MCRWPT system multi-target parameter optimization method based on improved dragonfly optimization algorithm

Abstract

The invention discloses an MCRWPT system multi-target parameter optimization method based on an improved dragonfly optimization algorithm, and relates to the technical field of magnetic coupling resonance wireless power transmission, and the optimization method comprises the steps: firstly, analyzing an MCR WPT system of a bilateral LCC type topological structure, and taking a load resistance value, a coil spacing and a number of turns as design variables; establishing a multi-objective optimized mathematical model by taking the transmission efficiency and the output power of the system as optimization objectives; secondly, a mixed mutation operator is introduced into the multi-target dragonfly algorithm, so that the diversity of populations is increased, and the algorithm is prevented from falling into local optimum; a dynamic archive maintenance strategy based on a crowding distance is adopted, so that the obtained Pareto optimal solution set has better distributivity and diversity, and an optimal solution is obtained; and finally, multi-target parameter optimization is realized, and the system output power, the transmission efficiency and the optimal transmission distance are improved.

Description

Technical Field

[0001] The invention relates to the technical field of magnetically coupled resonant wireless power transmission, and in particular to a multi-objective parameter optimization method for an MCR_WPT system based on an improved dragonfly optimization algorithm. Background Art

[0002] In recent years, magnetic coupling resonant wireless power transmission technology has developed rapidly. It is a technology that enables the transmission of electric energy between the transmitting coil and the receiving coil at a limited distance under the premise of generating resonance between coils with the same resonant frequency. The magnetic coupling mechanism is the energy carrier of the wireless power transmission system, and the design of its structure and parameters is related to the transmission performance of the entire system. When designing parameters, multiple performance indicators often cannot be optimized at the same time. Therefore, the design of the wireless power transmission system is a multi-parameter and multi-objective optimization problem.

[0003] When applied to the MCR_WPT system, the traditional dragonfly algorithm is often used as the basic dragonfly algorithm, which is a swarm intelligence algorithm based on the dragonfly group behavior. The dragonfly algorithm simulates two types of dragonfly cluster behaviors, static and dynamic, to achieve global search and local development; under static behavior, dragonflies form multiple small groups to forage everywhere, that is, to conduct global search in a specific solution space; under dynamic behavior, dragonflies aggregate into a large group and fly in the same direction to forage or avoid natural enemies, that is, to conduct local development in a certain area. Each behavior of dragonfly movement can be expressed as:

[0004] (1) Dispersal behavior: Dragonflies avoid collisions with neighboring individuals;

[0005] (2) Alignment behavior: adjacent individuals tend to maintain the same speed;

[0006] (3) Aggregation behavior: Dragonflies tend to gather toward the center of neighboring individuals;

[0007] (4) Foraging behavior: the attraction of food to dragonflies;

[0008] (5) Avoidance behavior: dragonflies’ repulsion to natural enemies;

[0009] At present, the MCR WPT system has problems such as low multi-objective optimization efficiency, limited transmission distance, and difficulty in taking into account both transmission efficiency and output power. Most of the existing research is based on a single factor of efficiency or power, and the overall performance of the system is poor. Therefore, a multi-objective parameter optimization method for the MCR_WPT system based on the improved dragonfly optimization algorithm is proposed. Summary of the invention

[0010] 1. Technical issues to be resolved

[0011] In view of the deficiencies in the prior art, the present invention provides a multi-objective parameter optimization method for an MCR_WPT system based on an improved dragonfly optimization algorithm, and solves the problems raised in the background technology by running the optimization method.

[0012] (II) Technical solution

[0013] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0014] The multi-objective parameter optimization method of the MCR_WPT system based on the improved dragonfly optimization algorithm includes the following steps: initializing mathematical model parameters: establishing a mathematical model of the MCR_WPT system and determining design variables;

[0015] Initialize population parameters: set the initial parameters of the dragonfly population, including population size and number of iterations;

[0016] Generate dragonfly population: Generate the initial dragonfly population according to the initial parameters, and the dragonfly individuals represent random parameter combinations;

[0017] Fitness calculation: Use simulation software to build a simulation model of the MCR-WPT system based on the initial parameters. On the basis of the simulation model, calculate the corresponding objective function value for each dragonfly individual, and evaluate it based on the objective function value to determine the fitness of each dragonfly individual;

