducing operating costs, and enhancing user experience.

[0005] In the prior art, the particle swarm optimization algorithm is usually used for spatial site selection planning of charging infrastructure, and the PSO algorithm has been widely used due to its few control parameters and fast convergence speed.

[0006] Although this layout method can quickly obtain a charging pile planning scheme, it still has the following defects:

[0007] 1. Most of the actual power system optimization problems are multi-objective, large-scale, mixed integer, non-linear and non-convex problems. Due to the large amount of data, the algorithm complexity is high and the model solving difficulty is large.

[0008] 2. The global search ability of the conventional particle swarm optimization algorithm is poor, and it is easy to fall into local optimum, resulting in low stability.

[0009] Disclosing the information of this background art section is only intended to increase the understanding of the overall background of the present application, and should not be regarded as an admission or any form of implication that this information constitutes the prior art already known to those of ordinary skill in the art. Summary of the Invention

[0010] The object of the present invention is to overcome the disadvantages of difficult solution and low stability in the prior art, and provide a layout method of energy storage vehicle charging piles based on an improved adaptive particle swarm algorithm with easy solution and high stability.

[0011] To achieve the above object, the technical solution of the present invention is as follows:

[0012] A layout method of energy storage vehicle charging piles based on an improved adaptive particle swarm algorithm, the optimization method comprising the following steps:

[0013] S1. Simplify the road nodes, select the central node with the smallest Manhattan distance to each road node for modeling calculation, and obtain the shortest distance matrix D n ;

[0014] S2. Construct an energy storage vehicle charging pile layout model with the minimum comprehensive cost as the objective function, and the comprehensive cost considers: charging loss, land cost, the sum of calculation investment cost and maintenance cost, and the sum of network loss and charging cost;

[0015] S3. Solve the energy storage vehicle charging pile layout model, and solve the energy storage vehicle charging pile layout model through an improved adaptive particle swarm algorithm to obtain an energy storage vehicle charging pile layout scheme.

[0016] In the above S1, the road nodes are simplified by the Floyd shortest path algorithm, and the weighted adjacency matrix A is used as the initial value of the distance matrix D to construct the Manhattan distance matrix between all road nodes:

[0017]

[0018] The element therein represents the length of the shortest path in the path from road node v i to road node v j where the intermediate points are allowed to be any road node in {v 1 , v 2 v 3 , …, v s}; when s = n, in is the length of the shortest path in the path from road node v i to road node v j where the intermediate points are allowed to be any road node in {v 1 , v 2 v 3 , …, v s}, and D n is the shortest distance matrix.

[0019] In S2, the objective function of the comprehensive cost is:

[0020]

[0021] In the above formula, f 1i is the objective function of the sum of the investment cost and maintenance cost in the i-th year, f 2i is the objective function of the sum of the network loss and charging cost in the i-th year, f 3i is the objective function of the charging loss in the i-th year, f 4i is the objective function of the land cost in the i-th year, I is the total number of years, and i is the i-th year;

[0022] The objective function f 1 of the sum of the investment cost and maintenance cost includes:

[0023]

[0024] In the above formula, r j is the number of transformers at the public charging pile site j, a is the unit price of the transformer for building the charging pile, p j is the number of charging piles at the public charging pile site j, b is the unit price of the charging pile, c j is the cost of the charging station at the public charging pile site j, r 0 is the discount rate, n is the number of years of use, and μ is the depreciation rate;

[0025] The objective function f 2 of the sum of the network loss and charging cost includes:

[0026] f 2 = r j · C 1 · T v · p c · 365 + p j · C 2 · k t · T v · e c · 365 + pQ · 365;

[0027] In the above formula, r j is the number of transformers at the public charging pile site j, C 1 is the steel loss, T v is the effective charging time of the charging station, p c is the electricity price of the power company, C 2 is the charging loss, k t is the simultaneous operation rate of multiple charging piles in the charging station, e c is the electricity price of the charging station, and pQ is the charging cost of the user;

[0028] The objective function f of the charging loss 3 includes:

[0029]

[0030] In the above formula, b 1 is the cost generated by the empty driving power consumption during the user's annual charging process, and b 2 is the indirect loss cost, m j is the number of public charging pile sites, and L j is the comprehensive distance from all charging demand points within the service area of the public charging pile site j to the public charging pile site j. g is the unit power consumption of the energy storage vehicle, v is the average speed of the energy storage vehicle, and p is the driving time cost;

[0031] The objective function f of the land cost 4 includes:

[0032]

[0033] In the above formula, x j is the floor area of the public charging pile site j, C t is the highest land use cost in the area, and λ i is the comprehensive factor of each public charging station.

