Chassis suspension parameter optimization method and device, electronic equipment, medium and product

Abstract

The invention discloses a chassis suspension parameter optimization method and device, electronic equipment, a medium and a product. The method comprises the following steps: acquiring a simulation working condition, a core performance index, a constraint condition and a weight distribution table according to chassis suspension demand data input by a user; according to the simulation working condition and the core performance index, obtaining a historical parameter combination and an initial suspension parameter combination; simulating the initial suspension parameter combination according to the simulation working condition to obtain a simulation result; performing performance index constraint checking on the initial suspension parameter combination according to a simulation result and a constraint condition to obtain a structured data table; and performing iterative optimization according to the weight distribution table, the structured data table and the historical parameter combination to obtain a parameter combination optimization result, and outputting the parameter combination optimization result. The method can solve the problems of low efficiency, difficulty in systematic balance of multiple performance targets and unsatisfactory optimization results caused by the fact that an existing chassis suspension parameter optimization method depends on manual experience trial and error.

Description

Technical Field

[0001] This application relates to the field of vehicle design technology, specifically to a method, apparatus, electronic device, readable storage medium, and computer program product for optimizing chassis suspension parameters. Background Technology

[0002] In the development of automotive chassis suspension systems, engineers face complex trade-offs between dozens of suspension parameters and multiple conflicting performance objectives. Currently, the industry commonly employs an experience-based trial-and-error optimization process. This involves engineers manually adjusting parameters such as lower control arm length and spring stiffness based on personal experience, then repeatedly verifying performance using multibody dynamics simulation software. While this method leverages engineers' intuition, in real-world applications, such as the development of family SUVs where comfort and handling must be balanced simultaneously, the experience-based trial-and-error approach often exhibits significant limitations. Its optimization process is inefficient, requiring weeks or even months of iterative iterations, and heavily relies on engineers' subjective experience. It struggles to systematically find the optimal solution within a complex multidimensional parameter space, and it fails to effectively balance the conflicts between multiple performance objectives, resulting in long development cycles and results that often fall short of achieving the best balance. Summary of the Invention

[0003] In view of the above problems, this application provides a method, apparatus, electronic device, readable storage medium and computer program product for optimizing chassis suspension parameters, which can solve the problems of low efficiency, difficulty in systematically balancing multiple performance objectives and unsatisfactory optimization results caused by relying on manual experience and trial and error in existing chassis suspension parameter optimization methods.

[0004] Firstly, this application provides a method for optimizing chassis suspension parameters, including: Based on the chassis suspension requirements data input by the user, obtain the simulation conditions, core performance indicators, constraints, and weight allocation table; Based on the simulation conditions and the core performance indicators, obtain the historical parameter combinations and the initial suspension parameter combinations; The initial suspension parameter combination was simulated based on the simulation conditions to obtain simulation results. Based on the simulation results and the constraints, the performance index of the initial suspension parameter combination is checked to obtain a structured data table; Iterative optimization is performed based on the weight allocation table, the structured data table, and the historical parameter combinations to obtain the parameter combination optimization result, and the parameter combination optimization result is output.

[0005] In the above technical solution, the method can first determine user needs and application scenarios, and obtain optimization base samples based on these, thereby obtaining high-quality basic information to ensure the pertinence of subsequent optimization; then, a structured data table that is easy to read and optimize is obtained through simulation and constraint verification; and finally, the optimal suspension parameter combination that meets user needs is obtained through optimization processing, so that users can put it into use directly.

[0006] In some implementations, obtaining the simulation conditions, core performance indicators, constraints, and weight allocation table based on user-input chassis suspension requirement data includes: Keyword extraction is performed on the chassis suspension requirement data input by the user to obtain a set of requirement keywords; The intent is determined by performing intent recognition on the keyword set. Entity recognition is performed on the stated demand intent to obtain the demand entity; The simulation conditions, core performance indicators, constraints, and weight allocation tables are matched according to the required entities.

[0007] In the above technical solution, the method can accurately map user needs into simulation conditions, performance indicators, constraints and weight allocation through keyword extraction, intent and entity recognition, so that subsequent design is directly aligned with core requirements, thereby improving the pertinence and application efficiency of the method.

