Inverse optimization method for solenoid valve parameters based on performance and correlation weight iteration
By using a performance and correlation weighted iterative method, and leveraging a positive performance prediction model and iterative search, the problems of high computational cost, reliance on human experience, and multiple solutions in the optimization of solenoid valve structural parameters are solved. This achieves efficient and automated multi-parameter optimization, outputting a combination of structural parameters that meets engineering constraints.
Patent Information
- Application Number
- CN202611244639.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-17
- Publication Date
- 2026-11-10
AI Technical Summary
Existing technologies have high computational costs, rely heavily on human experience, lack differentiated adjustments when optimizing multiple parameters, and produce non-unique and unreasonable reverse prediction results, making it difficult to meet engineering constraints.
A performance- and correlation-weighted iterative approach is adopted. By combining a positive performance prediction model with iterative search, a mapping relationship between structural parameters and traffic performance is established. Correlation weights and error feedback are used to update parameters to meet target traffic performance and engineering constraints.
It reduces computational costs, improves optimization efficiency and automation, reduces the number of multiple solutions, and ensures the rationality and manufacturability of optimization results.
Smart Images

Figure CN123069880A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the technical field of electromagnetic valve structural parameter optimization methods, specifically involving a reverse optimization method for electromagnetic valve parameters based on performance and correlation weight iteration. Background Technology
[0002] A solenoid valve is an actuator that uses electromagnetic force to control the on / off state, flow direction, and flow rate of fluid. It is widely used in hydraulic control, pneumatic control, automotive, aerospace, energy equipment, and industrial automation. With the increasing demands for fluid control accuracy and operational stability in related equipment, the flow performance of solenoid valves under different pressure or differential pressure conditions has become a crucial indicator in structural design. The flow performance of a solenoid valve is closely related to its internal flow channels, valve core structure, valve port clearance, and structural parameters such as step height, column height, and column base platform height.
[0003] In the design of solenoid valves, computational fluid dynamics (CFD) numerical simulation is an important means of obtaining structural performance data. By establishing fluid domain models under different combinations of structural parameters, setting working conditions such as pressure, and solving the flow field, corresponding flow performance data can be obtained. Further combining existing product data, experimental data, or numerical simulation data, a sample set consisting of structural parameters, pressure conditions, and corresponding flow rates can be formed. Then, data-driven methods such as machine learning can be used to establish a positive predictive relationship between structural parameters and flow performance. A patent with publication number CN111767663A discloses a convenient method for calculating valve flow coefficients based on CFD simulation. The main steps include: establishing flow field models of upstream pipes, valves, and downstream pipes based on a three-dimensional model of the valve assembly; dividing and optimizing the flow field model mesh; applying the re-optimized mesh to set boundary conditions for simulation calculations; collecting average pressure values from multiple cross-sections of the upstream pipe; determining the linear fitting order based on the axial pressure distribution diagram and performing linear fitting; combining the total pressure difference of the pipe with the valve obtained from the simulation calculation results; and calculating the valve flow coefficient using the valve pressure difference. Although the CFD simulation-based method for optimizing the structural parameters of solenoid valves can accurately obtain the flow performance under different structural schemes, this method still has the following shortcomings:
[0004] (1) The computational cost is high, making it difficult to meet the needs of rapid optimization of multiple parameters. CFD methods usually require rebuilding the geometric model, dividing the computational grid, and solving the flow field for different combinations of structural parameters. When there are many parameters to be optimized or multiple pressure conditions need to be considered at the same time, numerical simulations need to be performed on a large number of candidate structural schemes, resulting in a large number of calculations and a long optimization cycle, which makes it difficult to meet the needs of rapid structural design. The original disclosure also pointed out that if the intelligent optimization algorithm directly calls the numerical simulation model to evaluate each group of candidate structures, it will generate a large computational cost.
