A multi-objective optimization design method for solenoid valve structural parameters

Abstract

A multi-objective optimization design method for the structural parameters of a solenoid valve, steps: establishing a finite element simulation model of the solenoid valve; selecting the optimization parameters and optimization objectives of the solenoid valve, and determining the multi-objective optimization model of the solenoid valve; constructing a relationship model between the optimization parameters and optimization objectives of the solenoid valve by using an artificial neural network. It includes constructing a simulation experiment database and training the artificial neural network by using the simulation database to construct a relationship model between the optimization parameters and optimization objectives of the solenoid valve; performing multi-objective optimization on the structure of the solenoid valve by using a genetic algorithm. It includes constructing a fitness function of the genetic algorithm, using the genetic algorithm to find the optimal solution set, and constructing a decision function to make a decision on the optimal solution in the optimal solution set. The present invention takes the position of the magnetic isolation ring, the length of the moving iron core and the number of turns of the coil of the solenoid valve as the optimization parameters, and takes shortening the response time of the solenoid valve and increasing the steady-state electromagnetic force of the solenoid valve as the optimization objectives, and finally realizes the multi-objective optimization of the structural parameters of the solenoid valve, providing a method for the multi-objective optimization design of the structural parameters of the solenoid valve.

Description

Technical Field

[0001] The present invention belongs to the field of solenoid valve optimization methods, and relates to a multi-objective optimization design method for the structural parameters of solenoid valves. Background Technique

[0002] A solenoid valve is an industrial device controlled by electricity magnetism. It is a basic automation component for controlling fluids and belongs to an actuator. It is an important industrial product directly related to the safe production of enterprises and plays an indispensable role in many fields such as automobile manufacturing and industrial control. The performance of the solenoid valve directly affects the reliability of industrial equipment and affects our production and life. It is of great significance to conduct multi-objective optimization design on the solenoid valve to design a high-performance solenoid valve.

[0003] At present, the research at home and abroad mainly focuses on the optimization analysis and research of single parameters of solenoid valves. However, a solenoid valve is a non-linear system with multi-physical field coupling, having a complex structure and the parameters influencing each other. Optimizing only a single parameter will lead to an unsatisfactory optimization effect. In addition, during the multi-objective optimization design process, it is necessary to evaluate the performance of solenoid valves with a large number of different structural parameters. If only the finite element simulation method is used for evaluation, due to the large computational amount for dynamic simulation of solenoid valves, it will consume a lot of time, with a long optimization cycle and low efficiency.

[0004] There is less research on the multi-objective optimization design method for the structural parameters of solenoid valves at home and abroad, and the existing methods have the following problems. Patent 201910594428.6 provides a multi-objective optimization method for the dynamic response characteristics of high-speed solenoid valves, but the modeling method adopted by this method has low accuracy and a large computational amount, with a long optimization cycle. Patent CN202111405584.7 provides an optimization design method for solenoid valves, but this method can only complete the optimization of a single parameter and cannot consider the mutual influence relationship between multiple parameters. Patent 202110953593.3 provides a motor parameter design method based on multi-objective optimization, but this method is not applicable to the multi-objective optimization of solenoid valves. The purpose of this patent is to achieve the multi-objective optimization design of the solenoid valve structure. Summary of the Invention

[0005] The purpose of the present invention is to provide a multi-objective optimization design method for the structural parameters of solenoid valves to solve the problems raised in the above background technique.

[0006] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0007] A multi-objective optimization design method for the structural parameters of solenoid valves includes the following steps:

[0008] Step 1: Establish a finite element simulation model of the solenoid valve and experimentally verify the reliability of the simulation model.

[0009] According to the structural dimension parameters of the solenoid valve to be optimized, construct a two-dimensional finite element model of the solenoid valve, and verify the reliability of the model through experiments. Among them, the verification of the reliability of the solenoid valve finite element model includes verifying the steady-state electromagnetic force of the solenoid valve using a digital display force gauge and verifying the response time of the solenoid valve using a laser displacement sensor. If the error between the simulation value and the actual value is within 10%, the finite element simulation model is considered reliable; if the error exceeds 10%, the finite element simulation model is considered unreliable, and the model structure and the convergence conditions of the finite element simulation need to be adjusted until the error is reduced within 10%.

[0010] Step 2: Select the optimization parameters and optimization objectives of the solenoid valve, that is, determine the multi-objective optimization model of the solenoid valve.