[0008] In some implementations, obtaining the historical parameter combinations and the initial suspension parameter combinations based on the simulation conditions and the core performance indicators includes: Obtain target models that match the simulation conditions and core performance indicators from a preset historical vehicle database and a preset competitor parameter database; Obtain the actual value range of the core suspension parameters of the target vehicle model; The range of core suspension parameters is determined based on the actual value range. Determine the combination of historical parameters based on the pre-stored historical optimal parameters; Based on the preset geometric design specifications, the parameter restriction areas, the core parameter ranges of the suspension, and the historical parameter combinations, multiple initial suspension parameter combinations are generated. The initial suspension parameter combination is within the range of the core suspension parameters and does not exceed the restricted areas specified in the geometric design specifications.

[0009] In the above technical solution, the method can generate an appropriate initial parameter combination for the simulation conditions and core performance requirements, thereby providing high-quality and highly targeted basic samples for subsequent simulation verification and iterative optimization, which is conducive to improving the overall parameter optimization accuracy.

[0010] In some implementations, the step of simulating the initial suspension parameter combination based on the simulation conditions to obtain simulation results includes: A simulation task is generated for each of the initial suspension parameter groups based on the simulation conditions. If the simulation time of a target simulation task exceeds a preset time threshold during the simulation process, the target simulation task will be simplified, and the simulation will be performed again based on the simplified target simulation task to obtain the simulation result.

[0011] In the above technical solution, the method can generate a dedicated simulation task for each initial suspension parameter combination under the simulation conditions, thereby providing accurate simulation data support for subsequent performance constraint verification.

[0012] In some implementations, the step of performing performance index constraint verification on the initial suspension parameter combination based on the simulation results and the constraints to obtain a structured data table includes: Based on the simulation results, parameter combinations that failed in the simulation are marked as invalid samples, and an invalid sample marking table is generated based on the marking of the invalid samples; Based on the invalid sample labeling table and the simulation results, determine the combination of parameters to be optimized that was successfully simulated; Based on the simulation results of the parameter combinations to be optimized, obtain the specific values ​​of the performance indicators corresponding to each parameter combination to be optimized; The specific values ​​of the performance indicators are constrained and checked according to the constraints, and the constraint check results corresponding to each combination of parameters to be optimized are obtained. Based on the constraint verification results, the combination of parameters to be optimized that does not meet the constraint conditions in terms of specific performance index values ​​are marked as constraint violations, and a constraint violation mark table is generated based on the constraint violation marks. A structured data table is generated based on the constraint violation flag table and the specific values ​​of the performance indicators of the combination of parameters to be optimized.

[0013] In the above technical solution, the method can complete the screening of invalid samples and the verification of performance index constraints based on simulation results and constraints, thereby providing structured data support that has been verified by compliance for subsequent iterative optimization.

[0014] In some implementations, the iterative optimization based on the weight allocation table, the structured data table, and the historical parameter combinations to obtain the parameter combination optimization result includes: The basic population is determined based on the combination of historical parameters. The initial population is obtained by initializing the basic population and the combination of parameters to be optimized in the structured data table; The initial population is identified as the population to be optimized. The population to be optimized is screened according to the weight allocation table to obtain the screened target population; The target population is sorted to obtain a population sequence; Population genetic processing is performed on the population sequence to obtain new candidate populations; If the iteration termination condition is met, the parameter combination optimization result is generated based on the final candidate population.

[0015] In the above technical solution, the method can perform systematic iterative optimization based on the weight allocation table, structured data table and historical parameter combinations, thereby obtaining a candidate population for gradual optimization and a clear optimization path, which is conducive to improving the adaptability and reliability of the final parameter combination optimization results.

[0016] In some implementations, the step of filtering the population to be optimized according to the weight allocation table to obtain the filtered target population includes: The constraint violation degree of each individual solution in the population to be optimized is calculated according to the weight allocation table; wherein, the constraint violation degree is obtained by weighted calculation based on the weight allocation table and the constraint conditions; Individual solutions in the population to be optimized whose constraint violation degree exceeds the first preset threshold are eliminated to obtain the filtered target population.

[0017] In the above technical solution, the method can calculate the constraint violation degree of individual solutions based on the weight allocation table and eliminate individuals that exceed the standard, thereby obtaining the target population that meets the constraint requirements, which can improve the pertinence and efficiency of subsequent iterative optimization.

[0018] In some implementations, the sorting process for sorting the target population first sorts the individual solutions in the target population based on the constraint violation degree; for individual solutions with equal or all lower than a second preset threshold constraint violation degree, sorting is performed based on the Pareto dominance relationship of the individual solutions; for individual solutions belonging to the same non-dominance level, sorting is finally performed based on the crowding distance of the individual solutions.

