A Lightning Arrester Dynamic Optimization Algorithm and Device Based on Improved Genetic Algorithm

Abstract

The present invention provides a lightning arrester dynamic optimization algorithm and device based on an improved genetic algorithm, including initializing a population, calculating fitness, selection operation, crossover operation, mutation operation, recalculating fitness, judging termination conditions, and outputting the optimal solution; a lightning arrester dynamic optimization method based on an improved genetic algorithm: using the improved genetic algorithm to optimize the resistance, inductance, and capacitance parameters of the lightning arrester to improve the dynamic response and residual voltage control of the lightning arrester under different lightning current waveforms, so as to achieve the best overvoltage protection effect.

Description

Technical Field

[0001] The present invention relates to the field of power system protection, and particularly to a lightning arrester dynamic optimization algorithm and device based on an improved genetic algorithm. Background Art

[0002] Metal oxide lightning arresters, as key protection devices in power systems, are widely used in lightning overvoltage protection. The lightning arrester limits the overvoltage caused by lightning current or switching current to ensure the normal operation of power equipment. With the expansion of the scale of power systems and the diversification of grid equipment, the protection performance and response characteristics of lightning arresters become increasingly important. The dynamic model of the lightning arrester is used to describe its response characteristics under different current waveforms, which is crucial for evaluating and optimizing the performance of the lightning arrester.

[0003] Currently, commonly used dynamic models of lightning arresters include the IEEE model, the Pinceti model, and the Fernandez model. These models simulate the dynamic characteristics of lightning arresters through the combination of components such as non-linear resistors, capacitors, and inductors. These traditional models can simulate the behavior of lightning arresters under lightning current to a certain extent, but have the following deficiencies:

[0004] Model complexity and accuracy issues: When existing models simulate complex lightning waveforms (such as 1 / 10 μs steep-front lightning current, 8 / 20 μs lightning current, 30 / 60 μs switching current, etc.), they usually face high complexity and low simulation accuracy, resulting in the inability to meet high-precision requirements in practical applications.

[0005] Single-objective optimization: Traditional optimization methods (such as particle swarm optimization, genetic algorithm) usually use a single objective function for optimization, and cannot simultaneously take into account the optimization requirements under multiple test conditions, resulting in large errors under different current waveform conditions.

[0006] Low computational efficiency: When existing algorithms solve multi-objective optimization problems, they often require multiple iterations, with a slow convergence speed and high computational costs, and cannot achieve efficient real-time optimization.

[0007] Therefore, there is an urgent need for a new optimization method that can simultaneously optimize the residual voltage error under multiple current waveform conditions and improve the accuracy and computational efficiency of the model. Especially when facing complex current waveforms and multiple test conditions, the method for constructing a lightning arrester dynamic optimization model based on an improved genetic algorithm proposed by the present invention can effectively solve the problems of insufficient accuracy, single optimization objective, and low computational efficiency in existing methods. Summary of the Invention

[0008] The object of the present invention is to provide an arrester dynamic optimization algorithm and device based on an improved genetic algorithm in view of the deficiencies of the above-mentioned existing technologies, aiming to solve the problems of insufficient accuracy, single optimization objective, and low calculation efficiency in the existing arrester dynamic model optimization methods.

[0009] To achieve the above object, the present invention adopts the following technical solutions:

[0010] The present invention provides an algorithm for an arrester dynamic optimization model based on an improved genetic algorithm, including the following steps:

[0011] S1. Initialize the population: Set the parameters of the arrester dynamic optimization model as the initial population;

[0012] S2. Fitness calculation: Calculate the fitness value according to multiple objective functions for the parameter combinations of the arrester dynamic optimization model;

[0013] S3. Selection operation: Based on the fitness value, select the individuals with high fitness from the current population as the parents, and the parent individuals will be used to generate the next-generation individuals;

[0014] S4. Crossover operation: Through the crossover operation, combine the genes of the parent individuals to generate new offspring individuals;

[0015] S5. Mutation operation: Increase the diversity of the population to prevent the algorithm from falling into a local optimal solution;

[0016] S6. Recalculate fitness: After the crossover operation and the mutation operation, the newly generated individuals will continue to perform fitness calculation to ensure that the error of the new individuals is minimized under different current waveforms, so as to improve the protection performance of the arrester;

[0017] S7. Termination condition judgment: In each generation of optimization process, the algorithm will judge whether the preset termination condition is met;

[0018] S8. Output the optimal solution: When the termination condition is met, output the optimal parameters of the arrester dynamic optimization model.

