A method and device for designing and optimizing structure parameters of a solenoid valve

Abstract

The application discloses a kind of design optimization method and device of solenoid valve structure parameter, belong to modeling simulation field.Method includes: S1, the multidimensional design space of solenoid valve is sampled and simulated, obtains sample point set and performance dataset;S2, according to the sample point set and performance dataset of current iteration, constructs surrogate model;S3, according to the expected improvement of candidate sample point to surrogate model, and according to result, genetic optimization is carried out, obtains the new sample point of maximum expected improvement value;S4, simulation is carried out to new sample point, obtains corresponding real performance index value, and candidate sample point set and performance dataset are updated in turn;S5, whether current iteration is converged according to the real performance index value obtained by iteration, if yes, then end iteration and output optimal solution set, otherwise, steps S2-S5 are executed cyclically until the preset convergence condition is met.The application can solve the problems of long cycle, high cost and difficulty to find global optimum of traditional solenoid valve design method.

Description

Technical Field

[0001] This invention relates to the field of engineering modeling and simulation technology, and in particular to a method and apparatus for designing and optimizing the structural parameters of a solenoid valve. Background Technology

[0002] As a critical fluid control component, the performance of a solenoid valve (such as dynamic response speed, electromagnetic force, volumetric mass, and power consumption) directly determines the efficiency of the entire system. Solenoid valve structural design is the engineering design work that determines the shape, size, material, relative position, and assembly relationship of the core components of the solenoid valve. The core objective is to enable the solenoid valve to accurately control fluid flow while meeting key performance requirements such as pull-in speed, electromagnetic force, mass, power consumption, and sealing performance. This is the core step in transforming a solenoid valve from concept to a workable drawing.

[0003] Traditional solenoid valve design relies mainly on engineers' experience and a "design-simulation-trial and error" cycle. However, this approach not only has a long design cycle and makes it difficult to obtain the global optimal solution, but also requires tens of thousands of simulation calls for simulation-based parameter scanning or optimization algorithms, resulting in optimization processes that take weeks and have high computational costs, making it unsuitable for the needs of rapid product development.

[0004] Therefore, there is an urgent need for a design optimization method and device for the structural parameters of solenoid valves to solve the above-mentioned technical problems. Summary of the Invention

[0005] This invention provides a method and apparatus for designing and optimizing the structural parameters of a solenoid valve, which can solve the problems of long design cycles, high costs, and difficulty in finding the global optimum in traditional solenoid valve design methods. The technical solution is as follows: On the one hand, a method for designing and optimizing the structural parameters of a solenoid valve is provided, the method comprising: S1. Perform Latin hypercube sampling and finite element simulation on the multidimensional design space of the solenoid valve to obtain a set of sample points and a performance dataset corresponding to all sample points; wherein, the multidimensional design space is determined according to the design variables and the range of variable values ​​of the solenoid valve, and the design variables are determined by parametric modeling of the key structure of the solenoid valve. S2. Construct a Gaussian process regression surrogate model based on the sample point set and performance dataset of the current iteration; S3. Based on the surrogate model, perform expected improvement calculations on the candidate sample points generated by the Latin hypercube in the current iteration, and perform genetic optimization based on the calculation results to obtain the new sample point with the largest expected improvement value. S4. Perform the finite element simulation on the new sample points to obtain the true performance index values ​​corresponding to the new sample points, and update the candidate sample point set and the corresponding performance dataset in sequence according to the simulation results. S5. Determine whether the current iteration has converged based on the actual performance index values ​​obtained from the current iteration and the previous iteration. If so, end the iteration and output the updated candidate sample point set and performance dataset as the optimal solution set for the structural design. Otherwise, repeat steps S2-S5 until the preset convergence condition is met.

