A method and system for AC / DC distribution network planning based on preference multi-objective algorithm

Abstract

The present invention discloses an AC / DC distribution network planning method and system based on a preference multi-objective algorithm. The constructed AC / DC distribution network planning model not only takes into account the lowest economic cost and the highest reliability, but also takes into account the maximum new energy consumption, the minimum voltage deviation and the minimum network loss. It is closer to the AC / DC distribution network planning scenario with high penetration rate of new energy access. In the model solving process, the preference multi-objective optimization algorithm based on NSGAIII is used to obtain the optimal solution that meets user preferences for the AC / DC distribution network planning model, thereby improving the solution speed of the multi-objective planning and solving the technical problems that the existing AC / DC distribution network multi-objective planning model only considers economic cost and reliability, cannot be adapted to the AC / DC distribution network planning with high penetration rate of new energy access, and in the process of solving the multi-objective planning model, the Pareto solution is relatively scattered, the computing power cannot be concentrated on the decision maker's preference solution, and the optimization speed is slow.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network planning, and in particular to an AC / DC distribution network planning method and system based on a preference multi-objective algorithm. Background Art

[0002] As the penetration of renewable energy in distribution networks continues to increase, uncertainty in power balance continues to grow. Traditional distribution network planning methods, which use maximum load for power source and grid planning, are overly conservative, lack economic efficiency and reliability, and fail to achieve optimal resource allocation. Therefore, research on AC / DC distribution network planning methods suitable for high-penetration renewable energy access is a crucial topic.

[0003] The integration of renewable energy into AC / DC distribution networks affects voltage deviations, network losses, and renewable energy absorption characteristics. Existing multi-objective planning models for AC / DC distribution networks only consider economic costs and reliability, making them unsuitable for planning AC / DC distribution networks with high renewable energy penetration. Furthermore, when solving multi-objective planning models, the Pareto solutions are relatively dispersed, making it difficult to concentrate computing power on the decision-maker's preferred solution, resulting in slow optimization. Summary of the Invention

[0004] The present invention provides an AC / DC distribution network planning method and system based on a preference multi-objective algorithm, which is used to solve the technical problems that the existing AC / DC distribution network multi-objective planning model only considers economic cost and reliability, is not suitable for AC / DC distribution network planning with high penetration of new energy, and in the process of solving the multi-objective planning model, the Pareto solution is relatively scattered, and the computing power cannot be concentrated on the decision maker's preference solution, resulting in slow optimization speed.

[0005] In view of this, a first aspect of the present invention provides an AC / DC distribution network planning method based on a preference multi-objective algorithm, comprising:

[0006] Obtain the distribution network parameters of the area to be optimized;

[0007] Based on the distribution network parameters of the area to be optimized, an AC / DC distribution network planning model is constructed that takes into account the lowest economic cost, highest reliability, maximum new energy absorption, minimum voltage deviation, and minimum network loss.

[0008] The AC / DC distribution network planning model is solved according to the preference multi-objective optimization algorithm based on NSGAIII to obtain the optimal solution that satisfies user preferences. Among them, the preference multi-objective optimization algorithm based on NSGAIII introduces user preferences into the NSGAIII algorithm to limit the size of the objective space.

[0009] Optionally, the objective functions of the AC / DC distribution network planning model include an economic cost objective function, a reliability objective function, a new energy consumption objective function, a voltage deviation objective function, and a network loss objective function;

[0010] The economic cost objective function is:

[0011] min F1=C DG +C line +C vsc +C rec +C inv

[0012] Among them, F1 is the economic cost of the distribution network, C DG is the investment and construction cost of distributed power generation, C line is the line construction cost, C vsc is the converter investment cost, C rec is the investment cost of the rectifier, C inv The investment cost of the inverter;

[0013]

[0014] Among them, α W , α P and α ESS The investment and construction costs of wind power, photovoltaic and energy storage equipment are converted to annual costs, c W 、c P and c ESS are the investment and construction costs of wind power, photovoltaic and energy storage equipment per unit capacity, and are the capacities of wind power, photovoltaic power and energy storage equipment connected to node i, Ω W ,Ω P and Ω ESS are the collections of wind power, photovoltaic and energy storage nodes to be selected respectively;

[0015]

[0016] Among them, α LINE The cost of line construction converted to annual cost, c LINE is the construction cost per unit length of line, Is line i put into construction? The value is 1, if no investment is made The value is 0, Ω LINE is the set of routes to be selected;

[0017]

[0018] Among them, α VSC The investment and construction cost of the converter is converted into annual cost, cVSC is the investment and construction cost of the unit capacity converter, is the capacity of the converter, Ω VSC is a set of converters to be selected;

[0019]

[0020] Among them, α REC M is the annual cost of rectifier investment and construction. i is the bus type of node i, c REC is the investment and construction cost of the unit capacity rectifier, is the proportion of DC load in node i, P i LOAD is the load size of node i, Ω LOAD is the set of all load nodes;

[0021] C INV =α INV ∑(1-M i )c INV (1-d dc )P i LOAD

[0022] Among them, α INV The investment and construction cost of the inverter is converted into annual cost, c INV is the investment and construction cost of the inverter per unit capacity;

[0023] The reliability objective function is:

[0024] min F2=k1EENS+k2SAIDI+k3(1-ASAI)

[0025] Among them, F2 is the reliability of the distribution network, EENS, SAIDI and ASAI are the system expected power supply, system average power outage duration and system average power supply availability, respectively; k1, k2 and k3 are the target weights of the system expected power supply, system average power outage duration and system average power supply availability, respectively;

[0026] The objective function of new energy consumption is:

[0027]

[0028] Among them, F3 is the new energy consumption rate of the distribution network, is the maximum output of wind power at node i at time t in scenario s, In order to consider the actual output of wind power after the power flow constraint, is the maximum output of the photovoltaic power plant at node i at time t in scenario s, The actual output of photovoltaic power after considering the power flow constraint;

[0029] The voltage deviation objective function is:

[0030]

[0031] Among them, F4 is the voltage deviation of the distribution network, U i,s,t is the voltage of node i at time t in scenario s, U base is the voltage reference value;

[0032] The network loss objective function is:

[0033]

[0034] Among them, F5 is the network loss of the distribution network, is the loss of line k at time t in scenario s, is the loss of the kth converter at time t in scenario s.