[0019] Further, the specific content of S1 is: Set the population size and the initial value range of the parameters, and perform normal distribution initialization. Use the normal distribution to generate electrical parameters to ensure that the parameters have reasonable mean and variance;

[0020] Assume the resistance conforms to the normal distribution with a mean and a standard deviation :

[0021] ;

[0022] Among them, The th resistance value generated for initialization; the resistance R ∈ [ R min ， R max ] ; is the normal distribution random number generation function;

[0023] For the inductor , use the uniform distribution for initialization:

[0024] ;

[0025] where is the th inductor value generated for initialization; the inductor L ∈ [ L max , L min ] ; is the uniform distribution random number generation function;

[0026] The final initialization process is:

[0027] ;

[0028] where is the th capacitance value generated for initialization; the capacitance C ∈ [ C max , C min ] ; is the normal distribution random number generation function for the capacitance .

[0029] Furthermore, the specific S2 is:

[0030] S201. The reference voltage is:

[0031] ;

[0032] where is the standard lightning current waveform; is the non-linear resistor, whose value depends on the standard lightning current waveform ; is the function of the standard lightning current waveform ; is the time parameter;

[0033] The simulated voltage is:

[0034] ;

[0035] where is the input lightning strike current; is the non-linear resistor, whose value depends on the input lightning strike current ;

[0036] S202. Suppose there are multiple different lightning current waveforms, and the error calculation formula for each waveform is:

[0037] ;

[0038] where, is the number of individuals of the arrester model parameters; is the weight of the th lightning current waveform; and are the simulated voltage and the reference voltage under the th lightning current waveform respectively;

[0039] S203. The fitness function is:

[0040] .

[0041] Furthermore, the specific content of S3 is:

[0042] Suppose there are individuals of the arrester model parameters. The fitness of individual is , and the total fitness is , where, , then the selection probability of individual is:

[0043] ;

[0044] Calculate the cumulative fitness of the individuals and make a selection according to the cumulative probability. The cumulative fitness is:

[0045] ;

[0046] where, is the cumulative fitness value of individual ; When making a selection, generate a random number , and select the individual that satisfies .

[0047] Furthermore, the specific content of S5 is: The normal distribution makes the variation amount fluctuate around the zero value, and the random perturbation conforms to the Gaussian distribution:

[0048] ;

[0049] where, is the normal distribution, the mean value is , and the standard deviation is ; When the mean value Set to 0, indicating that the mutation is perturbed without offset; standard deviation Control the randomness and amplitude of the mutation amount;

[0050] The resistance value after mutation is:

[0051] ;

[0052] Among them, Is the scaling factor, used to adjust the mutation amplitude; Is the new resistance value generated after mutation; Is the original resistance value before mutation; Is the resistance difference before and after mutation.

[0053] Furthermore, the specific content of S6 is: Calculate the fitness of the new individuals after crossover and mutation, evaluate the residual voltage error under different current waveforms, use the fitness function calculation formula in S2 to re-evaluate the fitness values of the new individuals after crossover and mutation, update the population according to the new fitness values, and prepare to enter the selection operation of the next generation.

[0054] Furthermore, the specific content of S7 is:

[0055] Assume that the current iteration number is , and the maximum iteration number is , then the termination condition is:

[0056] ;

[0057] If the current iteration number reaches or exceeds the maximum iteration number , then return yes , and the algorithm stops; if the current iteration number does not reach or exceed the maximum iteration number , then return no , and continue the iteration;

[0058] Set a target fitness threshold , when the fitness of the optimal individual in the population reaches the threshold, stop the algorithm:

[0059] ;

[0060] If the fitness of the optimal individual reaches or exceeds the threshold, then return yes , and the algorithm stops; if the fitness of the optimal individual does not reach or exceed the threshold, then return no , and continue the iteration.

