Power distribution network dispatching model solving method and system based on improved MOPSO algorithm

Abstract

The invention discloses a power distribution network dispatching model solving method and system based on an improved MOPSO algorithm, and the method comprises the steps: obtaining a power distribution network structure, building a target function and constraint conditions according to the power distribution network structure, and obtaining a double-layer coordinated dispatching model of a power distribution network; optimizing multiple targets of the double-layer coordinated dispatching model of the power distribution network to obtain an optimal solution set, evaluating an evolution state of the optimal solution set, and adjusting parameters of a multi-target optimization algorithm to obtain a first multi-target optimization algorithm; and solving the double-layer coordinated scheduling model of the power distribution network by using the first target optimization algorithm to obtain a power distribution network coordinated control strategy, thereby realizing comprehensive optimization of power supply quality, economy and environmental protection.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network dispatching, and in particular to a distribution network dispatching model solving method and system based on an improved MOPSO algorithm. Background Art

[0002] Existing technologies have many deficiencies in distribution network dispatching: first, most of them adopt single-objective optimization, which cannot fully consider the comprehensive needs of power supply quality, economy and environmental protection; second, they adopt static optimization models, which lack the ability to dynamically adapt to rapid changes in the power supply side and the load side; finally, the network structure adjustment ability is limited and cannot be flexibly adjusted according to the actual operation situation, resulting in the real-time performance and stability of the system being affected when facing the uncertainty of intermittent energy and electric vehicle loads. Commonly used optimization algorithms such as traditional particle swarm optimization (PSO) and genetic algorithms (GA) are prone to fall into local optimality when dealing with multi-objective optimization problems and lack effective global search capabilities; in addition, these algorithms have slow convergence speed and low computational efficiency when dealing with high-dimensional and nonlinear problems, and parameter selection relies on experience and lacks adaptive adjustment mechanisms, which makes it difficult to obtain the optimal solution in a complex and changeable distribution network environment.

[0003] Solve the problems of insufficient multi-objective optimization, poor dynamic adaptability and limited network structure adjustment ability in the existing distribution network dispatching technology. By introducing the improved multi-objective particle swarm optimization (MOPSO) algorithm, the global search ability and convergence speed of the algorithm are improved, and the comprehensive optimization of power supply quality, economy and environmental protection is achieved. By dynamically adjusting the network structure, the real-time performance and stability of the system are enhanced when facing intermittent energy and electric vehicle load uncertainty. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a distribution network dispatching model solving method and system based on the improved MOPSO algorithm to solve the problems of insufficient multi-objective optimization, poor dynamic adaptability and limited network structure adjustment capability existing in the existing distribution network dispatching technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a method for solving a distribution network dispatching model based on an improved MOPSO algorithm, comprising:

[0008] Acquire the distribution network structure, establish the objective function and constraint conditions according to the distribution network structure, and obtain a two-layer coordinated dispatching model of the distribution network;

[0009] Optimizing the multiple objectives of the two-layer coordinated dispatch model of the distribution network to obtain an optimal solution set, evaluating the evolutionary state of the optimal solution set, adjusting the parameters of the multi-objective optimization algorithm, and obtaining a first multi-objective optimization algorithm;

[0010] The first objective optimization algorithm is used to solve the two-layer coordinated scheduling model of the distribution network to obtain a distribution network collaborative control strategy.

[0011] As a preferred solution of the distribution network dispatching model solving method based on the improved MOPSO algorithm described in the present invention, wherein: the optimal solution set obtained by optimizing the multi-objectives of the two-layer coordinated dispatching model of the distribution network includes:

[0012] Mapping the optimal solution set using mapping logic;

[0013] The mapping logic is expressed as the ratio of the value of the mth objective function corresponding to the Kth solution in the optimal solution set minus the minimum value of the mth objective function and the maximum value of the mth objective function minus the minimum value of the mth objective function multiplied by a constant, and the result is rounded.

[0014] As a preferred solution of the distribution network dispatching model solution method based on the improved MOPSO algorithm described in the present invention, it also includes:

[0015] The entropy calculation of the mapped optimal solution vector is expressed as the sum of the ratio of the number of components in the Kth row and the mth column after mapping to a constant and the number of components in the Kth row and the mth column after mapping divided by the logarithm of the constant;

[0016] The difference entropy is calculated as the entropy in the hth iteration and the entropy in the h-1th iteration.

[0017] As a preferred solution of the distribution network dispatching model solution method based on the improved MOPSO algorithm described in the present invention, wherein: evaluating the evolutionary state of the optimal solution set includes:

[0018] If the absolute value of the entropy in the optimal solution set is greater than the population evolution processing convergence state threshold, the current population is in the evolution convergence state;

[0019] If the absolute value of entropy in the optimal solution set is greater than the population stagnation threshold and less than the population evolution convergence threshold, the current population is in an evolutionary diversification state;

[0020] If the absolute value of the entropy in the optimal solution set is less than the population evolution processing convergence state threshold, the current population is in an evolutionary stagnation state.

