Spring adjustment piece matching optimization method and system based on suspension offset frequency

Abstract

The invention relates to the technical field of vehicle design optimization, in particular to a spring adjustment part matching optimization method and system based on suspension offset frequency, and the method comprises the steps: extracting spring structure parameters and adjustment part structure parameters associated with offset frequency response, and constructing a disturbance combination matrix; performing response analysis on the disturbance combination matrix based on a simulation model of the whole vehicle suspension system, obtaining first scale data and second scale data, and forming a multi-scale response data set; according to the multi-scale response data set, calculating the first sensitivity and the second sensitivity of the spring adjustment piece, and constructing a multi-scale sensitivity index system; based on a multi-scale sensitivity index system, the independent contribution degree of each adjustment piece to the target offset frequency is evaluated, key influence adjustment pieces are identified, a structure parameter-offset frequency response mapping model is constructed according to the independent contribution degrees, and the specific influence mode and influence degree of each adjustment piece are obtained; and optimizing according to a multi-conflict design target, and outputting a spring and adjustment part combination matching suggestion.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle design optimization, and in particular to a spring adjustment component matching optimization method and system based on suspension frequency deviation. Background Art

[0002] As a core component of a vehicle's vertical dynamics, the suspension system's frequency characteristics directly impact the vehicle's handling stability and ride comfort. The off-frequency response, the characteristic response frequency of the suspension system under typical excitation conditions, is often used as a key basis for dynamic performance evaluation and matching design. During the vehicle matching development process, designers typically need to adjust the parameter combination of the suspension system's springs and associated tuning components based on the target off-frequency indicator to achieve optimal matching of the vehicle's dynamic performance.

[0003] However, existing frequency offset matching processes mostly rely on system-level simulation modeling and experimental verification, focusing primarily on the mapping relationship between overall structural parameters and system responses. They lack the ability to model the action paths and decompose the responses of the components at the tuning component level. When multiple tuning components interact, the effects of each component parameter on the frequency offset exhibit nonlinear superposition and coupling characteristics. Existing models are unable to differentiate the sensitivity of each component parameter and quantify its impact. This makes it difficult to effectively track the source of frequency offset in engineering practice, resulting in unclear optimization directions, low computational efficiency, and coarse tuning granularity.

[0004] Therefore, there is an urgent need for an optimization method and system that can identify the sensitivity and contribution of tuning component-level parameters to suspension offset frequency, and construct an offset frequency response analysis model with granular controllable capabilities. This not only achieves traceable mapping from system frequency characteristics to component-level action paths, but also significantly improves the matching efficiency and tuning accuracy of the suspension system, thereby enhancing the interpretability, controllability and automation level of the tuning process.

[0005] Therefore, a spring adjustment parts matching optimization method and system based on suspension offset frequency is proposed. Summary of the Invention

[0006] The present invention aims to provide a method and system for optimizing spring tuning component matching based on suspension offset frequency. By identifying the primary effects of parameter changes on the offset frequency of each tuning component, this method achieves traceable mapping from system frequency characteristics to component-level action paths, thereby improving the matching efficiency and tuning accuracy of the suspension system. A disturbance combination matrix is constructed by extracting spring and tuning component structural parameters associated with the offset frequency response. A response analysis of the disturbance combination matrix is performed based on a full-vehicle suspension system simulation model, acquiring first-scale and second-scale data to form a multi-scale response dataset. Based on the multi-scale response dataset, the first and second sensitivities of the spring tuning components are calculated, and a multi-scale sensitivity index system is constructed. Based on the multi-scale sensitivity index system, the independent contribution of each tuning component to the target offset frequency is evaluated, critical influencing tuning components are identified, and a structural parameter-offset frequency response mapping model is constructed based on the independent contribution to determine the specific impact mode and degree of each tuning component. The structural parameter-offset frequency response mapping model and the independent contribution are optimized according to multi-conflicting design objectives, and a matching recommendation for the spring and tuning component combination is output.

[0007] To achieve the above object, the present invention provides the following technical solutions: A spring adjustment component matching optimization method based on suspension frequency deviation includes: Extract the spring structural parameters and tuning component structural parameters associated with the off-frequency response and construct the disturbance combination matrix; Performing response analysis on each disturbance combination in the disturbance combination matrix based on a vehicle suspension system simulation model, obtaining first-scale data and second-scale data, and forming a multi-scale response data set; Calculating a first sensitivity and a second sensitivity of a spring adjustment component according to the multi-scale response data set, and constructing a multi-scale sensitivity index system; Based on the multi-scale sensitivity index system, the independent contribution of each adjustment component to the target offset frequency is evaluated, key influencing adjustment components are identified, and a structural parameter-offset frequency response mapping model is constructed based on the independent contribution to obtain the specific impact mode and impact degree of each adjustment component; According to the specific influence mode and influence degree of each adjustment component, the structural parameter-offset frequency response mapping model and the independent contribution are optimized based on the multi-conflict design objectives, and a combination matching recommendation of the spring and the adjustment component is output.

