Dynamic cooperative wireless charging system control method based on genetic gradient optimization algorithm

Abstract

The invention discloses a dynamic cooperative wireless charging system control method based on a genetic gradient optimization algorithm, and the method comprises the steps: carrying out the global search based on a genetic algorithm, taking the transmitting power and phase angle of a transmitting coil of a group of transmitting arrays as particles for coding, and evaluating the adaptation degree of individuals in a population through building a fitness function. Iteration of the population is realized through roulette selection and self-adaptive crossover and variation; when the population diversity is insufficient, triggering gradient descent to carry out local deep optimization on the optimal individual; selecting excellent individuals of current genetic iteration, and performing gradient calculation and updating by taking phase angle parameters of the excellent individuals as particles; and combining the phase angle parameter with the power parameter to form a new population, sorting individuals in the new population according to fitness, retaining the individuals with the fitness being first 10%, continuing to enter genetic algorithm iteration, and starting a new round of circulation. The energy efficiency of the system is improved, and the convergence speed of the algorithm and the electromagnetic safety are optimized.

Description

Technical Field

[0001] The present invention relates to wireless power transfer (WPT) technology, and more specifically to a dynamic collaborative wireless charging system control method based on a genetic gradient optimization algorithm. The method is suitable for dynamic charging of multiple devices in crowded places such as shopping malls, airports, and stations. Background Art

[0002] Traditional wireless charging systems typically transmit energy using a single transmitter and a single receiver. Therefore, in practical applications, these systems are often affected by factors such as relative offset between the primary and secondary sides or changes in system parameters, resulting in reduced transmission power and efficiency, making them unable to meet demand. An array-based transmitter system can, to a certain extent, compensate for this relative offset, creating a larger and more evenly distributed magnetic field within the charging area. The array format also provides the system with a wider range of energy transmission channels.

[0003] In addition, optimizing the operating parameters of the wireless charging system through relevant algorithms can also help the system adapt to more complex operating scenarios. Regarding the use of algorithms, particle swarm optimization and genetic algorithm have certain advantages in global search, which can make the system have stronger global convergence. However, if applied to the entire array system, the computational complexity is high and it is easy to fall into the problem of local optimality. For local search, the gradient descent method is more appropriate, but this method has difficulty in handling multi-constraint conditions and has poor adaptability in more complex scenarios. The simple fusion of genetic algorithm and gradient descent method results in limited performance improvement, but cannot effectively solve the real-time requirements in dynamic environments. Summary of the Invention

[0004] The purpose of this invention is to propose a dynamic collaborative wireless charging system control method based on a genetic gradient optimization algorithm. This method combines a genetic algorithm with a gradient descent method to achieve a two-stage adaptive optimization of global exploration and local search, organically combining the advantages of both algorithms and applying them in their respective appropriate fields. This dynamic collaborative system design can not only improve the performance of wireless charging systems operating under multiple devices, but also is more suitable for scenarios with high-density device access in public areas, and has broad development prospects in the future.

[0005] The technical solution to realize the present invention is: a dynamic cooperative wireless charging system control method based on genetic gradient optimization algorithm, comprising the following steps:

[0006] Step 1: Build an array-type wireless power transmission system coupling unit. The array-type wireless power transmission system coupling unit is composed of a transmitting array of transmitting coils. Each transmitting coil is placed separately in a plane to ensure that the number of transmitting coils in the array is the same as the number of devices to be charged.

[0007] Step 2: Perform a global search based on a genetic algorithm. The transmission power and phase angle of the transmitting coils of a group of transmitting arrays are encoded as particles. A fitness function is established to evaluate the fitness of individuals in the population. Roulette wheel selection, adaptive crossover, and mutation are used to achieve population iteration.

[0008] Step 3: Calculate the population diversity. If the population diversity is insufficient, trigger the gradient descent to perform local deep optimization on the optimal individual. Otherwise, continue the genetic algorithm to perform global search.

[0009] Step 4: Select the top 10% of the best individuals in the current genetic iteration, use the phase angle parameters of these individuals as particles to perform gradient calculation and update, achieve local optimization of the best individuals in the global search, and obtain a set of phase angle parameter configurations for the transmitting array;

[0010] Step 5: Combine the phase angle parameters of the individuals that have passed the gradient descent algorithm with the power parameters of the individuals in the previous genetic algorithm to form a new population. Sort the individuals in the new population by fitness, retain the top 10% of the fitness, and continue to enter the genetic algorithm iteration to start a new cycle.