[0034] The above-mentioned S3 includes the following steps:

[0035] S3.1. Algorithm initialization. Set the maximum number of iterations K = 250 of the adaptive particle swarm algorithm, the initial number of iterations k = 1, the initial velocity matrix is 0, and use the pruning boundary processing method of modifying the out-of-limit dimension value of the particle to the boundary value through the boundary processing operator to ensure that all solutions are within the feasible domain;

[0036] S3.2. Enable the global optimal probability selection operator, update the elite solution library to merge the population and the elite solution library, and fill the updated Pareto front into the elite solution library;

[0037] S3.3. Iteratively solve the energy storage vehicle charging pile layout model through the adaptive particle swarm algorithm and make a judgment. The judgment conditions include:

[0038] a. Judge whether the optimal solution of the algorithm has not been updated for 20 consecutive generations. If so, enable the crossover operator and the mutation operator;

[0039] b. Judge whether the continuous stagnation time of the convergence accuracy of the algorithm exceeds 25 generations. If so, rotate the global optimal selection method, and the rotated global optimal selection method is the global optimal probability selection method inversely proportional to the crowding distance and the global optimal probability selection method inversely proportional to the number of controlled particles;

[0040] c. Determine whether the continuous stagnation time of the diversity of the judgment algorithm exceeds 25 generations. If so, rotate the boundary processing method, where the rotated boundary processing methods are the trimmed boundary processing method and the exponential distribution boundary processing method;

[0041] Proceed to S3.4;

[0042] S3.4 Update the individual optimal leader and the global optimal leader. The individual optimal leader is the optimal solution found by the particle solution z in the t-th iteration, and the global optimal leader is the optimal solution found by the entire particle swarm so far. All particles select the same optimal solution as the global leader;

[0043] S3.5 Update the velocity and position of the particle swarm, and use the boundary processing operator to ensure that the new velocity and position satisfy their respective boundary constraints. K = K + 1, and proceed to S3.3 for continued iteration until K ≥ Kmax or the elite solution library fails to update for 50 consecutive iterations. The obtained result is the optimal result.

[0044] In S3.2, the global optimal probability selection operator enables the global optimal probability selection method inversely proportional to the normalized crowding distance. The normalized crowding distance CD of the solution z in the population k can be expressed as:

[0045]

[0046] In the above formula, N d represents the number of objective functions; is the maximum value of the j-th objective function, is the minimum value of the j-th objective function; assuming that the values of the j-th objective function are sorted, the serial number of the solution z is i, N is the population size, and when 1 < i < N, f i+1,j -f i-1,j represents the crowding distance of the solution z in the objective function j. When the objective function value is at the boundary, that is, i ∈ [1, N], the crowding distance of the solution z in the objective function j is infinite.

[0047] In the global optimal leader selection method of S3.4, the minimum value of the sum of the objective functions is used to roughly evaluate the convergence accuracy of the algorithm. The rotated global optimal probability selection operator inversely proportional to the number of control particles can be described as:

[0048]

[0049] In the above formula, X a = {x ∈ X|a < x} represents the particle set controlled by the member a of the elite library A. When x n is controlled by A, the global leader G n is selected from the control particle x nSelected from the elite solution set A, which is beneficial to maintaining the continuity of the particle search direction; when x n is not controlled by A, G n is selected from the entire A, and the selection probability of each global guidance candidate a is inversely proportional to the size |X a | of the particle set X a controlled by a.

[0050] The update equation for the velocity of the id-th particle in the K-th iteration in S3.5 is:

[0051]

[0052] In the above formula, w, c 1 , c 2 are all inertia factors, is the velocity of the id-th particle in the (K - 1)-th iteration, pbest id is the personal extreme value of the id-th particle, gbest id is the global extreme value of the id-th particle, is the position of the id-th particle, and r1, r2 are both random numbers;

[0053] The inertia factor w is changed with the number of iterations:

[0054]

[0055] In the above formula, k is the current number of iterations, M is the total number of iterations, w max is the maximum value of the inertia factor, w min is the minimum value of the inertia factor;

[0056] The update equation for the position of the id-th particle in the K-th iteration is:

[0057]

[0058] In the above formula, χ ∈ [0, 1] is a contraction factor that controls the magnitude of the velocity, and χ = 1;

[0059] The boundary constraint formula for the velocity is:

[0060]

[0061] In the above formula, VMax and VMin respectively represent the upper and lower bounds of the particle velocity;

[0062] The boundary constraint formula for the position is:

[0063]

[0064] VarMax and VarMin respectively represent the upper and lower bounds of the particle position.