[0061] A lightning arrester dynamic optimization device based on an improved genetic algorithm is used to implement a lightning arrester dynamic optimization algorithm based on an improved genetic algorithm. It also includes a main circuit unit and a control circuit unit.

[0062] The main circuit unit is used to simulate the lightning current waveform, excite the lightning arrester model through the waveform, and evaluate the dynamic response and residual voltage, including

[0063] A fast pulse current transmitter: generates a simulated lightning current waveform;

[0064] A lightning arrester: absorbs the overvoltage brought by the lightning current to protect electrical equipment;

[0065] A voltage measurement unit: measures the residual voltage at both ends of the lightning arrester, records the voltage waveform, and transmits it to the control unit for analysis;

[0066] The control circuit unit: is used for the control and data processing of the entire system, ensures the normal operation of the main circuit unit, and performs optimization calculations, including:

[0067] A decision variable resistance, inductance, and capacitance data module extracts basic electrical parameters from the waveform-excited lightning arrester model as decision variables for the genetic algorithm for adjustment by the optimization algorithm;

[0068] An improved genetic algorithm module: uses an improved genetic algorithm to optimize the parameters of the lightning arrester, and through multiple iterations, finds the parameter combination that makes the lightning arrester perform best under different lightning current waveforms;

[0069] A simulated voltage and fitness function value module: is responsible for simulating the voltage response of the lightning arrester under different current waveforms and calculating the fitness function value based on the simulation results;

[0070] A termination condition judgment module: judges whether the termination condition of the optimization is met;

[0071] An optimal solution output module: outputs the optimized parameters of the waveform-excited lightning arrester model. The ultimate goal is to find a parameter combination of the lightning arrester that shows the best performance under different lightning current waveforms.

[0072] The beneficial effects of the present invention are as follows: A lightning arrester dynamic optimization method based on an improved genetic algorithm: uses an improved genetic algorithm to optimize the resistance R, inductance L, and capacitance C parameters of the lightning arrester to improve the dynamic response and residual voltage control of the lightning arrester under different lightning current waveforms, thereby achieving the best overvoltage protection effect;

[0073] Consider the influence of multiple lightning current waveforms: simultaneously consider the influence of three lightning current waveforms, namely, 1 / 10 μs steep wave lightning current, 8 / 20 μs lightning current, and 30 / 60 μs switching current, on the arrester to ensure that the arrester can provide effective protection under various lightning current conditions;

[0074] Introduce the analysis of residual voltage pulses in fitness calculation: evaluate the performance of the arrester under different working conditions by calculating the residual voltage of the arrester under lightning current waveforms with different front times and comparing it with the reference voltage, and provide a more accurate fitness evaluation criterion;

[0075] Closed-loop feedback mechanism in the optimization process: through the parameter update and feedback unit, feedback the optimal solution obtained by optimization into the arrester model, update the electrical parameters in real time, and conduct simulation verification to ensure that the solution after each optimization can improve the performance of the arrester;

[0076] Improve the performance of the arrester and the reliability of the system: by continuously optimizing the parameters of the arrester, the present invention improves the response speed and stability of the arrester under various lightning current waveforms, reduces the damage of overvoltage to power equipment, and enhances the robustness and safety of the system. Description of the Drawings

[0077] Figure 1 is a flowchart of an arrester dynamic optimization algorithm based on an improved genetic algorithm;

[0078] Figure 2 is a comparison chart of the reference residual voltage and the simulation residual voltage. Detailed Embodiment

[0079] In order to make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0080] An algorithm for an arrester dynamic optimization model based on an improved genetic algorithm includes the following steps:

[0081] S1. Initialize the population: Set the parameters of the arrester dynamic optimization model as the initial population;

[0082] S2. Calculate the fitness: Calculate the fitness value according to multiple objective functions for the parameter combinations of the arrester dynamic optimization model;

[0083] During the fitness calculation process, consider the influence of lightning currents with different front times on the dynamic response of the arrester and conduct the analysis of residual voltage pulses;

[0084] Specifically, the lightning current waveforms of 1 / 10 μs steep wave lightning current, 8 / 20 μs lightning current, and 30 / 60 μs switching current are simulated to excite the dynamic optimization model of the arrester, and the residual voltage under each waveform is calculated; by calculating the error between the residual voltage and the reference voltage, the performance of the arrester under different working conditions is evaluated, and finally, the parameters of the arrester are adjusted through an optimization algorithm to achieve the best protection effect under various current waveforms.