[0011] Step 6: Determine whether to terminate the search and optimization by judging the population diversity index, maximum number of iterations and fitness value.

[0012] Furthermore, the circular spiral coil is made of copper.

[0013] Furthermore, the transmitting coil adopts a circular spiral coil.

[0014] Furthermore, in step 2, a global search is performed based on a genetic algorithm. The transmission power and phase angle of the transmitting coils of a group of transmitting arrays are encoded as particles. The fitness of individuals in the population is evaluated by establishing a fitness function. The population iteration is achieved through roulette wheel selection and adaptive crossover and mutation. The specific method is as follows:

[0015] Step 2.1, chromosome encoding design and population initialization

[0016] The chromosomes are encoded with real numbers and the initial population is randomly generated. The chromosome structure is:

[0017] [P1,θ1,P2,θ2,…P i ,θ i ]

[0018] Power value P i ∈[0,2000W]

[0019] Phase angle θ i ∈[0,2π]

[0020] The maximum number of iterations of the genetic algorithm is 50, and the population size is set to 100;

[0021] Each chromosome represents a complete parameter configuration of a set of transmit arrays, including transmit power and phase angle. Real number coding is used to avoid the precision loss of binary coding for continuous parameter optimization.

[0022] Step 2.2, establish fitness function

[0023] Establish a fitness function that includes system energy efficiency, delay penalty terms, and electromagnetic interference terms:

[0024]

[0025] Energy efficiency items Among them, P i,out is the transmission power of the i-th transmitting coil, θ i is the phase angle of the i-th transmitting coil, P total The total power limit of the wireless charging system;

[0026] Delay penalty where t max is the maximum delay of the device to be charged, P i,in is the actual received power of the i-th device to be charged;

[0027] Electromagnetic interference Among them, M safe is the electromagnetic safety threshold;

[0028] Considering the maximization of energy efficiency and the suppression of delay and electromagnetic interference, the weights ω1, ω2, and ω3 of the three parts are set to 0.6, 0.3, and 0.1 respectively;

[0029] The fitness function is used to calculate the fitness of individuals in the population, thereby measuring the quality of individuals and driving the population to evolve towards a better solution;

[0030] Step 2.3, selection and crossover mutation operations in genetic algorithms

[0031] According to roulette selection, individuals in the population are screened according to the fitness ratio, and the top 20% of individuals in fitness enter the next generation.

[0032] Set adaptive crossover mutation:

[0033] Using single-point crossover, crossover probability: Where max(F) is the maximum value of the population fitness, min(F) is the minimum value of the population fitness, and Current(F) is the fitness of the current individual. The crossover rate is dynamically adjusted to enhance the transmission of high-quality genes and reduce the combination of inefficient genes, thereby giving priority to retaining high-quality genes.

[0034] Add adaptive Gaussian mutation, mutation probability: Where N is the number of devices to be charged in the wireless charging system. When the population size increases, the mutation rate decreases to avoid excessive disturbance, and a minimum mutation rate of 0.01 is set to ensure minimum diversity retention.

[0035] Furthermore, in step 3, the population diversity is calculated. When the population diversity is insufficient, the gradient descent is triggered to perform local deep optimization on the optimal individual. Otherwise, the genetic algorithm continues to perform global search. The population diversity is measured by calculating the standard deviation of the population fitness, specifically:

[0036]

[0037] Among them, DI is the population diversity index; σ(F) is the standard deviation of population fitness, which measures the degree of dispersion of individual fitness within the population;

[0038] When DI < 0.25 and the number of iterations > 10, the top 10% of the individuals ranked by fitness in the genetic algorithm iteration are marked to enter the gradient descent. If the above two conditions cannot be met at the same time, the genetic algorithm iteration continues.

[0039] Furthermore, in step 4, the top 10% of excellent individuals in the current genetic iteration are selected, and the phase angle parameters of these individuals are used as particles for gradient calculation and update, thereby realizing local optimization of the excellent individuals in the global search and obtaining a configuration of the phase angle parameters of a set of emission arrays. The specific method is as follows:

[0040] Step 4.1, calculate the gradient of the phase angle

[0041] Perform gradient descent processing on the excellent individuals marked in the iterative process of the genetic algorithm, and use the finite difference method to calculate the gradient of the phase angle in the excellent individuals:

[0042]

[0043] The perturbation step size δ = 1×10 -5 , and set the number of iterations of gradient descent to 8, by applying positive and negative perturbations to the phase angle of the i-th device to be charged and performing gradient approximation calculation;

[0044] 4.2 Gradient Descent Update Strategy

[0045] Phase angle update calculation method:

[0046]

[0047] in is the new phase angle of individual i, is the old phase angle of individual i, α is the learning rate, which is set to 0.01 to balance the convergence speed and stability;

[0048] By calculating and updating the gradient of the phase angle, local optimization of the excellent individuals in the global search is achieved, and a configuration of the phase angle parameters of a set of transmitting arrays is obtained.