[0018] Screening operation: According to the comparison results between fitness and preset standard threshold, elite individuals are screened out and the external file G is updated;

[0019] Determine operation: determine whether the real-time capacity of the updated external file G exceeds the limit;

[0020] If it exceeds, the dynamic maintenance strategy of the external archive based on the crowding distance removes the redundant values ​​and adds the new non-inferior solution to the external archive G; if it does not exceed, the new non-inferior solution is directly added to the external archive G;

[0021] Introduce a hybrid mutation operator: update the dragonfly individuals newly added to the external archive G, and update the positions of the dragonfly individuals based on the hybrid mutation operation;

[0022] Iterative optimization: Determine whether the set termination condition is met based on the result of introducing the hybrid mutation operator;

[0023] If so, output the current Pareto optimal solution set;

[0024] If not, the operation of generating the dragonfly population is executed again until the set termination condition is met.

[0025] Furthermore, the determined setting variables include: load resistance, coil spacing, and number of turns.

[0026] Furthermore, the simulation software used is any one of Multisim, MATLAB and PSpice.

[0027] Furthermore, the objective function values ​​include the transmission efficiency and output power corresponding to each dragonfly individual;

[0028] Among them, the transmission efficiency is obtained by calculating the ratio of input power to output power;

[0029] The output power is obtained from the simulation results.

[0030] Furthermore, the process of evaluation based on the objective function value is as follows:

[0031]

[0032] Q = a*max(0,η min -η); (2)

[0033] In formula (1), F represents the fitness of the corresponding dragonfly individual, α and β are the weighted coefficients of transmission efficiency and output power, respectively, and their value ranges are [0, 1], η is the transmission efficiency, η min represents the minimum transmission efficiency threshold, Po represents the output power, and Po_ref represents the reference output power;

[0034] In formula (2), Q represents the penalty term, a represents the penalty coefficient, and k>0.

[0035] Furthermore, when the fitness of the corresponding dragonfly individual exceeds the standard threshold, it is indicated as an elite individual and is screened out, and the original external file G is updated, that is, the elite individual is entered into the original external file G.

[0036] Furthermore, the process of removing redundant values ​​based on the dynamic maintenance strategy of external archives based on crowding distance is as follows:

[0037] All optimal solutions are sorted from small to large according to the objective function value. The position information of the sorting number of the optimal solution x on the kth objective dimension is represented as P(x,k), and the objective function value corresponding to the optimal solution is represented as f P(x,k) According to the definition of crowding distance, when the optimal solution with position information P(x, k) on the k-th target dimension is deleted, it will affect the crowding distance components of the two optimal solutions with position information P(x, k)+1 and P(x, k)-1 on the k-th target dimension. Therefore, it is only necessary to recalculate the crowding distance components of the two adjacent affected optimal solutions on the k-th target dimension.

[0038] Furthermore, the mixed mutation operations include: random average mutation and Gaussian mutation;

[0039] Among them, random average mutation is a mutation operation that generates new individuals by introducing random perturbations;

[0040] The specific operations are as follows: select a variant individual: randomly select an individual from the current population as a variant individual; generate a random perturbation: generate a random perturbation according to the preset variable step length; apply the perturbation: add the generated random perturbation to the gene of the variant individual to generate a new individual;

[0041] Gaussian mutation is a mutation operation that uses the normal distribution characteristics to generate new individuals;

[0042] The specific operations are as follows: select mutant individuals: randomly select an individual from the current population as a mutant individual; generate normal random numbers: generate a random number that obeys the normal distribution according to the preset mean and standard deviation; apply perturbations: add the generated normal random numbers to the genes of the mutant individual to generate a new individual.

[0043] (III) Beneficial effects

[0044] The present invention provides a multi-objective parameter optimization method for an MCR_WPT system based on an improved dragonfly optimization algorithm, which has the following beneficial effects:

[0045] (1) This scheme constructs the MCR-WPT system model through simulation software, and optimizes the parameter combination based on the dragonfly algorithm, realizing the comprehensive evaluation of transmission efficiency and output power. The design of the fitness function makes the optimization goal clear, that is, maximizing the fitness value to obtain the optimal parameter combination. This scheme solves the complex problems in the parameter optimization of the MCR-WPT system, improves the transmission efficiency and output power of the system, and ensures the accuracy and reliability of the simulation results. By setting the penalty term, it ensures that the transmission efficiency is not lower than the minimum threshold, further improving the performance of the system.