[0008] Preferably, the spring structure parameters include: stiffness characteristic curve, number of effective turns, wire diameter, median diameter and free length; The structural parameters of the adjustment component include: the stiffness characteristic curve and intervention timing of the auxiliary spring, the material properties, height, effective area and shape characteristics of the buffer block, the diameter, arm length and connection point hardness of the stabilizer bar, and the radial stiffness, axial stiffness and torsional stiffness of the bushing; The process of constructing the disturbance combination matrix is as follows: determining the value range and level number of the spring structural parameters and the adjustment component structural parameters, using the orthogonal experimental design method to generate an experimental scheme containing multiple parameter level combinations to form the disturbance combination matrix.

[0009] Preferably, the vehicle suspension system simulation model includes: a suspension system simulation unit, a frequency response analysis unit and a multi-scale data acquisition unit; The suspension system simulation unit simulates the vehicle suspension system based on the spring structural parameters and the adjustment component structural parameters to obtain a three-dimensional solid model of the vehicle suspension system; The frequency response analysis unit is integrated into the suspension system simulation unit and is used to perform frequency response analysis on each disturbance combination in the input disturbance combination matrix; The multi-scale data acquisition unit is used to obtain first-scale data and second-scale data from the frequency response analysis results; wherein, the first-scale data is macro-scale data, including the pitch modal frequency of the whole vehicle, the roll modal frequency of the whole vehicle, the vertical first-order main frequency of the whole vehicle, and the vertical second-order main frequency of the whole vehicle; the second-scale data is micro-scale data, including the local response frequency of the subassembly.

[0010] Preferably, the process of constructing the multi-scale sensitivity index system is: Calculating sensitivity values of the spring structure parameters and the adjustment component structure parameters at a first scale and a second scale respectively to obtain the first sensitivity and the second sensitivity; performing normalization processing on the first sensitivity and the second sensitivity; fusing the normalized first sensitivity and the second sensitivity using a nonlinear fusion algorithm to obtain a fused sensitivity value; The fused sensitivity values of all structural parameters are aggregated to form the multi-scale sensitivity index system.

[0011] Preferably, the process of identifying the key influencing adjustment components is: quantifying the influence relationship between each of the structural parameters of the adjustment component and the offset frequency response based on a multi-scale sensitivity index system; Attribution analysis is used to evaluate the independent contribution of each adjustment component to the target frequency deviation; Setting an independent contribution threshold, and marking the adjustment components corresponding to the structural parameters above the independent contribution threshold as the key influencing adjustment components; Based on the key influencing adjustment components, with the structural parameters of the key influencing adjustment components as input and the offset frequency response as output, a structural parameter-offset frequency response mapping model is established using a graph neural network to output the specific impact mode and impact degree of each key influencing adjustment component.

[0012] Preferably, the structural parameter-offset frequency response mapping model includes: a structural graph construction unit, a graph neural network modeling unit, an off-set frequency response prediction unit, and an impact mode and degree analysis unit; The structure graph construction unit constructs the spring and each adjustment component into a structure graph, wherein each node represents a different adjustment component, the feature of each node represents the disturbance value corresponding to each adjustment component in the disturbance combination matrix, and each edge represents the correlation between the structural parameters of the adjustment components; the correlation between the structural parameters of the adjustment components is obtained through correlation analysis; The graph neural network modeling unit performs feature propagation and graph convolution calculation based on the structural graph to learn the influence relationship between the structural parameters of each adjustment component and the frequency response of the suspension system; The bias frequency response prediction unit decodes the high-dimensional embedded features output by the graph neural network modeling unit to obtain a predicted target bias frequency response value; The influence mode and degree analysis unit performs a slight disturbance on the parameter value of the key influencing adjustment component and observes the change amount and change direction of the predicted target offset frequency response value to obtain the specific influence mode and the influence degree.

[0013] Preferably, a spring adjustment component matching optimization system based on suspension frequency deviation includes: The parameter perturbation generation module extracts the spring structure parameters and the tuning component structure parameters associated with the offset frequency response and constructs the perturbation combination matrix; a multi-scale response analysis module, which performs response analysis on each disturbance combination in the disturbance combination matrix based on a vehicle suspension system simulation model, obtains first-scale data and second-scale data, and forms a multi-scale response data set; a multi-scale sensitivity calculation module, which calculates the first sensitivity and the second sensitivity of the spring adjustment component according to the multi-scale response data set and constructs a multi-scale sensitivity index system; A tuning component contribution assessment module, based on the multi-scale sensitivity index system, evaluates the independent contribution of each tuning component to the target offset frequency, identifies key influencing tuning components, and constructs a structural parameter-offset frequency response mapping model based on the independent contribution to determine the specific impact mode and impact degree of each tuning component; The spring adjustment component matching optimization module optimizes the structural parameter-offset frequency response mapping model and the independent contribution based on the specific impact mode and impact degree of each adjustment component and multi-conflicting design objectives, and outputs a combination matching recommendation for the spring and adjustment component.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention adopts the orthogonal experimental design method to construct a disturbance combination matrix, and systematically arranges the combined disturbances of the parameters of the spring and various adjustment parts. Compared with the global disturbance, this method greatly improves the structural resolution of the action paths between the adjustment parts while ensuring the disturbance coverage. By constructing the disturbance combination matrix through the orthogonal experimental design method, the influence of each structural parameter on the off-frequency response can be clearly separated under a design without mutual interference, realizing the explicitness of the independent action paths of the adjustment parts. This strategy provides a more interpretable structural foundation for sensitivity evaluation and subsequent contribution extraction, and also provides solid support for the accurate modeling of the subsequent structural parameter-off-frequency response mapping model.