[0049] Furthermore, in step 6, by judging the population diversity index, the maximum number of iterations and the fitness value, it is determined whether to terminate the search and optimization. The specific method is as follows:

[0050] 1. The upper limit of the number of iterations is reached;

[0051] 2. No significant improvement in fitness over five consecutive generations;

[0052] 3. Population diversity index DI<0.1.

[0053] A dynamic cooperative wireless charging system based on a genetic gradient optimization algorithm is implemented. The control method of the dynamic cooperative wireless charging system based on the genetic gradient optimization algorithm is implemented to realize the control of the dynamic cooperative wireless charging system based on the genetic gradient optimization algorithm, including:

[0054] Module 1: Build an array-type wireless power transmission system coupling unit. The array-type wireless power transmission system coupling unit is composed of a transmitting array of circular spiral coils. Each circular spiral coil is placed separately in a plane to ensure that the number of transmitting coils in the array is the same as the number of devices to be charged;

[0055] Module 2 performs a global search based on a genetic algorithm. The transmission power and phase angle of the transmitting coils of a group of transmitting arrays are encoded as particles. The fitness level of individuals in the population is evaluated by establishing a fitness function. The population is iterated through roulette wheel selection and adaptive crossover and mutation.

[0056] Module 3 calculates population diversity. When the population diversity is insufficient, gradient descent is triggered to perform local deep optimization of the optimal individual. Otherwise, the genetic algorithm continues to perform global search.

[0057] Module 4 selects the top 10% of the best individuals in the current genetic iteration, uses the phase angle parameters of these individuals as particles for gradient calculation and update, and achieves local optimization of the best individuals in the global search, obtaining a set of phase angle parameter configurations for the emission array.

[0058] Module 5: Combine the phase angle parameters of the individuals that have passed the gradient descent algorithm with the power parameters of the individuals in the previous genetic algorithm to form a new population. Then sort the individuals in the new population by fitness, retain the top 10% of the fitness, and continue to enter the genetic algorithm iteration to start a new cycle.

[0059] Module 6 determines whether to terminate the search and optimization by judging the population diversity index, the maximum number of iterations and the fitness value.

[0060] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for controlling a dynamic cooperative wireless charging system based on a genetic gradient optimization algorithm is implemented to realize control of a dynamic cooperative wireless charging system based on a genetic gradient optimization algorithm.

[0061] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the method for controlling a dynamic cooperative wireless charging system based on a genetic gradient optimization algorithm is implemented to realize control of a dynamic cooperative wireless charging system based on a genetic gradient optimization algorithm.

[0062] Compared with the existing technology, the significant advantages of the present invention are: combining genetic algorithm and gradient descent method to complete two-stage adaptive optimization of global exploration and local search, organically combining the advantages of the two algorithms, and applying them in their respective appropriate fields. The design of this dynamic collaborative system can, on the one hand, improve the performance of the wireless charging system under the operation of multiple devices, and on the other hand, it is more suitable for scenarios with high-density device access in public areas, and has broad development prospects in the future. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 Schematic diagram of the simulation model of the array wireless power transmission system.

[0064] Figure 2 Flowchart for the implementation of the genetic gradient optimization algorithm.

[0065] Figure 3 This is a curve diagram of the genetic gradient dynamic collaborative optimization effect. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0067] A dynamic cooperative wireless charging system control method based on genetic gradient optimization algorithm, the steps are as follows:

[0068] Step 1: Build the coupling unit of the array wireless power transmission system

[0069] Design an m*n array of circular spiral coils for transmission. Place each coil separately in a plane. Ensure that the number of transmitting coils in the array is the same as the number of devices to be charged, which is N. Set the radius r of the coil and determine the spacing between adjacent coils based on the wavelength of the system. Copper, a common metal used in wireless charging coils, was used as the coil material. The array coil was modeled using Ansys Maxwell simulation software to ensure basic energy transmission capabilities. In practical applications, this software can be used for electromagnetic simulation to optimize parameters such as coil radius and spacing in the transmitting array.