[0035] Optionally, the constraints of the AC / DC distribution network planning model include an upper limit constraint on investment capacity, a power flow constraint, a node voltage constraint, and a line transmission capacity constraint;

[0036] The upper limit of investment and construction capacity is constrained as follows:

[0037]

[0038] in, and are the wind power capacity, photovoltaic capacity, energy storage capacity and converter capacity of node i, and are the upper limits of wind power capacity, photovoltaic capacity, energy storage capacity, and converter capacity at node i, respectively;

[0039] The node voltage constraints are:

[0040]

[0041] Among them, U i,s,t is the voltage of node i at time t in scenario s, and are the minimum and maximum voltages allowed at node i, respectively;

[0042] The line transmission capacity constraint is:

[0043]

[0044]

[0045]

[0046]

[0047] in, is the maximum capacity allowed for line k, is the active power of line k at time t in scenario s, is the reactive power of line k at time t in scenario s;

[0048] The power flow constraint is:

[0049]

[0050] in, and are the energy storage charging power and discharging power respectively, is the active power of the i-th node at time t in scenario s, is the reactive power of the i-th node at time t in scenario s, U j,s,t is the voltage of node j at time t in scenario s, Ω ij is the set of lines with the first and last nodes being i and j, Ω ji is the set of lines with the first and last nodes being j and i, is the active power of the line with the first node i and the last node j at time t in scenario s, is the reactive power of the line with the first node i and the last node j at time t in scenario s, X k is the reactance of line k.

[0051] Optionally, the AC / DC distribution network planning model is solved according to a preference multi-objective optimization algorithm based on NSGAIII to obtain an optimal solution that satisfies user preferences, including:

[0052] S1. Randomly generate an initial parent population of 2N individuals and calculate the target value of the initial parent population according to the objective function of the AC / DC distribution network planning model, where N is an arbitrary positive integer.

[0053] S2. Perform non-dominated sorting on all individuals and divide them into several non-dominated layers. Individuals in each layer are mutually non-dominated, and any individual in the previous layer dominates all individuals in the next layer.

[0054] S3. Construct a descendant population with a population size of N. Start selecting individuals from the top non-dominated layer. If the number of individuals in the current layer plus the number of individuals already selected is less than N, randomly select individuals from the next non-dominated layer until the number of individuals in the descendant population is equal to N.

[0055] S4. Calculate the maximum and minimum values ​​of all individuals in the offspring population on all targets, and normalize all individuals according to the maximum and minimum values;

[0056] S5. Generate uniformly distributed reference vectors in the space of target dimension, calculate the angle between the reference vector and the preference vector, sort the angles between all reference vectors and the preference vector, and select the first M reference vectors > 0 as the benchmark for subsequent individual selection;

[0057] S6. Calculate the distance between all individuals in the offspring population and the reference vector, associate each individual with the reference vector closest to the individual, and record the number of individuals associated with each reference vector;

[0058] S7, construct a new population, starting from the reference vector with the least number of associated individuals, select individuals and add them to the new population until the number of individuals in the new population is equal to N;

[0059] S8. Perform crossover and mutation on the new population to generate the next generation population, and determine whether the iteration number requirement is met. If so, recalculate the target values ​​of all individuals in the optimal population to obtain the optimal solution that meets the user's preference. Otherwise, return to step S2.

[0060] Optionally, the formula for calculating the angle between the reference vector and the preference vector is:

[0061]

[0062] Among them, θ is the angle between the reference vector and the preference vector, λ ref is the reference vector, λ love is the preference vector.

[0063] A second aspect of the present invention provides an AC / DC distribution network planning system based on a preference multi-objective algorithm, comprising:

[0064] Parameter acquisition module, used to obtain the distribution network parameters of the area to be optimized;

[0065] The planning model construction module is used to construct an AC / DC distribution network planning model based on the distribution network parameters of the area to be optimized, which takes into account the lowest economic cost, highest reliability, maximum new energy absorption, minimum voltage deviation and minimum network loss;

[0066] The model solving module is used to solve the AC / DC distribution network planning model according to the NSGAIII-based preference multi-objective optimization algorithm to obtain the optimal solution that satisfies user preferences. The NSGAIII-based preference multi-objective optimization algorithm introduces user preferences into the NSGAIII algorithm to limit the size of the target space.

[0067] Optionally, the objective functions of the AC / DC distribution network planning model include an economic cost objective function, a reliability objective function, a new energy consumption objective function, a voltage deviation objective function, and a network loss objective function;

[0068] The economic cost objective function is:

[0069] min F1=C DG +C line +C vsc +C rec +C inv

[0070] Among them, F1 is the economic cost of the distribution network, C DG is the investment and construction cost of distributed power generation, C line is the line construction cost, C vsc is the converter investment cost, C rec is the investment cost of the rectifier, C inv The investment cost of the inverter;

[0071]

[0072] Among them, α W , α P and α ESS The investment and construction costs of wind power, photovoltaic and energy storage equipment are converted to annual costs, c W 、c P and c ESS are the investment and construction costs of wind power, photovoltaic and energy storage equipment per unit capacity, and are the capacities of wind power, photovoltaic power and energy storage equipment connected to node i, Ω W ,Ω P and Ω ESS are the collections of wind power, photovoltaic and energy storage nodes to be selected respectively;

[0073]

[0074] Among them, α LINE The cost of line construction converted to annual cost, c LINE is the construction cost per unit length of line, Is line i put into construction? The value is 1, if no investment is made The value is 0, Ω LINE is the set of routes to be selected;

[0075]

[0076] Among them, α VSC The investment and construction cost of the converter is converted into annual cost, cVSC is the investment and construction cost of the unit capacity converter, is the capacity of the converter, Ω VSC is a set of converters to be selected;

[0077]

[0078] Among them, α REC M is the annual cost of rectifier investment and construction. i is the bus type of node i, c REC is the investment and construction cost of the unit capacity rectifier, is the proportion of DC load in node i, P i LOAD is the load size of node i, Ω LOAD is the set of all load nodes;

[0079] C INV =α INV ∑(1-M i )c INV (1-d dc )P i LOAD

[0080] Among them, α INV The investment and construction cost of the inverter is converted into annual cost, c INV is the investment and construction cost of the inverter per unit capacity;

[0081] The reliability objective function is:

[0082] min F2=k1EENS+k2SAIDI+k3(1-ASAI)

[0083] Among them, F2 is the reliability of the distribution network, EENS, SAIDI and ASAI are the system expected power supply, system average power outage duration and system average power supply availability, respectively; k1, k2 and k3 are the target weights of the system expected power supply, system average power outage duration and system average power supply availability, respectively;

[0084] The objective function of new energy consumption is:

[0085]

[0086] Among them, F3 is the new energy consumption rate of the distribution network, is the maximum output of wind power at node i at time t in scenario s, In order to consider the actual output of wind power after the power flow constraint, is the maximum output of the photovoltaic power plant at node i at time t in scenario s, The actual output of photovoltaic power after considering the power flow constraint;

[0087] The voltage deviation objective function is:

[0088]

[0089] Among them, F4 is the voltage deviation of the distribution network, U i,s,t is the voltage of node i at time t in scenario s, U base is the voltage reference value;

[0090] The network loss objective function is:

[0091]

[0092] Among them, F5 is the network loss of the distribution network, is the loss of line k at time t in scenario s, is the loss of the kth converter at time t in scenario s.