[0065] An energy storage vehicle charging pile layout system based on an improved adaptive particle swarm optimization algorithm, the system is used to execute the aforementioned energy storage vehicle charging pile layout method based on the improved adaptive particle swarm optimization algorithm, and specifically includes: a node reduction module, a model construction module and a model solution module;

[0066] The node reduction module is used to reduce the road nodes, select the central node with the smallest Manhattan distance to each road node for modeling calculation, and obtain the shortest distance matrix D n ;

[0067] The model construction module is used to construct an energy storage vehicle charging pile layout model with the minimum comprehensive cost as the objective function. The comprehensive cost considers: charging loss, land cost, the sum of calculation investment cost and maintenance cost, and the sum of network loss and charging cost;

[0068] The model solution module is used to solve the energy storage vehicle charging pile layout model, and solve the energy storage vehicle charging pile layout model through the improved adaptive particle swarm optimization algorithm to obtain the energy storage vehicle charging pile layout plan.

[0069] An energy storage vehicle charging pile layout device based on an improved adaptive particle swarm optimization algorithm, including a memory and a processor. The memory is used to store computer program code and transmit the computer program code to the processor;

[0070] The processor is used to execute the aforementioned energy storage vehicle charging pile layout method based on the instructions in the computer program code.

[0071] A computer-readable storage medium stores a computer program, and the computer program is executed by the processor to perform the aforementioned energy storage vehicle charging pile layout method based on the improved adaptive particle swarm optimization algorithm.

[0072] Compared with the prior art, the beneficial effects of the present invention are:

[0073] 1. In the energy storage vehicle charging pile layout method based on the improved adaptive particle swarm optimization algorithm of the present invention, first, the Floyd shortest path algorithm is used to reduce the location selection of the charging pile nodes to be allocated to obtain the shortest distance matrix D n , so as to minimize the distance from the center to each node, ensure scientific location selection under the condition of minimum time cost, and thus achieve the optimal economic benefit. Therefore, this design can obtain the shortest distance matrix through the Floyd shortest path algorithm, effectively improving the economic benefit of the layout method.

[0074] 2. In the layout method of energy storage vehicle charging piles based on the improved adaptive particle swarm algorithm of the present invention, an improved adaptive multi-objective particle swarm algorithm is adopted. The algorithm accelerates the convergence speed by adopting the global optimal probability selection inversely proportional to the number of controlled particles, improves the diversity of solutions by adopting the global optimal probability selection inversely proportional to the crowding distance, searches for solutions located at the boundary by adopting trimmed boundary processing, searches for solutions located near the boundary by adopting exponential distribution boundary processing, and takes into account the advantages of all the above operations through adaptive adjustment strategies. Therefore, this design can effectively reduce the difficulty of solving problems through the improved adaptive multi-objective particle swarm algorithm.

[0075] 3. In the layout method of energy storage vehicle charging piles based on the improved adaptive particle swarm algorithm of the present invention, when the optimal solution of the algorithm has not been updated for 20 consecutive generations, the crossover mutation operator is enabled to continue iterative calculation, and the new particles generated by the crossover and mutation operations each account for 10% of the population. Half of the parent generation of the crossover operation comes from the elite library, and half comes from the population, both of which are randomly selected. Two solutions are randomly selected from the entire parent generation for single-point crossover operation, and the intersection point is also randomly selected. The crossover mutation operator enables the algorithm to jump out of the local optimum and improves the global search capability. Therefore, this design can improve the global search capability of the algorithm through the crossover mutation operator, and effectively improve the stability of the layout method. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 is a flow chart of the method of the present invention.

[0077] Figure 2 It is a structural diagram of the system described in the present invention.

[0078] Figure 3 It is a structural diagram of the device described in the present invention.

[0079] Figure 4 It is a schematic diagram of the method of the present invention.