[0070] Step 2: Genetic algorithm performs global search

[0071] The complete parameter configuration of a set of transmitting arrays is encoded as particles, including the transmitting power and phase angle of each transmitting coil. The fitness of individuals in the population is evaluated by establishing a fitness function. The iteration of the population is achieved through roulette wheel selection and adaptive crossover and mutation.

[0072] 2.1 Chromosome encoding design and population initialization

[0073] The chromosomes are encoded with real numbers and the initial population is randomly generated. The chromosome structure is:

[0074] [P1,θ1,P2,θ2,…P i ,θ i ]

[0075] Power value P i ∈[0,2000W]

[0076] Phase angle θ i ∈[0,2π]

[0077] The maximum number of iterations of the genetic algorithm is 50, and the population size is set to 100.

[0078] Each chromosome represents a complete parameter configuration of a set of transmit arrays, where the parameters include transmit power and phase angle; real number coding is used to avoid the precision loss of binary coding for continuous parameter optimization.

[0079] 2.2 Establishing fitness function

[0080] Establish a fitness function that includes system energy efficiency, delay penalty terms, and electromagnetic interference terms:

[0081]

[0082] Energy efficiency items Among them, P i,out is the transmission power of the i-th transmitting coil, θ i is the phase angle of the i-th transmitting coil, P total The total power limit of the wireless charging system;

[0083] Delay penalty where t max is the maximum delay of the device to be charged, P i,in is the actual received power of the i-th device to be charged;

[0084] Electromagnetic interference Among them, M safe is the electromagnetic safety threshold;

[0085] To maximize energy efficiency and minimize latency and electromagnetic interference, the weights ω1, ω2, and ω3 of the three components are set to 0.6, 0.3, and 0.1, respectively. In practical applications, appropriate adjustments can be made to ensure that ω1 + ω2 + ω3 = 1.

[0086] The fitness function is used to calculate the fitness of individuals in the population, thereby measuring the quality of individuals and driving the population to evolve towards a better solution.

[0087] 2.3 Selection and crossover mutation operations in genetic algorithms

[0088] According to roulette wheel selection, individuals in the population are screened according to their fitness ratio, and the top 20% of individuals in fitness enter the next generation.

[0089] Set adaptive crossover mutation:

[0090] Using single-point crossover, crossover probability: Where max(F) is the maximum population fitness (optimal fitness), min(F) is the minimum population fitness (worst fitness), and current(F) is the fitness of the current individual. Dynamically adjust the crossover rate to enhance the transmission of high-quality genes and reduce the combination of inefficient genes, thereby prioritizing the retention of high-quality genes.

[0091] Add adaptive Gaussian mutation, mutation probability: Where N is the number of devices to be charged in the wireless charging system. As the population size increases, the mutation rate decreases to avoid excessive perturbations, and a minimum mutation rate of 0.01 is set to ensure a minimum level of diversity.

[0092] Step 3: Dynamic collaborative detection of the system

[0093] The dynamic collaborative detection of the system is completed by judging the population diversity. When the population diversity is insufficient, the gradient descent is triggered to perform local deep optimization of the optimal individual.

[0094] The diversity of the population is measured by calculating the standard deviation of the population fitness:

[0095]

[0096] DI is the core indicator for subsequent dynamic collaborative testing; σ(F) is the standard deviation of population fitness, which measures the degree of dispersion of individual fitness within the population. The DI indicator is used to determine whether the population has fallen into a local optimum and to avoid early misjudgments in the algorithm. When DI < 0.25 and the number of iterations > 10, the system triggers gradient descent intervention to ensure stable algorithm operation. The system determines whether to enter the gradient descent optimization phase by combining the assessment of population diversity with the current number of genetic algorithm iterations.

[0097] When DI < 0.25 and the number of iterations > 10, the top 10% of the individuals ranked by fitness in the genetic algorithm iteration are marked to enter the gradient descent. If the above two conditions cannot be met at the same time, the genetic algorithm iteration will continue.

[0098] Step 4: Gradient descent method performs local optimization on excellent individuals

[0099] After iterations of the genetic algorithm, based on the system's dynamic collaborative detection scheme, the top 10% of individuals in the current genetic iteration are fed into the gradient descent module when conditions are met. The phase angle parameters of these individuals are used as particles for gradient calculation and solution update. Based on the global search of the genetic algorithm, local optimization improves the accuracy of the solution.