[0046] (2) This scheme analyzes the MCR WPT system with bilateral LCC topology, takes load resistance, coil spacing and number of turns as design variables, and takes the system transmission efficiency and output power as optimization targets, and establishes a mathematical model for multi-objective optimization; secondly, a hybrid mutation operator is introduced into the multi-objective dragonfly algorithm to increase the diversity of the population and avoid the algorithm from falling into the local optimum. A dynamic archive maintenance strategy based on crowding distance is adopted to make the obtained Pareto optimal solution set have better distribution and diversity, and obtain the optimal solution; finally, multi-objective parameter optimization is achieved, and the system output power, transmission efficiency and optimal transmission distance are all improved; in summary, under the multi-objective optimization of the MCR-WPT system, the transmission distance is improved, and the transmission efficiency and output power are taken into account at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic diagram of the bilateral LCC compensation network structure in the present invention;

[0048] Figure 2 Schematic diagram of the bilateral LCC mutual inductance coupling model in the present invention;

[0049] Figure 3 This is a flow chart of the improved multi-objective dragonfly algorithm in the present invention. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in 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 creative work are within the scope of protection of the present invention.

[0051] Embodiment 1:

[0052] See also Figures 1 to 3 This embodiment provides a multi-objective parameter optimization method for the MCR_WPT system based on the improved dragonfly optimization algorithm, introduces the improved multi-objective dragonfly algorithm to optimize the system, improves the system output power and transmission efficiency, and improves the optimal transmission distance of the system. The steps of the optimization method are as follows:

[0053] S1. Initialize mathematical model parameters:

[0054] Establish the mathematical model of the MCR_WPT system and determine the design variables;

[0055] Among them, the MCR WPT system with bilateral LCC topology was analyzed, and a multi-objective optimization mathematical model was established with load resistance, coil spacing and number of turns as design variables and system transmission efficiency and output power as optimization targets.

[0056] Reference Figure 1 and Figure 2 These are the structures and models corresponding to bilateral LCC;

[0057] Therefore, the specific setting variables determined are: load resistance, coil spacing and number of turns.

[0058] S2. Initialize population parameters:

[0059] Set the initial parameters of the dragonfly population, including the population size (i.e. the number of dragonfly individuals) and the number of iterations;

[0060] Among them, the initial parameters mentioned above will determine the scale and depth of the optimization process;

[0061] It should be noted that in this step, setting the initial parameters of the dragonfly population is a key link in the algorithm preparation stage, which directly determines the scale and depth of the optimization process; specifically, the initial parameters that need to be set mainly include the population size (that is, the number of dragonfly individuals) and the number of iterations. The following is a specific description of these two parameters:

[0062] 1. Population size (number of dragonflies):

[0063] Definition: Population size refers to the total number of dragonfly individuals that participate in evolution at the same time during the optimization process.

[0064] Impact: Diversity: A larger population size can increase the diversity of the population, because more individuals mean more genetic combinations and phenotypic differences, which helps the algorithm explore more widely in the search space; Computational cost: However, the larger the population size, the higher the computational cost required for each iteration, because each individual needs to be calculated and evaluated; Convergence speed: A moderate population size can ensure that the algorithm converges to the optimal solution at a reasonable speed while maintaining diversity;

[0065] Selection: The choice of population size should be determined based on the complexity of the problem and the computational resources; for complex problems, a larger population size may be more advantageous; whereas for situations with limited computational resources, a smaller population size may be more appropriate;

[0066] 2. Number of iterations:

[0067] Definition: The number of iterations refers to the total number of rounds of the algorithm, that is, the number of times the dragonfly population performs evolution, selection, crossover, and mutation operations;

[0068] Impact: Optimization depth: A higher number of iterations allows the algorithm to explore the search space more deeply, and may find a better solution; Computation time: However, the more iterations, the longer the computation time required; Termination condition: In practical applications, the number of iterations is usually combined with some termination condition, such as reaching a preset fitness value or the rate of change of the solution being less than a certain threshold, to avoid unnecessary computation waste; Selection: The choice of the number of iterations should be based on the difficulty of the problem, the size of the population, and the available computation time; for difficult problems, more iterations may be required to find a satisfactory solution;

[0069] In summary, setting the initial parameters of the dragonfly population (population size and number of iterations) is a process that requires a trade-off between diversity, computational cost, and optimization depth; reasonable parameter settings can significantly improve the performance and efficiency of the algorithm and help the algorithm find better solutions under limited computing resources; in practical applications, these parameters usually need to be determined through experiments or experience, and may need to be adjusted according to the specific circumstances of the problem, and the initial parameters are a range of values.