[0093] Optionally, the constraints of the AC / DC distribution network planning model include an upper limit constraint on investment capacity, a power flow constraint, a node voltage constraint, and a line transmission capacity constraint;

[0094] The upper limit of investment and construction capacity is constrained as follows:

[0095]

[0096] in, and are the wind power capacity, photovoltaic capacity, energy storage capacity and converter capacity of node i, and are the upper limits of wind power capacity, photovoltaic capacity, energy storage capacity, and converter capacity at node i, respectively;

[0097] The node voltage constraints are:

[0098]

[0099] Among them, U i,s,t is the voltage of node i at time t in scenario s, and are the minimum and maximum voltages allowed at node i, respectively;

[0100] The line transmission capacity constraint is:

[0101]

[0102]

[0103]

[0104]

[0105] in, is the maximum capacity allowed for line k, is the active power of line k at time t in scenario s, is the reactive power of line k at time t in scenario s;

[0106] The power flow constraint is:

[0107]

[0108] in, and are the energy storage charging power and discharging power respectively, is the active power of the i-th node at time t in scenario s, is the reactive power of the i-th node at time t in scenario s, U j,s,t is the voltage of node j at time t in scenario s, Ω ij is the set of lines with the first and last nodes being i and j, Ω ji is the set of lines with the first and last nodes being j and i, is the active power of the line with the first node i and the last node j at time t in scenario s, is the reactive power of the line with the first node i and the last node j at time t in scenario s, X k is the reactance of line k.

[0109] Optionally, the model solving module is specifically configured to perform the following steps:

[0110] S1. Randomly generate an initial parent population of 2N individuals and calculate the target value of the initial parent population according to the objective function of the AC / DC distribution network planning model, where N is an arbitrary positive integer.

[0111] S2. Perform non-dominated sorting on all individuals and divide them into several non-dominated layers. Individuals in each layer are mutually non-dominated, and any individual in the previous layer dominates all individuals in the next layer.

[0112] S3. Construct a descendant population with a population size of N. Start selecting individuals from the top non-dominated layer. If the number of individuals in the current layer plus the number of individuals already selected is less than N, randomly select individuals from the next non-dominated layer until the number of individuals in the descendant population is equal to N.

[0113] S4. Calculate the maximum and minimum values ​​of all individuals in the offspring population on all targets, and normalize all individuals according to the maximum and minimum values;

[0114] S5. Generate uniformly distributed reference vectors in the space of target dimension, calculate the angle between the reference vector and the preference vector, sort the angles between all reference vectors and the preference vector, and select the first M reference vectors > 0 as the benchmark for subsequent individual selection;

[0115] S6. Calculate the distance between all individuals in the offspring population and the reference vector, associate each individual with the reference vector closest to the individual, and record the number of individuals associated with each reference vector;

[0116] S7, construct a new population, starting from the reference vector with the least number of associated individuals, select individuals and add them to the new population until the number of individuals in the new population is equal to N;

[0117] S8. Perform crossover and mutation on the new population to generate the next generation population, and determine whether the iteration number requirement is met. If so, recalculate the target values ​​of all individuals in the optimal population to obtain the optimal solution that meets the user's preference. Otherwise, return to step S2.

[0118] Optionally, the formula for calculating the angle between the reference vector and the preference vector is:

[0119]

[0120] Among them, θ is the angle between the reference vector and the preference vector, λ ref is the reference vector, λ love is the preference vector.

[0121] From the above technical solutions, it can be seen that the AC / DC distribution network planning method and system based on the preference multi-objective algorithm provided by the present invention have the following advantages:

[0122] The AC / DC distribution network planning method based on the preference multi-objective algorithm provided by the present invention constructs an AC / DC distribution network planning model, which takes into account the lowest economic cost and the highest reliability, as well as the maximum new energy consumption, the minimum voltage deviation and the minimum network loss. It is closer to the AC / DC distribution network planning scenario with high penetration rate of new energy access. In the model solving process, the preference multi-objective optimization algorithm based on NSGAIII is used to obtain the optimal solution that meets user preferences for the AC / DC distribution network planning model, thereby improving the solution speed of the multi-objective planning and solving the technical problems that the existing AC / DC distribution network multi-objective planning model only considers economic cost and reliability, cannot be adapted to the AC / DC distribution network planning with high penetration rate of new energy access, and in the process of solving the multi-objective planning model, the Pareto solution is relatively scattered, the computing power cannot be concentrated on the decision maker's preference solution, and the optimization speed is slow.

[0123] The AC / DC distribution network planning system based on the preference multi-objective algorithm provided by the present invention is used to execute the AC / DC distribution network planning method based on the preference multi-objective algorithm provided by the present invention. Its principles and technical effects are the same as those of the AC / DC distribution network planning method based on the preference multi-objective algorithm provided by the present invention, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0124] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying any creative work.

[0125] Figure 1 A schematic diagram of a flow chart of an AC / DC distribution network planning method based on a preference multi-objective algorithm provided in the present invention;

[0126] Figure 2 A logic block diagram of solving an AC / DC distribution network planning model using the NSGAIII-based preference multi-objective optimization algorithm provided in the present invention;

[0127] Figure 3 This is a structural diagram of an AC / DC distribution network planning system based on a preference multi-objective algorithm provided in the present invention. DETAILED DESCRIPTION

[0128] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0129] For easier understanding, see Figure 1 The present invention provides an embodiment of an AC / DC distribution network planning method based on a preference multi-objective algorithm, comprising:

[0130] Step 101: Obtain the distribution network parameters of the area to be optimized.