[0011] According to the optimization parameters and optimization objectives of the solenoid valve, construct the multi-objective optimization model of the solenoid valve as follows:

[0012]

[0013] Among them, T is the response time of the solenoid valve; F is the steady-state electromagnetic force of the moving iron core of the solenoid valve; X1,…,X n are the optimization parameters of the solenoid valve; is the minimum value of the i-th optimization parameter; is the maximum value of the i-th optimization parameter.

[0014] Step 3: Use an artificial neural network to construct a relationship model between the optimization parameters and optimization objectives of the solenoid valve

[0015] 3.1 Construct a simulation experiment database

[0016] On the premise of ensuring the reasonable structure of the solenoid valve, a certain value (the floating value is determined according to experience and actual situation) is used to float the optimization parameters up and down as its value range. A series of different values are taken within the value range of the optimization parameters to conduct several simulation experiments. In each simulation experiment, record the optimization objective values corresponding to each group of optimization parameters to construct a simulation experiment database.

[0017] 3.2 Use the simulation database to train the artificial neural network to construct a relationship model between the optimization parameters and optimization objectives of the solenoid valve

[0018] The artificial neural network mentioned is a BP artificial neural network with a single hidden layer. The number of neurons in the input layer is equal to the number of optimization parameters; the more neurons in the hidden layer, the higher the accuracy of the established model, but the corresponding computational workload will increase; the neurons in the output layer correspond to the optimization objectives of the solenoid valve.

[0019] Select 75% of the data in the simulation database as the training set to train the network. After training, the remaining 25% of the data is used as the test set to test the prediction accuracy of the network. If the error between the network prediction result and the simulation result is within 5%, this network is used as the relationship model between the optimized parameters and the optimization objectives of the solenoid valve. If the error exceeds 5%, the network should be retrained.

[0020] Step 4: Use the genetic algorithm to perform multi-objective optimization on the solenoid valve structure.

[0021] 4.1 Construct the fitness function of the genetic algorithm

[0022] Use the relationship model obtained in Step 3 to construct the fitness function of the genetic algorithm as follows:

[0023]

[0024] where θ i is the fitness of the i-th individual in the population; T i ’ and F i ′ are the predicted values of the steady-state electromagnetic force and response time of the i-th individual in the population, and this predicted value is calculated through the neural network model established in Step 3.

[0025] The smaller the value of the described θ i is, the greater the steady-state electromagnetic force, the shorter the response time, the higher the performance of the corresponding solenoid valve, and the higher the fitness of the individual.

[0026] 4.2 Use the genetic algorithm to find the optimal solution set

[0027] Within the value range of the optimization parameters, use a random algorithm to generate n individuals as the initial population. Calculate the fitness of each individual in the population. And sort the individuals according to the fitness, eliminate the individuals with low fitness, and retain the individuals with high fitness.

[0028] The retained individuals generate a new population through crossover and mutation. Among them, the crossover process is achieved by exchanging a certain optimization parameter value of two individuals, and the mutation process is achieved by selecting an optimization parameter of an individual and replacing it with a new value generated by a random algorithm. The crossover probability and mutation probability can be adjusted according to requirements.

[0029] Repeat the above process until the number of population iterations reaches the preset value, and output the finally generated population as the optimal solution set.

[0030] 4.3 Construct a decision function to decide the optimal solution in the optimal solution set

[0031] The decision function constructed according to the optimization objective is:

[0032]

[0033] Wherein: a and b are decision parameters, and the weights of the electromagnetic force and the response time in the final decision can be adjusted by changing the values of a and b; T0 is the response time of the solenoid valve before optimization, and T i is the response time of the i-th individual in the optimal solution set; F0 is the magnitude of the steady-state electromagnetic force of the solenoid valve before optimization, and F i is the magnitude of the steady-state electromagnetic force of the i-th individual in the optimal solution set.

[0034] Calculate the σ value of each group of parameters in the optimal solution set in Step 4.2, and select the parameter group with the largest σ value as the final result of the multi-objective optimization.

[0035] In summary, the beneficial effects of the present invention are as follows:

[0036] (1) The present invention provides a multi-objective-based optimal design method for the structural parameters of a solenoid valve, which overcomes the limitations of the previous single-objective optimization, takes into account the mutual influence relationship between various parameters during the optimization process, and has better optimization performance compared with the single-objective optimization.