[0070] S3. Generate dragonfly population:

[0071] Generate an initial dragonfly population based on the initial parameters, where each dragonfly individual represents a random parameter combination;

[0072] Each dragonfly individual represents a possible parameter combination, which will be used in the subsequent algorithm optimization process. This step is immediately after initializing the population parameters. The corresponding values ​​are randomly selected from the range values ​​corresponding to the above population size and number of iterations for combination.

[0073] It should be noted that:

[0074] Generating the initial dragonfly population is a key step in the algorithm optimization process. By randomly generating diverse dragonfly individuals, the algorithm can explore possible solutions in a wide search space and provide a good foundation for subsequent optimization processes. In practical applications, the process of generating the initial population may need to be adjusted and optimized according to the specific needs of the problem and prior knowledge, which will not be elaborated here.

[0075] S4, fitness calculation:

[0076] Using simulation software or tools, build a simulation model of the MCR-WPT system based on the initial parameters. This step provides a basic platform for subsequent calculations and optimizations. Based on the simulation model, calculate the corresponding objective function value for each dragonfly individual (i.e., each set of parameter combinations), evaluate based on the objective function value, and determine the fitness of each dragonfly individual.

[0077] Commonly used simulation software includes Multisim, MATLAB / Simulink and PSpice; these software provide a rich circuit component library and powerful simulation functions, which can meet the simulation requirements of the MCR-WPT system;

[0078] Incidental Notes on Building a System Model:

[0079] According to the working principle of the MCR-WPT system, a circuit model including power supply, inverter, resonant coil, rectifier, load and other components is constructed; the parameters of circuit components, such as coil inductance, compensation capacitor, load resistance, etc., are set based on the parameter combination of the individual dragonfly; simulation parameters such as simulation time and step size are set to ensure the accuracy and reliability of the simulation results; in addition, different simulation modes can be configured according to needs, such as steady-state analysis, transient analysis, etc.

[0080] Among them, the objective function value includes the transmission efficiency and output power corresponding to each dragonfly individual;

[0081] The transmission efficiency is obtained by calculating the ratio of input power to output power. The output power can be directly obtained from the simulation results, or obtained by calculating the product of the voltage and current across the load. In this embodiment, the simulation results shall prevail, but in actual operation, the output power A is obtained from the simulation results.

[0082] Calculate the product of the voltage and current across the load to obtain the output power B;

[0083] If the absolute value of the difference between the output power A and B exceeds the set error value, the simulation software or tool should be replaced to ensure the validity and accuracy of the simulation results. The error value set here is usually 1% of the output power A.

[0084] For each dragonfly individual (i.e., each set of parameter combinations), set the corresponding parameter values ​​in the simulation model;

[0085] Run simulation to obtain the system’s output power and transmission efficiency;

[0086] The output power and transmission efficiency are calculated as the objective function values;

[0087] The process of evaluation based on the objective function value is as follows:

[0088]

[0089] Q = a*max(0,η min -η); (2)

[0090] In formula (1), F represents the fitness of the corresponding dragonfly individual, α and β are the weighted coefficients of transmission efficiency and output power, respectively, both in the range of [0, 1], η is the transmission efficiency (%), η min represents the minimum acceptable transmission efficiency threshold (%), Po represents the output power (W), and Po_ref represents the reference output power (W, used for normalization or setting a desired output level);

[0091] In formula (2), Q represents the penalty term, a represents the penalty coefficient, which is used to control the intensity of the penalty, and k>0;

[0092] The weight coefficient is determined by the coefficient of variation method, which is a method of assigning weights to each indicator according to the degree of variation between the current value of each evaluation indicator and the target value. If the numerical difference of a certain indicator is large, it can clearly distinguish the evaluated objects, indicating that the indicator has rich information for distinguishing, and thus the indicator should be given a larger weight. On the contrary, if the numerical difference of each evaluated object on a certain indicator is small, then the ability of this indicator to distinguish the evaluated objects is weak, and thus the indicator should be given a smaller weight. This method directly uses the information contained in each indicator to obtain the weight of the indicator through calculation, so it is objective.