[0015] 2. This invention introduces a dual-scale response analysis mechanism. By acquiring macroscale and microscale data, it constructs a multiscale response dataset. A sensitivity index system is then constructed using a normalization and nonlinear fusion algorithm, enabling unified modeling of structural component-system responses. This dual-scale response analysis mechanism significantly improves the ability to express the impact of local adjustments on the main modal frequency deviation. This effectively addresses the problem of focusing solely on system response while ignoring local transmission paths and the difficulty in capturing the microscopic driving paths of local structural components on system frequency deviation. It effectively reflects how local structural changes are transmitted to the overall vehicle main modal, significantly reducing modeling errors and significantly enhancing the model's expressiveness and stability.

[0016] 3. Based on a multi-scale sensitivity index system, this invention evaluates the independent contribution of each tuning component to offset frequency and utilizes a graph neural network to construct a structural parameter-offset frequency response mapping model, accurately quantifying the mode and extent of influence of key tuning components. Furthermore, by guiding the combinatorial optimization process through contribution ranking, this method effectively circumvents the implementation difficulties of traditional "black box" intelligent optimization approaches, ensuring controllable and interpretable tuning solution recommendations, significantly improving suspension system matching efficiency and tuning accuracy in practical engineering applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic flow chart of a spring adjustment component matching optimization method based on suspension offset frequency provided by an embodiment of the present invention; Figure 2 A schematic structural diagram of a spring adjustment component matching optimization system based on suspension frequency deviation provided by an embodiment of the present invention; Figure 3 A working principle diagram of a vehicle suspension system simulation model provided by an embodiment of the present invention; Figure 4 A schematic structural diagram of a structural parameter-offset frequency response mapping model provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0019] The present invention proposes a spring tuning component matching optimization method based on suspension offset frequency. This method, applied to a spring tuning component matching optimization system based on suspension offset frequency, identifies the primary effect of parameter changes on the offset frequency, enabling a traceable mapping from system frequency characteristics to component-level action paths, thereby improving the matching efficiency and tuning accuracy of the suspension system. To demonstrate the effectiveness of the present method in identifying the primary effect of parameter changes on the offset frequency and achieving a traceable mapping from system frequency characteristics to component-level action paths, the following two examples will illustrate the effectiveness of the present invention.

[0020] Example 1 In this embodiment, the method proposed by the present invention is used to identify the main effects of parameter changes on the offset frequency of each tuning component, thereby achieving a traceable mapping from the system frequency characteristics to the component-level action path, thereby improving the matching efficiency and tuning accuracy of the suspension system. This embodiment applies the method of the present invention during the parameter tuning process of the front suspension system of a Class B sedan. Figure 1 This is a specific flow chart of the method of the present invention, including: extracting spring structural parameters and adjustment component structural parameters associated with the off-frequency response to construct a disturbance combination matrix; performing response analysis on the disturbance combination matrix based on a vehicle suspension system simulation model to obtain first-scale data and second-scale data to form a multi-scale response data set; calculating the first and second sensitivities of the spring adjustment components based on the multi-scale response data set, and constructing a multi-scale sensitivity index system; based on the multi-scale sensitivity index system, evaluating the independent contribution of each adjustment component to the target off-frequency, identifying key influencing adjustment components, and constructing a structural parameter-off-frequency response mapping model based on the independent contribution to obtain the specific impact mode and impact degree of each adjustment component; and optimizing according to multi-conflicting design objectives to output a combination matching recommendation of springs and adjustment components. Figure 2 The following is a structural diagram of a spring adjustment component matching optimization system based on suspension frequency deviation. Figure 1 and Figure 2 The following content is described: Extract the spring structural parameters and tuning component structural parameters associated with the off-frequency response and construct the disturbance combination matrix; The spring structural parameters include: stiffness characteristic curve, effective number of coils, wire diameter, median diameter and free length; The structural parameters of the adjustment component include: the stiffness characteristic curve and intervention timing of the auxiliary spring, the material properties, height, effective area and shape characteristics of the buffer block, the diameter, arm length and connection point hardness of the stabilizer bar, and the radial stiffness, axial stiffness and torsional stiffness of the bushing; The process of constructing the disturbance combination matrix is: The value ranges and level numbers of the spring structural parameters and the adjustment component structural parameters are determined, and an orthogonal experimental design method is used to generate an experimental scheme containing multiple parameter level combinations to form the disturbance combination matrix.