[0131] It should be noted that in this embodiment of the present invention, the parameters of the regional distribution network to be optimized are first obtained, including multi-objective optimization parameters and constraint parameters. The multi-objective optimization parameters include the population size and the maximum number of iterations, while the constraint parameters are the parameters of the constraint conditions required in the AC / DC distribution network planning model.

[0132] Step 102: Based on the distribution network parameters of the area to be optimized, an AC / DC distribution network planning model is constructed that takes into account the lowest economic cost, the highest reliability, the maximum absorption of new energy, the smallest voltage deviation, and the smallest network loss.

[0133] It should be noted that after new energy is connected to the distribution network, the new energy absorption characteristics, voltage deviation and network loss of the distribution network will be affected. Therefore, when constructing a multi-objective planning model for the area to be optimized, it is necessary to consider the goals of maximizing new energy absorption, minimizing voltage deviation and minimizing network loss. The AC / DC distribution network planning model includes two parts: objective function and constraint conditions. Therefore, the objective function of the AC / DC distribution network planning model includes economic cost objective function, reliability objective function, new energy absorption objective function, voltage deviation objective function and network loss objective function. Specifically, the economic cost objective function is:

[0134] min F1=C DG +C line +C vsc +C rec +C inv

[0135] Among them, F1 is the economic cost of the distribution network, C DG is the investment and construction cost of distributed power generation, C line is the line construction cost, C vsc is the converter investment cost, C rec is the investment cost of the rectifier, C inv The investment cost of the inverter;

[0136]

[0137] Among them, α W , α P and α ESS The investment and construction costs of wind power, photovoltaic and energy storage equipment are converted to annual costs, c W 、c P and c ESS are the investment and construction costs of wind power, photovoltaic and energy storage equipment per unit capacity, and are the capacities of wind power, photovoltaic power and energy storage equipment connected to node i, Ω W ,Ω P and Ω ESS are the collections of wind power, photovoltaic and energy storage nodes to be selected respectively;

[0138]

[0139] Among them, α LINE The cost of line construction converted to annual cost, c LINEis the construction cost per unit length of line, Is line i put into construction? The value is 1, if no investment is made The value is 0, Ω LINE is the set of routes to be selected;

[0140]

[0141] Among them, α VSC The investment and construction cost of the converter is converted into annual cost, c VSC is the investment and construction cost of the unit capacity converter, is the capacity of the converter, Ω VSC is a set of converters to be selected;

[0142]

[0143] Among them, α REC M is the annual cost of rectifier investment and construction. i is the bus type of node i. If the bus type is AC bus, then M i The value is 1. If the bus type is DC bus, then M i The value is 0, c REC is the investment and construction cost of the unit capacity rectifier, is the proportion of DC load in node i, P i LOAD is the load size of node i, Ω LOAD is the set of all load nodes;

[0144] C INV =α INV ∑(1-M i )c INV (1-d dc )P i LOAD

[0145] Among them, α INV The investment and construction cost of the inverter is converted into annual cost, c INV is the investment and construction cost of the inverter per unit capacity;

[0146] The reliability objective function is:

[0147] min F2=k1EENS+k2SAIDI+k3(1-ASAI)

[0148] Among them, F2 is the reliability of the distribution network, EENS, SAIDI and ASAI are the system expected power supply, system average power outage duration and system average power supply availability, respectively; k1, k2 and k3 are the target weights of the system expected power supply, system average power outage duration and system average power supply availability, respectively;

[0149] The calculation formula for the system average power outage duration SAIDI is:

[0150]

[0151] The calculation formula for the system expected power supply EENS is:

[0152]

[0153] The calculation formula of the system average power supply availability rate ASAI is:

[0154]

[0155] Among them, λ l is the failure rate of line l, CID i is the average annual power outage time of node i, the unit of SAIDI is hours / household*year, NC i is the number of users carried by node i. The unit of EENS is megawatt / year. 8760 means the number of hours in a year (i.e., one year = 8760 hours).

[0156] The objective function of new energy consumption is:

[0157]

[0158] Among them, F3 is the new energy consumption rate of the distribution network, is the maximum output of wind power at node i at time t in scenario s, In order to consider the actual output of wind power after the power flow constraint, is the maximum output of the photovoltaic power plant at node i at time t in scenario s, The actual output of photovoltaic power after considering the power flow constraint;

[0159] The voltage deviation objective function is:

[0160]

[0161] Among them, F4 is the voltage deviation of the distribution network, U i,s,t is the voltage of node i at time t in scenario s, U base is the voltage reference value;

[0162] The network loss objective function is:

[0163]

[0164] Among them, F5 is the network loss of the distribution network, is the loss of line k at time t in scenario s, is the loss of the kth converter at time t in scenario s.

[0165] The constraints of the AC / DC distribution network planning model include the upper limit of investment and construction capacity, power flow constraints, node voltage constraints, and line transmission capacity constraints.

[0166] The upper limit of investment and construction capacity is constrained as follows:

[0167]

[0168] in, and are the wind power capacity, photovoltaic capacity, energy storage capacity and converter capacity of node i, and are the upper limits of wind power capacity, photovoltaic capacity, energy storage capacity, and converter capacity at node i, respectively;

[0169] The node voltage constraints are:

[0170]

[0171] Among them, U i,s,t is the voltage of node i at time t in scenario s, and are the minimum and maximum voltages allowed at node i, respectively;

[0172] The line transmission capacity constraint is:

[0173]

[0174]

[0175]

[0176]

[0177] in, is the maximum capacity allowed for line k, is the active power of line k at time t in scenario s, is the reactive power of line k at time t in scenario s;

[0178] The power flow constraint is:

[0179]

[0180] in, and are the energy storage charging power and discharging power respectively, is the active power of the i-th node at time t in scenario s, is the reactive power of the i-th node at time t in scenario s, U j,s,t is the voltage of node j at time t in scenario s, Ω ij is the set of lines with the first and last nodes being i and j, Ω ji is the set of lines with the first and last nodes being j and i, is the active power of the line with the first node i and the last node j at time t in scenario s, is the reactive power of the line with the first node i and the last node j at time t in scenario s, X k is the reactance of line k.

[0181] Step 103: Solve the AC / DC distribution network planning model according to the NSGAIII-based preference multi-objective optimization algorithm to obtain an optimal solution that satisfies user preferences.