[0093] Logic description:

[0094] This part normalizes the transmission efficiency from a possible range (such as 0% to 100%) to a new range (0 to 1). This is done to ensure that the comparison between different transmission efficiencies is relative and to make the effect of transmission efficiency on fitness more linear; This part normalizes the output power to the reference output power Po_ref. This is done to make the comparison between different output powers consistent, especially when the range of output power may be large; Q: This is a penalty term used to penalize transmission efficiency below the minimum threshold η min If the transmission efficiency is lower than η min , then the penalty term is positive, otherwise it is 0;

[0095] In the design of the fitness function, the larger the fitness F, the better. This is because the fitness value represents the performance of an individual (in this scenario, the dragonfly individual, that is, each set of parameter combinations) in the optimization problem; the larger the fitness, the better the individual's performance on the objective function (such as the comprehensive evaluation of transmission efficiency and output power).

[0096] S5. Screening operation:

[0097] According to the comparison results between fitness and preset standard threshold, elite individuals are selected and external archive G is updated;

[0098] Among them, when the fitness of the corresponding dragonfly individual exceeds the standard threshold, it is considered an elite individual and is screened out, and the original external file G is updated, that is, the elite individual is recorded in the original external file G;

[0099] The comparison process between fitness and preset standard threshold is the dominance relationship;

[0100] It should be noted that the definition of dominance relationship: In multi-objective optimization problems, dominance relationship is used to determine whether one solution is better than another solution; if one solution performs better than another solution on all objective functions, the former solution is said to dominate the latter solution; comparison process and dominance relationship: In this problem, the comparison process of fitness and the preset standard threshold is actually a special dominance relationship judgment; elite individuals (individuals whose fitness exceeds the standard threshold) are considered to perform better than other individuals in the optimization problem (that is, dominate other individuals).

[0101] S6. Determination operation:

[0102] Determine whether the real-time capacity of the updated external file G exceeds the limit;

[0103] If it exceeds, the dynamic maintenance strategy of the external archive based on the crowding distance removes the redundant values ​​and adds the new non-inferior solution to the external archive G; if it does not exceed, the new non-inferior solution is directly added to the external archive G;

[0104] Among them, the process of removing redundant values ​​based on the dynamic maintenance strategy of external archives based on crowding distance is as follows:

[0105] All optimal solutions are sorted from small to large according to the objective function value. The position information of the sorting number of the optimal solution x on the kth objective dimension is represented as P(x,k), and the objective function value corresponding to the optimal solution is represented as f P(x,k) , from the definition of crowding distance, when the optimal solution with position information P(x, k) on the kth target dimension is deleted, it will affect the crowding distance components of the two optimal solutions with position information P(x, k)+1 and P(x, k)-1 on the kth target dimension. Therefore, it is only necessary to recalculate the crowding distance components of the two adjacent affected optimal solutions on the kth target dimension;

[0106] The present invention adopts a dynamic maintenance strategy of external archives based on crowding distance to maintain the diversity of Pareto solution set;

[0107] In multi-objective optimization problems, external archives (or elite pools) are used to store the currently found non-inferior solutions (Pareto optimal solutions); however, with the iteration of the algorithm, the number of non-inferior solutions in the external archives may continue to increase, resulting in increased computational complexity and a possible reduction in the diversity of the solution set; therefore, an effective strategy is needed to dynamically maintain the external archives to maintain the diversity of the solution set and the efficiency of the algorithm; the dynamic maintenance strategy of the external archive based on crowding distance is such a method;

[0108] The steps to implement the strategy are as follows:

[0109] Calculate crowding distance: Sort all optimal solutions (i.e., non-inferior solutions in the external archive) from small to large according to each objective function value; for each objective dimension, record the sorting number (position information) of each optimal solution and the objective function value corresponding to the optimal solution; calculate the crowding distance of each optimal solution according to the definition of crowding distance. Crowding distance describes the density around a certain optimal solution and is an important indicator for evaluating the diversity of optimal solutions;