[0100] 4.1 Gradient calculation of phase angle

[0101] Perform gradient descent on the excellent individuals marked during the iteration of the genetic algorithm. Use the finite difference method to calculate the gradient of the phase angle in the excellent individuals and update the solution:

[0102]

[0103] The perturbation step size δ = 1×10 -5 , and set the number of iterations of gradient descent to 8. By applying positive and negative perturbations to the phase angle of the i-th device to be charged and performing gradient approximation calculation.

[0104] 4.2 Gradient Descent Update Strategy

[0105] The calculation method for setting the updated phase angle in the gradient descent module is:

[0106]

[0107] in is the new phase angle of individual i, is the old phase angle of individual i, and α is the learning rate, which is set to 0.01 to balance the convergence speed and stability. Only the phase angle is calculated to avoid perturbations in the power allocation and thus prevent the violation of the power constraint.

[0108] By calculating and updating the gradient of the phase angle, local optimization of the excellent individuals in the global search is achieved, and a configuration of the phase angle parameters of a set of transmitting arrays is obtained.

[0109] Step 5: Population merging and elite retention strategy

[0110] Through iteration of the genetic algorithm and further optimization of the gradient descent algorithm, the individuals (phase angle parameters) obtained by gradient descent are combined with the power parameters of the individuals in the previous genetic algorithm iteration to form a new population.

[0111] The individuals in the new population are sorted by fitness, and the top 10% of the individuals are retained, thereby retaining the high-quality solutions generated by gradient descent and avoiding the loss of historical optimal genes due to genetic operations.

[0112] Step 6: Termination condition judgment

[0113] The algorithm is terminated by judging whether it has converged or whether the algorithm has reached the maximum number of iterations:

[0114] 1. The upper limit of the number of iterations is reached;

[0115] 2. No significant improvement in fitness over five consecutive generations;

[0116] 3. Population diversity index DI<0.1.

[0117] When any of the above points is met, it is considered that the termination condition has been reached and the algorithm terminates immediately.

[0118] If the system keeps running, other individuals continue to enter the genetic algorithm iteration and start a new cycle.

[0119] The present invention also proposes a dynamic cooperative wireless charging system based on a genetic gradient optimization algorithm. The dynamic cooperative wireless charging system control method based on the genetic gradient optimization algorithm is implemented to realize the dynamic cooperative wireless charging system control based on the genetic gradient optimization algorithm, including:

[0120] Module 1: Build an array-type wireless power transmission system coupling unit. The array-type wireless power transmission system coupling unit is composed of a transmitting array of circular spiral coils. Each circular spiral coil is placed separately in a plane to ensure that the number of transmitting coils in the array is the same as the number of devices to be charged;

[0121] Module 2 performs a global search based on a genetic algorithm. The transmission power and phase angle of the transmitting coils of a group of transmitting arrays are encoded as particles. The fitness level of individuals in the population is evaluated by establishing a fitness function. The population is iterated through roulette wheel selection and adaptive crossover and mutation.

[0122] Module 3 calculates population diversity. When the population diversity is insufficient, gradient descent is triggered to perform local deep optimization of the optimal individual. Otherwise, the genetic algorithm continues to perform global search.

[0123] Module 4 selects the top 10% of the best individuals in the current genetic iteration, uses the phase angle parameters of these individuals as particles for gradient calculation and update, and achieves local optimization of the best individuals in the global search, obtaining a set of phase angle parameter configurations for the emission array.

[0124] Module 5: Combine the phase angle parameters of the individuals that have passed the gradient descent algorithm with the power parameters of the individuals in the previous genetic algorithm to form a new population. Then sort the individuals in the new population by fitness, retain the top 10% of the fitness, and continue to enter the genetic algorithm iteration to start a new cycle.

[0125] Module 6 determines whether to terminate the search and optimization by judging the population diversity index, the maximum number of iterations and the fitness value.

[0126] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for controlling a dynamic cooperative wireless charging system based on a genetic gradient optimization algorithm is implemented to realize control of a dynamic cooperative wireless charging system based on a genetic gradient optimization algorithm.

[0127] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the method for controlling a dynamic cooperative wireless charging system based on a genetic gradient optimization algorithm is implemented to realize control of a dynamic cooperative wireless charging system based on a genetic gradient optimization algorithm.