[0019] In the above technical solution, the method can obtain an ordered and high-quality population sequence through hierarchical sorting logic of constraint violation degree, Pareto dominance relationship and crowding distance, thereby improving the targeting and accuracy of subsequent population genetic processing and optimization direction.

[0020] In some implementations, after outputting the optimized result of the parameter combination, the method further includes: If an adjustment instruction is received for the optimization result of the parameter combination, the weight allocation table is adjusted according to the adjustment instruction to obtain the adjusted target weight allocation table, and iterative optimization is performed again according to the target weight allocation table, the structured data table and the historical parameter combinations to obtain the optimized optimal parameter combination result; An optimization report is generated and output based on the optimal parameter combination results.

[0021] In the above technical solution, the method can respond to the adjustment instructions of the parameter combination optimization results, carry out iterative optimization again by adjusting the weight allocation table, obtain the optimized optimal parameter combination results and generate an optimization report, thereby improving the flexibility and adaptability of parameter optimization.

[0022] Secondly, this application provides a chassis suspension parameter optimization device, comprising: The first acquisition unit is used to acquire simulation conditions, core performance indicators, constraints and weight allocation tables based on the chassis suspension requirement data input by the user. The second acquisition unit is used to acquire historical parameter combinations and initial suspension parameter combinations based on the simulation conditions and the core performance indicators. The simulation unit is used to simulate the initial suspension parameter combination according to the simulation conditions and obtain simulation results. The constraint verification unit is used to perform performance index constraint verification on the initial suspension parameter combination based on the simulation results and the constraint conditions, and obtain a structured data table. The iterative optimization unit is used to perform iterative optimization based on the weight allocation table, the structured data table, and the historical parameter combinations to obtain the parameter combination optimization result; The output unit is used to output the optimization result of the parameter combination.

[0023] In the above technical solution, the device can first determine user needs and application scenarios, and obtain optimization basic samples based on these, thereby obtaining high-quality basic information to ensure the pertinence of subsequent optimization; then, it obtains a structured data table that is easy to read and optimize through simulation and constraint verification; and finally, it obtains the optimal suspension parameter combination that meets user needs through optimization processing, so that users can put it into use directly.

[0024] Thirdly, this application provides an electronic device, the electronic device including a memory and a processor, the memory for storing a computer program, the processor running the computer program to cause the electronic device to perform the chassis suspension parameter optimization method as described in any one of the first aspects.

[0025] Fourthly, this application provides a readable storage medium storing a computer program, which, when executed by a processor, performs the chassis suspension parameter optimization method described in any one of the first aspects.

[0026] Fifthly, this application provides a computer program product, which includes a computer program that, when executed by a processor, performs the chassis suspension parameter optimization method described in any one of the first aspects.

[0027] The beneficial effects of this application are: it enables chassis engineers to complete optimization without mastering algorithm and simulation expertise through natural language interaction and automatic target parsing driven by a large language model, thereby effectively improving the average efficiency of optimization task processing per person; It can also avoid manual intervention throughout the entire process from demand input to parameter output. Combined with simulation anomaly handling and intelligent parameter range setting, it can significantly reduce the optimization cycle of suspension parameters for new models and adapt to rapid R&D needs. It can also effectively improve the "performance compliance rate" and "feasibility compliance rate" of the optimization results by optimizing the algorithm with constraint awareness and setting parameter range driven by benchmark data, thereby reducing the workload of engineers in subsequent parameter adjustments; It can also receive engineers' experience and judgment through interactive re-optimization technology, thereby avoiding the limitations of the algorithm being "purely data-driven" and making the final parameters more in line with actual production needs, thereby improving the performance compliance rate of the vehicle trial production stage. It can also meet the requirements of "reproducibility and traceability" in automotive R&D through a full-process data archiving mechanism, so that historical optimization data can be quickly retrieved when subsequent model iterations or problem investigations are conducted, thus shortening the time for locating R&D problems. Furthermore, through continuous learning of large language models, the system can accumulate and inherit the design experience of outstanding engineers, thereby significantly improving design effectiveness. Attached Figure Description

[0028] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a flowchart illustrating the chassis suspension parameter optimization method in some embodiments of this application; Figure 2 This is a system flowchart illustrating the application of the chassis suspension parameter optimization method in some embodiments of this application; Figure 3 This is a schematic diagram of the chassis suspension parameter optimization device in some embodiments of this application; Figure 4 This is a schematic diagram of the structure of an electronic device in some embodiments of this application. Detailed Implementation

[0030] The embodiments of the technical solution of this application will now be described in detail with reference to the accompanying drawings. These embodiments are only used to more clearly illustrate the technical solution of this application and are therefore merely examples, and should not be used to limit the scope of protection of this application.