[0021] As a preferred solution of the distribution network dispatching model solution method based on the improved MOPSO algorithm described in the present invention, it also includes:

[0022] When the population is in the evolving state, the random operator is not enabled;

[0023] When the population is in a state of evolutionary stagnation, add random operators and increase the step size;

[0024] When the population is in the state of evolutionary convergence, a negative random operator is added to randomly reduce the step size.

[0025] As a preferred solution of the distribution network dispatching model solution method based on the improved MOPSO algorithm described in the present invention, wherein: adjusting the parameters of the multi-objective optimization algorithm includes:

[0026] For the inertia weight, if the population is in a state of evolutionary stagnation, the inertia weight is the inertia weight of the last iteration;

[0027] If the population is in a state of evolutionary convergence, the inertia weight is the previous inertia weight minus the adjustment factor of twice the inertia weight multiplied by 1 plus the absolute value of the entropy in the optimal solution set;

[0028] If the population is in an evolutionary diversified state, the inertia weight is the previous inertia weight plus the adjustment factor of the inertia weight multiplied by the absolute value of the entropy in the optimal solution set;

[0029] For the learning factor, if the population is in a state of evolutionary stagnation, the learning factor is the learning factor of the last iteration;

[0030] If the population is in a state of evolutionary convergence, the learning factor is the previous learning factor minus twice the adjustment factor of the corresponding learning factor multiplied by 1 plus the absolute value of the entropy in the optimal solution set;

[0031] If the population is in an evolutionary diversified state, the learning factor is the previous learning factor plus the adjustment factor of the corresponding learning factor multiplied by the absolute value of the entropy in the optimal solution set.

[0032] As a preferred solution of the distribution network dispatching model solution method based on the improved MOPSO algorithm described in the present invention, wherein:

[0033] Solving the two-layer coordinated dispatch model of the distribution network by using the first objective optimization algorithm includes:

[0034] Enter the initialization schedule period;

[0035] Read the basic load of the distribution network, initialize parameters and constraints;

[0036] Initialize the external archive and save the initial optimal release into the archive;

[0037] Calculate the entropy and differential entropy of the optimal solution set, obtain the population evolution state, and calculate the next iteration parameters;

[0038] Calculate the fitness of each particle;

[0039] Update external archives and individual optimality;

[0040] The solver is used to solve the dispatching layer of the two-layer coordinated dispatching model of the distribution network to obtain the output level of the corresponding power generation resources;

[0041] The optimality of the new particles is liberated to the external archive, and the first judgment on the overflow of the archive is made;

[0042] Perform a second judgment on the number of iterations;

[0043] If the second judgment is met, a third judgment is performed to output the output level of the optimal power generation resources of the two-layer coordinated dispatch model of the distribution network on that day.

[0044] In a second aspect, the present invention provides a distribution network dispatching model solving system based on an improved MOPSO algorithm, comprising:

[0045] A model building module is used to obtain the distribution network structure, establish the objective function and constraint conditions according to the distribution network structure, and obtain a two-layer coordinated dispatch model of the distribution network;

[0046] An adjustment module, used for optimizing the multiple objectives of the two-layer coordinated dispatch model of the distribution network to obtain an optimal solution set, evaluating the evolution state of the optimal solution set, adjusting the parameters of the multi-objective optimization algorithm, and obtaining a first multi-objective optimization algorithm;

[0047] A solution module is used to use the first objective optimization algorithm to solve the two-layer coordinated scheduling model of the distribution network to obtain a distribution network collaborative control strategy.

[0048] In a third aspect, the present invention provides a computing device, comprising:

[0049] Memory and processor;

[0050] The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the distribution network dispatching model solving method based on the improved MOPSO algorithm are implemented.

[0051] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for solving the distribution network dispatching model based on the improved MOPSO algorithm.

[0052] Compared with the prior art, the present invention has the following beneficial effects: the present invention introduces the concept of Pareto entropy by using an improved multi-objective particle swarm optimization algorithm, effectively evaluates the evolutionary state of the population, and guides the adjustment of algorithm parameters, thereby enhancing the global search capability of the algorithm in the search space, and helps the algorithm to more comprehensively explore possible solutions when facing complex multi-objective optimization problems. By dynamically adjusting the inertia weights and learning factors of the particles, the algorithm can converge to the optimal solution faster, and helps the algorithm to more effectively balance exploration and development during the iteration process, thereby accelerating the convergence speed and achieving comprehensive optimization of power supply quality, economy, and environmental protection. By dynamically adjusting the network structure, the real-time performance and stability of the system in the face of intermittent energy and electric vehicle load uncertainty are enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0054] Figure 1 A schematic diagram of the overall process logic of a method for solving a distribution network dispatching model based on an improved MOPSO algorithm according to an embodiment of the present invention;

[0055] Figure 2 The energy change of an electric vehicle cluster before and after a certain scheduling period of a distribution network scheduling model solving method based on an improved MOPSO algorithm described in an embodiment of the present invention. DETAILED DESCRIPTION

[0056] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0057] Example 1

[0058] Reference Figure 1-Figure 2 , is an embodiment of the present invention, and provides a method for solving a distribution network dispatching model based on an improved MOPSO algorithm, comprising:

[0059] S100: Obtain the distribution network structure, establish the objective function and constraint conditions according to the distribution network structure, and obtain a two-layer coordinated dispatching model of the distribution network;

[0060] Specifically, Figure 2 As shown in the figure, the energy change of electric vehicle cluster before and after a certain scheduling period is expressed as:

[0061] N t+1 S EV SOC t+1 =N t S EV SOC t +(n ch,t -n dis,t ) EV Δt+ΔE EV,t

[0062]

[0063] Among them, SOC t is the average state of charge of the electric vehicle cluster at time t, SOC t+1 is the average state of charge of the electric vehicle cluster at time t+1, S EV is the rated capacity of the electric vehicle battery, n ch,t is the number of cars that choose to charge in the electric vehicle cluster at time t, n dis,t is the number of electric vehicles that choose to discharge in the electric vehicle cluster at time t, P EV is the charging and discharging power of a single electric vehicle, N t is the number of electric vehicles in the electric vehicle cluster at time t, N t+1 is the number of electric vehicles in the electric vehicle cluster at time t+1, ΔE EV,t is the change in the electric vehicle cluster power caused by the entry and exit of electric vehicles at time t, SOC i SOC is the state of charge of the electric vehicle at the time of entering the vehicle at time t. j is the charge state of the electric vehicle at the time of leaving during period t, N A,t is the number of electric vehicles entering the electric vehicle cluster during period t, N D,t is the number of electric vehicles leaving the electric vehicle cluster during period t.

[0064] When controlling the charging and discharging behavior of an electric vehicle cluster, the constraints are expressed as:

[0065] SOC min ≤SOC t ≤SOC max

[0066] n ch,t +n dis,t ≤N t

[0067]

[0068] Among them, SOCmin =0.3, SOC max =0.95, H is the charge and discharge state of an electric vehicle, H ＝1 is the charging state, H ＝-1 It is in discharge state.

[0069] In the embodiment of the present application, the two-layer coordinated dispatching model of the distribution network includes a new distribution network dispatching layer optimization model and a new distribution network network layer coordinated optimization model;

[0070] Specifically, in addition to electric vehicle clusters, the new distribution network dispatching layer optimization model considers coordinated control of the output of distributed photovoltaics, flexible loads and traditional energy storage in the new distribution network;

[0071] Distributed photovoltaics, as a distributed power source with adjustable output, participate in the coordinated control of the new distribution network, which can be expressed as:

[0072] P MT (t)≤P MT,max

[0073] P MT (t)-P MT (t-1)≤ΔP MT

[0074] Among them, P MT,max is the maximum output of distributed photovoltaics, ΔP MT is the ramp output constraint of distributed photovoltaics, P MT (t) is the output of distributed photovoltaic at time t, P MT (t-1) is the output of distributed photovoltaic power generation at time t-1;

[0075] The flexible loads in the new distribution network can be divided into two types: curtailable loads and transferable loads. According to the output plan of each flexible load user, the response output per unit time and the control cost of the flexible load are expressed as:

[0076]

[0077] C FL,t =|U i,t |×P FL,t ×e I

[0078] Among them, U i,l is the response state of user i at time t, 1 means increasing the load, -1 means reducing the load, and 0 means not participating. FL,t is the control cost of the flexible load at time t, e I is the compensation cost for flexible loads participating in coordinated control, P FL,t is all the flexible loads involved in the control at time t, P i,lis the output that user i can respond to at time t, satisfying the following constraints:

[0079] P i,t ≤P Li,t ×τ

[0080] Among them, P Li,t is the maximum load of user i at time t, τ is the response ratio of the user at time t;

[0081] For the transferable load in the flexible load, the following constraints must also be satisfied:

[0082]

[0083] Among them, U I,t is the voltage of node I at time t, P l,t is the power of node I at time t, and Δt is the time step.

[0084] For energy storage in the new distribution network, when participating in coordinated control, its costs include call cost and low storage and high generation. The control cost is expressed as:

[0085]

[0086] Among them, C ESS,1 is the call cost of energy storage, |P t | is the power change of energy storage during period t, C ESS,2 is the benefit of low storage and high generation of energy storage, C o (t) is the peak and valley time-of-use electricity price.

[0087] When calling energy storage, the constraints are expressed as:

[0088] SOC min ≤SOC t ≤SOC max

[0089] P min ≤P ESS,t ×P max

[0090] Among them, SOC min =0.1, SOC max =0.9, P ESS,t is the power output of the energy storage system at time t, P min , P max It is the minimum power requirement and the maximum power requirement.

[0091] The scheduling layer is optimized with the lowest operating cost as the optimization goal, and the objective function is expressed as:

[0092] minf1=CESS +C FL +C EV +C MT +C grid

[0093] Among them, C ESS is the control cost of energy storage, C FL is the control cost of flexible load, C EV is the control cost of the controllable electric vehicle cluster, C MT is the control cost of the gas unit, C grid The cost of purchasing electricity from the upper power grid.