[0006] On the other hand, a device for designing and optimizing the structural parameters of a solenoid valve is provided, the device comprising: The sampling module is used to perform Latin hypercube sampling and finite element simulation on the multidimensional design space of the solenoid valve to obtain a set of sample points and a performance dataset corresponding to all sample points; wherein, the multidimensional design space is determined according to the design variables and the range of variable values ​​of the solenoid valve, and the design variables are determined by parametric modeling of the key structures of the solenoid valve. The modeling module is used to construct a Gaussian process regression surrogate model based on the sample point set and performance dataset of the current iteration; The calculation module is used to calculate the expected improvement of the candidate sample points generated by the Latin hypercube in the current iteration according to the surrogate model, and to perform genetic optimization based on the calculation results to obtain the new sample point with the largest expected improvement value. The simulation module is used to perform the finite element simulation on the new sample points to obtain the real performance index values ​​corresponding to the new sample points, and to update the candidate sample point set and the corresponding performance dataset in sequence according to the simulation results. The convergence module is used to determine whether the current iteration has converged based on the actual performance index values ​​obtained from the current iteration and the previous iteration. If it has converged, the iteration ends and the updated candidate sample point set and performance dataset are output as the optimal solution set for the structural design. Otherwise, the functions of the modeling module and the convergence module are executed repeatedly until the preset convergence condition is met.

[0007] On the other hand, a computer device is provided, the computer device including a memory and a processor, the memory for storing computer programs, and the processor for executing the computer programs stored in the memory to implement the steps of the above-described method for designing and optimizing the structural parameters of the solenoid valve.

[0008] On the other hand, a computer-readable storage medium is provided, wherein a computer program is stored therein, and when the computer program is executed by a processor, the steps of the above-described method for designing and optimizing the structural parameters of the solenoid valve are implemented.

[0009] On the other hand, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the above-described method for designing and optimizing the structural parameters of a solenoid valve.

[0010] The technical solution provided by this invention offers at least the following beneficial effects: First, through parametric modeling and finite-sampling simulation, a surrogate model capable of rapid performance evaluation is constructed, laying the foundation for efficient analysis. Next, utilizing this model and the desired improvement criterion, the most promising new design point can be intelligently located, thus avoiding the massive simulations required in traditional optimization and shortening the design cycle from weeks to hours. Then, through a closed-loop iteration of "simulation verification - model update," the optimization process is ensured to converge towards the true optimal direction. Finally, by outputting the optimal solution set that satisfies the convergence condition, the trade-offs between multiple objectives such as absorption time, mass, and power are intuitively displayed, enabling designers to select the optimal solution that meets specific requirements based on clear criteria. The entire process, while ensuring or even improving performance, significantly reduces reliance on expert experience, achieving global, automatic, and efficient optimization. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart of a method for designing and optimizing the structural parameters of a solenoid valve according to an embodiment of the present invention; Figure 2 This is a schematic diagram of sample points in a multidimensional design space provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the three-dimensional Pareto front distribution at the time of optimization termination provided in an embodiment of the present invention; Figure 4 This is a structural diagram of a device for designing and optimizing the structural parameters of a solenoid valve according to an embodiment of the present invention; Figure 5 This is a hardware architecture diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0014] Please refer to Figure 1The present invention provides a method for designing and optimizing the structural parameters of a solenoid valve, the method comprising: Step S1: Perform Latin hypercube sampling and finite element simulation on the multidimensional design space of the solenoid valve to obtain a set of sample points and a performance dataset corresponding to all sample points; wherein, the multidimensional design space is determined according to the design variables and the range of variable values ​​of the solenoid valve, and the design variables are determined by parametric modeling of the key structure of the solenoid valve. Step S2: Construct a Gaussian process regression surrogate model based on the current set of sample points and performance dataset; Step S3: Calculate the expected improvement of the candidate sample points generated by the Latin hypercube in the current iteration according to the surrogate model, and perform genetic optimization based on the calculation results to obtain the new sample point with the largest expected improvement value. Step S4: Perform the finite element simulation on the new sample points to obtain the true performance index values ​​corresponding to the new sample points, and update the candidate sample point set and the corresponding performance dataset in sequence according to the simulation results. Step S5: Determine whether the current iteration has converged based on the actual performance index values ​​obtained from the current iteration and the previous iteration. If so, end the iteration and output the updated candidate sample point set and performance dataset as the optimal solution set for the structural design. Otherwise, repeat steps S2-S5 until the preset convergence condition is met.