[0110] Eliminate redundant non-inferior solutions: Sort the non-inferior solutions from large to small according to the crowding distance, retain the non-inferior solutions with larger crowding distance, and eliminate the non-inferior solutions with smaller crowding distance until the size of the external archive is restored to the set value; it should be noted that when a non-inferior solution is eliminated, the crowding distance components of its two adjacent non-inferior solutions on the corresponding target dimension will be affected; therefore, it is only necessary to recalculate the crowding distance components of the two adjacent optimal solutions affected on the corresponding target dimension;

[0111] Add new non-inferior solutions: Add the new non-inferior solutions generated during the algorithm iteration to the updated external file G;

[0112] The strategic advantages are as follows:

[0113] Maintain diversity. By calculating the crowding distance and eliminating non-inferior solutions with a small crowding distance, the diversity of the solution set in the external archive can be effectively maintained. Improve algorithm efficiency. By limiting the size of the external archive, the increase in computational complexity of the algorithm due to storing too many non-inferior solutions can be avoided. Strong adaptability. This strategy is suitable for a variety of multi-objective optimization problems and has strong adaptability and robustness.

[0114] S7. Introduce hybrid mutation operator:

[0115] Update the dragonfly individuals newly added to the external file G, and update the positions of the dragonfly individuals based on the mixed mutation operation;

[0116] Among them, the mixed mutation operations include: random average mutation and Gaussian mutation;

[0117] In this step, the importance of introducing mixed mutation operators (such as random average mutation and Gaussian mutation) is particularly emphasized; this is to further improve the diversity of the population and avoid the algorithm falling into the local optimum;

[0118] Introducing a hybrid mutation operator to avoid falling into local optimality:

[0119] In order to improve the diversity of the MODA algorithm population and avoid falling into the local optimum, mutation operations are added to the algorithm. Since uniform mutation is prone to degeneration, the performance of the algorithm is reduced. It should be noted that random average mutation is a mutation operation that generates new individuals by introducing random disturbances.

[0120] The specific operations are as follows:

[0121] Select a mutant individual: randomly select an individual from the current population as a mutant individual; generate a random perturbation: generate a random perturbation according to the preset mutation step length or mutation range; this perturbation can be a uniformly distributed random number or other distribution forms; apply perturbation: add the generated random perturbation to the gene of the mutant individual to generate a new individual;

[0122] Gaussian mutation is a mutation operation that uses the characteristics of normal distribution to generate new individuals;

[0123] The specific operations are as follows:

[0124] Select a mutant individual: randomly select an individual from the current population as a mutant individual; generate a normal random number: generate a random number that obeys the normal distribution according to the preset mean and standard deviation; apply perturbation: add the generated normal random number to the gene of the mutant individual to generate a new individual;

[0125] For the mixed probability, it is usually set to 50% probability using random average variation and 50% probability using Gaussian variation; in this embodiment, a probability setting method is also provided, which comprehensively considers multiple factors to construct a method based on the required parameters to calculate the probability value;

[0126] 1. Considerations

[0127] Algorithm stage:

[0128] At different stages of the algorithm, the demand for mutation operators may be different; for example, in the early stages of the algorithm, more global search capabilities may be required, so mutation operators with stronger exploration capabilities (such as Gaussian mutation) may be selected; while in the later stages of the algorithm, more local search capabilities may be required, so mutation operators with stronger development capabilities (such as random average mutation) may be selected;

[0129] Population diversity:

[0130] The diversity of the population is an important indicator to measure whether the algorithm is trapped in the local optimum. When the population diversity is low, the probability of using mutation operators with stronger exploration capabilities can be increased to promote the generation of new individuals and improve the diversity of the population.

[0131] Fitness distribution:

[0132] The fitness distribution reflects the quality of individuals in the population. When the fitness distribution is relatively concentrated, it can be considered that the algorithm may have fallen into a local optimum. At this time, the probability of using a mutation operator with stronger exploration ability can be increased;

[0133] Performance of mutation operator:

[0134] During the actual operation of the algorithm, the performance of each mutation operator (such as the quality of generating new individuals, the impact on population diversity, etc.) can be recorded, and the probability of using various mutation operators can be dynamically adjusted based on these performances;

[0135] 2. Methods for constructing required parameters

[0136] Based on the above considerations, a probability value for calculating which mutation operator to adopt based on the required parameters can be constructed;

[0137] Here's one possible way to build it:

[0138] Defining diversity metrics:

[0139] Define the metric for measuring population diversity, i.e., standard deviation based on the distance between individuals or information entropy based methods;

[0140] Define the fitness concentration index:

[0141] Define a metric to measure the concentration of the fitness distribution, i.e., the standard deviation of the fitness values ​​or the maximum difference between the fitness values;

[0142] Set weights:

[0143] Set different weights for diversity indicators and fitness concentration indicators according to the algorithm stage and actual needs;

[0144] Calculate the probability value:

[0145] According to the actual values ​​of the diversity index and the fitness concentration index and the set weights, a comprehensive index is calculated; then, the probability value of which mutation operator to adopt is set according to this comprehensive index; for example, when the comprehensive index is high (indicating that the population diversity is low or the fitness distribution is more concentrated), the probability of adopting a mutation operator with stronger exploration ability can be increased; conversely, the probability of adopting a mutation operator with stronger development ability can be increased;

[0146] Dynamic Adjustment:

[0147] During the operation of the algorithm, the calculation method of weights or comprehensive indicators is dynamically adjusted according to certain conditions (such as population diversity is below a certain threshold) to better meet the needs of the algorithm;

[0148] 3. Examples

[0149] Assume that at a certain moment, the population diversity index is 0.5 (lower), the fitness concentration index is 0.3 (higher), and the weights are set to 0.6 (the emphasis on the diversity index) and 0.4 (the emphasis on the fitness concentration index);

[0150] Comprehensive index = 0.6 × 0.5 + 0.4 × 0.3 = 0.42;

[0151] A threshold is set (such as 0.5). When the comprehensive index is lower than the threshold, it is considered that the population diversity is low and the fitness distribution is more concentrated. Therefore, the probability of using the Gaussian mutation operator is increased (such as setting it to 0.7), and the probability of using the random average mutation operator is reduced (such as setting it to 0.3). In this way, the probability values ​​of using various mutation operators can be dynamically adjusted to meet the needs of the algorithm at different stages and conditions.

[0152] S8, iterative optimization:

[0153] Determine whether the set termination condition is met according to the result of S7;

[0154] If so, output the current Pareto optimal solution set;

[0155] If not, the operation of S3 is executed again until the set termination condition is met;

[0156] Among them, the termination condition can be set according to actual needs, and the principle is the same as setting the threshold, so it will not be elaborated here; In summary, the purpose of the present invention is to propose an improved multi-objective dragonfly optimization algorithm to optimize system parameters to improve system performance, aiming at the problems that the current wireless power transmission system has low multi-objective optimization efficiency, limited transmission distance, and difficulty in taking into account both transmission efficiency and output power.

[0157] Firstly, by analyzing the MCR WPT system with bilateral LCC topology, taking load resistance, coil spacing and number of turns as design variables, and taking the system transmission efficiency and output power as optimization targets, a mathematical model of multi-objective optimization was established; secondly, a hybrid mutation operator was introduced into the multi-objective dragonfly algorithm to increase the diversity of the population and avoid the algorithm from falling into the local optimum. A dynamic archive maintenance strategy based on crowding distance was adopted to make the obtained Pareto optimal solution set have better distribution and diversity, and obtain the optimal solution; finally, multi-objective parameter optimization was achieved, and the system output power, transmission efficiency and optimal transmission distance were all improved;

[0158] At present, the MCR WPT system has problems such as low multi-objective optimization efficiency, limited transmission distance, and difficulty in taking into account both transmission efficiency and output power at the same time. Most of the existing research is carried out from a single factor of efficiency or power, and the overall performance of the system is poor. Based on the above research, this application constructs a system equivalent model, takes load resistance, coil spacing and number of turns as design variables, analyzes the output power and transmission efficiency of the system, and uses an improved multi-objective dragonfly algorithm to optimize the system parameters, so as to increase the system transmission distance and improve the system performance.