[0021] Specifically, the effective number of turns of the main spring: the value range is 5-7 turns, the level number is 3, which are 5, 6, and 7 respectively; Stabilizer bar diameter: the value range is 20-24mm, the level number is 3, namely 20mm, 22mm, and 24mm; Buffer block height: the value range is 50-70mm, the level number is 3, which are 50mm, 60mm, and 70mm respectively; Lower control arm rear bushing radial stiffness: The value range is 300-500N / mm, with a level of 3, namely 300N / m, 400N / mm, and 500N / mm. The above data are selected as the four main disturbance parameters. Using L9(3 4 ) orthogonal table to generate nine parameter perturbation combinations. For example, one of these combinations is: 6 active turns, 22mm stabilizer bar diameter, 50mm bump stop height, and 400N / mm bushing stiffness. Each row represents a specific suspension parameter configuration.

[0022] Table 1 shows the impact of different perturbation design methods on the ability to resolve structural parameter paths. The accuracy of identifying the independent impact of key parameters is calculated by comparing the consistency between the simulation response results and the actual parameter impact relationship under different perturbation design methods, and then calculating the proportion of correctly identified key structural parameters. The efficiency of microscale problem location is calculated by recording the average computational time required for each perturbation design method to complete an effective microscale frequency anomaly location.

[0023] Table 1 Effects of different perturbation design methods on the path resolution of structural parameters

[0024] By employing an orthogonal perturbation modeling mechanism, the traditional global perturbation method, which may involve parameter coupling effects, is transformed into an orthogonalized, high-structural-resolution tuning parameter perturbation strategy. This perturbation strategy explicitly identifies the independent influence paths of each tuning component (such as springs, stabilizer bars, bump stops, and bushings), laying a solid foundation for subsequent contribution evaluation and optimization. For example, when evaluating the impact of bump stop height on the end-impact characteristics of the suspension, orthogonal design effectively eliminates the confounding effects of other parameters such as the number of spring coils or stabilizer bar diameter, thereby more accurately quantifying the net contribution of the single factor, bump stop height.

[0025] Furthermore, based on the vehicle suspension system simulation model, a response analysis is performed on each disturbance combination in the disturbance combination matrix to obtain first scale data and second scale data to form a multi-scale response data set; The vehicle suspension system simulation model includes: a suspension system simulation unit, a frequency response analysis unit and a multi-scale data acquisition unit; Figure 3 ; The suspension system simulation unit simulates the vehicle suspension system based on the spring structural parameters and the adjustment component structural parameters to obtain a three-dimensional solid model of the vehicle suspension system; The frequency response analysis unit is integrated into the suspension system simulation unit and is used to perform frequency response analysis on each disturbance combination in the input disturbance combination matrix; The multi-scale data acquisition unit is used to obtain first-scale data and second-scale data from the frequency response analysis results; wherein, the first-scale data is macro-scale data, including the pitch modal frequency of the whole vehicle, the roll modal frequency of the whole vehicle, the vertical first-order main frequency of the whole vehicle, and the vertical second-order main frequency of the whole vehicle; the second-scale data is micro-scale data, including the local response frequency of the subassembly.

[0026] Specifically, the suspension system simulation unit uses a multi-body dynamics modeling tool (such as Adams Car or Simcenter Amesim) to construct a three-dimensional solid model of the vehicle suspension system containing detailed geometric and physical properties; based on the standard vehicle body-suspension-wheel system structure, the structural parameters of the springs and various adjustment parts are defined and replaced; the suspension system simulation unit simulates the vehicle suspension system based on the input spring structural parameters and adjustment part structural parameters; each set of disturbance parameter inputs corresponds to a new set of simulation scenarios.

[0027] The frequency response analysis unit arranges sensors on the vehicle's pitch axis, vertical axis, and roll axis to output vehicle body modal acceleration; and arranges displacement sensors and acceleration sensors at subassembly nodes to obtain local response curves; the subassembly nodes include auxiliary spring nodes, stabilizer bar end nodes, and bushing center nodes; After each set of disturbance simulation, the complete frequency response curve is output and the corresponding frequency peak point is derived.

[0028] The multi-scale data acquisition unit performs spectrum analysis on the vehicle pitch angle and the vehicle body vertical displacement, identifies the main peak position, and extracts the local response frequency of the subassembly to obtain first scale data and second scale data; The first scale data is macro-scale data, including: Pitch modal frequency: the frequency of the vehicle body's maximum acceleration response around the transverse axis; Roll modal frequency: the maximum response frequency of the vehicle body around the longitudinal axis; Vertical first-order main frequency: the first main frequency point of the vehicle body in the vertical direction; Vertical second-order main frequency: the second significant vertical response frequency of the vehicle body; The second scale data is microscale data, including the local response frequency of the subcomponent.

[0029] By establishing a high-precision simulation model of the entire vehicle suspension system and performing multi-scale response analysis, this method achieves a precise mapping from component parameters to system performance. Compared to traditional single-scale analysis methods, multi-scale data acquisition efficiently captures the frequency responses of the entire vehicle and substructures under a wide range of parameter combinations, capturing both system- and component-level response characteristics and enhancing the sophistication of subsequent modeling. It also improves the interpretability of the sources of frequency deviation and enhances the traceability of the relationship between structural parameters and target frequencies.