[0015] In this embodiment of the invention, a surrogate model for rapid performance evaluation is first constructed through parametric modeling and finite-sampling simulations, laying the foundation for efficient analysis. Then, using this model and the desired improvement criterion, the most promising new design point can be intelligently located, thus avoiding the massive simulations required in traditional optimization and shortening the design cycle from weeks to hours. Next, a closed-loop iteration of "simulation verification - model update" ensures that the optimization process converges towards the true optimal direction. Finally, the optimal solution set satisfying the convergence condition is output, intuitively demonstrating the trade-offs between multiple objectives such as absorption time, mass, and power, enabling designers to select the optimal solution that meets specific requirements based on clear criteria. The entire process, while ensuring or even improving performance, significantly reduces reliance on expert experience, achieving global, automatic, and efficient optimization.

[0016] The following description Figure 1 The execution method of each step is shown.

[0017] First, for step S1, Latin hypercube sampling and finite element simulation are performed on the multidimensional design space of the solenoid valve to obtain the sample point set and the performance dataset corresponding to all sample points.

[0018] In this embodiment of the invention, the sample point set and the performance dataset corresponding to all sample points are obtained as follows: random and uniform Latin hypercube sampling is performed in the multidimensional design space to obtain the sample point set; finite element transient multiphysics coupling is performed on each sample point in the initial sample point set to obtain the key performance indicators of each sample point; wherein, the key performance indicators include pull-in time, total mass and maximum power; the key performance indicators of all sample points are summarized to generate the corresponding performance dataset.

[0019] Specifically, the key structures of the solenoid valve (such as the armature, valve body, and coil) are parametrically modeled to determine... N Each variable is a performance-sensitive design variable (coil turns n (500~1500), wire diameter d_wire (0.1mm~1mm), core diameter d_core (10mm~20mm), coil length L (15mm~30mm), valve port flow diameter d_f (2.5mm~4mm), etc.), and the value range of each variable is set, which together constitute an N-dimensional design space.

[0020] Using the Latin hypercube sampling method, (11) samples were selected within the design space. N -1) an initial sample point dataset X (with good spatial distribution representativeness) ),like Figure 2 As shown.

[0021] For each sample point, a finite element transient multiphysics coupling simulation was performed to calculate a series of performance index values, including: pull-in time T1, mass M, and maximum power P, forming an initial performance dataset Y. ).

[0022] Then, for step S2, a Gaussian process regression surrogate model is constructed based on the sample point set and performance dataset of the current iteration.

[0023] In this embodiment of the invention, the Gaussian process regression surrogate model is constructed as follows: the sample point set is used as input and the performance dataset after logarithmic transformation is used as output to train a preset initial model, resulting in three mutually independent Gaussian process regression surrogate models.

[0024] Specifically, to ensure positive predictions, the three performance target values ​​(T1, M, P) are first logarithmically transformed. Then, using the sample point dataset as input and the transformed performance target values ​​as outputs, three Gaussian process regression models are trained independently.

[0025] For example, based on the sample point dataset X, the set of objective functions for absorption time Quality objective function set Power objective function set Three Gaussian process regression (GPR) models were constructed respectively. .

[0026] The GPR model not only provides the predicted mean but also the predicted uncertainty (variance), a characteristic that lays the foundation for subsequent sampling strategies based on "expectation improvement." Three trained Gaussian process regression surrogate models were used to predict the logarithm of the pull-in time, mass, and power, as well as their uncertainties, for any combination of design variables.

[0027] For step S3, the expected improvement of the candidate sample points generated by the Latin hypercube in the current iteration is calculated according to the surrogate model, and genetic optimization is performed based on the calculation results to obtain the new sample point with the largest expected improvement value.