[0159] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product. A person of ordinary skill in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0160] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0161] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. The multi-objective parameter optimization method of MCR_WPT system based on the improved dragonfly optimization algorithm is characterized by: The steps included are: initializing mathematical model parameters: establishing the mathematical model of the MCR_WPT system and determining the design variables; Initialize population parameters: set the initial parameters of the dragonfly population, including population size and number of iterations; Generate dragonfly population: Generate the initial dragonfly population according to the initial parameters, and the dragonfly individuals represent random parameter combinations; Fitness calculation: Use simulation software to build a simulation model of the MCR-WPT system based on the initial parameters. On the basis of the simulation model, calculate the corresponding objective function value for each dragonfly individual, and evaluate it based on the objective function value to determine the fitness of each dragonfly individual; Screening operation: According to the comparison results between fitness and preset standard threshold, elite individuals are screened out and the external file G is updated; Determine operation: determine whether the real-time capacity of the updated external file G exceeds the limit; If it exceeds, the dynamic maintenance strategy of the external archive based on the crowding distance removes the redundant values ​​and adds the new non-inferior solution to the external archive G; if it does not exceed, the new non-inferior solution is directly added to the external archive G; Introduce a hybrid mutation operator: update the dragonfly individuals newly added to the external archive G, and update the positions of the dragonfly individuals based on the hybrid mutation operation; Iterative optimization: Determine whether the set termination condition is met based on the result of introducing the hybrid mutation operator; If so, output the current Pareto optimal solution set; If not, the operation of generating the dragonfly population is executed again until the set termination condition is met.

2. The multi-objective parameter optimization method of the MCR_WPT system based on the improved dragonfly optimization algorithm according to claim 1 is characterized in that: The set variables determined include: load resistance, coil spacing and number of turns.

3. The multi-objective parameter optimization method of the MCR_WPT system based on the improved dragonfly optimization algorithm according to claim 1 is characterized in that: The simulation software used is any one of Multisim, MATLAB and PSpice.

4. The multi-objective parameter optimization method of the MCR_WPT system based on the improved dragonfly optimization algorithm according to claim 1 is characterized in that: The objective function values ​​include the transmission efficiency and output power corresponding to each dragonfly individual; Among them, the transmission efficiency is obtained by calculating the ratio of input power to output power; The output power is obtained from the simulation results.

5. The multi-objective parameter optimization method of the MCR_WPT system based on the improved dragonfly optimization algorithm according to claim 4 is characterized in that: The process of evaluation based on the objective function value is as follows: Q=a*max(0,n min -h); (2) In formula (1), F represents the fitness of the corresponding dragonfly individual, α and β are the weighted coefficients of transmission efficiency and output power, respectively, and their value ranges are [0, 1], η is the transmission efficiency, η min represents the minimum transmission efficiency threshold, Po represents the output power, and Po_ref represents the reference output power; In formula (2), Q represents the penalty term, a represents the penalty coefficient, and k>0.

6. The multi-objective parameter optimization method of the MCR_WPT system based on the improved dragonfly optimization algorithm according to claim 1 is characterized in that: When the fitness of the corresponding dragonfly individual exceeds the standard threshold, it is considered an elite individual and is screened out, and the original external file G is updated, that is, the elite individual is entered into the original external file G.

7. The multi-objective parameter optimization method of the MCR_WPT system based on the improved dragonfly optimization algorithm according to claim 1 is characterized in that: The process of removing redundant values ​​by the dynamic maintenance strategy of external archives based on crowding distance is as follows: All optimal solutions are sorted from small to large according to the objective function value. The position information of the sorting number of the optimal solution x on the kth objective dimension is represented as P(x,k), and the objective function value corresponding to the optimal solution is represented as f P(x,k) According to the definition of crowding distance, when the optimal solution with position information P(x, k) on the k-th target dimension is deleted, it will affect the crowding distance components of the two optimal solutions with position information P(x, k)+1 and P(x, k)-1 on the k-th target dimension. Therefore, it is only necessary to recalculate the crowding distance components of the two adjacent affected optimal solutions on the k-th target dimension.

8. The multi-objective parameter optimization method of the MCR_WPT system based on the improved dragonfly optimization algorithm according to claim 1 is characterized in that: Mixed mutation operations include: random average mutation and Gaussian mutation; Among them, random average mutation is a mutation operation that generates new individuals by introducing random perturbations; The specific operations are as follows: select a variant individual: randomly select an individual from the current population as a variant individual; generate a random perturbation: generate a random perturbation according to the preset variable step length; apply the perturbation: add the generated random perturbation to the gene of the variant individual to generate a new individual; Gaussian mutation is a mutation operation that uses the normal distribution characteristics to generate new individuals; The specific operations are as follows: select mutant individuals: randomly select an individual from the current population as a mutant individual; generate normal random numbers: generate a random number that obeys the normal distribution according to the preset mean and standard deviation; apply perturbations: add the generated normal random numbers to the genes of the mutant individual to generate a new individual.

Images