[0080] Figure 5 It is a schematic diagram of the process of the crossover mutation operator of the present invention.

[0081] Figure 6 This is a schematic diagram of the result of using the traditional MOPSO algorithm to solve the ZDT1 test problem in Example 2.

[0082] Figure 7 It is a schematic diagram of the result of solving the ZDT1 test problem using the method of the present invention in Example 2.

[0083] Figure 8 This is a schematic diagram of the result of using the traditional MOPSO algorithm to solve the Viennet2 test problem in Example 2.

[0084] Figure 9It is a schematic diagram of the result of using the method described in the present invention to solve the Viennet2 test problem in Example 2. Detailed implementation manners

[0085] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation manners.

[0086] Example 1:

[0087] Refer to Figure 1 and Figure 4 , a layout method for energy storage vehicle charging piles based on an improved adaptive particle swarm optimization algorithm, the optimization method includes the following steps:

[0088] S1. Reduce the road nodes, select the central node with the smallest Manhattan distance to each road node for modeling and calculation, and obtain the shortest distance matrix D n ;

[0089] S2. Construct a layout model for energy storage vehicle charging piles with the minimum comprehensive cost as the objective function, and the comprehensive cost considers: charging loss, land cost, the sum of calculation investment cost and maintenance cost, and the sum of network loss and charging cost;

[0090] S3. Solve the layout model for energy storage vehicle charging piles, and solve the layout model for energy storage vehicle charging piles through an improved adaptive particle swarm optimization algorithm to obtain a layout plan for energy storage vehicle charging piles.

[0091] In the above S1, the road nodes are reduced by the Floyd shortest path algorithm, and the weighted adjacency matrix A is used as the initial value of the distance matrix D to construct the Manhattan distance matrix between all road nodes:

[0092]

[0093] The element represents the length of the shortest path from the road node v i to the road node v j in the path, and the intermediate point is allowed to be any road node in {v 1 , v 2 v 3 , …, v s}; when s = n, in is the length of the shortest path from the road node v i to the road node v j in the path, and the intermediate point is allowed to be any road node in {v 1 , v 2 v 3 , …, v s}, that is, it is to find the shortest path from v i to vj The shortest path length in a path where any vertex can be inserted, D n That is the shortest distance matrix.

[0094] In the above-mentioned S2, the objective function of the comprehensive cost is:

[0095]

[0096] In the above formula, f 1i is the objective function of the sum of the investment cost and maintenance cost in the i-th year, f 2i is the objective function of the sum of the network loss and charging cost in the i-th year, f 3i is the objective function of the charging loss in the i-th year, f 4i is the objective function of the land cost in the i-th year, I is the total number of years, and i is the i-th year;

[0097] The objective function f of the sum of the investment cost and maintenance cost 1 includes:

[0098]

[0099] In the above formula, r j is the number of transformers at the public charging pile site j, a is the unit price of the transformer for building the charging pile, p j is the number of charging piles at the public charging pile site j, b is the unit price of the charging pile, c j is the cost of the charging station at the public charging pile site j, r 0 is the discount rate, n is the number of years of use, and μ is the depreciation rate;

[0100] The objective function f of the sum of the network loss and charging cost 2 includes:

[0101] f 2 =r j ·C 1 ·T v ·p c ·365 + p j ·C 2 ·k t ·T v ·e c ·365 + pQ·365;

[0102] In the above formula, r j is the number of transformers at the public charging pile site j, C 1 is the steel loss, T v is the effective charging time of the charging station, p c is the electricity price of the power company, C 2 is the charging loss, k tis the simultaneous operation rate of multiple charging piles in the charging station, e c is the electricity price of the charging station, and pQ is the charging cost of users;

[0103] The objective function f of the charging loss 3 includes:

[0104]

[0105] In the above formula, b 1 is the cost generated by the empty driving electricity consumption during the annual charging process of users, b 2 is the indirect loss cost, m j is the number of public charging pile sites, L j is the comprehensive distance from all charging demand points in the service area of the public charging pile site j to the public charging pile site j, g is the unit power consumption of the energy storage vehicle, v is the average speed of the energy storage vehicle, and p is the driving time cost;

[0106] The objective function f of the land cost 4 includes:

[0107]

[0108] In the above formula, x j is the floor area of the public charging pile site j, C t is the highest land use cost in the area, λ i is the comprehensive factor of each public charging station.