[0094] The calculation of each cost is expressed as:

[0095]

[0096] Among them, k is the control price of distributed photovoltaic, P grid,t is the amount of electricity purchased from the upper power grid during period t, The price for purchasing electricity from the superior power grid.

[0097] When coordinating and controlling the new distribution network, the following constraints need to be met:

[0098]

[0099] Among them, P Grid,t P is the amount of electricity purchased from the main grid during period t, RL,t is the basic load, P DG,t is the output of distributed power sources such as photovoltaics, wind turbines and gas turbines during period t, P FL,t is the output of all flexible load users in period t, P ESS,t It is the energy storage load, which is positive when charged by the grid and negative when discharged to the grid;

[0100] Specifically, the objective function of the new distribution network layer coordination optimization model is expressed as:

[0101] minf1=C ESS +C FL +C EV +C MT +C switch +C grid +C loss

[0102]

[0103] Among them, C ESS is the control cost of energy storage, C FL is the flexible load control cost, C EV To control the cost of EV cluster, CMT To control the cost of gas turbines, C switch In order to control the grid-side switch operation cost, C grid is the cost of purchasing electricity from the upper grid, C loss is the network loss cost, S n (t) is the switch state of the contact switch n during period t, S n (t-1) is the switch state of tie switch n in period t+1, N is the number of tie switches in the network, e swich is the cost of controlling the tie switch once, P loss(t) is the loss in the network during period t, C p The network loss price.

[0104] The minimum standard deviation of the system load curve is expressed as:

[0105]

[0106] Among them, P load,t is the system base load during period t, H k is the state of the kth electric car, N t is the number of electric vehicles participating in the control during period t, P FI,t is the total amount of flexible load involved in control during period t, P ESS,t is the output status of the energy storage participating in the control during period t, P WT,t , P PV,t is the wind power output in period t, P PV,t is the photovoltaic output during period t, P MT,t is the output of the gas turbine during period t.

[0107] Taking the minimization of the cost of electric vehicle clusters as one of the goals, its charging and discharging behavior is controlled and guided. The objective function is expressed as:

[0108]

[0109] Among them, n ch.t The number of electric vehicles charged for period t, n dis.t is the number of electric vehicles discharged during period t, C D (t) is the time-of-use electricity price during period t;

[0110] In order to ensure the reliability of power grid operation, the following constraints need to be met:

[0111] Node voltage constraint: U i,min ≤U i ≤U i,max , where U i,min is the upper limit of the voltage amplitude at node i, and is the lower limit of the voltage amplitude at node i;

[0112] Branch power constraint: P L,t ≤P L,max , where P L,t is the power transmitted on branch L during period t, P L,max is the limit value that line L can transmit;

[0113] Power conservation constraint: P Grid.t =P RL.t -P DG,t +P FL,t +P ESS.t +P EV,t +P loss.t

[0114] Among them, P Grid.t P is the amount of electricity purchased from the main grid during period t, RL.t is the basic load, P DG,t is the output of distributed power sources such as photovoltaics, wind turbines and gas turbines during period t, P FL,t is the output of all flexible load users in period t, P ESS.t is the energy storage load, P EV,t is the electric vehicle cluster load, P loss.t is the loss in the network during period t.

[0115] State of Charge Constraint: SOC min ≤SOC t ≤SOC max , where SOC t SOC is the state of charge of the energy storage or electric vehicle during period t. min is the lower limit of the state of charge, SOC max The upper limit of the state of charge.

[0116] Controllable DG output constraint: P MT,t ≤P MT,max , where P MT,t is the distributed photovoltaic output during period t, P MT,max is the maximum value of the distributed output during period t.

[0117] It should be noted that by establishing objective functions and constraints, it is possible to achieve optimal allocation and scheduling of resources such as distributed photovoltaics, energy storage systems, flexible loads, and electric vehicle clusters. By dynamically adjusting the network structure and resource allocation, it is helpful to improve the real-time performance and stability of the power grid when facing intermittent energy and electric vehicle load uncertainties. With the lowest operating cost as the optimization goal, the operating cost of the power grid can be effectively reduced by coordinating and controlling distributed photovoltaics, energy storage, flexible loads, etc., taking into account the climbing output constraints of distributed photovoltaics and the response state of flexible loads, so that the power grid can better adapt to the fluctuations of renewable energy and load changes.

[0118] S200: Optimizing multiple objectives of the two-layer coordinated dispatch model of the distribution network to obtain an optimal solution set, evaluating the evolution state of the optimal solution set, adjusting parameters of the multi-objective optimization algorithm, and obtaining a first multi-objective optimization algorithm;

[0119] S300: Solve the two-layer coordinated dispatch model of the distribution network using the first objective optimization algorithm to obtain a coordinated control strategy for the distribution network.