[0028] In this embodiment of the invention, new sample points are obtained by re-sampling Latin hypercube within the multidimensional design space to obtain a candidate sample set; Perform multi-objective Pareto front analysis on the performance dataset of the current iteration to obtain a front point set containing multiple optimal solutions, and calculate a reference point to characterize the performance level that is inferior to the current optimal boundary based on the front point set. Each candidate sample point in the candidate sample set is input into the Gaussian process regression surrogate model, and the predicted values ​​of three performance indicators for each candidate sample point in the original target space are output; wherein, the predicted performance indicator values ​​include the performance prediction mean and the performance prediction variance; Based on the reference point and the predicted performance index values, the first expected improvement values ​​of the three performance indices of the candidate sample points relative to the reference point are calculated. Based on the current set of leading edge points and the first expected improvement value, the second expected improvement value of the candidate sample point relative to the set of leading edge points is calculated; Using the initial candidate sample set as the initial population and the second expected improvement value as the fitness function, genetic optimization is performed in the multidimensional design space to obtain new sample points that maximize the second expected improvement value.

[0029] Specifically, firstly, through the Latin hypercube design, in N Randomly generated in 3D design space A uniformly distributed sample point.

[0030] Next, a multi-objective Pareto front analysis is performed on the performance dataset to compare the performance of each design scheme on the three objectives (T1, M, P). The non-dominated solutions that are better on at least one objective and are not dominated on the other objectives are identified, thus obtaining a set containing m optimal solutions (m ≤ 11N-1).P f ( It visually demonstrates the optimal trade-offs that can be achieved between multiple conflicting performance metrics in the initial design space.

[0031] Next, based on the current cutting-edge P f The reference point Ref_point is dynamically set. Each coordinate value of this reference point is taken as 1.1 times the maximum value of the corresponding objective function on the current Pareto front, that is: Ref_point= This reference point represents a performance level slightly worse than the current optimal boundary.

[0032] For each sample point in the candidate population, its design variables are input into the three Gaussian process regression surrogate models M1, M2, and M3 constructed in step one. Each model outputs the predicted mean of its corresponding performance index (absorption time, mass, power) in logarithmic space. and prediction variance Then, an inverse logarithmic transform is performed to obtain the predicted mean and variance of the original objective function space.

[0033] Specifically The first expected improvement value was then calculated using the following formula. : In the formula, The mean of the performance predictions; The variance of the performance prediction; The reference point is mentioned above; is the cumulative probability distribution function of the standard normal distribution; The standard normal distribution probability density function is used; i=1, 2, and 3 are three performance indicators.

[0034] The second expected improvement value is calculated using the following formula. : In the formula, , , These are the three performance metrics of the candidate sample points relative to the frontier point set. a Expected improvement value at each frontier point , where m is the total number of front points in the front point set.

[0035] Finally, using the aforementioned candidate sample points as the initial population, and the calculated second expected improvement value as the fitness function (which needs to be maximized), the genetic algorithm is run to iteratively evolve the population through selection, crossover, mutation, and other operations, searching for points in the continuous design space that maximize the second expected improvement value.

[0036] This point is considered the design scheme most likely to significantly improve the Pareto frontier in this iteration and will be sent to the next step for high-precision simulation verification.

[0037] For step S4, the finite element simulation process is performed on the new sample points to obtain the true performance index values ​​corresponding to the new sample points, and the candidate sample point set and the corresponding performance dataset are updated sequentially according to the simulation results.

[0038] The new sample points were simulated using finite element transient multiphysics coupling simulation with the same accuracy level as in step S1. Simulations were performed to obtain the corresponding real performance indicators. The new sample points and performance metrics are updated and saved to the end of the corresponding set.

[0039] For step S5, determine whether the current iteration has converged based on the actual performance index values ​​obtained from the current iteration and the previous iteration. If it has, end the iteration and output the updated candidate sample point set and performance dataset as the optimal solution set for the structural design. Otherwise, repeat steps S2-S5 until the preset convergence condition is met.