[0109] The said S3 includes the following steps:

[0110] S3.1. Algorithm initialization, set the maximum number of iterations K = 250 of the adaptive particle swarm algorithm, the initial number of iterations K = 1, the initial velocity matrix is 0, and use the pruning boundary processing method of modifying the out-of-bounds dimension value of the particle to the boundary value through the boundary processing operator to ensure that all solutions are within the feasible domain;

[0111] S3.2. Enable the global optimal probability selection operator, update the elite solution library to merge the population and the elite solution library, and fill the updated Pareto front into the elite solution library;

[0112] S3.3. Iteratively solve the energy storage vehicle charging pile layout model through the adaptive particle swarm algorithm and make judgments. The judgment conditions include:

[0113] a. Judge whether the optimal solution of the algorithm has not been updated for 20 consecutive generations. If so, enable the crossover operator and the mutation operator;

[0114] b. Determine whether the continuous stagnation time of the convergence accuracy of the algorithm exceeds 25 generations. If so, rotate the global optimal selection method, where the rotated global optimal selection method is the global optimal probability selection method inversely proportional to the crowding distance and the global optimal probability selection method inversely proportional to the number of controlled particles.

[0115] c. Determine whether the continuous stagnation time of the diversity of the algorithm exceeds 25 generations. If so, rotate the boundary handling method, where the rotated boundary handling method is the pruning boundary handling method and the exponential distribution boundary handling method.

[0116] Proceed to S3.4.

[0117] S3.4 Update the individual optimal leader and the global optimal leader. The individual optimal leader is the optimal solution found by the particle solution z in the t-th iteration, and the global optimal leader is the optimal solution found by the entire particle swarm so far. All particles select the same optimal solution as the global leader.

[0118] S3.5 Update the velocity and position of the particle swarm. Use the boundary handling operator to ensure that the new velocity and position satisfy their respective boundary constraints. K = K + 1, and proceed to S3.3 for continued iteration until K ≥ Kmax or the elite solution library fails to update for 50 consecutive iterations. The obtained result is the optimal result.

[0119] See Figure 5 , in S3.3, let the new particles generated by the crossover and mutation operations each account for 10% of the population. Half of the parents for the crossover operation come from the elite library and half come from the population, both randomly selected. Two solutions are randomly selected from the entire set of parents for single-point crossover operation, and the crossover point is also randomly selected. The parents for the mutation operation are only selected from the elite library. The mutation probability for each dimension is 0.1, and the mutation value does not exceed 5% of the value range. Finally, the new particles obtained from the crossover and mutation operations randomly replace the same number of old particles in the population.

[0120] In S3.3, the diversity of the algorithm is evaluated by the average crowding distance of the particles. The rotated exponential distribution boundary handling operator pulls the position of the out-of-bounds dimension back to between the old position before update and the out-of-bounds boundary. The new particles fall near the boundary with a higher probability and far from the boundary with a lower probability, and this probability increases exponentially as the new particles approach the boundary.

[0121] In S3.2, the global optimal probability selection operator enables the global optimal probability selection method inversely proportional to the normalized crowding distance. The normalized crowding distance CD of the solution z in the population k can be expressed as:

[0122]

[0123] In the above formula, N dIndicates the number of objective functions; is the maximum value of the j-th objective function, is the minimum value of the j-th objective function; assuming that the values of the j-th objective function are sorted, the sequence number of the solution z is i, N is the population size, when 1 < i < N, f i+1,j -f i-1,j represents the crowding distance of the solution z in the objective function j. When the objective function value is at the boundary, that is, i ∈ [1, N], the crowding distance of the solution z in the objective function j is infinite.

[0124] In the global optimal leader selection method of S3.4, the minimum value of the sum of objective functions is used to roughly evaluate the convergence accuracy of the algorithm. The global optimal probability selection operator that is inversely proportional to the number of control particles in rotation can be described as:

[0125]

[0126] In the above formula, X a = {x ∈ X|a < x} represents the particle set controlled by the member a of the elite library A. When x n is controlled by A, the global leader G n is selected from the elite solution set A of the control particle x n , which is beneficial to maintaining the continuity of the particle search direction; when x n is not controlled by A, G n is selected from the entire A. The selection probability of each global leader candidate a is inversely proportional to the size |X a | of the particle set X a controlled by a.