[0128] Example

[0129] To verify the effectiveness of the proposed dynamic cooperative wireless charging system based on the genetic gradient optimization algorithm, a wireless charging system capable of powering eight devices to be charged was built, and Matlab R2019a software was used to verify the effectiveness of the proposed dynamic cooperative wireless charging system based on the genetic gradient optimization algorithm. The specific steps are as follows:

[0130] Construction of the transmitting coil model and setting of system parameters

[0131] Ansys Maxwell electromagnetic field simulation software was used to establish a circular spiral coil transmitting array with a number of 8 (N=8) matching the number of devices to be charged. Each coil was evenly distributed in the plane. The radius of the coil was set to 10 cm and the coil spacing was set to (λ is the wavelength, the system operating frequency is set to 250 kHz, and the spacing is 12.56 cm at this time), and the material of the wire is copper.

[0132] Set the total power limit P of the system total =2000W, the threshold of electromagnetic safety in the area M safe =0.05T, the maximum number of iterations of the genetic algorithm is 50, and the population size is set to 100.

[0133] Using improved genetic algorithm to search the system globally

[0134] Generate the initial population: Real number encoding of chromosomes to obtain the following individuals:

[0135] [P1,θ1,P2,θ2,P3,θ3,P4,θ4,P5,θ5,P6,θ6,P7,θ7,P8,θ8]

[0136] P i ∈[0,2000W]

[0137] θ i ∈[0,2π]

[0138] Among them, P i is the power of the i-th device to be charged, θ i is the phase angle of the i-th device.

[0139] The fitness function is established based on the comprehensive energy efficiency term, delay penalty term, and electromagnetic interference term according to the weights:

[0140]

[0141] The meaning of each part of the expression in the function is as follows:

[0142] Calculating energy efficiency The weight is 0.6;

[0143] Calculating delay penalties The weight is 0.3;

[0144] Calculation of electromagnetic interference terms The weight is 0.1.

[0145] The genetic algorithm performs selection and adaptive crossover and mutation operations. Based on roulette wheel selection, the parent generation individuals are selected according to the fitness ratio, and the top 20% of individuals in fitness enter the next generation. And the adaptive crossover and mutation are set:

[0146] Single-point crossover: crossover probability

[0147] Adaptive Gaussian mutation: mutation probability

[0148] Dynamic collaborative detection and gradient descent calculation of the system

[0149] When the population diversity is insufficient, gradient descent local optimization is triggered. The diversity of the population is measured by calculating the standard deviation of the population fitness:

[0150]

[0151] When DI < 0.3 and the current number of iterations is greater than 10, the system triggers the gradient descent intervention. The finite difference method is used to calculate the gradient of the phase angle and update the solution.

[0152]

[0153] where δ = 1 × 10 -5 .

[0154] Merging populations and elite retention

[0155] Keep the historical optimal solution to avoid the loss of excellent individuals

[0156] The proposed genetic gradient optimization algorithm is used to dynamically coordinate the optimization of the wireless charging system with set parameters. The dynamic coordinated wireless charging system implemented based on the proposed genetic gradient optimization algorithm is run through the software Matlab R2019a according to the system parameters and algorithm flow.

[0157] Result analysis: Based on the genetic gradient optimization algorithm of the present invention, the wireless charging system with the set parameters is optimized to obtain the optimized system energy efficiency, convergence algebra, and electromagnetic interference term. The optimization effect is compared with that of the wireless charging system optimized by the traditional genetic algorithm under the same parameter settings.

[0158] Table 1 Optimal solution of genetic gradient optimization algorithm

[0159]

[0160] Table 2 Comparison of the effects of traditional genetic algorithm and genetic gradient optimization algorithm

[0161]

[0162] In summary, the present invention combines two optimization algorithms: the genetic algorithm and the gradient descent method. During the system's global search phase, the genetic algorithm uses selection, crossover, and mutation operations to explore potential optimal regions in the solution space and retain high-quality individuals. Dynamic collaborative detection and control then switches to the gradient descent algorithm for further optimization. Within the optimal regions found by the genetic algorithm, a detailed local search is performed to find the optimal solution, accelerating the system's convergence process.