[0040] Specifically, Compare the results with the last row of the performance set before the update, and calculate the relative rate of change for each objective function. ,like (Where i = 1, 2, 3), optimization terminates. The rate of change is calculated as follows: In the formula, The change rate is denoted as .

[0041] If the rate of change for all three objectives is less than the threshold of 0.05, the optimization process is considered to have converged, and there is no further significant room for improvement. At this point, the optimization loop terminates, and the optimization result is as follows: Figure 3 As shown, the process proceeds to the final decision.

[0042] If any rate of change is greater than or equal to 0.05, the optimization is deemed to still have potential for improvement. At this point, the process returns to step S2, using the latest updated surrogate model and Pareto front to begin the next iteration of "expected improvement-based optimization," searching for the next potentially promising sample point. .

[0043] Please refer to Figure 4 This invention provides a device for designing and optimizing the structural parameters of a solenoid valve, the device comprising: The sampling module 400 is used to perform Latin hypercube sampling and finite element simulation on the multidimensional design space of the solenoid valve to obtain a set of sample points and a performance dataset corresponding to all sample points; wherein, the multidimensional design space is determined according to the design variables and the range of variable values ​​of the solenoid valve, and the design variables are determined by parametric modeling of the key structure of the solenoid valve. Modeling module 402 is used to construct a Gaussian process regression surrogate model based on the sample point set and performance dataset of the current iteration; The calculation module 404 is used to perform expected improvement calculation on the candidate sample points generated by the Latin hypercube in the current iteration according to the surrogate model, and perform genetic optimization according to the calculation results to obtain the new sample point with the largest expected improvement value. The simulation module 406 is used to perform the finite element simulation processing on the new sample points to obtain the real performance index values ​​corresponding to the new sample points, and update the candidate sample point set and the corresponding performance dataset in sequence according to the simulation results. The convergence module 408 is used to determine whether the current iteration has converged based on the actual performance index values ​​obtained from the current iteration and the previous iteration. If it has converged, the iteration ends and the updated candidate sample point set and performance dataset are output as the optimal solution set for the structural design. Otherwise, the functions of the modeling module to the convergence module are executed repeatedly until the preset convergence condition is met.

[0044] In this embodiment of the invention, the Latin hypercube sampling and finite element simulation processing of the multidimensional design space of the solenoid valve to obtain a sample point set and a performance dataset corresponding to all sample points includes: Random and uniform Latin hypercube sampling is performed within the multidimensional design space to obtain a sample point set; For each sample point in the initial sample point set, a finite element transient multiphysics coupling method is applied to obtain the key performance indicators for each sample point; wherein, the key performance indicators include pull-in time, total mass, and maximum power; The key performance indicators of all sample points are summarized to generate the corresponding performance dataset.

[0045] In this embodiment of the invention, the Gaussian process regression surrogate model is constructed in the following manner: The sample point set is used as input and the performance dataset after logarithmic transformation is used as output to train the preset initial model, resulting in three independent Gaussian process regression surrogate models.

[0046] In this embodiment of the invention, the step of calculating the expected improvement of candidate sample points generated by the Latin hypercube in the current iteration according to the surrogate model, and performing genetic optimization based on the calculation results to obtain a new sample point with the largest expected improvement value, includes: Latin hypercube sampling is performed again within the multidimensional design space to obtain a candidate sample set; Perform multi-objective Pareto front analysis on the performance dataset of the current iteration to obtain a front point set containing multiple optimal solutions, and calculate a reference point to characterize the performance level that is inferior to the current optimal boundary based on the front point set. Each candidate sample point in the candidate sample set is input into the Gaussian process regression surrogate model, and the predicted values ​​of three performance indicators for each candidate sample point in the original target space are output; wherein, the predicted performance indicator values ​​include the performance prediction mean and the performance prediction variance; Based on the reference point and the predicted performance index values, the first expected improvement values ​​of the three performance indices of the candidate sample points relative to the reference point are calculated. Based on the current set of leading edge points and the first expected improvement value, the second expected improvement value of the candidate sample point relative to the set of leading edge points is calculated; Using the initial candidate sample set as the initial population and the second expected improvement value as the fitness function, genetic optimization is performed in the multidimensional design space to obtain new sample points that maximize the second expected improvement value.