[0127] The update equation of the velocity of the id-th particle in the K-th iteration in S3.5 is:

[0128]

[0129] In the above formula, w, c 1 , c 2 are all inertia factors, is the velocity of the id-th particle in the (K - 1)-th iteration, pbest id is the personal extreme value of the id-th particle, gbest id is the global extreme value of the id-th particle, is the position of the id-th particle, r1 and r2 are both random numbers;

[0130] Make the inertia factor w change with the number of iterations:

[0131]

[0132] In the above formula, k is the current iteration number, M is the total number of iterations, wmax is the maximum value of the inertia factor, w min is the minimum value of the inertia factor;

[0133] The update equation for the position of the id-th particle in the K-th iteration is:

[0134]

[0135] In the above formula, χ∈[0,1] is the contraction factor that controls the magnitude of the velocity, and χ = 1;

[0136] The boundary constraint formula for the velocity is:

[0137]

[0138] In the above formula, VMax and VMin respectively represent the upper and lower bounds of the particle velocity;

[0139] The boundary constraint formula for the position is:

[0140]

[0141] VarMax and VarMin respectively represent the upper and lower bounds of the particle position.

[0142] Example 2:

[0143] In this example, for the ZDT1 and Viennet2 test problems, the traditional MOPSO algorithm is used as a comparative example to solve with the improved adaptive multi-objective particle swarm algorithm proposed by the present invention. The population size of the two algorithms is set to 200, and the maximum number of iterations is K = 250, and the comparison results are obtained;

[0144] See Figures 6 to 9 , the improved adaptive MOPSO algorithm proposed in this paper can search for more effective solutions, and the solution distribution is uniform.

[0145] Example 3:

[0146] See Figure 2 , a layout system for energy storage vehicle charging piles based on an improved adaptive particle swarm algorithm, the system is used to execute the layout method of energy storage vehicle charging piles based on the improved adaptive particle swarm algorithm as described in Example 1, and specifically includes: a node reduction module, a model construction module, and a model solving module;

[0147] The node reduction module is used to reduce the road nodes, select the central node with the smallest Manhattan distance to each road node for modeling calculation, and obtain the shortest distance matrix D n ;

[0148] The model construction module is used to construct a layout model of energy storage vehicle charging piles with the minimum comprehensive cost as the objective function. The comprehensive cost takes into account: charging loss, land cost, the sum of calculation investment cost and maintenance cost, the sum of network loss and charging cost;

[0149] The model solving module is used to solve the layout model of energy storage vehicle charging piles. The layout model of energy storage vehicle charging piles is solved by an improved adaptive particle swarm optimization algorithm to obtain a layout scheme of energy storage vehicle charging piles.

[0150] Embodiment 4:

[0151] See Figure 3 , a layout device of energy storage vehicle charging piles based on an improved adaptive particle swarm optimization algorithm, including a memory and a processor. The memory is used to store computer program codes and transmit the computer program codes to the processor;

[0152] The processor is used to execute the layout method of energy storage vehicle charging piles based on the improved adaptive particle swarm optimization algorithm as described in Embodiment 1 according to the instructions in the computer program codes.

[0153] A computer-readable storage medium stores a computer program, and the computer program is executed by a processor to perform the layout method of energy storage vehicle charging piles based on the improved adaptive particle swarm optimization algorithm as described in Embodiment 1.

[0154] The above is only the preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. Any equivalent modification or change made by those of ordinary skill in the art according to the disclosed content of the present invention shall be included in the protection scope recorded in the claims.

Claims

1. A method for arranging charging piles for energy storage vehicles based on an improved adaptive particle swarm algorithm, characterized in that: The optimization method comprises the following steps: S1. Simplify the road nodes and select the central node with the smallest Manhattan distance to each road node for modeling calculation to obtain the shortest distance matrix D n ; S2. Constructing a layout model of energy storage vehicle charging piles with the minimum comprehensive cost as the objective function. The comprehensive cost takes into account: charging loss, land cost, the sum of calculation investment cost and maintenance cost, and the sum of network loss and charging cost; S3. Solve the energy storage vehicle charging pile layout model, and solve the energy storage vehicle charging pile layout model through the improved adaptive particle swarm algorithm to obtain the energy storage vehicle charging pile layout plan.