[0030] Further, according to the multi-scale response data set, a first sensitivity and a second sensitivity of the spring adjustment component are calculated, and a multi-scale sensitivity index system is constructed; The process of constructing the multi-scale sensitivity index system is as follows: Calculating sensitivity values of the spring structure parameters and the adjustment component structure parameters at a first scale and a second scale respectively to obtain the first sensitivity and the second sensitivity; performing normalization processing on the first sensitivity and the second sensitivity; fusing the normalized first sensitivity and the second sensitivity using a nonlinear fusion algorithm to obtain a fused sensitivity value; The fused sensitivity values of all structural parameters are aggregated to form the multi-scale sensitivity index system.

[0031] Specifically, kernel regression is used to fit the relationship between structural parameters and macro-scale indicators, and the local perturbation slope is extracted as the sensitivity value at the first scale; the same regression analysis is performed on micro-scale indicators to obtain the sensitivity value of each parameter at the second scale.

[0032] Table 2 compares the impact of different sensitivity analysis methods on modeling accuracy. The effect of local parameter adjustment on the offset-frequency prediction error is calculated by adjusting a single structural parameter using different sensitivity analysis methods, predicting the offset-frequency change, and then comparing the relative error with the actual simulation results. The local response consistency score is obtained by comparing the response trend of the predicted model under local parameter perturbations with the consistency of multi-scale simulation data.

[0033] Table 2 Comparison of the impact of different sensitivity analysis methods on modeling accuracy

[0034] The multi-scale sensitivity fusion algorithm employed in this embodiment enables consistent modeling and quantitative evaluation of local structural component effects and system-level macroscopic responses, effectively resolving the issues of "local effects failing to effectively transmit to the system's main mode" or unclear transmission paths. Through normalization and nonlinear fusion strategies, comprehensive consideration of response effects of varying physical significance and dimensions is achieved, enabling engineers to balance local detail optimization with overall performance enhancement within a unified framework. This provides a scientific basis for the subsequent identification and optimized design of key tuning components, significantly improving the efficiency of suspension tuning.

[0035] Furthermore, based on the multi-scale sensitivity index system, the independent contribution of each adjustment component to the target offset frequency is evaluated, key influencing adjustment components are identified, and a structural parameter-offset frequency response mapping model is constructed based on the independent contribution to obtain the specific impact mode and impact degree of each adjustment component; The process of identifying the key influencing adjustments is: quantifying the influence relationship between each of the structural parameters of the adjustment component and the offset frequency response based on a multi-scale sensitivity index system; Using the attribution analysis method based on the influence relationship, the independent contribution of each adjustment component to the target frequency deviation is evaluated; Setting an independent contribution threshold, and marking the adjustment components corresponding to the structural parameters above the independent contribution threshold as the key influencing adjustment components; Based on the key influencing adjustment components, with the structural parameters of the key influencing adjustment components as input and the offset frequency response as output, a structural parameter-offset frequency response mapping model is established using a graph neural network to output the specific impact mode and impact degree of each key influencing adjustment component.

[0036] Specifically, in this embodiment, the independent contribution threshold is the first few parameters that account for 80% of the total influence.

[0037] By evaluating contribution and identifying key tuning components, this implementation focuses analysis resources and optimization efforts on a small number of parameters that are crucial to performance, significantly improving tuning efficiency. Furthermore, the use of graph neural networks to construct parameter-response mappings allows for deeper exploration of nonlinear coupling effects between parameters. Furthermore, the identification of key tuning components provides engineers with clear guidance on tuning priorities, avoiding the time and cost associated with blindly adjusting noncritical parameters.

[0038] Furthermore, the structural parameter-bias frequency response mapping model includes: a structural graph construction unit, a graph neural network modeling unit, a bias frequency response prediction unit and an impact mode and degree analysis unit; Figure 4 ; The structure graph construction unit constructs the spring and each adjustment component into a structure graph, wherein each node represents a different adjustment component, the feature of each node represents the disturbance value corresponding to each adjustment component in the disturbance combination matrix, and each edge represents the correlation between the structural parameters of the adjustment components; the correlation between the structural parameters of the adjustment components is obtained through correlation analysis; The graph neural network modeling unit performs feature propagation and graph convolution calculation based on the structural graph to learn the influence relationship between the structural parameters of each adjustment component and the frequency response of the suspension system; The bias frequency response prediction unit decodes the high-dimensional embedded features output by the graph neural network modeling unit to obtain a predicted target bias frequency response value; The influence mode and degree analysis unit performs a slight disturbance on the parameter value of the key influencing adjustment component and observes the change amount and change direction of the predicted target offset frequency response value to obtain the specific influence mode and the influence degree.

[0039] In this embodiment, the main springs and the identified key influencing adjustment parts in the front suspension system of the B-class car are constructed as a structural diagram.