[0120] It should be noted that through the improved multi-objective particle swarm optimization (MOPSO) algorithm, the comprehensive optimization of power supply quality, economy and environmental protection is achieved, the global search ability of the algorithm is improved, and the risk of falling into the local optimal solution is reduced, so as to obtain the optimal solution in the complex and changeable distribution network environment. By dynamically adjusting the network structure, the dynamic adaptability of the distribution network in the face of rapid changes in the power supply side and the load side is enhanced, especially in the case of intermittent energy and electric vehicle load uncertainty, the real-time performance and stability of the distribution network are improved, which is crucial to ensure the reliable operation of the power grid.

[0121] In the embodiment of the present application, the above step S200 includes the following sub-steps A1-A4;

[0122] In A1: the optimal solution set is mapped using mapping logic;

[0123] In A2: The mapping logic is expressed as the ratio of the value of the mth objective function corresponding to the Kth solution in the optimal solution set minus the minimum value of the mth objective function and the maximum value of the mth objective function minus the minimum value of the mth objective function multiplied by a constant, and the result is rounded.

[0124] In A3: the entropy calculation of the mapped optimal solution vector is expressed as the sum of the ratio of the number of components in the Kth row and the mth column after mapping to a constant and the logarithm of the number of components in the Kth row and the mth column after mapping divided by the constant;

[0125] In A4: the difference entropy is calculated as the entropy in the hth iteration and the entropy in the h-1th iteration.

[0126] Specifically, there are two common methods to solve multi-objective optimization problems. One is to transform the multi-objective problem into a single-objective problem for solution; the other is to directly use a multi-objective optimization algorithm. The former combines the objective functions by weight, but this process is highly subjective and objectives of different dimensions are difficult to integrate; the latter, such as the multi-objective particle swarm optimization algorithm, can optimize multiple objectives at the same time and obtain a set of Pareto optimal solutions.

[0127] In the field of model optimization, the multi-objective optimization problem is expressed as:

[0128] min F(x)=[f1(x),f2(x),...,fM (x)]

[0129] x=[x1,x2,...,x n ]

[0130]

[0131] Among them, x is the decision variable, n is the number of decision variables, g i (x) is an equality constraint that specifies the feasible domain of the control variable, h j (x) is an inequality constraint that specifies the feasible domain of the control variable;

[0132] When M ≥ 2, the model is a multi-objective optimization model, and the objective functions f1(x), f2(x), ..., f M (x) constitutes the M-dimensional target space;

[0133] Specifically, Pareto dominance is expressed as:

[0134]

[0135] For any two solution vectors u and v in the solution space Ω, if they satisfy the above equation, then u is called dominant over v, denoted as u<v, indicating that the solution vector u is not inferior to v in all objective functions;

[0136] Pareto optimal solution: x * is a solution in the solution space Ω. For any solution x in the solution space, if The solution vector x is called * It is the Pareto optimal solution;

[0137] Pareto optimal solution set or Pareto front end: The solution set consisting of all the above optimal solutions x is called the Pareto optimal solution set, and the area formed by the corresponding objective function values ​​is called the Pareto front end.

[0138] In the embodiment of the present application, in order to calculate the distribution entropy of the multi-dimensional Pareto solution set and overcome the difficulty of direct calculation in the Cartesian coordinate system, the Pareto solution set can be first mapped from the Cartesian coordinate system to a two-dimensional parallel grid coordinate system (Parallel Cell Coordinates Systems, PCCS), and the mapping logic is expressed as:

[0139]

[0140] in, is a rounding function, the return value is an integer, f k,m is the value of the mth objective function corresponding to the kth solution in the optimal solution set, is the minimum value of the mth objective function, is the maximum value of the mth objective function;

[0141] After mapping the optimal solution vector in the Cartesian coordinate system to PCCS, the entropy calculation of all the optimal solution vectors in the external archive can be expressed as:

[0142]

[0143] Among them, E(t) is the entropy of the Pareto solution set in the previous external archive, p k,m (t) is the number of components that fall in the kth row and mth column of PCCS after mapping;

[0144] The calculated differential entropy is expressed as: ΔE = E(h) - E(h-1), where ΔE is the differential entropy, E(h) is the Pareto entropy in the h-th iteration, and E(h-1) is the entropy in the h-1th iteration;

[0145] In summary, for a Pareto optimal solution set of K solution vectors, the number of optimized objective functions is m, then the maximum Pareto entropy of the solution set is logKM, and the minimum Pareto entropy is logM.

[0146] In the embodiment of the present application, after completing steps A1-A4 in the above step S200, the following steps A5-A7 are also included;

[0147] In A5: If the absolute value of the entropy in the optimal solution set is greater than the population evolution processing convergence state threshold, the current population is in the evolution convergence state;

[0148] In A6: If the absolute value of the entropy in the optimal solution set is greater than the population stagnation threshold and less than the population evolution convergence threshold, the current population is in an evolutionary diversification state;

[0149] In A7: If the absolute value of the entropy in the optimal solution set is less than the population evolution processing convergence state threshold, the current population is in an evolutionary stagnation state;

[0150] Specifically, the entropy of the Pareto solution set is used to describe the dispersion of solutions in the solution set and reflect the evolutionary state of the current population, which is expressed as:

[0151]

[0152] Among them, H is the number of solution vectors in the solution set after the dominated old solution is kicked out, δc is the threshold for the population evolution to be in a convergence state, and δ s is the threshold of population evolution stagnation.