[0005] (2) Structural parameter adjustment is highly dependent on human experience. Existing methods usually involve manually modifying relevant structural dimensions based on the deviation between the simulated flow rate and the target flow rate of the current structural scheme. Since there is a nonlinear and parameter coupling relationship between the structural parameters and flow performance of the solenoid valve, the degree and direction of the influence of different parameters on the flow rate vary. Therefore, the parameter adjustment process is easily affected by the designer's experience, making it difficult to form a unified, stable and repeatable structural optimization process.
[0006] (3) Difficulty in fully utilizing the differences in the impact of different structural parameters on flow performance. When multiple structural parameters participate in optimization simultaneously, the degree of influence of each structural parameter on flow performance is not the same. Some parameters are more sensitive to changes in flow, while others have a smaller impact. Existing parameter search or manual adjustment methods usually lack a mechanism for differentiated updates based on the degree and direction of influence of each parameter, which can easily lead to problems such as unreasonable parameter adjustment range, inaccurate adjustment direction, and invalid search. The original disclosure document identified the significant differences in the degree and direction of influence of different parameters as an important feature that needs to be addressed in reverse optimization.
[0007] Therefore, although existing CFD simulation-based methods for optimizing the structural parameters of solenoid valves can obtain relatively accurate flow performance evaluation results, they still have problems such as high computational cost, reliance on manual experience for parameter adjustment, and lack of differentiated optimization for different structural parameters, making it difficult to meet the needs of rapid reverse design and automated optimization of multiple parameters of solenoid valves.
[0008] With the application of machine learning technology in the field of engineering design, existing technologies can use a data-driven approach to establish a direct inverse mapping relationship between the target flow performance of solenoid valves and structural parameters, thereby reducing the computational workload of repeated numerical simulations in the traditional structural optimization process.
[0009] This method first collects structural parameters, pressure conditions, and corresponding flow performance data for different solenoid valves, forming a structural parameter-operating condition-flow sample set. Structural parameters may include step height, column height, column base platform height, and other relevant flow channel dimensions, while flow performance corresponds to the flow results obtained under different pressure or differential pressure conditions. The original disclosure also used structural parameters, pressure conditions, and corresponding flow data as the data foundation for machine learning modeling.
[0010] After obtaining sample data, the target flow rate or target flow rates under multiple pressure conditions are used as model input, and the corresponding structural parameters are used as model output. A machine learning regression model is then used to learn the inverse mapping relationship between flow performance and structural parameters. After the model is trained, the target flow performance can be input into a new structural design task, and the combination of structural parameters that meets the performance requirements can be directly predicted without repeatedly performing CFD flow field calculations for each set of candidate structures.
[0011] Although machine learning-based direct inverse prediction methods can quickly obtain solenoid valve structural parameters using existing sample data and reduce repetitive numerical simulation calculations, this method still has the following shortcomings:
[0012] (1) There is a one-to-many relationship between the target flow rate and structural parameters, and direct reverse mapping can easily lead to non-unique predictions. The flow performance of a solenoid valve is affected by multiple structural parameters, and different combinations of structural parameters may yield the same or similar flow performance. Therefore, directly predicting structural parameters from the target flow rate is not a strict one-to-one correspondence, and the same target performance may correspond to multiple sets of feasible structural parameter combinations. When using a single machine learning model to directly learn this reverse relationship, it is easy to encounter problems such as non-unique output parameters or unstable prediction results. The original disclosure clearly pointed out that the same target flow rate may correspond to multiple sets of different structural parameter combinations, making the direct derivation of structural parameters from flow performance obviously ambiguous.
[0013] (2) The predicted combination of structural parameters may lack reasonable parameter coordination relationships. The direct reverse prediction model mainly learns the data mapping relationship between the target flow and structural parameters based on the training samples. When there is a coupling relationship between multiple structural parameters, although the values of each parameter predicted by the model may be within the range of the training data, their combination results may not be able to form a reasonable solenoid valve structure scheme, and there may be problems with unreasonable parameter combinations. The original disclosure document has already listed "unreasonable structural combination" as a possible problem of direct reverse prediction.