[0040] In this B-class sedan's structural parameter-to-deviational frequency response mapping model, the key tuning components are the main springs, stabilizer bar, and lower control arm rear bushing. The input features are their normalized parameter perturbations, and edges are defined based on their mechanical connections within the McPherson strut suspension. A two-layer GAT network is employed, with each layer followed by a Reluctant Unit (ReLU) activation function. After graph embedding learning, an average pooling layer is used to obtain a global representation of the graph, followed by two fully connected layers to predict the pitch and roll modal frequencies, respectively. After training, for example, with the stabilizer bar diameter and bushing stiffness fixed and the number of main spring effective turns varied from -1 (lowest level) to +1 (highest level), the pitch modal frequency curves output by the structural parameter-to-deviational frequency response mapping model are observed to determine the nonlinear effect of the effective turns on the pitch frequency and the specific degree of influence in different ranges, which is referred to as the slope.

[0041] Table 3 shows the impact of different optimization mechanisms on the performance and interpretability of the tuning scheme. The recommended accuracy is calculated by comparing the tuning scheme output by each optimization method with the optimal solution for actual off-frequency performance, and then statistically analyzing the matching accuracy. The parameter controllability score is calculated based on the feasibility and engineering feasibility of the adjustment range of each structural parameter in the output scheme, using expert scoring and normalization to a range of 0-1. The interpretability score is calculated based on whether the optimization mechanism can clearly indicate the path and direction of the parameter's influence on the off-frequency response. It is scored by multiple experts and normalized by taking the average value.

[0042] Table 3. Impact of different optimization mechanisms on the performance and interpretability of the tuning scheme.

[0043] By constructing a structural parameter-offset frequency response mapping model based on a graph neural network, this invention enables in-depth exploration and precise modeling of the complex, nonlinear coupling relationships between the parameters of various tuning components in complex suspension systems. This model not only accurately predicts the off-set frequency response after parameter adjustment (with a prediction error of less than 5%), but also better captures the combined impact of parameters on the final system response through multiple paths and levels of transmission. The impact mode and degree analysis unit provides clear, quantified parameter-response relationships, providing strong technical support for subsequent optimization and tuning decisions.

[0044] Furthermore, according to the specific influence mode and influence degree of each adjustment component, the structural parameter-offset frequency response mapping model and the independent contribution are optimized based on the multi-conflict design objectives, and a combination matching recommendation of the spring and the adjustment component is output.

[0045] In this embodiment, the multi-conflicting design objectives include a bias frequency target, a ride comfort target, a handling stability target, a durability target, and a weight consideration.

[0046] By systematically constructing parameter perturbations, conducting multi-scale simulation response analysis, and building a multi-scale sensitivity index system that integrates macro- and micro-effects, combined with an advanced structural parameter-offset frequency response mapping model to accurately assess the independent contribution of each tuning component and its specific impact mode and degree, a clear and traceable mapping relationship is achieved from the target frequency characteristics of the entire vehicle to the action path of specific component parameters. This not only significantly improves the efficiency and tuning accuracy of suspension parameter matching design when addressing multiple conflicting performance objectives, but also outputs optimized spring and tuning component configuration recommendations with greater engineering interpretability and controllability, effectively shortening the vehicle development cycle and improving the ultimate vehicle dynamic performance.

[0047] Example 2 In Example 1, the proposed method successfully identified the primary effects of parameter changes on the offset frequency of each tuning component, achieving a traceable mapping from system frequency characteristics to component-level action paths, thereby improving the matching efficiency and tuning accuracy of the suspension system. To further verify the effectiveness of the present invention, parameter tuning optimization was performed on the front suspension system of another sedan in this example.

[0048] A spring adjustment component matching optimization system based on suspension frequency deviation includes: The parameter perturbation generation module extracts the spring structure parameters and the tuning component structure parameters associated with the offset frequency response and constructs the perturbation combination matrix; The spring structural parameters include: stiffness characteristic curve, number of effective coils, wire diameter, median diameter, and free length; the adjustment component structural parameters include: stiffness characteristic curve and intervention timing of the auxiliary spring, material properties, height, effective area, and shape characteristics of the buffer block, diameter, arm length, and connection point hardness of the stabilizer bar, and radial stiffness, axial stiffness, and torsional stiffness of the bushing; The process of constructing the disturbance combination matrix is as follows: determining the value range and level number of the spring structural parameters and the adjustment component structural parameters, using the orthogonal experimental design method to generate an experimental scheme containing multiple parameter level combinations to form the disturbance combination matrix.

[0049] a multi-scale response analysis module, which performs response analysis on each disturbance combination in the disturbance combination matrix based on a vehicle suspension system simulation model, obtains first-scale data and second-scale data, and forms a multi-scale response data set; The vehicle suspension system simulation model includes: a suspension system simulation unit, a frequency response analysis unit and a multi-scale data acquisition unit; The suspension system simulation unit simulates the whole vehicle suspension system based on the spring structure parameters and the adjustment component structure parameters to obtain a three-dimensional solid model of the whole vehicle suspension system; the frequency response analysis unit is integrated in the suspension system simulation unit, and is used to perform frequency response analysis on each group of disturbance combinations in the input disturbance combination matrix; the multi-scale data acquisition unit is used to obtain first-scale data and second-scale data from the frequency response analysis results; wherein, the first-scale data is macro-scale data, including the whole vehicle pitch modal frequency, the whole vehicle roll modal frequency, the whole vehicle vertical first-order main frequency and the whole vehicle vertical second-order main frequency; the second-scale data is micro-scale data, including the local response frequency of the sub-component.