[0037] (2) During the multi-objective optimal design process, it is necessary to evaluate the performance of solenoid valves with a large number of different structural parameters. If only the finite element simulation method is used for evaluation, due to the large computational amount of dynamic simulation of the solenoid valve, it will consume a lot of time. The present invention provides a method for establishing a relationship model between the optimized parameters and the optimization objectives of the solenoid valve, which can quickly predict the steady-state electromagnetic force and the response time of the solenoid valve according to the solenoid valve parameters. Applying this model to the performance evaluation of the solenoid valve in the multi-objective optimization process can avoid the above problems and greatly shorten the optimization cycle. Description of the Drawings

[0038] Figure 1 is a flowchart of a multi-objective-based optimal design method for the structural parameters of a solenoid valve according to the present application.

[0039] Figure 2 is a two-dimensional finite element model diagram of the solenoid valve constructed in the embodiment of the present application.

[0040] Figure 3 is a schematic diagram of the neural network structure constructed in the embodiment of the present application.

[0041] Figure 4 is the optimal solution set found in the embodiment of the present invention.

[0042] Figure 5 is a comparison of the response times of the solenoid valve before and after optimization in the embodiment of the present invention.

[0043] Figure 6This is the comparison of the steady-state electromagnetic force of the solenoid valve before and after optimization in the embodiments of the present invention. Specific embodiments

[0044] The following further elaborates on this application in conjunction with embodiments. It should be understood that the specific embodiments described herein are merely used to explain the present invention and are not used to limit the present invention.

[0045] Figure 1 This is the flow chart of the optimization design method for the structural parameters of the solenoid valve based on multiple objectives according to this application, including the following steps:

[0046] The first step: Establish a finite element simulation model of the solenoid valve and experimentally verify the reliability of the simulation model.

[0047] Construct a two-dimensional finite element model of the solenoid valve according to the structural dimension parameters of the solenoid valve to be optimized. The established model is as Figure 2 shown, where 1 is the magnetic isolation tube, 2 is the magnetic conductive sleeve, 3 is the moving iron core, 4 is the coil, 5 is the magnetic isolation ring, and 6 is the magnetic conductive plate. The moving iron core 3 is installed in the magnetic isolation tube 1, the magnetic isolation tube 1 is installed in the inner hole of the magnetic conductive sleeve 2, the magnetic isolation ring 5 is installed in the corresponding groove of the magnetic conductive sleeve 2, the coil 4 is wound around the magnetic conductive sleeve 2, and the magnetic conductive sleeve 2 is installed inside the magnetic conductive plate 6.

[0048] Use the established finite element model to conduct finite element simulation experiments, and obtain the simulation values of the steady-state electromagnetic force and response time of the solenoid valve through finite element experiments. Then, use a digital display force measuring instrument to measure the actual value of the steady-state electromagnetic force of the solenoid valve, and use a laser displacement sensor to measure the actual value of the response time of the solenoid valve. Calculate the deviation between the simulation values of the steady-state electromagnetic force and response time and the actual values. In this embodiment, the deviation of the electromagnetic force is 6.5%, and the deviation of the response time is 7.8%, indicating that the finite element simulation model is reliable.

[0049] The second step: Select the optimization parameters and optimization objectives of the solenoid valve, that is, determine the multi-objective optimization model of the solenoid valve.

[0050] According to the optimization parameters and optimization objectives of the solenoid valve, construct the multi-objective optimization model of the solenoid valve as follows:

[0051]

[0052] where T is the response time of the solenoid valve; F is the steady-state electromagnetic force of the moving iron core of the solenoid valve; X is the position of the magnetic isolation ring of the solenoid valve, L is the length of the moving iron core of the solenoid valve, and N is the number of turns of the coil of the solenoid valve.

[0053] The third step: Use an artificial neural network to construct a relationship model between the optimization parameters and optimization objectives of the solenoid valve

[0054] 3.1 Construct a simulation experiment database

[0055] The optimized parameters are fluctuated up and down by 10% as their value ranges, and a series of different values are taken within the value ranges of the optimized parameters to conduct several simulation experiments. In each simulation experiment, the optimized target values corresponding to each group of optimized parameters are recorded to construct a simulation experiment database. In this embodiment, a database with a sample size of 125 is constructed through finite element simulation experiments.