[0014] (3) It is difficult to fully consider structural and manufacturing constraints in the direct prediction process. In addition to meeting the target flow performance, the structural parameters of the solenoid valve also need to meet engineering constraints such as parameter value range, geometric assembly relationship, structural interference, and processing and manufacturing. When the structural parameters are directly output by the machine learning model, if these constraints are not effectively incorporated into the prediction process, although the data may be close to the target performance, the combination of structural parameters may not be feasible to process, assemble, or implement in practice. Therefore, the original disclosure document incorporated the parameter range, geometric assembly relationship, and manufacturing constraints into the subsequent optimization process.
[0015] Therefore, although existing methods based on machine learning to directly predict structural parameters have high prediction efficiency, they still have problems such as multiple solutions in the inverse mapping, insufficient rationality of the combination of structural parameters, and difficulty in directly guaranteeing engineering constraints. It is difficult to stably obtain a solenoid valve design scheme that simultaneously meets the target flow performance and actual structural constraints. Summary of the Invention
[0016] This invention addresses the problems existing in the prior art by designing a reverse optimization method for solenoid valve parameters based on performance and correlation weight iteration. The purpose of this invention is to establish a reverse optimization method for solenoid valve structural parameters based on forward performance prediction, correlation weight, and error feedback iteration, so as to automatically solve the combination of solenoid valve structural parameters that meet performance requirements and engineering constraints from the target flow performance.
[0017] I. Addressing the issue of high computational load and long design cycle caused by repeated CFD simulations during the optimization of multiple structural parameters of solenoid valves. This is achieved by establishing a machine learning performance prediction model using existing product data, experimental data, or numerical simulation data. This performance prediction model replaces the repeated CFD flow field solutions for each candidate structure during the optimization iteration process, enabling rapid evaluation of the flow performance of candidate structures and reducing computational costs during structural parameter optimization.
[0018] II. Addressing the issue of excessive reliance on human experience in structural parameter adjustment. To address the problem that traditional methods primarily rely on designers' experience to adjust structural parameters, an automatic parameter iteration mechanism based on feedback from the error between target and predicted flow rates is established. This mechanism enables structural parameters to be automatically updated according to current performance deviations, forming a repeatable closed-loop optimization process.
[0019] Third, addressing the challenge of targeted adjustments for multiple structural parameters with varying degrees and directions of influence. This involves analyzing historical sample data to understand the correlation between different structural parameters and flow performance. The magnitude of this correlation determines the adjustment weight of each structural parameter, and the positive or negative correlation determines the adjustment direction, allowing for differentiated updates of different structural parameters based on their actual impact on flow performance. The original disclosure explicitly proposed using the correlation between structural parameters and flow to determine parameter influence weights and adjustment directions.
[0020] Fourth, this invention addresses the problem of unstable direct inverse prediction results caused by the one-to-many relationship between target flow and structural parameters. To address the issue that the same target flow may correspond to multiple combinations of structural parameters, this invention does not directly establish a single inverse mapping model from target flow to structural parameters. Instead, it establishes a forward prediction model from structural parameters and pressure conditions to flow performance, and obtains the combination of structural parameters that satisfies the target performance through iterative search, thereby avoiding multiple solution conflicts that arise during direct inverse prediction.
[0021] Fifth, addressing the issue that the combination of structural parameters obtained through reverse engineering may not meet actual engineering constraints. During the structural parameter iteration process, constraints such as allowable parameter ranges, geometric assembly relationships, structural interference, and manufacturing conditions are introduced simultaneously. The updated candidate structural parameters are then verified and corrected to ensure that the final output scheme meets both the target flow performance and the actual structural design and manufacturing requirements.