[0163] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0164] The above-described embodiment merely represents one embodiment of the present application. While the description is relatively specific and detailed, it should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and such modifications and improvements are all within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

[0165] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0166] The above-described embodiment merely represents one embodiment of the present application. While the description is relatively specific and detailed, it should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and such modifications and improvements are all within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A dynamic cooperative wireless charging system control method based on genetic gradient optimization algorithm, characterized in that: The steps include: Step 1: Build an array-type wireless power transmission system coupling unit. The array-type wireless power transmission system coupling unit is composed of a transmitting array of transmitting coils. Each transmitting coil is placed separately in a plane to ensure that the number of transmitting coils in the array is the same as the number of devices to be charged. Step 2: Perform a global search based on a genetic algorithm. The transmission power and phase angle of the transmitting coils of a group of transmitting arrays are encoded as particles. A fitness function is established to evaluate the fitness of individuals in the population. Roulette wheel selection, adaptive crossover, and mutation are used to achieve population iteration. Step 3: Calculate the population diversity. If the population diversity is insufficient, trigger the gradient descent to perform local deep optimization on the optimal individual. Otherwise, continue the genetic algorithm to perform global search. Step 4: Select the top 10% of the best individuals in the current genetic iteration, use the phase angle parameters of these individuals as particles to perform gradient calculation and update, achieve local optimization of the best individuals in the global search, and obtain a set of phase angle parameter configurations for the transmitting array; Step 5: Combine the phase angle parameters of the individuals that have passed the gradient descent algorithm with the power parameters of the individuals in the previous genetic algorithm to form a new population. Sort the individuals in the new population by fitness, retain the top 10% of the fitness, and continue to enter the genetic algorithm iteration to start a new cycle. Step 6: Determine whether to terminate the search and optimization by judging the population diversity index, maximum number of iterations and fitness value.

2. The dynamic cooperative wireless charging system control method based on genetic gradient optimization algorithm according to claim 1 is characterized in that: The material of the circular spiral coil is copper.

3. The dynamic cooperative wireless charging system control method based on genetic gradient optimization algorithm according to claim 1 is characterized in that: The transmitting coil adopts a circular spiral coil, and the spacing between adjacent coils is Determined according to the working wavelength λ.

4. The dynamic cooperative wireless charging system control method based on genetic gradient optimization algorithm according to claim 1 is characterized in that: Step 2: Perform a global search based on a genetic algorithm. Encode the transmission power and phase angle of the transmitting coils of a group of transmitting arrays as particles. Establish a fitness function to evaluate the fitness of individuals in the population. Implement population iteration through roulette wheel selection and adaptive crossover and mutation. The specific method is as follows: Step 2.1, chromosome encoding design and population initialization The chromosomes are encoded with real numbers and the initial population is randomly generated. The chromosome structure is: [P1,θ1,P2,θ2,…P i ,i i ] Power value P i ∈[0,2000W] Phase angle θ i ∈[0,2π] The maximum number of iterations of the genetic algorithm is 50, and the population size is set to 100; Each chromosome represents a complete parameter configuration of a set of transmit arrays, including transmit power and phase angle. Real number coding is used to avoid the precision loss of binary coding for continuous parameter optimization. Step 2.2, establish fitness function Establish a fitness function that includes system energy efficiency, delay penalty terms, and electromagnetic interference terms: Energy efficiency items Among them, P i,out is the transmission power of the i-th transmitting coil, θ i is the phase angle of the i-th transmitting coil, P total The total power limit of the wireless charging system; Delay penalty where t max is the maximum delay of the device to be charged, P i,in is the actual received power of the i-th device to be charged; Electromagnetic interference Among them, M safe is the electromagnetic safety threshold; Considering the maximization of energy efficiency and the suppression of delay and electromagnetic interference, the weights ω1, ω2, and ω3 of the three parts are set to 0.6, 0.3, and 0.1 respectively; The fitness function is used to calculate the fitness of individuals in the population, thereby measuring the quality of individuals and driving the population to evolve towards a better solution; Step 2.3, selection and crossover mutation operations in genetic algorithms According to roulette selection, individuals in the population are screened according to the fitness ratio, and the top 20% of individuals in fitness enter the next generation. Set adaptive crossover mutation: Using single-point crossover, crossover probability: Where max(F) is the maximum value of the population fitness, min(F) is the minimum value of the population fitness, and Current(F) is the fitness of the current individual. The crossover rate is dynamically adjusted to enhance the transmission of high-quality genes and reduce the combination of inefficient genes, thereby giving priority to retaining high-quality genes. Add adaptive Gaussian mutation, mutation probability: Where N is the number of devices to be charged in the wireless charging system. When the population size increases, the mutation rate decreases to avoid excessive disturbance, and a minimum mutation rate of 0.01 is set to ensure minimum diversity retention.