[0085] S3. Selection operation: Based on the fitness value, select individuals with high fitness from the current population as the parent generation, and the parent generation individuals will be used to generate the next generation of individuals.

[0086] S4. Crossover operation: Through the crossover operation, the genes of the parent generation individuals are combined to generate new offspring individuals.

[0087] S5. Mutation operation: Increase the diversity of the population to prevent the algorithm from falling into a local optimal solution.

[0088] S6. Fitness recalculation: After the crossover operation and the mutation operation, the newly generated individuals will continue to perform fitness calculation to ensure that the error of the new individuals is minimized under different current waveforms, so as to improve the protection performance of the arrester.

[0089] S7. Termination condition judgment: In each generation of optimization process, the algorithm will judge whether the preset termination condition is met.

[0090] S8. Output the optimal solution: When the termination condition is met, output the optimal parameters of the dynamic optimization model of the arrester.

[0091] Specifically, S1 is as follows: Set the initial value range of the population size and parameters, initialize with a normal distribution, and use the normal distribution to generate electrical parameters to ensure that the parameters have reasonable means and variances.

[0092] Assume the resistance complies with a normal distribution with a mean and a standard deviation :

[0093] ;

[0094] Among them, is the th resistance value generated by initialization; the resistance R ∈ [ R min ， R max ] ; is the normal distribution random number generation function;

[0095] For the inductance , use a uniform distribution for initialization:

[0096] ;

[0097] Among them, is the th inductance value generated by initialization; the inductance L ∈ [ L max , L min ] ; is a uniformly distributed random number generation function;

[0098] The final initialization process is as follows:

[0099] ;

[0100] Among them, is the th capacitance value generated by initialization; the capacitance C ∈ [ C max , C min ] ; is the normal distribution random number generation function of the capacitance.

[0101] Specifically, S2 is as follows:

[0102] S201. The reference voltage is:

[0103] ;

[0104] Among them, is the standard lightning current waveform; is a non-linear resistor, and its value depends on the standard lightning current waveform ; is the standard lightning current waveform function; is the time parameter;

[0105] The simulated voltage is:

[0106] ;

[0107] Among them, is the input lightning strike current; is a non-linear resistor, and its value depends on the input lightning strike current ;

[0108] Specifically, the simulated voltage is obtained through measurement. In the case of multiple lightning current waveforms, the error calculation does not only consider a single waveform, but needs to comprehensively consider the errors of all waveforms. Therefore, weighted errors are introduced to calculate the errors under each lightning current waveform respectively, and weighted according to the importance of each waveform.

[0109] S202. Assume there are multiple different lightning current waveforms, and the error calculation formula for each waveform is:

[0110] ;

[0111] Among them, is the number of individuals of the arrester model parameters; is the weight of the th lightning current waveform; and are the simulated voltage and the reference voltage under the th lightning current waveform, respectively;

[0112] S203, fitness function is:

[0113] .

[0114] The specific S3 is:

[0115] Suppose there are individuals of the arrester model parameters, and the fitness of individual is , and the total fitness is . Among them, , then the probability that individual is selected is:

[0116] ;

[0117] Calculate the cumulative fitness of the individuals and make selections according to the cumulative probability. The cumulative fitness is:

[0118] ;

[0119] Among them, the cumulative fitness value of individual ; when making selections, generate a random number , and select the individual that satisfies .

[0120] Specifically, the mutation operation is to increase the diversity of the population, prevent the algorithm from falling into a local optimal solution, and produce new solutions by randomly changing some gene values of the individuals.