[0050] The multi-scale sensitivity calculation module calculates the first sensitivity and the second sensitivity of the spring adjustment component according to the multi-scale response data set, and constructs a multi-scale sensitivity index system. The process of constructing the multi-scale sensitivity index system is as follows: The sensitivity values of the spring structure parameters and the adjustment component structure parameters at the first scale and the second scale are calculated respectively to obtain the first sensitivity and the second sensitivity; the first sensitivity and the second sensitivity are normalized; the normalized first sensitivity and the second sensitivity are fused using a nonlinear fusion algorithm to obtain a fused sensitivity value; and the fused sensitivity values of all structural parameters are aggregated to form the multi-scale sensitivity index system.

[0051] The tuning component contribution assessment module evaluates the independent contribution of each tuning component to the target offset frequency based on the multi-scale sensitivity index system, identifies key influencing tuning components, and constructs a structural parameter-offset frequency response mapping model based on the independent contribution to obtain the specific impact mode and impact degree of each tuning component. The process of identifying the key influencing tuning components is as follows: Based on a multi-scale sensitivity index system, the influence relationship between the structural parameters of each adjustment component and the offset frequency response is quantified; according to the influence relationship, the independent contribution of each adjustment component to the target offset frequency is evaluated using an attribution analysis method; an independent contribution threshold is set, and the adjustment components corresponding to structural parameters above the independent contribution threshold are marked as the key-influencing adjustment components; based on the key-influencing adjustment components, a structural parameter-offset frequency response mapping model is established using a graph neural network with the structural parameters of the key-influencing adjustment components as input and the offset frequency response as output, and the specific influence mode and influence degree of each key-influencing adjustment component are output.

[0052] The structural parameter-bias frequency response mapping model includes: a structural graph construction unit, a graph neural network modeling unit, a bias frequency response prediction unit, and an impact mode and degree analysis unit; The structure graph construction unit constructs the spring and each adjustment component into a structure graph, wherein each node represents a different adjustment component, the feature of each node represents the disturbance value corresponding to each adjustment component in the disturbance combination matrix, and each edge represents the correlation between the structural parameters of the adjustment components; the correlation between the structural parameters of the adjustment components is obtained through correlation analysis; The graph neural network modeling unit performs feature propagation and graph convolution calculation based on the structural graph to learn the influence relationship between the structural parameters of each adjustment component and the frequency response of the suspension system; The bias frequency response prediction unit decodes the high-dimensional embedded features output by the graph neural network modeling unit to obtain a predicted target bias frequency response value; The influence mode and degree analysis unit performs a slight disturbance on the parameter value of the key influencing adjustment component and observes the change amount and change direction of the predicted target offset frequency response value to obtain the specific influence mode and the influence degree.

[0053] The spring adjustment component matching optimization module optimizes the structural parameter-offset frequency response mapping model and the independent contribution based on the specific impact mode and impact degree of each adjustment component and multi-conflicting design objectives, and outputs a combination matching recommendation for the spring and adjustment component.

[0054] Table 4 provides a comparison of the technical innovations and verifications of this solution. Among them, multi-scale sensitivity fusion technology combines macro- and micro-scale response characteristics; graph neural network path modeling technology constructs a graph structure prediction model from structural parameters to offset frequency responses, effectively capturing the complex relationships between parameters; and the multi-objective combined optimization mechanism generates tuning combination recommendations that are executable and engineering controllable.

[0055] Table 4 Comparison of technical innovations and verifications of this solution

[0056] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A spring adjustment component matching optimization method based on suspension frequency deviation, characterized in that: include: Extract the spring structural parameters and tuning component structural parameters associated with the off-frequency response and construct the disturbance combination matrix; Performing response analysis on each disturbance combination in the disturbance combination matrix based on a vehicle suspension system simulation model, obtaining first-scale data and second-scale data, and forming a multi-scale response data set; Calculating a first sensitivity and a second sensitivity of a spring adjustment component according to the multi-scale response data set, and constructing a multi-scale sensitivity index system; Based on the multi-scale sensitivity index system, the independent contribution of each adjustment component to the target offset frequency is evaluated, key influencing adjustment components are identified, and a structural parameter-offset frequency response mapping model is constructed based on the independent contribution to obtain the specific impact mode and impact degree of each adjustment component; According to the specific influence mode and influence degree of each adjustment component, the structural parameter-offset frequency response mapping model and the independent contribution are optimized based on the multi-conflict design objectives, and a combination matching recommendation of the spring and the adjustment component is output.