[0022] The objective of this invention is achieved through the following technical solution: a method for inverse optimization of solenoid valve parameters based on performance and correlation weight iteration, comprising the following steps: Step 1: Establishing a sample set of structural parameters and flow performance; collecting different structural parameters, pressure conditions, and their corresponding flow data to form a training sample set; Step 2: Inputting target performance and structural constraints; inputting target flow curves under different pressure conditions and setting the value range of each structural parameter; Step 3: Calculating the correlation weight of structural parameters; calculating the Pearson correlation coefficient between the i-th structural parameter and the flow rate based on the sample data, and normalizing it according to the absolute value of the correlation coefficient to obtain the influence weight of each structural parameter; Step 4: Determining... Step 5: Determine the initial structural parameters based on existing structural schemes, historical samples, or parameter ranges; Step 6: Predict the flow rate and calculate the performance error; Input the current structural parameters and various pressure conditions into the performance prediction model to obtain the predicted flow rate curve; Calculate the comprehensive error between the target flow rate and the predicted flow rate; Step 7: Update the structural parameters based on error feedback; Calculate the weighted relative deviation between the target flow rate and the predicted flow rate, and update the i-th structural parameter according to the correlation weight and error direction; Step 8: Perform parameter constraint verification and perform boundary correction on the updated structural parameters; Step 9: Re-input the updated structural parameters into the performance prediction model and repeat the flow rate prediction, error calculation, and parameter update process.
[0023] As a preferred approach, in step one, data on different structural parameters, pressure conditions, and corresponding flow rates are collected to form a training sample set. The structural parameters to be optimized are represented as follows:
[0024]
[0025] in, Let be the i-th structural parameter, including step height, column height, column base platform height, and other related structural dimensions.
[0026] As a preferred option, the target flow rate curve in step two is:
[0027]
[0028] The value ranges for each structural parameter are set as follows:
[0029]
[0030] It also sets the geometric assembly relationships, machining range, and manufacturing constraints between structural parameters.
[0031] Preferably, the formula for calculating the Pearson correlation coefficient between the i-th structural parameter and the flow rate in step three is:
[0032]
[0033] The influence weights of each structural parameter are obtained by normalizing the absolute value of the correlation coefficient:
[0034]
[0035] The absolute value of the correlation coefficient is used to determine the magnitude of parameter adjustment, while the sign of the correlation coefficient is used to determine the direction of parameter adjustment.
[0036] As a preferred option, the initial structural parameter combination in step four is as follows:
[0037]
[0038] The initial parameter combination should meet the structural dimension range and geometric assembly constraints.
[0039] As a preferred option, the predicted flow curve in step five is as follows:
[0040]
[0041] The combined error between the target flow and the predicted flow is:
[0042]
[0043] in, Let be the weight of the j-th pressure condition, and satisfy: .
[0044] As a preferred embodiment, the weighted relative deviation between the target flow and the predicted flow in step six is:
[0045]
[0046] Update the i-th structure parameter based on the correlation weight and error direction:
[0047]
[0048] in, This is the iteration step size.
[0049] As a preferred option, the boundary correction formula for the updated structural parameters in step seven is as follows:
[0050]
[0051] At the same time, it verifies whether the updated parameter combination meets the requirements of geometric assembly, structural interference, machining dimensions and manufacturing process; parameter combinations that do not meet the constraints are corrected or eliminated.
[0052] Preferably, the iteration in step eight stops when the following condition is met:
[0053]
[0054] or:
[0055]
[0056] Finally, the combination of structural parameters with the smallest comprehensive error is selected from the candidate schemes that satisfy the structural constraints.
[0057]
[0058] Output the optimal combination of structural parameters, the corresponding predicted flow curve, and the performance error results under various pressure conditions.
[0059] This invention is based on an indirect inverse solution method for solenoid valve structural parameters that combines forward performance prediction with iterative search. Instead of directly establishing an inverse mapping relationship from target flow rate to structural parameters, it establishes a forward prediction model from structural parameters, pressure conditions to flow performance, and iteratively solves for structural parameters through the error between target flow rate and predicted flow rate. This reduces the multiple solutions and parameter instability problems caused by one-to-many mapping in direct inverse prediction.