[0121] For each individual, perform mutation with a certain mutation probability, and the mutation operation randomly changes some parameter values of the individual; for example, in order to increase the complexity of the mutation, assume that the mutation amount follows a normal distribution, which will increase the randomness of the mutation and make the change range of the mutation more flexible.

[0122] The specific S5 is: The normal distribution makes the mutation amount fluctuate around zero, conforming to the random perturbation of the Gaussian distribution:

[0123] ;

[0124] Among them, is a normal distribution with a mean of and a standard deviation of ; when the mean is set to 0, it means that the variation is perturbed without deviation; the standard deviation controls the randomness and amplitude of the variation amount;

[0125] The resistance value after variation is:

[0126] ;

[0127] Among them, is a scaling factor used to adjust the variation amplitude; is the newly generated resistance value after variation; is the original resistance value before variation; is the resistance difference before and after variation.

[0128] Specifically, S6 is: Calculate the fitness of the new individuals after crossover and mutation, evaluate the residual voltage error under different current waveforms, use the fitness function calculation formula in S2 to re-evaluate the fitness values of the new individuals after crossover and mutation, update the population according to the new fitness values, and prepare to enter the selection operation of the next generation.

[0129] Specifically, S7 is:

[0130] Assume that the current iteration number is and the maximum iteration number is , then the termination condition is:

[0131] ;

[0132] If the current iteration number reaches or exceeds the maximum iteration number , then return yes , and the algorithm stops; if the current iteration number does not reach or exceed the maximum iteration number , then return no , and continue the iteration;

[0133] Set a target fitness threshold , when the fitness of the optimal individual in the population reaches the threshold, stop the algorithm:

[0134] ;

[0135] If the fitness of the optimal individual reaches or exceeds the threshold, return yes , the algorithm stops; if the fitness of the optimal individual does not reach or exceed the threshold, return no , and continue the iteration.

[0136] A lightning arrester dynamic optimization device based on an improved genetic algorithm, which is used to implement a lightning arrester dynamic optimization algorithm based on an improved genetic algorithm, and further includes: a main circuit unit and a control circuit unit

[0137] The main circuit unit is used to simulate the lightning current waveform, and excite the lightning arrester model through the waveform to evaluate the dynamic response and residual voltage, including

[0138] Fast pulse current transmitter: Generate a simulated lightning current waveform

[0139] Lightning arrester: Absorb the overvoltage brought by the lightning current to protect the power equipment

[0140] Voltage measurement unit: Measure the residual voltage across the lightning arrester, record the voltage waveform, and transmit it to the control unit for analysis

[0141] The control circuit unit: Used for the control and data processing of the entire system, ensuring the normal operation of the main circuit unit, and performing optimization calculations, including

[0142] Decision variable resistance, inductance and capacitance data module, which extracts basic electrical parameters from the waveform-excited lightning arrester model to be used as decision variables for the genetic algorithm for adjustment by the optimization algorithm

[0143] Improved genetic algorithm module: Use the improved genetic algorithm to optimize the parameters of the lightning arrester, and thus through multiple iterations, find the parameter combination that makes the lightning arrester perform best under different lightning current waveforms

[0144] Simulated voltage and fitness function value module: Responsible for simulating the voltage response of the lightning arrester under different current waveforms, and calculating the fitness function value according to the simulation results

[0145] Termination condition judgment module: Judge whether the termination condition of the optimization is met

[0146] Output optimal solution module: Output the optimized parameters of the waveform-excited lightning arrester model. The ultimate goal is to find a parameter combination of the lightning arrester that shows the best performance under different lightning current waveforms

[0147] Example 1:

[0148] In this example, the following initial parameter ranges are selected:

[0149] Resistance (R): [20, 40] Ω, with an initial mean value of 30 Ω and a standard deviation = 5;

[0150] Inductance (L): [0.02, 0.08] mH, with an initial mean value of 0.05 mH, uniformly distributed;

[0151] Capacitance (C): [300, 600] pF, with an initial mean value of 450 pF and a standard deviation = 50.