2. The spring adjustment component matching optimization method based on suspension frequency deviation according to claim 1, characterized in that: The spring structural parameters include: stiffness characteristic curve, effective number of turns, steel wire diameter, mean diameter and free length; the adjustment part structural parameters include: stiffness characteristic curve and intervention timing of the auxiliary spring, material properties, height, effective area and shape characteristics of the buffer block, diameter, arm length and connection point hard point of the stabilizer bar, and radial stiffness, axial stiffness and torsional stiffness of the bushing; the process of constructing the disturbance combination matrix is: determining the value range and level number of the spring structural parameters and the adjustment part structural parameters, using the orthogonal experimental design method to generate an experimental scheme containing multiple parameter level combinations to form the disturbance combination matrix.

3. The spring adjustment component matching optimization method based on suspension frequency deviation according to claim 1, characterized in that: The whole vehicle suspension system simulation model includes: a suspension system simulation unit, a frequency response analysis unit and a multi-scale data acquisition unit; the suspension system simulation unit simulates the whole vehicle suspension system based on the spring structure parameters and the adjustment component structure parameters to obtain a three-dimensional solid model of the whole vehicle suspension system; the frequency response analysis unit is integrated in the suspension system simulation unit, and is used to perform frequency response analysis on each group of disturbance combinations in the input disturbance combination matrix; the multi-scale data acquisition unit is used to obtain first scale data and second scale data from the frequency response analysis results; wherein, the first scale data is macro-scale data, including the whole vehicle pitch modal frequency, the whole vehicle roll modal frequency, the whole vehicle vertical first-order main frequency and the whole vehicle vertical second-order main frequency; the second scale data is micro-scale data, including the local response frequency of the sub-component.

4. The spring adjustment component matching optimization method based on suspension frequency deviation according to claim 1, characterized in that: The process of constructing the multi-scale sensitivity index system is as follows: calculating the sensitivity values of the spring structure parameters and the adjustment component structure parameters at the first scale and the second scale respectively to obtain the first sensitivity and the second sensitivity; Normalizing the first sensitivity and the second sensitivity; fusing the normalized first sensitivity and the second sensitivity using a nonlinear fusion algorithm to obtain a fused sensitivity value; The fused sensitivity values of all structural parameters are aggregated to form the multi-scale sensitivity index system.

5. The spring adjustment component matching optimization method based on suspension frequency deviation according to claim 1, characterized in that: The process of identifying the key influencing adjustment components is as follows: based on a multi-scale sensitivity index system, quantifying the influence relationship between the structural parameters of each adjustment component and the offset frequency response; and using an attribution analysis method to evaluate the independent contribution of each adjustment component to the target offset frequency based on the influence relationship; An independent contribution threshold is set, and the adjustment parts corresponding to the structural parameters higher than the independent contribution threshold are marked as the key influencing adjustment parts; based on the key influencing adjustment parts, with the structural parameters of the key influencing adjustment parts as input and the offset frequency response as output, a structural parameter-offset frequency response mapping model is established using a graph neural network to output the specific impact mode and impact degree of each of the key influencing adjustment parts.

6. The spring adjustment component matching optimization method based on suspension frequency deviation according to claim 5, characterized in that: The structural parameter-offset frequency response mapping model includes: a structural graph construction unit, a graph neural network modeling unit, an off-set frequency response prediction unit and an influence mode and degree analysis unit; the structural graph construction unit constructs the spring and each adjustment part into a structural graph, wherein the nodes represent different adjustment parts, the features of the nodes represent the disturbance values corresponding to each adjustment part in the disturbance combination matrix, and the edges represent the correlation between the structural parameters of the adjustment parts; the correlation between the structural parameters of the adjustment parts is obtained through correlation analysis; the graph neural network modeling unit performs feature propagation and graph convolution calculation based on the structural graph to learn the influence relationship of the structural parameters of each adjustment part on the off-set frequency response of the suspension system; the off-set frequency response prediction unit decodes the influence relationship to obtain the predicted target off-set frequency response value; the influence mode and degree analysis unit performs a small disturbance on the parameter value of the key influencing adjustment part, and observes the change amount and change direction of the predicted target off-set frequency response value to obtain the specific influence mode and the influence degree.

7. A spring adjustment component matching optimization system based on suspension frequency deviation, characterized in that: include: The parameter perturbation generation module extracts the spring structure parameters and the tuning component structure parameters associated with the offset frequency response and constructs the perturbation combination matrix; a multi-scale response analysis module, which performs response analysis on each disturbance combination in the disturbance combination matrix based on a vehicle suspension system simulation model, obtains first-scale data and second-scale data, and forms a multi-scale response data set; a multi-scale sensitivity calculation module, which calculates the first sensitivity and the second sensitivity of the spring adjustment component according to the multi-scale response data set and constructs a multi-scale sensitivity index system; A tuning component contribution assessment module, based on the multi-scale sensitivity index system, evaluates the independent contribution of each tuning component to the target offset frequency, identifies key influencing tuning components, and constructs a structural parameter-offset frequency response mapping model based on the independent contribution to determine the specific impact mode and impact degree of each tuning component; The spring adjustment component matching optimization module optimizes the structural parameter-offset frequency response mapping model and the independent contribution based on the specific impact mode and impact degree of each adjustment component and multi-conflicting design objectives, and outputs a combination matching recommendation for the spring and adjustment component.

Images