5. The dynamic cooperative wireless charging system control method based on genetic gradient optimization algorithm according to claim 1 is characterized in that: Step 3: Calculate the population diversity. When the population diversity is insufficient, trigger the gradient descent to perform local deep optimization on the optimal individual. Otherwise, continue the genetic algorithm to perform global search. The population diversity is measured by calculating the standard deviation of the population fitness. Specifically: Among them, DI is the population diversity index; σ(F) is the standard deviation of population fitness, which measures the degree of dispersion of individual fitness within the population; When DI < 0.25 and the number of iterations > 10, the top 10% of the individuals ranked by fitness in the genetic algorithm iteration are marked to enter the gradient descent. If the above two conditions cannot be met at the same time, the genetic algorithm iteration continues.

6. The dynamic cooperative wireless charging system control method based on genetic gradient optimization algorithm according to claim 1 is characterized in that: Step 4: Select the top 10% of the best individuals in the current genetic iteration, use the phase angle parameters of these individuals as particles to perform gradient calculation and update, and achieve local optimization of the best individuals in the global search to obtain a set of phase angle parameter configurations for the emission array. The specific method is as follows: Step 4.1, calculate the gradient of the phase angle Perform gradient descent processing on the excellent individuals marked in the iterative process of the genetic algorithm, and use the finite difference method to calculate the gradient of the phase angle in the excellent individuals: The perturbation step size δ = 1×10 -5 , and set the number of iterations of gradient descent to 8, by applying positive and negative perturbations to the phase angle of the i-th device to be charged and performing gradient approximation calculation; 4.2 Gradient Descent Update Strategy Phase angle update calculation method: in is the new phase angle of individual i, is the old phase angle of individual i, α is the learning rate, which is set to 0.01 to balance the convergence speed and stability; By calculating and updating the gradient of the phase angle, local optimization of the excellent individuals in the global search is achieved, and a configuration of the phase angle parameters of a set of transmitting arrays is obtained.

7. The dynamic cooperative wireless charging system control method based on genetic gradient optimization algorithm according to claim 1 is characterized in that: Step 6: Determine whether to terminate the search and optimization by judging the population diversity index, maximum number of iterations, and fitness value. The specific method is as follows:

1. The upper limit of the number of iterations is reached; 2. No significant improvement in fitness over five consecutive generations; 3. Population diversity index DI<0.

1.

8. A dynamic cooperative wireless charging system based on genetic gradient optimization algorithm, characterized in that: Implementing the dynamic cooperative wireless charging system control method based on the genetic gradient optimization algorithm according to any one of claims 1 to 7 to realize the dynamic cooperative wireless charging system control based on the genetic gradient optimization algorithm, comprising: Module 1: Build an array-type wireless power transmission system coupling unit. The array-type wireless power transmission system coupling unit is composed of a transmitting array of circular spiral coils. Each circular spiral coil is placed separately in a plane to ensure that the number of transmitting coils in the array is the same as the number of devices to be charged; Module 2 performs a global search based on a genetic algorithm. The transmission power and phase angle of the transmitting coils of a group of transmitting arrays are encoded as particles. The fitness level of individuals in the population is evaluated by establishing a fitness function. The population is iterated through roulette wheel selection and adaptive crossover and mutation. Module 3 calculates population diversity. When the population diversity is insufficient, gradient descent is triggered to perform local deep optimization of the optimal individual. Otherwise, the genetic algorithm continues to perform global search. Module 4 selects the top 10% of the best individuals in the current genetic iteration, uses the phase angle parameters of these individuals as particles for gradient calculation and update, and achieves local optimization of the best individuals in the global search, obtaining a set of phase angle parameter configurations for the emission array. Module 5: Combine the phase angle parameters of the individuals that have passed the gradient descent algorithm with the power parameters of the individuals in the previous genetic algorithm to form a new population. Then sort the individuals in the new population by fitness, retain the top 10% of the fitness, and continue to enter the genetic algorithm iteration to start a new cycle. Module 6 determines whether to terminate the search and optimization by judging the population diversity index, the maximum number of iterations and the fitness value.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for controlling a dynamic cooperative wireless charging system based on a genetic gradient optimization algorithm according to any one of claims 1 to 7 is implemented to realize control of a dynamic cooperative wireless charging system based on a genetic gradient optimization algorithm.

10. A computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for controlling a dynamic cooperative wireless charging system based on a genetic gradient optimization algorithm according to any one of claims 1 to 7 is implemented to realize control of a dynamic cooperative wireless charging system based on a genetic gradient optimization algorithm.

Images