[0152] Generate 100 individuals, and the parameters of each individual are randomly generated according to the normal distribution or the uniform distribution. The formula is as in S1:

[0153]

[0154] Specifically, select three typical lightning current waveforms, and the input parameters are as follows:

[0155] 1 / 10 μs steep wave lightning current: peak value 10 kA, decay constant τ = 1 μs, frequency f = 100 kHz;

[0156] 8 / 20 μs lightning current: peak value 5 kA, decay constant τ = 5 μs, frequency f = 20 kHz;

[0157] 8 / 20 μs lightning current: peak value 2 kA, decay constant τ = 30 μs, frequency f = 10 kHz;

[0158] As Figure 2 shown, in this embodiment, the initial parameters are: R = 30 Ω, L = 0.05 mH, C = 450 pF. Then, according to the above waveform parameters, input them into step S2 to calculate the reference residual voltage and the simulation residual voltage.

[0159] After continuous parameter iteration and update of the total fitness parameter through steps S2 - S7, Table 1 is obtained:

[0160] Table 1 Simulation parameters

[0161]

[0162] Through simulation, the reference voltage and the simulation voltage parameters are continuously iterated so that the simulation voltage continuously approaches the reference voltage. However, when reaching coincidence and then performing iteration, the error will become larger. Finally, the parameters that are closest to the coincidence after 16 iterations are returned to the arrester parameter model to obtain the optimal parameter model.

[0163] The above-described embodiments merely represent the implementation modes of the present invention. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the patent for the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all fall within the protection scope of the present invention. Therefore, the protection scope of the patent for the present invention shall be based on the appended claims.

Claims

1. An algorithm for a dynamic optimization model of a lightning arrester based on an improved genetic algorithm, characterized in that, It includes the following steps: S1. Initialize the population: Set the parameters of the lightning arrester dynamic optimization model as the initial population; S2. Fitness calculation: Calculate the fitness value according to multiple objective functions for the parameter combination of the lightning arrester dynamic optimization model; S3. Selection operation: Based on the fitness value, select individuals with high fitness from the current population as the parents, and the parent individuals will be used to generate the next generation of individuals; S4. Crossover operation: Through the crossover operation, combine the genes of the parent individuals to generate new offspring individuals; S5. Mutation operation: Increase the diversity of the population to prevent the algorithm from falling into a local optimal solution; S6. Recalculate fitness: After the crossover operation and the mutation operation, the newly produced individuals will continue to perform fitness calculation to ensure that the error of the new individuals under different current waveforms is minimized, so as to improve the protection performance of the lightning arrester; S7. Termination condition judgment: In each generation of optimization process, the algorithm will judge whether the preset termination condition is met; S8. Output the optimal solution: When the termination condition is met, output the optimal parameters of the lightning arrester dynamic optimization model; The specific content of S1 is: Set the initial value range of the population size and parameters, initialize with a normal distribution, and use the normal distribution to generate electrical parameters to ensure that the parameters have reasonable means and variances; Assume resistance that follows a normal distribution with a mean and a standard deviation: ； Among them, is the th resistance value generated by initialization; the resistance ; is a normal distribution random number generation function; For the inductor , initialize using a uniform distribution: ； Among them, is the th inductance value generated by initialization; the inductance ; is a uniformly distributed random number generation function; The final initialization process is: ； Among them, is the th capacitance value generated by initialization; capacitance ; is the normal distribution random number generation function of capacitance .

2. The algorithm of a lightning arrester dynamic optimization model based on an improved genetic algorithm according to claim 1, wherein The specific content of S2 is: S201. Reference voltage is as follows: ； Among them, is the standard lightning current waveform; is a non-linear resistor, the value of which depends on the standard lightning current waveform ; is the standard lightning current waveform is a function of; is the time parameter; Simulated voltage is as follows: ； Among them, is the input lightning strike current; is a non-linear resistor, whose value depends on the input lightning strike current ; S202. Assume there are multiple different lightning current waveforms, and the error calculation formula for each waveform is: ； Among them, is the number of individuals of the arrester model parameters; is the weight of the th lightning current waveform; and are the simulated voltage and the reference voltage under the th lightning current waveform, respectively; S203, Fitness Function is as follows: 。 3. The algorithm of a lightning arrester dynamic optimization model based on an improved genetic algorithm according to claim 2, wherein The specific content of S3 is: Suppose there are individuals of the arrester model parameters, and the fitness of individual is , and the total fitness is . Among them, , then the probability that individual is selected is: ； Calculate the cumulative fitness of an individual and make a selection based on the cumulative probability. The cumulative fitness is as follows: ； Among them, individual cumulative fitness value; when selecting, generate a random number , and select the individual that satisfies .