[0047] In this embodiment of the invention, the first desired improvement value It is calculated using the following formula: In the formula, The mean of the performance predictions; The variance of the performance prediction; The reference point is mentioned above; is the cumulative probability distribution function of the standard normal distribution; The standard normal distribution probability density function is used; i=1, 2, and 3 are three performance indicators.

[0048] In this embodiment of the invention, the second desired improvement value It is calculated using the following formula: In the formula, , , These are the three performance metrics of the candidate sample points relative to the frontier point set. a Expected improvement value at each frontier point , where m is the total number of front points in the front point set.

[0049] It should be noted that the solenoid valve structural parameter design optimization device provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the solenoid valve structural parameter design optimization device and the solenoid valve structural parameter design optimization method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process is detailed in the method embodiments, which will not be repeated here.

[0050] Embodiments of this application also provide a computer device, please refer to... Figure 5 The computer device includes a processor and a memory, the memory storing at least one instruction, at least one program, code set or instruction set, the at least one instruction, at least one program, code set or instruction set being loaded and executed by the processor to implement the design optimization method for the solenoid valve structural parameters provided in the above-described method embodiments.

[0051] Embodiments of this application also provide a computer-readable storage medium storing at least one instruction, at least one program, code set, or instruction set, wherein the at least one instruction, at least one program, code set, or instruction set is loaded and executed by a processor to implement the design optimization method for the solenoid valve structural parameters provided in the above-described method embodiments.

[0052] Embodiments of this application also provide a computer program product, which includes a computer program. A processor of a computer device reads the computer program from a computer-readable storage medium and executes the computer program, causing the computer device to perform the design optimization method for the solenoid valve structural parameters described in any of the above embodiments.

[0053] For ease of description, the above systems or devices are described separately as various modules or units based on their functions. Of course, in implementing this application, the functions of each unit can be implemented in one or more software and / or hardware components.

[0054] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0055] Finally, it should be noted that in this document, relational terms such as first, second, third, and fourth are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0056] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for designing and optimizing the structural parameters of a solenoid valve, characterized in that, The method includes: S1. Perform Latin hypercube sampling and finite element simulation on the multidimensional design space of the solenoid valve to obtain a set of sample points and a performance dataset corresponding to all sample points; wherein, the multidimensional design space is determined according to the design variables and the range of variable values ​​of the solenoid valve, and the design variables are determined by parametric modeling of the key structure of the solenoid valve. S2. Construct a Gaussian process regression surrogate model based on the sample point set and performance dataset of the current iteration; S3. Based on the surrogate model, perform expected improvement calculations on the candidate sample points generated by the Latin hypercube in the current iteration, and perform genetic optimization based on the calculation results to obtain the new sample point with the largest expected improvement value. S4. Perform the finite element simulation on the new sample points to obtain the true performance index values ​​corresponding to the new sample points, and update the candidate sample point set and the corresponding performance dataset in sequence according to the simulation results. S5. Determine whether the current iteration has converged based on the actual performance index values ​​obtained from the current iteration and the previous iteration. If so, end the iteration and output the updated candidate sample point set and performance dataset as the optimal solution set for the structural design. Otherwise, repeat steps S2-S5 until the preset convergence condition is met.

2. The method as described in claim 1, characterized in that, The process of performing Latin hypercube sampling and finite element simulation on the multidimensional design space of the solenoid valve yields a set of sample points and a performance dataset corresponding to all sample points, including: Random and uniform Latin hypercube sampling is performed within the multidimensional design space to obtain a sample point set; A finite element transient multiphysics coupling method is applied to each sample point in the sample point set to obtain the key performance indicators for each sample point; wherein, the key performance indicators include pull-in time, total mass and maximum power; The key performance indicators of all sample points are summarized to generate the corresponding performance dataset.