[0153] In the embodiment of the present application, after completing steps A5-A7 in the above step S200, the following steps A8-A10 are also included;

[0154] In A8: When the population is in the evolving state, the random operator is not enabled;

[0155] In A9: When the population is in a state of evolutionary stagnation, add random operators and increase the step size;

[0156] In A10: When the population is in the evolutionary convergence state, a negative random operator is added to randomly reduce the step size.

[0157] Specifically, in order to prevent the multi-objective optimization algorithm (such as the multi-objective particle swarm optimization algorithm MOPSO) from falling into the local optimum, the exploration ability of the algorithm can be enhanced by randomly perturbing the velocity operator. The specific perturbation strategy is expressed as:

[0158] V i (h+1)=wV i (h)+c1r1(P best -X i (h))+c2r2(G best -X i (h))+v rd (h)

[0159]

[0160] Among them, v rd (h) is the random speed operator, a and b are random numbers with values ​​of (0.001, 0.01), and z(h) is the random operator startup state;

[0161] It should be noted that when the population is in a state of evolutionary stagnation, random operators are added to increase the step size so that the algorithm can improve its ability to escape from the dead zone. When the population is in a state of evolutionary convergence, negative random operators are added to randomly reduce the step size and increase the convergence speed.

[0162] In the embodiment of the present application, after completing steps A8-A10 in the above step S200, the following steps A11-A16 are also included;

[0163] In A11: For the inertia weight, if the population is in a state of evolutionary stagnation, the inertia weight is the inertia weight of the last iteration;

[0164] In A12: If the population is in a state of evolutionary convergence, the inertia weight is the previous inertia weight minus the adjustment factor of twice the inertia weight multiplied by 1 plus the absolute value of the entropy in the optimal solution set;

[0165] In A13: If the population is in an evolutionary diversified state, the inertia weight is the previous inertia weight plus the adjustment factor of the inertia weight multiplied by the absolute value of the entropy in the optimal solution set;

[0166] In A14: For the learning factor, if the population is in a state of evolutionary stagnation, the learning factor is the learning factor of the last iteration;

[0167] In A15: If the population is in a state of evolutionary convergence, the learning factor is the previous learning factor minus twice the adjustment factor of the corresponding learning factor multiplied by 1 plus the absolute value of the entropy in the optimal solution set;

[0168] In A16: If the population is in an evolutionary diversified state, the learning factor is the previous learning factor plus the adjustment factor of the corresponding learning factor multiplied by the absolute value of the entropy in the optimal solution set.

[0169] Specifically, in order to improve the global search capability of the algorithm and reduce the local optimal situation, the particle inertia weight and learning factor optimization strategy is expressed as:

[0170]

[0171] Among them, h is the number of contemporary iterations, l w , is the adjustment factor, w(h) is the inertia weight, C1(h) and C2(h) are the learning factors;

[0172] In the embodiment of the present application, the dense distance of the calculated particles in the external file is expressed as:

[0173]

[0174] Among them, A is the external archive set. The smaller the dense distance, the denser the area where the solution is located. The larger the dense distance, the sparser the area where the solution is located.

[0175] Individual optimal position P best Update: If the position X of particle i in a certain iteration i Dominate P best , then update P best For X i Otherwise, the one with the highest degree of dominance is still selected as P best ;

[0176] The global optimal position G best Update: In multi-objective optimization algorithms, external archives are used to store non-dominated solutions (Pareto optimal solutions) to maintain the diversity and quality of solutions. When the external archive reaches its maximum capacity, proper management is required to ensure that the archive stores the most valuable solutions.

[0177] If the external archive has not reached its maximum capacity or the new solution dominates some of the solutions in the external archive, the new solution is stored in the external archive. If the external archive has reached its maximum capacity, the density distance of the new solution in the archive is calculated and sorted in descending order by density distance. Only when the density distance of the new solution is greater than the density distance of at least one member in the archive, the new solution is stored in the archive and the member with the smallest density distance is removed. At the same time, the global optimal position G is selected. best ;

[0178] It should be noted that entropy, as an indicator of the degree of disorder in a system, can also be used to represent the diversity of optimal solutions in the Pareto set. In the process of multi-objective optimization, a higher Pareto entropy value means that the current solution set has a higher diversity and distribution range, reflecting that the population has explored more extensively in the search space. Conversely, a lower entropy value indicates that the solution set is more concentrated and may have approached or reached a local optimal area. By monitoring the changes in Pareto entropy, the evolutionary state of the population can be effectively evaluated, the adjustment of algorithm parameters can be guided, and the balance between global search and local development can be promoted.

[0179] Using the concept of Pareto entropy to visualize the evolution of the population can provide a scientific basis for adjusting algorithm parameters and avoid the blindness of traditional adaptive methods when adjusting parameters. According to the evolutionary state of the population, a random velocity operator and a dynamic adjustment strategy of weight factors are introduced to prevent the algorithm from falling into the local optimum too early and improve the convergence accuracy of the final solution. This method not only enhances the exploration ability of the algorithm, but also improves the quality of the solution.