[0060] A differentiated parameter update method based on the correlation weights and influence directions of structural parameters and flow performance is proposed. The correlation between each structural parameter and flow performance is calculated using sample data. The adjustment weights of different structural parameters are determined by the degree of correlation, and the direction of parameter increase or decrease is determined by the direction of correlation. This allows for differentiated updates of different structural parameters according to their impact on flow performance, improving the targeting of parameter adjustments and optimizing convergence efficiency.
[0061] A closed-loop iterative optimization mechanism for structural parameters based on target performance error feedback is employed. The current candidate structural parameters are input into a forward performance prediction model to obtain the predicted flow rate. The structural parameters are continuously corrected through error feedback between the predicted and target flow rates, forming a closed-loop iterative process of "performance prediction – error calculation – parameter update – re-prediction" until the set performance requirements are met.
[0062] Compared with the prior art, the present invention has the following beneficial effects:
[0063] (1) Reduce the amount of repetitive calculations in the structural optimization process and improve the efficiency of reverse design. This invention uses existing product data, experimental data or numerical simulation data to train a forward flow performance prediction model. During the parameter iteration process, the flow performance of the candidate structure can be quickly obtained through the prediction model. There is no need to repeatedly perform geometric modeling, mesh generation and CFD flow field solution for each set of candidate parameters, thereby reducing the amount of computation in the structural optimization process and improving the efficiency of multi-parameter reverse design of solenoid valves.
[0064] (2) Reduce reliance on human experience and improve the automation of the parameter optimization process. The present invention automatically updates the structural parameters based on the error between the target flow and the predicted flow, forming a closed-loop iterative process of "performance prediction - error calculation - parameter adjustment - constraint verification", which enables the structural parameters to be continuously adjusted according to the performance deviation, reducing the problem of relying on the designer's experience to repeatedly adjust the structural dimensions in the traditional design process, and improving the uniformity and repeatability of the optimization process.
[0065] (3) Improve the targeting and optimization stability of multi-structural parameter adjustment. This invention determines the influence weight and adjustment direction of each parameter based on the correlation between structural parameters and flow performance, so that structural parameters with a greater impact on flow are adjusted first, while parameters with a smaller impact on flow are adjusted by a smaller amount. The direction of parameter increase or decrease is determined according to the correlation, thereby reducing blind search and ineffective adjustment in the multi-parameter optimization process and improving the targeting, convergence efficiency and optimization stability of parameter updates.
[0066] (4) Reduce parameter uncertainty caused by one-to-many mapping in direct inverse prediction. This invention does not directly establish a single inverse mapping relationship from target flow to structural parameters. Instead, it adopts an indirect inverse solution method that combines a forward performance prediction model with iterative search. Based on the error between the predicted performance of the candidate structure and the target performance, it gradually finds the required combination of structural parameters, thereby reducing the problem of multiple solution conflicts and unreasonable parameter combinations caused by multiple sets of structural parameters corresponding to the same target flow in the direct inverse prediction process.
[0067] (5) Improve the engineering feasibility of the optimization results. In the process of parameter optimization, this invention considers engineering constraints such as the allowable range of structural parameters, geometric assembly relationship, structural interference and processing and manufacturing, and performs constraint verification on the parameter combination after each round of updates, so that the final structural scheme can not only approximate the target flow performance under multiple pressure conditions, but also meet the actual structural design and manufacturing requirements, thereby improving the manufacturability and engineering application value of the optimization results. Attached Figure Description
[0068] Figure 1 This is a flowchart illustrating the overall implementation of the present invention; Detailed Implementation
[0069] The present invention will be further described below with reference to the embodiments illustrated in the accompanying drawings:
[0070] like Figure 1 As shown, this embodiment discloses a method for inverse optimization of solenoid valve parameters based on performance and correlation weight iteration, including the following steps: Step 1, establishing a sample set of structural parameters and flow performance; collecting different structural parameters, pressure conditions, and their corresponding flow data to form a training sample set. The structural parameters to be optimized are represented as:
[0071]
[0072] in, Let be the i-th structural parameter, including step height, column height, column base platform height, and other related structural dimensions. A machine learning performance prediction model is trained using sample data to establish a positive mapping relationship between structural parameters, pressure conditions, and flow performance.