2. According to claim 1, a method for distributing energy storage vehicle charging piles based on an improved adaptive particle swarm algorithm is characterized in that: In S1, the road nodes are simplified by the Floyd shortest path algorithm, and the weighted adjacency matrix A is used as the initial value of the distance matrix D to construct the Manhattan distance matrix between all road nodes: The elements (s=1,2,3...N) represents the number of nodes from the road node v i To road node v j In the path, the intermediate points are allowed to be {v1,v2 v3,…,v s } the length of the shortest path to any road node; when s = n, middle From the road node v i To road node v j In the path, the intermediate points are allowed to be {v1,v2v3,…,v s The length of the shortest path to any road node in}, D n This is the shortest distance matrix.

3. The energy storage vehicle charging pile layout method based on improved adaptive particle swarm algorithm according to claim 1 is characterized in that: In S2, the objective function of the comprehensive cost is: In the above formula, f 1i is the objective function of the sum of investment cost and maintenance cost in the ith year, f 2i is the objective function of the sum of network loss and charging cost in the i-th year, f 3i is the objective function of charging loss in the i-th year, f 4i is the objective function of the land cost in the i-th year, I is the total years, and i is the i-th year; The objective function f1 of the sum of the investment cost and the maintenance cost includes: In the above formula, r j is the number of transformers at public charging pile site j, a is the unit price of transformers for building charging piles, p j is the number of charging piles at public charging pile station j, b is the unit price of charging piles, c is j is the cost of the charging station at the public charging pile site j, r0 is the discount rate, n is the year of use, and μ is the depreciation rate; The objective function f2 of the sum of network loss and charging cost includes: f2=r j ·C1·T v ·p c ·365+p j ·C2·k t ·T v ·e c ·365+pQ·365; In the above formula, r j is the number of transformers at public charging station j, C1 is the steel loss, T v The effective charging time of the charging station, p c is the electricity price of the power company, C2 is the charging loss, k t is the simultaneous operation rate of multiple charging piles in the charging station, e c is the electricity price of the charging station, pQ is the charging cost for users; The objective function f3 of the charging loss includes: In the above formula, b1 is the cost generated by the user's annual idle power during charging, b2 is the indirect loss cost, and m j is the number of public charging pile sites, L j is the comprehensive distance from all charging demand points in the service area of ​​public charging station j to public charging station j, g is the unit power consumption of the energy storage vehicle, v is the average speed of the energy storage vehicle, and p is the driving time cost; The objective function f4 of the land cost includes: In the above formula, x j is the area occupied by the public charging station j, C t is the highest land use cost in the area, λ i Comprehensive factors for each public charging station.

4. The energy storage vehicle charging pile layout method based on improved adaptive particle swarm algorithm according to claim 1 is characterized in that: The S3 comprises the following steps: S3.1, algorithm initialization, set the maximum number of iterations of the adaptive particle swarm algorithm K = 250, the initial number of iterations K = 1, the initial velocity matrix is ​​0, and the boundary processing operator is used to modify the particle over-limit dimension value to the boundary value to ensure that all solutions are within the feasible domain; S3.2, enable the global optimal probability selection operator, update the elite solution library, merge the population and the elite solution library, and fill the updated Pareto front into the elite solution library; S3.

3. Iteratively solve the energy storage vehicle charging pile layout model through an adaptive particle swarm algorithm and make a judgment, wherein the judgment conditions include: a. Determine whether the optimal solution of the algorithm has not been updated for 20 consecutive generations. If so, enable the crossover operator and mutation operator; b. Determine whether the continuous stagnation time of the convergence accuracy of the algorithm exceeds 25 generations. If so, rotate the global optimal selection method, wherein the rotating global optimal selection method is a global optimal probability selection method inversely proportional to the crowding distance and a global optimal probability selection method inversely proportional to the number of controlled particles; c. Determine whether the continuous stagnation time of the diversity of the algorithm exceeds 25 generations. If so, rotate the boundary processing method, which includes the trimming boundary processing method and the exponential distribution boundary processing method; Enter S3.4; S3.4 updates the individual optimal guide and the global optimal guide, the individual optimal guide is the optimal solution found by particle solution z in the tth cycle, the global optimal guide is the optimal solution found by the entire particle group so far, and all particles select the same optimal solution as the global guide; S3.5 updates the speed and position of the particle swarm, and uses the boundary processing operator to ensure that the new speed and position meet their respective boundary constraints, K = K + 1, and enters S3.3 to continue iterating until K ≥ Kmax or the elite solution library fails to update after 50 consecutive iterations, and the result obtained is the optimal result.