[0180] In the embodiment of the present application, the above step S300 includes the following sub-steps B1-B10;

[0181] In B1: enter the initialization schedule period;

[0182] In B2: read the basic load of the distribution network, initialize parameters and constraints;

[0183] In B3: Initialize the external archive and save the initial optimal release into the archive;

[0184] In B4: Calculate the entropy and differential entropy of the optimal solution set, obtain the population evolution state, and calculate the next iteration parameters;

[0185] In B5: calculate the fitness of each particle;

[0186] In B6: Update external archive and individual optimality;

[0187] In B7: using a solver to solve the dispatching layer of the two-layer coordinated dispatching model of the distribution network, the output level of the corresponding power generation resources is obtained;

[0188] In B8: the optimality of the new particles is liberated to the external archive, and the first judgment on the overflow of the archive is made;

[0189] In B9: a second determination is made on the number of iterations;

[0190] In B10: If the second judgment is met, a third judgment is performed to output the output level of the optimal power generation resources of the two-layer coordinated scheduling model of the distribution network on that day.

[0191] Specifically, when solving the coordinated control model of the new distribution network, the output of the energy storage ESS involved in the dispatching layer is a continuous variable, and the state of the interconnection switch involved in the network layer is a discrete variable. When solving, the cplex solver is used for the dispatching layer, and the improved MOPSO algorithm is used for the network layer.

[0192] The first judgment includes: if the Pareto optimality of the new particle is released to the external archive and the archive overflows, then the optimal position P of the individual is selected according to the dense distance best and the global optimal position G best , if the file does not overflow, it is directly stored;

[0193] The second judgment includes: if the solution process reaches the number of iterations, then output the optimal switch combination and energy storage, FL and other outputs during this period, if not, then continue to iterate;

[0194] The third judgment includes: if the current scheduling period is greater than or equal to the total time scheduling length, then the optimal switch combination and energy storage, FL and other outputs for the day are output; if the current scheduling period is less than the total time scheduling length, then the solution is re-calculated.

[0195] It should be noted that through hierarchical scheduling and optimization, continuous and discrete variables in the distribution network can be handled more effectively, the efficiency of the scheduling process can be improved, and by optimizing the output of power generation resources, including energy storage, flexible loads and gas turbines, the rational allocation and utilization of resources can be achieved, and operating costs can be reduced, which will help improve the utilization rate of renewable energy and reduce dependence on traditional energy.

[0196] The above is a schematic scheme of a distribution network dispatching model solving method based on an improved MOPSO algorithm in this embodiment. It should be noted that the technical scheme of the distribution network dispatching model solving system based on the improved MOPSO algorithm and the technical scheme of the distribution network dispatching model solving method based on the improved MOPSO algorithm belong to the same concept. The details not described in detail in the technical scheme of the distribution network dispatching model solving system based on the improved MOPSO algorithm in this embodiment can be found in the description of the technical scheme of the distribution network dispatching model solving method based on the improved MOPSO algorithm.

[0197] The distribution network dispatching model solving system based on the improved MOPSO algorithm in this embodiment includes:

[0198] A model building module is used to obtain the distribution network structure, establish the objective function and constraint conditions according to the distribution network structure, and obtain a two-layer coordinated dispatch model of the distribution network;

[0199] An adjustment module, used for optimizing the multiple objectives of the two-layer coordinated dispatch model of the distribution network to obtain an optimal solution set, evaluating the evolution state of the optimal solution set, adjusting the parameters of the multi-objective optimization algorithm, and obtaining a first multi-objective optimization algorithm;

[0200] A solution module is used to use the first objective optimization algorithm to solve the two-layer coordinated scheduling model of the distribution network to obtain a distribution network collaborative control strategy.

[0201] This embodiment also provides a computing device, which is suitable for solving the distribution network dispatching model based on the improved MOPSO algorithm, including:

[0202] Memory and processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the distribution network dispatching model solving method based on the improved MOPSO algorithm as proposed in the above embodiment.

[0203] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for solving the distribution network scheduling model based on the improved MOPSO algorithm as proposed in the above embodiment is implemented.

[0204] The storage medium proposed in this embodiment and the distribution network dispatching model solving method based on the improved MOPSO algorithm proposed in the above embodiment belong to the same inventive concept. The technical details not described in detail in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0205] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course can also be implemented by hardware. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ReadOnly, Memory, ROM), random access memory (Random Access Memory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.

Claims

1. A method for solving a distribution network dispatching model based on an improved MOPSO algorithm, characterized in that: include: Acquire the distribution network structure, establish the objective function and constraint conditions according to the distribution network structure, and obtain a two-layer coordinated dispatching model of the distribution network; Optimizing the multiple objectives of the two-layer coordinated dispatch model of the distribution network to obtain an optimal solution set, evaluating the evolutionary state of the optimal solution set, adjusting the parameters of the multi-objective optimization algorithm, and obtaining a first multi-objective optimization algorithm; The first objective optimization algorithm is used to solve the two-layer coordinated scheduling model of the distribution network to obtain a distribution network collaborative control strategy.