3. The method as described in claim 1, characterized in that, The Gaussian process regression surrogate model is constructed in the following manner: The sample point set is used as input and the performance dataset after logarithmic transformation is used as output to train the preset initial model, resulting in three independent Gaussian process regression surrogate models.

4. The method as described in claim 2, characterized in that, The step of calculating the expected improvement of candidate sample points generated by the Latin hypercube in the current iteration according to the surrogate model, and performing genetic optimization based on the calculation results to obtain a new sample point with the largest expected improvement value, includes: Latin hypercube sampling is performed again within the multidimensional design space to obtain a candidate sample set; Perform multi-objective Pareto front analysis on the performance dataset of the current iteration to obtain a front point set containing multiple optimal solutions, and calculate a reference point to characterize the performance level that is inferior to the current optimal boundary based on the front point set. Each candidate sample point in the candidate sample set is input into the Gaussian process regression surrogate model, and the predicted values ​​of three performance indicators for each candidate sample point in the original target space are output; wherein, the predicted performance indicator values ​​include the performance prediction mean and the performance prediction variance; Based on the reference point and the predicted performance index values, the first expected improvement values ​​of the three performance indices of the candidate sample points relative to the reference point are calculated. Based on the current set of leading edge points and the first expected improvement value, the second expected improvement value of the candidate sample point relative to the set of leading edge points is calculated; Using the initial candidate sample set as the initial population and the second expected improvement value as the fitness function, genetic optimization is performed in the multidimensional design space to obtain new sample points that maximize the second expected improvement value.

5. The method as described in claim 4, characterized in that, First expected improvement value It is calculated using the following formula: In the formula, The mean of the performance predictions; The variance of the performance prediction; The reference point is mentioned above; is the cumulative probability distribution function of the standard normal distribution; The standard normal distribution probability density function is used; i=1, 2, and 3 are three performance indicators.

6. The method as described in claim 4, characterized in that, Second expected improvement value It is calculated using the following formula: In the formula, , , These are the three performance metrics of the candidate sample points relative to the frontier point set. a Expected improvement value at each frontier point , where m is the total number of front points in the front point set.

7. A device for designing and optimizing the structural parameters of a solenoid valve, characterized in that, The device includes: The sampling module is used to perform Latin hypercube sampling and finite element simulation on the multidimensional design space of the solenoid valve to obtain a set of sample points and a performance dataset corresponding to all sample points; wherein, the multidimensional design space is determined according to the design variables and the range of variable values ​​of the solenoid valve, and the design variables are determined by parametric modeling of the key structures of the solenoid valve. The modeling module is used to construct a Gaussian process regression surrogate model based on the sample point set and performance dataset of the current iteration; The calculation module is used to calculate the expected improvement of the candidate sample points generated by the Latin hypercube in the current iteration according to the surrogate model, and to perform genetic optimization based on the calculation results to obtain the new sample point with the largest expected improvement value. The simulation module is used to perform the finite element simulation on the new sample points to obtain the real performance index values ​​corresponding to the new sample points, and to update the candidate sample point set and the corresponding performance dataset in sequence according to the simulation results. The convergence module is used to determine whether the current iteration has converged based on the actual performance index values ​​obtained from the current iteration and the previous iteration. If it has converged, the iteration ends and the updated candidate sample point set and performance dataset are output as the optimal solution set for the structural design. Otherwise, the functions of the modeling module and the convergence module are executed repeatedly until the preset convergence condition is met.

8. A computer device, characterized in that, The computer device includes a memory and a processor. The memory is used to store computer programs, and the processor is used to execute the computer programs stored in the memory to implement the steps of the method according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the steps of the method described in any one of claims 1-6.

10. A computer program product, characterized in that, Includes a computer program, which, when executed by a processor, implements the steps of the method according to any one of claims 1-6.

Images