[0073] Step 2: Input target performance and structural constraints; input target flow rate curves under different pressure conditions:
[0074]
[0075] Simultaneously set the value range for each structural parameter:
[0076]
[0077] It also sets the geometric assembly relationships, machining range, and manufacturing constraints between structural parameters.
[0078] Step 3: Calculate the correlation weights of structural parameters; calculate the Pearson correlation coefficient between the i-th structural parameter and the flow rate based on the sample data:
[0079]
[0080] The influence weights of each structural parameter are obtained by normalizing the absolute value of the correlation coefficient:
[0081]
[0082] The absolute value of the correlation coefficient is used to determine the magnitude of parameter adjustment, while the sign of the correlation coefficient is used to determine the direction of parameter adjustment.
[0083] Step 4: Determine the initial structural parameters; based on existing structural schemes, historical samples, or parameter ranges, determine the combination of initial structural parameters:
[0084]
[0085] The initial parameter combination should meet the structural dimension range and geometric assembly constraints.
[0086] Step 5: Predict flow rate and calculate performance error; input the current structural parameters and various pressure conditions into the performance prediction model to obtain the predicted flow rate curve:
[0087]
[0088] Calculate the combined error between the target flow and the predicted flow:
[0089]
[0090] in, Let be the weight of the j-th pressure condition, and satisfy:
[0091]
[0092] Step 6: Update the structural parameters based on error feedback; calculate the weighted relative deviation between the target flow and the predicted flow:
[0093]
[0094] Update the i-th structure parameter based on the correlation weight and error direction:
[0095]
[0096] in, This is the iteration step size.
[0097] When the predicted flow is lower than the target flow, adjust the relevant structural parameters to increase the flow; when the predicted flow is higher than the target flow, adjust the relevant structural parameters to decrease the flow.
[0098] Step 7: Perform parameter constraint verification; perform boundary correction on the updated structural parameters.
[0099]
[0100] Simultaneously, verify whether the updated parameter combinations meet the requirements for geometric assembly, structural interference, machining dimensions, and manufacturing processes. Parameter combinations that do not meet the constraints are corrected or eliminated.
[0101] Step 8: Determine convergence and output the results; re-input the updated structural parameters into the performance prediction model and repeat the flow prediction, error calculation, and parameter update.
[0102] Iteration stops when the following condition is met:
[0103]
[0104] or:
[0105]
[0106] Finally, the combination of structural parameters with the smallest comprehensive error is selected from the candidate schemes that satisfy the structural constraints.
[0107]
[0108] Output the optimal combination of structural parameters, the corresponding predicted flow curve, and the performance error results under various pressure conditions.
[0109] The performance prediction model in this invention is not limited to a specific machine learning algorithm. As long as it can predict the corresponding flow performance based on candidate structural parameters and pressure conditions, it can be used as a positive performance prediction model. In different implementations, different types of machine learning regression models can be selected according to the sample size, number of parameters, and prediction accuracy requirements, or multiple prediction models can be combined to form a performance prediction model. The core of the original technical solution lies in establishing a positive mapping relationship between structural parameters and pressure conditions to flow performance, rather than limiting the specific model form.
[0110] In this invention, the preferred approach is to use Pearson correlation coefficient analysis to determine the correlation between structural parameters and flow performance, and to determine the parameter adjustment weights based on the absolute value of the correlation coefficient and the adjustment direction based on the sign of the correlation coefficient.
[0111] In other implementations, other correlation evaluation or parameter sensitivity evaluation methods that can reflect the strength and direction of the influence of structural parameters on flow performance can be used to obtain the corresponding parameter influence weights, and the obtained weights can be used for subsequent structural parameter differential updates.
[0112] The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to replace them, without departing from the spirit of the invention or exceeding the scope defined by the appended claims.