[0056] 3.2 Training an artificial neural network using the simulation database to construct a relationship model between the optimized parameters of the solenoid valve and the optimized target

[0057] In this embodiment, the structure of the artificial neural network constructed is as Figure 3 shown. The input layer neurons are 3, corresponding to the position of the magnetic isolation ring, the length of the moving iron core, and the number of turns of the coil respectively; the hidden layer neurons are 6; the output layer neuron is 1, corresponding to the steady-state electromagnetic force of the solenoid valve or the response time of the solenoid valve. Randomly select 93 data in the simulation database as the training set to train the network. After training, the remaining 32 groups of data are used as the test set to test the prediction accuracy of the network. The error between the network prediction result and the simulation result is within 5%, reaching the expected accuracy. This network is used as the relationship model between the optimized parameters of the solenoid valve and the optimized target.

[0058] Step 4: Using the genetic algorithm to conduct multi-objective optimization on the solenoid valve structure.

[0059] 4.1 Constructing the fitness function of the genetic algorithm

[0060] Using the relationship model obtained in the third step to construct the fitness function of the genetic algorithm as follows:

[0061]

[0062] Among them, θ i is the fitness of the i-th individual in the population; T i ’ and F i ′ are the predicted values of the steady-state electromagnetic force and the response time of the i-th individual in the population, and this predicted value is calculated through the neural network model established in step three. The smaller the θ i value, the greater the steady-state electromagnetic force and the shorter the response time corresponding to this individual, the higher the performance of the corresponding solenoid valve, and the higher the fitness of the individual.

[0063] 4.2 Using the genetic algorithm to find the optimal solution set

[0064] Within the value ranges of the optimized parameters, a random algorithm is used to generate 50 individuals as the initial population. The random algorithm used to generate individuals in this embodiment is:

[0065]

[0066] Among them: is the minimum value of the i-th optimization parameter in the individual; is the maximum value of the i-th optimization parameter in the individual. L i is the step size of the i-th optimization parameter in the individual; t is a random number generated by the random function, and the value of the random number is between 0 and and.

[0067] Calculate the fitness of each individual in the population. And sort the individuals according to the fitness, eliminate the individuals with high fitness, and retain the individuals with high fitness. The retained individuals are crossed and mutated to generate a new population. Among them, the crossover process is realized by exchanging a certain optimization parameter value of two individuals, and the mutation process is realized by selecting an optimization parameter of an individual and replacing it with a new value generated by a random algorithm. The crossover probability and mutation probability can be adjusted according to requirements.

[0068] Continuously iterate the population, and use the finally generated population as the optimal solution set, as Figure 4 shown.

[0069] 4.3 Construct a decision function to make a decision on the optimal solution in the optimal solution set

[0070] The decision function constructed according to the optimization objective is:

[0071]

[0072] where: a and b are decision parameters, and the weights of the electromagnetic force and response time in the final decision can be adjusted by changing the values of a and b; T0 is the response time of the solenoid valve before optimization, and T i is the response time of the i-th individual in the optimal solution set; F0 is the steady-state electromagnetic force magnitude of the solenoid valve before optimization, and F i is the steady-state electromagnetic force magnitude of the i-th individual in the optimal solution set. In this embodiment, more attention is paid to the response time of the solenoid valve during the design process, so the coefficient a is taken as 0.7 and the coefficient b is taken as 0.3.

[0073] Calculate the σ value of each group of parameters in the optimal solution set, and select the parameter group with the largest σ value as the final result of the multi-objective optimization. In this embodiment, the optimized position of the magnetic isolation ring is -0.68 mm, the length of the moving iron core is 19.76 mm, and the number of turns of the coil is 3050 turns. The performance comparison of the solenoid valve before and after optimization is as Figure 5 shown.

[0074] The applicant of the present invention has made a detailed description and illustration of the embodiments of the present invention in conjunction with the accompanying drawings. However, those skilled in the art should understand that the above embodiments are only the preferred implementation schemes of the present invention, and the detailed description is only to help readers better understand the spirit of the present invention, rather than a limitation on the protection scope of the present invention. On the contrary, any improvement or modification made based on the spirit of the present invention should fall within the protection scope of the present invention.