[0182] It should be noted that the preference multi-objective optimization algorithm based on NSGAIII is an algorithm that introduces user preferences into the NSGAIII algorithm to limit the size of the target space. The NSGAIII algorithm is an intelligent algorithm based on non-dominated sorting. In each iteration, it determines the next generation of population based on the reference vector and whether each individual is dominated by other individuals. This algorithm can provide a multi-objective Pareto solution set for a specific problem, and each solution in the solution set is better than all other solutions in a certain goal. However, in a high-dimensional target space, non-dominated sorting is very time-consuming, and searching the entire target space will consume a huge amount of computational examples. In fact, planners have certain requirements for the importance of each target, which is called user preference. If user preferences are introduced into the algorithm and the size of the target space is limited, the solution speed will be greatly improved, and a solution that satisfies the user can be provided. Therefore, in an embodiment of the present invention, user preferences are introduced into the NSGAIII algorithm to limit the size of the target space, so as to improve the speed of obtaining the optimal solution of user preferences. Specifically, the process of solving the AC / DC distribution network planning model according to the preference multi-objective optimization algorithm based on NSGAIII and obtaining the optimal solution that meets user preferences is as follows: Figure 2 As shown, the following steps are included:

[0183] S1. Randomly generate an initial parent population of 2N individuals and calculate the initial parent population T according to the objective function of the AC / DC distribution network planning model. t The target value of , N is any positive integer.

[0184] S2. Perform non-dominated sorting on all individuals and divide them into several non-dominated layers (U1, U2, ... from top to bottom). The individuals in each layer are mutually non-dominated, and any individual in the previous layer dominates all individuals in the next layer.

[0185] S3, construct a descendant population S with a population size of N t , start selecting individuals from the top (U1) non-dominated layer. If the number of individuals in the current layer plus the number of individuals that have been selected is less than N, randomly select from the next non-dominated layer until the offspring population S t The number of individuals in is equal to N.

[0186] S4. Calculate the offspring population S t The maximum value F of all individuals on all targets max and the minimum value F min , and according to the maximum value F max and the minimum value F min Normalize all individuals.

[0187] For the minimum target (such as economic cost, voltage deviation and network loss), the normalization formula is:

[0188]

[0189] Among them, F n is the data after normalization, and F is the data before normalization.

[0190] For the maximum target (such as reliability and new energy consumption rate), the normalization formula is:

[0191]

[0192] S5. Generate uniformly distributed reference vectors in the space of the target dimension, calculate the angle between the reference vector and the preference vector, sort the angles between all reference vectors and the preference vector, and select the first M reference vectors > 0 as the basis for subsequent individual selection.

[0193] The formula for calculating the angle between the reference vector and the preference vector is:

[0194]

[0195] Among them, θ is the angle between the reference vector and the preference vector, λ ref is the reference vector, λ love is the preference vector.

[0196] The closer the value of cosθ is to 1, the closer the reference vector is to the preference vector, and the greater the probability of it being selected.

[0197] S6. Calculate the offspring population S t The distance between all individuals and the reference vector is calculated, each individual is associated with the reference vector closest to the individual, and the number of individuals associated with each reference vector is recorded.

[0198] S7. Construct a new population T t+1 , starting from the reference vector with the least number of associated individuals, select individuals and add them to the new population T t+1 until the number of individuals in the new population is equal to N.

[0199] S8. Perform crossover and mutation on the new population to generate the next generation population, and determine whether the iteration number requirement is met. If so, recalculate the target values ​​of all individuals in the optimal population to obtain the optimal solution that meets the user's preference. Otherwise, return to step S2.

[0200] The AC / DC distribution network planning method based on the preference multi-objective algorithm provided by the present invention constructs an AC / DC distribution network planning model, which takes into account the lowest economic cost and the highest reliability, as well as the maximum new energy consumption, the minimum voltage deviation and the minimum network loss. It is closer to the AC / DC distribution network planning scenario with high penetration rate of new energy access. In the model solving process, the preference multi-objective optimization algorithm based on NSGAIII is used to obtain the optimal solution that meets user preferences for the AC / DC distribution network planning model, thereby improving the solution speed of the multi-objective planning and solving the technical problems that the existing AC / DC distribution network multi-objective planning model only considers economic cost and reliability, cannot be adapted to the AC / DC distribution network planning with high penetration rate of new energy access, and in the process of solving the multi-objective planning model, the Pareto solution is relatively scattered, the computing power cannot be concentrated on the decision maker's preference solution, and the optimization speed is slow.

[0201] For easier understanding, see Figure 3 The present invention provides an embodiment of an AC / DC distribution network planning system based on a preference multi-objective algorithm, comprising:

[0202] Parameter acquisition module, used to obtain the distribution network parameters of the area to be optimized;

[0203] The planning model construction module is used to construct an AC / DC distribution network planning model based on the distribution network parameters of the area to be optimized, which takes into account the lowest economic cost, highest reliability, maximum new energy absorption, minimum voltage deviation and minimum network loss;

[0204] The model solving module is used to solve the AC / DC distribution network planning model according to the NSGAIII-based preference multi-objective optimization algorithm to obtain the optimal solution that satisfies user preferences. The NSGAIII-based preference multi-objective optimization algorithm introduces user preferences into the NSGAIII algorithm to limit the size of the target space.

[0205] The objective functions of the AC / DC distribution network planning model include economic cost objective function, reliability objective function, new energy consumption objective function, voltage deviation objective function and network loss objective function;

[0206] The economic cost objective function is:

[0207] min F1=C DG +C line +C vsc +C rec +C inv

[0208] Among them, F1 is the economic cost of the distribution network, C DG is the investment and construction cost of distributed power generation, C line is the line construction cost, C vsc is the converter investment cost, C rec is the investment cost of the rectifier, C inv The investment cost of the inverter;

[0209]

[0210] Among them, α W , α P and α ESS The investment and construction costs of wind power, photovoltaic and energy storage equipment are converted to annual costs, c W 、c P and c ESS are the investment and construction costs of wind power, photovoltaic and energy storage equipment per unit capacity, and are the capacities of wind power, photovoltaic power and energy storage equipment connected to node i, Ω W ,Ω P and Ω ESS are the collections of wind power, photovoltaic and energy storage nodes to be selected respectively;

[0211]

[0212] Among them, α LINE The cost of line construction converted to annual cost, c LINE is the construction cost per unit length of line, Is line i put into construction? The value is 1, if no investment is made The value is 0, Ω LINE is the set of routes to be selected;