Claims
1. A method for inverse optimization of solenoid valve parameters based on performance and correlation weight iteration, characterized in that, The process includes the following steps: Step 1: Establish a sample set of structural parameters and flow performance; collect different structural parameters, pressure conditions and their corresponding flow data to form a training sample set; Step 2: Input target performance and structural constraints; input the target flow curves under different pressure conditions and set the value range of each structural parameter; Step 3: Calculate the correlation weight of structural parameters. Calculate the Pearson correlation coefficient between the i-th structural parameter and the flow rate based on the sample data, and normalize it according to the absolute value of the correlation coefficient to obtain the influence weight of each structural parameter; Step 4: Determine the initial structural parameters by determining the combination of initial structural parameters based on existing structural schemes, historical samples, or parameter ranges; Step 5: Predict the flow rate and calculate the performance error; input the current structural parameters and various pressure conditions into the performance prediction model to obtain the predicted flow rate curve; calculate the comprehensive error between the target flow rate and the predicted flow rate; Step 6: Update the structural parameters based on error feedback; Calculate the weighted relative deviation between the target flow and the predicted flow, and update the i-th structural parameter according to the correlation weight and error direction; Step 7: Perform parameter constraint verification and perform boundary correction on the updated structural parameters; Step 8: Re-input the updated structural parameters into the performance prediction model and repeat the flow prediction, error calculation and parameter update.
2. The method for inverse optimization of solenoid valve parameters based on performance and correlation weight iteration according to claim 1, characterized in that, In step one, data on different structural parameters, pressure conditions, and corresponding flow rates are collected to form a training sample set. The structural parameters to be optimized are represented as follows: in, Let be the i-th structural parameter, including step height, column height, column base platform height, and other related structural dimensions.
3. The method for inverse optimization of solenoid valve parameters based on performance and correlation weight iteration according to claim 1, characterized in that, The target flow rate curve in step two is as follows: The value ranges for each structural parameter are set as follows: It also sets the geometric assembly relationships, machining range, and manufacturing constraints between structural parameters.
4. The method for inverse optimization of solenoid valve parameters based on performance and correlation weight iteration according to claim 1, characterized in that, The formula for calculating the Pearson correlation coefficient between the i-th structural parameter and the flow rate in step three is as follows: The influence weights of each structural parameter are obtained by normalizing the absolute value of the correlation coefficient: The absolute value of the correlation coefficient is used to determine the magnitude of parameter adjustment, while the sign of the correlation coefficient is used to determine the direction of parameter adjustment.
5. The method for inverse optimization of solenoid valve parameters based on performance and correlation weight iteration according to claim 1, characterized in that, The initial structural parameter combination in step four is as follows: The initial parameter combination should meet the structural dimension range and geometric assembly constraints.
6. The method for inverse optimization of solenoid valve parameters based on performance and correlation weight iteration according to claim 1, characterized in that, The predicted flow curve in step five is as follows: The combined error between the target flow and the predicted flow is: in, Let be the weight of the j-th pressure condition, and satisfy: .
7. The method for inverse optimization of solenoid valve parameters based on performance and correlation weight iteration according to claim 1, characterized in that, The weighted relative deviation between the target flow and the predicted flow in step six is: Update the i-th structure parameter based on the correlation weight and error direction: in, This is the iteration step size.
8. The method for inverse optimization of solenoid valve parameters based on performance and correlation weight iteration according to claim 1, characterized in that, The boundary correction formula for the updated structural parameters in step seven is as follows: At the same time, it verifies whether the updated parameter combination meets the requirements of geometric assembly, structural interference, machining dimensions and manufacturing process; parameter combinations that do not meet the constraints are corrected or eliminated.
9. The method for inverse optimization of solenoid valve parameters based on performance and correlation weight iteration according to claim 1, characterized in that, In step eight, the iteration stops when the following condition is met: or: Finally, the combination of structural parameters with the smallest comprehensive error is selected from the candidate schemes that satisfy the structural constraints. Output the optimal combination of structural parameters, the corresponding predicted flow curve, and the performance error results under various pressure conditions.
Citation Information
Patent Citations
Valve flow coefficient convenient calculation method based on CFD simulation
CN111767663A