5. The energy storage vehicle charging pile layout method based on improved adaptive particle swarm algorithm according to claim 4 is characterized in that: In S3.2, the global optimal probability selection operator enables a global optimal probability selection method that is inversely proportional to the normalized crowding distance, solving the crowding distance normalized CD of z in the population k It can be expressed as: In the above formula, N d represents the number of objective functions; is the maximum value of the jth objective function, is the minimum value of the jth objective function; Assume that the objective function values of the j-th objective function are sorted, the serial number of the solution z is i, N is the population size. When 1 < i < N, f i+1,j - f i-1,j represents the crowding distance of the solution z in the objective function j. When the objective function value is at the boundary, that is, i ∈ [1, N], the crowding distance of the solution z in the objective function j is infinite.

6. The energy storage vehicle charging pile layout method based on improved adaptive particle swarm algorithm according to claim 4 is characterized in that: In the global optimal leader selection method of S3.4, the minimum value of the sum of objective functions is used to roughly evaluate the convergence accuracy of the algorithm. The global optimal probability selection operator that is inversely proportional to the number of controlled particles can be described as: In the above formula, X a ={x∈X|a<x} represents the particle set controlled by member a of elite library A. When x n When controlled by A, the global leader G n From the control particle x n Selecting from the elite solution set A is conducive to maintaining the continuity of the particle search direction; when x n When not controlled by A, G n Select from the entire A, the probability of each global bootstrap candidate a being selected is inversely proportional to the particle set X controlled by a a Size|X a |.

7. The energy storage vehicle charging pile layout method based on improved adaptive particle swarm algorithm according to claim 4 is characterized in that: The update equation for the velocity of the idth particle in the Kth iteration in S3.5 is: In the above formula, w, c1, c2 are inertia factors, is the velocity of the idth particle in the K-1th iteration, pbest id is the individual extreme value of the idth particle, gbest id is the global extreme value of the idth particle, is the position of the idth particle, r1 and r2 are both random numbers; Make the inertia factor w change with the number of iterations: In the above formula, k is the current iteration number, M is the total iteration number, and w max is the maximum value of the inertia factor, w min is the minimum value of the inertia factor; The update equation for the position of the idth particle in the Kth iteration is: In the above formula, χ∈[0,1] is the contraction factor that controls the speed, and χ=1; The boundary constraint formula for the velocity is: In the above formula, VMax and VMin represent the upper and lower bounds of the particle velocity, respectively; The boundary constraint formula for the position is: VarMax and VarMin represent the upper and lower bounds of the particle position, respectively.

8. A charging pile layout system for energy storage vehicles based on an improved adaptive particle swarm algorithm, characterized in that: The system is used to execute the energy storage vehicle charging pile layout method based on the improved adaptive particle swarm algorithm as described in any one of claims 1 to 7, specifically comprising: a node simplification module, a model construction module and a model solution module; The node reduction module is used to reduce the road nodes, select the central node with the smallest Manhattan distance to each road node for modeling calculation, and obtain the shortest distance matrix D n ; The model building module is used to build a layout model of energy storage vehicle charging piles with the minimum comprehensive cost as the objective function. The comprehensive cost takes into account: charging loss, land cost, the sum of calculation investment cost and maintenance cost, and the sum of network loss and charging cost; The model solving module is used to solve the energy storage vehicle charging pile layout model, and solves the energy storage vehicle charging pile layout model through an improved adaptive particle swarm algorithm to obtain an energy storage vehicle charging pile layout plan.

9. A charging pile layout device for energy storage vehicles based on an improved adaptive particle swarm algorithm, characterized in that: The method comprises a memory and a processor, wherein the memory is used to store computer program code and transmit the computer program code to the processor; The processor is used to execute the energy storage vehicle charging pile layout method based on the improved adaptive particle swarm algorithm as described in any one of claims 1 to 7 according to the instructions in the computer program code.

10. A computer storable medium, wherein a computer program is stored in the computer storable medium, characterized in that: The computer program is executed by the processor as the energy storage vehicle charging pile layout method based on the improved adaptive particle swarm algorithm as described in any one of claims 1 to 7.

Images