[0213]

[0214] Among them, α VSCThe investment and construction cost of the converter is converted into annual cost, c VSC is the investment and construction cost of the unit capacity converter, is the capacity of the converter, Ω VSC is a set of converters to be selected;

[0215]

[0216] Among them, α REC M is the annual cost of rectifier investment and construction. i is the bus type of node i, c REC is the investment and construction cost of the unit capacity rectifier, is the proportion of DC load in node i, P i LOAD is the load size of node i, Ω LOAD is the set of all load nodes;

[0217] C INV =α INV ∑(1-M i )c INV (1-d dc )P i LOAD

[0218] Among them, α INV The investment and construction cost of the inverter is converted into annual cost, c INV is the investment and construction cost of the inverter per unit capacity;

[0219] The reliability objective function is:

[0220] min F2=k1EENS+k2SAIDI+k3(1-ASAI)

[0221] Among them, F2 is the reliability of the distribution network, EENS, SAIDI and ASAI are the system expected power supply, system average power outage duration and system average power supply availability, respectively; k1, k2 and k3 are the target weights of the system expected power supply, system average power outage duration and system average power supply availability, respectively;

[0222] The objective function of new energy consumption is:

[0223]

[0224] Among them, F3 is the new energy consumption rate of the distribution network, is the maximum output of wind power at node i at time t in scenario s, In order to consider the actual output of wind power after the power flow constraint, is the maximum output of the photovoltaic power plant at node i at time t in scenario s, The actual output of photovoltaic power after considering the power flow constraint;

[0225] The voltage deviation objective function is:

[0226]

[0227] Among them, F4 is the voltage deviation of the distribution network, U i,s,t is the voltage of node i at time t in scenario s, U base is the voltage reference value;

[0228] The network loss objective function is:

[0229]

[0230] Among them, F5 is the network loss of the distribution network, is the loss of line k at time t in scenario s, is the loss of the kth converter at time t in scenario s.

[0231] The constraints of the AC / DC distribution network planning model include the upper limit of investment and construction capacity, power flow constraints, node voltage constraints, and line transmission capacity constraints.

[0232] The upper limit of investment and construction capacity is constrained as follows:

[0233]

[0234] in, and are the wind power capacity, photovoltaic capacity, energy storage capacity and converter capacity of node i, and are the upper limits of wind power capacity, photovoltaic capacity, energy storage capacity, and converter capacity at node i, respectively;

[0235] The node voltage constraints are:

[0236]

[0237] Among them, U i,s,t is the voltage of node i at time t in scenario s, and are the minimum and maximum voltages allowed at node i, respectively;

[0238] The line transmission capacity constraint is:

[0239]

[0240]

[0241]

[0242]

[0243] in, is the maximum capacity allowed for line k, is the active power of line k at time t in scenario s, is the reactive power of line k at time t in scenario s;

[0244] The power flow constraint is:

[0245]

[0246] in, and are the energy storage charging power and discharging power respectively, is the active power of the i-th node at time t in scenario s, is the reactive power of the i-th node at time t in scenario s, U j,s,t is the voltage of node j at time t in scenario s, Ω ij is the set of lines with the first and last nodes being i and j, Ω ji is the set of lines with the first and last nodes being j and i, is the active power of the line with the first node i and the last node j at time t in scenario s, is the reactive power of the line with the first node i and the last node j at time t in scenario s, X k is the reactance of line k.

[0247] The model solving module is specifically used to perform the following steps:

[0248] S1. Randomly generate an initial parent population of 2N individuals and calculate the target value of the initial parent population according to the objective function of the AC / DC distribution network planning model, where N is an arbitrary positive integer.

[0249] S2. Perform non-dominated sorting on all individuals and divide them into several non-dominated layers. Individuals in each layer are mutually non-dominated, and any individual in the previous layer dominates all individuals in the next layer.

[0250] S3. Construct a descendant population with a population size of N. Start selecting individuals from the top non-dominated layer. If the number of individuals in the current layer plus the number of individuals already selected is less than N, randomly select individuals from the next non-dominated layer until the number of individuals in the descendant population is equal to N.

[0251] S4. Calculate the maximum and minimum values ​​of all individuals in the offspring population on all targets, and normalize all individuals according to the maximum and minimum values;

[0252] S5. Generate uniformly distributed reference vectors in the space of target dimension, calculate the angle between the reference vector and the preference vector, sort the angles between all reference vectors and the preference vector, and select the first M reference vectors > 0 as the benchmark for subsequent individual selection;

[0253] S6. Calculate the distance between all individuals in the offspring population and the reference vector, associate each individual with the reference vector closest to the individual, and record the number of individuals associated with each reference vector;

[0254] S7, construct a new population, starting from the reference vector with the least number of associated individuals, select individuals and add them to the new population until the number of individuals in the new population is equal to N;

[0255] S8. Perform crossover and mutation on the new population to generate the next generation population, and determine whether the iteration number requirement is met. If so, recalculate the target values ​​of all individuals in the optimal population to obtain the optimal solution that meets the user's preference. Otherwise, return to step S2.

[0256] The formula for calculating the angle between the reference vector and the preference vector is:

[0257]

[0258] Among them, θ is the angle between the reference vector and the preference vector, λ ref is the reference vector, λ love is the preference vector.

[0259] The AC / DC distribution network planning system based on the preference multi-objective algorithm provided by the present invention is used to execute the AC / DC distribution network planning method based on the preference multi-objective algorithm provided by the present invention. Its principle and technical effects are the same as those of the AC / DC distribution network planning method based on the preference multi-objective algorithm provided by the present invention, and will not be repeated here.