2. The method for solving the distribution network dispatching model based on the improved MOPSO algorithm according to claim 1, characterized in that: The optimal solution set obtained by optimizing the multi-objectives of the two-layer coordinated dispatch model of the distribution network includes: Mapping the optimal solution set using mapping logic; The mapping logic is expressed as the ratio of the value of the mth objective function corresponding to the Kth solution in the optimal solution set minus the minimum value of the mth objective function and the maximum value of the mth objective function minus the minimum value of the mth objective function multiplied by a constant, and the result is rounded.

3. The method for solving the distribution network dispatching model based on the improved MOPSO algorithm according to claim 2, characterized in that: Also includes: The entropy calculation of the mapped optimal solution vector is expressed as the sum of the ratio of the number of components in the Kth row and the mth column after mapping to a constant and the number of components in the Kth row and the mth column after mapping divided by the logarithm of the constant; The difference entropy is calculated as the entropy in the hth iteration and the entropy in the h-1th iteration.

4. The method for solving the distribution network dispatching model based on the improved MOPSO algorithm according to claim 1 or 3, characterized in that: Evaluating the evolutionary state of the optimal solution set includes: If the absolute value of the entropy in the optimal solution set is greater than the population evolution processing convergence state threshold, the current population is in the evolution convergence state; If the absolute value of entropy in the optimal solution set is greater than the population stagnation threshold and less than the population evolution convergence threshold, the current population is in an evolutionary diversification state; If the absolute value of the entropy in the optimal solution set is less than the population evolution processing convergence state threshold, the current population is in an evolutionary stagnation state.

5. The method for solving the distribution network dispatching model based on the improved MOPSO algorithm according to claim 4, characterized in that: Also includes: When the population is in the evolving state, the random operator is not enabled; When the population is in a state of evolutionary stagnation, add random operators and increase the step size; When the population is in the state of evolutionary convergence, a negative random operator is added to randomly reduce the step size.

6. The method for solving the distribution network dispatching model based on the improved MOPSO algorithm according to claim 5, characterized in that: Adjusting the parameters of the multi-objective optimization algorithm includes: For the inertia weight, if the population is in a state of evolutionary stagnation, the inertia weight is the inertia weight of the last iteration; If the population is in a state of evolutionary convergence, the inertia weight is the previous inertia weight minus the adjustment factor of twice the inertia weight multiplied by 1 plus the absolute value of the entropy in the optimal solution set; If the population is in an evolutionary diversified state, the inertia weight is the previous inertia weight plus the adjustment factor of the inertia weight multiplied by the absolute value of the entropy in the optimal solution set; For the learning factor, if the population is in a state of evolutionary stagnation, the learning factor is the learning factor of the last iteration; If the population is in a state of evolutionary convergence, the learning factor is the previous learning factor minus twice the adjustment factor of the corresponding learning factor multiplied by 1 plus the absolute value of the entropy in the optimal solution set; If the population is in an evolutionary diversified state, the learning factor is the previous learning factor plus the adjustment factor of the corresponding learning factor multiplied by the absolute value of the entropy in the optimal solution set.

7. The method for solving the distribution network dispatching model based on the improved MOPSO algorithm according to claim 6, characterized in that: Solving the two-layer coordinated dispatch model of the distribution network by using the first objective optimization algorithm includes: Enter the initialization schedule period; Read the basic load of the distribution network, initialize parameters and constraints; Initialize the external archive and save the initial optimal release into the archive; Calculate the entropy and differential entropy of the optimal solution set, obtain the population evolution state, and calculate the next iteration parameters; Calculate the fitness of each particle; Update external archives and individual optimality; The solver is used to solve the dispatching layer of the two-layer coordinated dispatching model of the distribution network to obtain the output level of the corresponding power generation resources; The optimality of the new particles is liberated to the external archive, and the first judgment on the overflow of the archive is made; Perform a second judgment on the number of iterations; If the second judgment is met, a third judgment is performed to output the output level of the optimal power generation resources of the two-layer coordinated scheduling model of the distribution network on that day.

8. A system using the distribution network dispatching model solving method based on the improved MOPSO algorithm as described in any one of claims 1 to 7, characterized in that: include: A model building module is used to obtain the distribution network structure, establish the objective function and constraint conditions according to the distribution network structure, and obtain a two-layer coordinated dispatch model of the distribution network; An adjustment module, used to optimize the multiple objectives of the two-layer coordinated dispatch model of the distribution network to obtain an optimal solution set, evaluate the evolution state of the optimal solution set, adjust the parameters of the multi-objective optimization algorithm, and obtain a first multi-objective optimization algorithm; A solution module is used to use the first objective optimization algorithm to solve the two-layer coordinated scheduling model of the distribution network to obtain a distribution network collaborative control strategy.

9. An electronic device, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the distribution network dispatching model solving method based on the improved MOPSO algorithm as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for solving the distribution network dispatching model based on the improved MOPSO algorithm as described in any one of claims 1 to 7.

Images