4. The algorithm of a lightning arrester dynamic optimization model based on an improved genetic algorithm according to claim 3, characterized in that The specific content of S5 is: The normal distribution makes the mutation amount fluctuate around zero, which is a random perturbation conforming to the Gaussian distribution: ； Among them, is a normal distribution with a mean of and a standard deviation of ; when the mean is set to 0, it means that the variation is perturbed without deviation; the standard deviation controls the randomness and amplitude of the variation amount. The mutated resistance value is: ； Among them, is the scaling factor, which is used to adjust the mutation amplitude; is the new resistance value generated after mutation; is the original resistance value before mutation; is the resistance difference before and after mutation.

5. The algorithm of a lightning arrester dynamic optimization model based on an improved genetic algorithm according to claim 4, characterized in that The specific content of S6 is: Calculate the fitness of the new individuals after crossover and mutation, evaluate the residual voltage error under different current waveforms, use the fitness function calculation formula in S2, re-evaluate the fitness value of the new individuals after crossover and mutation, update the population according to the new fitness value, and prepare to enter the selection operation of the next generation.

6. The algorithm of a lightning arrester dynamic optimization model based on an improved genetic algorithm according to claim 5, characterized in that, The specific content of S7 is: Assume that the current iteration number is , and the maximum iteration number is , then the termination condition is: ； If the current iteration count reaches or exceeds the maximum iteration count , then return yes , and the algorithm stops; if the current iteration count does not reach or exceed the maximum iteration count , then return no , and continue the iteration; Set a target fitness threshold , when the fitness of the optimal individual in the population reaches the threshold, stop the algorithm: ； If the fitness of the optimal individual reaches or exceeds the threshold, return yes , the algorithm stops; if the fitness of the optimal individual does not reach or exceed the threshold, return no , and continue the iteration.

7. An apparatus for a lightning arrester dynamic optimization model based on an improved genetic algorithm, characterized in that: The algorithm for implementing a lightning arrester dynamic optimization model based on an improved genetic algorithm as described in any one of claims 1 to 6 further includes: a main circuit unit and a control circuit unit, The main circuit unit is used to simulate the lightning current waveform, and through the waveform to excite the lightning arrester model, evaluate the dynamic response and the residual voltage, including Fast pulse current transmitter: Generate a simulated lightning current waveform; Lightning arrester: Absorb the overvoltage brought by the lightning current to protect the power equipment; Voltage measurement unit: Measure the residual voltage at both ends of the lightning arrester, record the voltage waveform, and transmit it to the control unit for analysis; The control circuit unit: Used for the control and data processing of the whole system, ensure the normal operation of the main circuit unit, and perform optimization calculations, including: Decision variable resistance, inductance and capacitance data module, extract the basic electrical parameters from the waveform exciting the lightning arrester model as the decision variables of the genetic algorithm for the adjustment of the optimization algorithm; Improved genetic algorithm module: Use the improved genetic algorithm to optimize the parameters of the lightning arrester, so as to find the parameter combination that makes the lightning arrester perform best under different lightning current waveforms through multiple iterations; Analog voltage and fitness function value module: Responsible for simulating the voltage response of the surge arrester under different current waveforms and calculating the fitness function value based on the simulation results; Termination condition judgment module: Judge whether the termination conditions for optimization are met; Optimal solution output module: Output the waveform excitation surge arrester model parameters after optimization. The ultimate goal is to find a combination of surge arrester parameters that exhibits the best performance under different lightning current waveforms.

Images