[0260] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for AC / DC distribution network planning based on a preference multi-objective algorithm, characterized in that: include: Obtain the distribution network parameters of the area to be optimized; Based on the distribution network parameters of the area to be optimized, an AC / DC distribution network planning model is constructed that takes into account the lowest economic cost, highest reliability, maximum new energy absorption, minimum voltage deviation, and minimum network loss. The AC / DC distribution network planning model is solved using a preference multi-objective optimization algorithm based on NSGAIII to obtain the optimal solution that satisfies user preferences. The preference multi-objective optimization algorithm based on NSGAIII introduces user preferences into the NSGAIII algorithm to limit the size of the objective space. The objective functions of the AC / DC distribution network planning model include economic cost objective function, reliability objective function, new energy consumption objective function, voltage deviation objective function and network loss objective function; The economic cost objective function is: minF1=C DG +C LINE +C VSC +C REC +C INV Among them, F1 is the economic cost of the distribution network, C DG is the investment and construction cost of distributed power generation, C LINE is the line construction cost, C VSC is the converter investment cost, C REC is the investment cost of the rectifier, C INV The investment cost of the inverter; Among them, α W , α P and α ESS The investment and construction costs of wind power, photovoltaic and energy storage equipment are converted to annual costs, c W 、c P and c ESS are the investment and construction costs of wind power, photovoltaic and energy storage equipment per unit capacity, and are the capacities of wind power, photovoltaic power and energy storage equipment connected to node i, Ω W ,Ω P and Ω ESS are the collections of wind power, photovoltaic and energy storage nodes to be selected respectively; Among them, α LINE The cost of line construction converted to annual cost, c LINE is the construction cost per unit length of line, Is line i put into construction? The value is 1, if no investment is made The value is 0, Ω LINE is the set of routes to be selected; Among them, α VSC The investment and construction cost of the converter is converted into annual cost, c VSC is the investment and construction cost of the unit capacity converter, is the capacity of the converter, Ω VSC is a set of converters to be selected; Among them, α REC M is the annual cost of rectifier investment and construction. i is the bus type of node i, c REC is the investment and construction cost of the unit capacity rectifier, is the proportion of DC load in node i, is the load size of node i, Ω LOAD is the set of all load nodes; Among them, α INV The investment and construction cost of the inverter is converted into annual cost, c INV is the investment and construction cost of the inverter per unit capacity; The reliability objective function is: minF2=k1EENS+k2SAIDI+k3(1-ASAI) Among them, F2 is the reliability of the distribution network, EENS, SAIDI and ASAI are the system expected power supply, system average power outage duration and system average power supply availability, respectively; k1, k2 and k3 are the target weights of the system expected power supply, system average power outage duration and system average power supply availability, respectively; The objective function of new energy consumption is: Among them, F3 is the new energy consumption rate of the distribution network, is the maximum output of wind power at node i at time t in scenario s, In order to consider the actual output of wind power after the power flow constraint, is the maximum output of the photovoltaic power plant at node i at time t in scenario s, The actual output of photovoltaic power after considering the power flow constraint; The voltage deviation objective function is: Among them, F4 is the voltage deviation of the distribution network, U i,s,t is the voltage of node i at time t in scenario s, U base is the voltage reference value; The network loss objective function is: Among them, F5 is the network loss of the distribution network, is the loss of line k at time t in scenario s, is the loss of the kth converter at time t in scenario s.

2. The AC / DC distribution network planning method based on the preference multi-objective algorithm according to claim 1 is characterized in that: The constraints of the AC / DC distribution network planning model include the upper limit of investment and construction capacity, power flow constraints, node voltage constraints, and line transmission capacity constraints. The upper limit of investment and construction capacity is constrained as follows: in, and are the wind power capacity, photovoltaic capacity, energy storage capacity and converter capacity of node i, and are the upper limits of wind power capacity, photovoltaic capacity, energy storage capacity, and converter capacity at node i, respectively; The node voltage constraints are: Among them, U i,s,t is the voltage of node i at time t in scenario s, and are the minimum and maximum voltages allowed at node i, respectively; The line transmission capacity constraint is: in, is the maximum capacity allowed for line k, is the active power of line k at time t in scenario s, is the reactive power of line k at time t in scenario s; The power flow constraint is: in, and are the energy storage charging power and discharging power respectively, is the active power of the i-th node at time t in scenario s, is the reactive power of the i-th node at time t in scenario s, U j,s,t is the voltage of node j at time t in scenario s, Ω ij is the set of lines with the first and last nodes being i and j, Ω ji is the set of lines with the first and last nodes being j and i, is the active power of the line with the first node i and the last node j at time t in scenario s, is the reactive power of the line with the first node i and the last node j at time t in scenario s, X k is the reactance of line k.

3. The AC / DC distribution network planning method based on the preference multi-objective algorithm according to claim 2 is characterized in that: The AC / DC distribution network planning model is solved using the NSGAIII-based preference multi-objective optimization algorithm to obtain the optimal solution that satisfies user preferences, including: S1. Randomly generate an initial parent population of 2N individuals and calculate the target value of the initial parent population according to the objective function of the AC / DC distribution network planning model, where N is an arbitrary positive integer. S2. Perform non-dominated sorting on all individuals and divide them into several non-dominated layers. Individuals in each layer are mutually non-dominated, and any individual in the previous layer dominates all individuals in the next layer. S3. Construct a descendant population with a population size of N. Start selecting individuals from the top non-dominated layer. If the number of individuals in the current layer plus the number of individuals already selected is less than N, randomly select individuals from the next non-dominated layer until the number of individuals in the descendant population is equal to N. S4. Calculate the maximum and minimum values ​​of all individuals in the offspring population on all targets, and normalize all individuals according to the maximum and minimum values; S5. Generate uniformly distributed reference vectors in the space of target dimension, calculate the angle between the reference vector and the preference vector, sort the angles between all reference vectors and the preference vector, and select the first M reference vectors > 0 as the benchmark for subsequent individual selection; S6. Calculate the distance between all individuals in the offspring population and the reference vector, associate each individual with the reference vector closest to the individual, and record the number of individuals associated with each reference vector; S7, construct a new population, starting from the reference vector with the least number of associated individuals, select individuals and add them to the new population until the number of individuals in the new population is equal to N; S8. Perform crossover and mutation on the new population to generate the next generation population, and determine whether the iteration number requirement is met. If so, recalculate the target values ​​of all individuals in the optimal population to obtain the optimal solution that meets the user's preference. Otherwise, return to step S2.

4. The AC / DC distribution network planning method based on the preference multi-objective algorithm according to claim 3 is characterized in that: The formula for calculating the angle between the reference vector and the preference vector is: Among them, θ is the angle between the reference vector and the preference vector, λ ref is the reference vector, λ love is the preference vector.