Claims

1. A multi-objective-based optimal design method for the structural parameters of a solenoid valve, characterized in that, It includes the following steps: The first step: Establish a finite element simulation model of the solenoid valve and experimentally verify the reliability of the simulation model; According to the structural dimension parameters of the solenoid valve to be optimized, construct a two-dimensional finite element model of the solenoid valve and verify the reliability of the model through experiments; The second step: Select the optimization parameters and optimization objectives of the solenoid valve, that is, determine the multi-objective optimization model of the solenoid valve; According to the optimization parameters and optimization objectives of the solenoid valve, construct the multi-objective optimization model of the solenoid valve as follows: Among them, T is the response time of the solenoid valve; F is the steady-state electromagnetic force of the moving iron core of the solenoid valve; X1, ..., X n are the optimization parameters of the solenoid valve; is the minimum value of the i-th optimization parameter; is the maximum value of the i-th optimization parameter; The third step: Use an artificial neural network to construct a relationship model between the optimization parameters and optimization objectives of the solenoid valve 3.1 Construct a simulation experiment database Take a series of different values within the value range of the optimization parameters to conduct several simulation experiments. Record the optimization objective values corresponding to each group of optimization parameters in each simulation experiment to construct a simulation experiment database; 3.2 Use the simulation database to train the artificial neural network and construct a relationship model between the optimization parameters and optimization objectives of the solenoid valve The artificial neural network mentioned is a BP artificial neural network with a single hidden layer; the number of neurons in the input layer is equal to the number of optimization parameters; the neurons in the output layer correspond to the optimization objectives of the solenoid valve; select 75% of the data in the simulation database as the training set to train the network. After training, the remaining 25% of the data is used as the test set to test the prediction accuracy of the network. If the error between the network prediction result and the simulation result is within 5%, then this network is used as the relationship model between the optimization parameters and optimization objectives of the solenoid valve; if the error exceeds 5%, the network should be retrained; The fourth step: Use the genetic algorithm to perform multi-objective optimization on the solenoid valve structure; 4.1 Construct the fitness function of the genetic algorithm Use the relationship model obtained in the third step to construct the fitness function of the genetic algorithm as follows: Among them, θ i is the fitness of the i-th individual in the population; T i ’ and F i ′ are the predicted values of the steady-state electromagnetic force and response time of the i-th individual in the population, and this predicted value is calculated by the neural network model established in step three; The described θ i The smaller the value, the greater the steady-state electromagnetic force corresponding to the individual, the shorter the response time, the higher the performance of the corresponding solenoid valve, and the higher the fitness of the individual; 4.2 Use the genetic algorithm to find the optimal solution set Within the value range of the optimization parameters, generate n individuals as the initial population, calculate the fitness of each individual in the population; and sort the individuals according to the fitness, eliminate the individuals with low individual fitness, and retain the individuals with high individual fitness; the retained individuals generate a new population through crossover and mutation; repeat the above process until the population iteration number reaches the preset value, and output the finally generated population as the optimal solution set; 4.3 Construct a decision function and make a decision on the optimal solution in the optimal solution set The decision function constructed according to the optimization objective is: Where: a and b are decision parameters, and the weights of electromagnetic force and response time in the final decision are adjusted by changing the values of a and b; T0 is the response time of the solenoid valve before optimization, and T i is the response time of the i-th individual in the optimal solution set; F0 is the steady-state electromagnetic force of the solenoid valve before optimization, and F i is the steady-state electromagnetic force of the i-th individual in the optimal solution set; Calculate the σ value of each group of parameters in the optimal solution set in step 4.2, and select the parameter group with the largest σ value as the final result of multi-objective optimization.

2. The multi-objective-based optimal design method for the structural parameters of a solenoid valve according to claim 1, characterized in that, In the first step mentioned above, the verification of the reliability of the solenoid valve finite element model includes verifying the steady-state electromagnetic force of the solenoid valve using a digital display force gauge and verifying the response time of the solenoid valve using a laser displacement sensor. If the error between the simulation value and the actual value is within 10%, it is considered that the finite element simulation model is reliable; if the error exceeds 10%, it is considered that the finite element simulation model is unreliable, and the model structure and the convergence conditions of the finite element simulation need to be adjusted until the error is reduced within 10%.

3. The multi-objective-based optimal design method for the structural parameters of a solenoid valve according to claim 1, characterized in that, In the step 4.2, the crossover process is achieved by swapping a certain optimization parameter value of two individuals, and the mutation process is achieved by generating a new value to replace an optimization parameter of an individual after selection; the crossover probability and the mutation probability are adjusted according to requirements.

Images