5. An AC / DC distribution network planning system based on a preference multi-objective algorithm, characterized in that: include: Parameter acquisition module, used to obtain the distribution network parameters of the area to be optimized; The planning model construction module is used to construct an AC / DC distribution network planning model based on the distribution network parameters of the area to be optimized, which takes into account the lowest economic cost, highest reliability, maximum new energy absorption, minimum voltage deviation and minimum network loss; A model solving module is used to solve the AC / DC distribution network planning model based on the NSGAIII-based preference multi-objective optimization algorithm to obtain the optimal solution that satisfies user preferences. The NSGAIII-based preference multi-objective optimization algorithm introduces user preferences into the NSGAIII algorithm to limit the size of the objective space. The objective functions of the AC / DC distribution network planning model include economic cost objective function, reliability objective function, new energy consumption objective function, voltage deviation objective function and network loss objective function; minF1=C DG +C LINE +C VSC +C REC +C INV Among them, F1 is the economic cost of the distribution network, C DG is the investment and construction cost of distributed power generation, C LINE is the line construction cost, C VSC is the converter investment cost, C REC is the investment cost of the rectifier, C INV The investment cost of the inverter; Among them, α W , α P and α ESS The investment and construction costs of wind power, photovoltaic and energy storage equipment are converted to annual costs, c W 、c P and c ESS are the investment and construction costs of wind power, photovoltaic and energy storage equipment per unit capacity, and are the capacities of wind power, photovoltaic power and energy storage equipment connected to node i, Ω W ,Ω P and Ω ESS are the collections of wind power, photovoltaic and energy storage nodes to be selected respectively; Among them, α LINE The cost of line construction converted to annual cost, c LINE is the construction cost per unit length of line, Is line i put into construction? The value is 1, if no investment is made The value is 0, Ω LINE is the set of routes to be selected; Among them, α VSC The investment and construction cost of the converter is converted into annual cost, c VSC is the investment and construction cost of the unit capacity converter, is the capacity of the converter, Ω VSC is a set of converters to be selected; Among them, α REC M is the annual cost of rectifier investment and construction. i is the bus type of node i, c REC is the investment and construction cost of the unit capacity rectifier, is the proportion of DC load in node i, is the load size of node i, Ω LOAD is the set of all load nodes; Among them, α INV The investment and construction cost of the inverter is converted into annual cost, c INV is the investment and construction cost of the inverter per unit capacity; The reliability objective function is: minF2=k1EENS+k2SAIDI+k3(1-ASAI) Among them, F2 is the reliability of the distribution network, EENS, SAIDI and ASAI are the system expected power supply, system average power outage duration and system average power supply availability, respectively; k1, k2 and k3 are the target weights of the system expected power supply, system average power outage duration and system average power supply availability, respectively; The objective function of new energy consumption is: Among them, F3 is the new energy consumption rate of the distribution network, is the maximum output of wind power at node i at time t in scenario s, In order to consider the actual output of wind power after the power flow constraint, is the maximum output of the photovoltaic power plant at node i at time t in scenario s, The actual output of photovoltaic power after considering the power flow constraint; The voltage deviation objective function is: Among them, F4 is the voltage deviation of the distribution network, U i,s,t is the voltage of node i at time t in scenario s, U base is the voltage reference value; The network loss objective function is: Among them, F5 is the network loss of the distribution network, is the loss of line k at time t in scenario s, is the loss of the kth converter at time t in scenario s.

6. The AC / DC distribution network planning system based on the preference multi-objective algorithm according to claim 5, characterized in that: The constraints of the AC / DC distribution network planning model include the upper limit of investment and construction capacity, power flow constraints, node voltage constraints, and line transmission capacity constraints. The upper limit of investment and construction capacity is constrained as follows: in, and are the wind power capacity, photovoltaic capacity, energy storage capacity and converter capacity of node i, and are the upper limits of wind power capacity, photovoltaic capacity, energy storage capacity, and converter capacity at node i, respectively; The node voltage constraints are: Among them, U i,s,t is the voltage of node i at time t in scenario s, and are the minimum and maximum voltages allowed at node i, respectively; The line transmission capacity constraint is: in, is the maximum capacity allowed for line k, is the active power of line k at time t in scenario s, is the reactive power of line k at time t in scenario s; The power flow constraint is: in, and are the energy storage charging power and discharging power respectively, is the active power of the i-th node at time t in scenario s, is the reactive power of the i-th node at time t in scenario s, U j,s,t is the voltage of node j at time t in scenario s, Ω ij is the set of lines with the first and last nodes being i and j, Ω ji is the set of lines with the first and last nodes being j and i, is the active power of the line with the first node i and the last node j at time t in scenario s, is the reactive power of the line with the first and last nodes i and j at time t in scenario s, X k is the reactance of line k.

7. The AC / DC distribution network planning system based on the preference multi-objective algorithm according to claim 6, characterized in that: The model solving module is specifically used to perform the following steps: S1. Randomly generate an initial parent population of 2N individuals and calculate the target value of the initial parent population according to the objective function of the AC / DC distribution network planning model, where N is an arbitrary positive integer. S2. Perform non-dominated sorting on all individuals and divide them into several non-dominated layers. Individuals in each layer are mutually non-dominated, and any individual in the previous layer dominates all individuals in the next layer. S3. Construct a descendant population with a population size of N. Start selecting individuals from the top non-dominated layer. If the number of individuals in the current layer plus the number of individuals already selected is less than N, randomly select individuals from the next non-dominated layer until the number of individuals in the descendant population is equal to N. S4. Calculate the maximum and minimum values ​​of all individuals in the offspring population on all targets, and normalize all individuals according to the maximum and minimum values; S5. Generate uniformly distributed reference vectors in the space of target dimension, calculate the angle between the reference vector and the preference vector, sort the angles between all reference vectors and the preference vector, and select the first M reference vectors > 0 as the benchmark for subsequent individual selection; S6. Calculate the distance between all individuals in the offspring population and the reference vector, associate each individual with the reference vector closest to the individual, and record the number of individuals associated with each reference vector; S7, construct a new population, starting from the reference vector with the least number of associated individuals, select individuals and add them to the new population until the number of individuals in the new population is equal to N; S8. Perform crossover and mutation on the new population to generate the next generation population, and determine whether the iteration number requirement is met. If so, recalculate the target values ​​of all individuals in the optimal population to obtain the optimal solution that meets the user's preference. Otherwise, return to step S2.

8. The AC / DC distribution network planning system based on the preference multi-objective algorithm according to claim 7, characterized in that: The formula for calculating the angle between the reference vector and the preference vector is: Among them, θ is the angle between the reference vector and the preference vector, λ ref is the reference vector, λ love is the preference vector.

Images