[0031] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0032] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more (including two), similarly, "multiple sets" refers to two or more sets (including two sets), and "multiple pieces" refers to two or more pieces (including two pieces) unless otherwise explicitly defined.

[0033] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0034] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0035] As a key component affecting core performance aspects such as vehicle comfort, handling, and safety, the automotive chassis suspension system involves dozens of parameters in its design, including spring stiffness and damping coefficient. Multiple conflicting performance objectives must be balanced within a limited development cycle. Currently, mainstream suspension parameter optimization methods in the industry include empirical trial-and-error, orthogonal experimental design, response surface methodology, and multi-objective optimization algorithms (such as NSGA-II and MOPSO). While empirical trial-and-error is simple and direct, it is inefficient and highly subjective; orthogonal experimental design is more systematic but limited by the number of parameters; response surface methodology is computationally efficient but has limited fitting accuracy for nonlinear problems; and multi-objective optimization algorithms can only generate one Pareto optimal solution in a single optimization process, requiring engineers to make subsequent performance trade-offs.

[0036] To address the aforementioned technical issues, this application provides a chassis suspension parameter optimization method. This method utilizes a large language model for natural language interaction and target parsing, thereby lowering the barrier to chassis suspension parameter optimization and allowing chassis engineers to independently complete optimizations without needing expertise in algorithms and simulation. Furthermore, this method automates the entire process from requirement input to parameter output, eliminating the need for manual intervention. Combined with simulation anomaly handling and intelligent parameter range setting, it significantly shortens the new vehicle model suspension parameter optimization cycle, adapting to rapid development needs. Moreover, relying on constraint-aware optimization algorithms and benchmarking experience, it significantly improves the performance compliance rate and feasibility of optimization results, greatly reducing the repetitive workload of engineers in subsequent parameter adjustments. Finally, interactive re-optimization technology allows engineers to incorporate experiential judgment, avoiding algorithmic limitations and ensuring that the final parameters better meet actual production needs, further improving vehicle performance during the prototyping stage. Finally, by establishing a full-process record archiving mechanism, it meets the requirements of reproducibility and traceability in automotive R&D, providing historical optimization information for subsequent model iterations or problem investigation, effectively shortening the time required for problem localization.

[0037] like Figure 1 As shown, some embodiments of this application provide a method for optimizing chassis suspension parameters, which includes: S101. Obtain simulation conditions, core performance indicators, constraints, and weight allocation tables based on the chassis suspension requirements data input by the user. S102. Based on the simulation conditions and core performance indicators, obtain the historical parameter combinations and the initial suspension parameter combinations; S103. Simulate the initial suspension parameter combination according to the simulation conditions and obtain the simulation results; S104. Based on the simulation results and constraints, the performance index of the initial suspension parameter combination is checked to obtain a structured data table. S105. Perform iterative optimization based on the weight allocation table, structured data table, and historical parameter combinations to obtain the parameter combination optimization results, and output the parameter combination optimization results.

[0038] In some embodiments, chassis suspension requirement data refers to the requirement information related to chassis suspension design input by users (such as R&D engineers of car companies). It can be provided through natural language description (such as "family SUV, comfort first") or preset option selection (such as selecting vehicle type, core performance priority), and includes key content such as vehicle positioning, performance preferences, and constraint requirements.

[0039] In some embodiments, simulation conditions refer to specific test scenarios that are transformed based on user needs and can be used for multibody dynamics simulation, such as ISOB / C road spectrum driving, serpentine test, double lane change test, etc., which are used to simulate the actual driving state of the vehicle to verify the suspension performance.

[0040] In some embodiments, core performance indicators refer to quantitative parameters that measure key performance of the chassis suspension, such as the root mean square value of the vehicle's vertical acceleration reflecting comfort, the yaw rate gain reflecting handling, and the peak roll angle reflecting safety, which are the core evaluation criteria for optimization results.

[0041] In some embodiments, constraints refer to the limitations that the chassis suspension design must meet, including geometric constraints (such as suspension travel ≥70mm), manufacturing process constraints, geometric interference taboos, regulatory requirements, etc., to ensure the feasibility and compliance of the optimization results.

[0042] In some embodiments, a weight allocation table refers to a performance indicator importance allocation file formulated according to the priority of user needs, which clarifies the weight ratio of each core performance indicator (such as comfort, handling, and safety) (e.g., comfort weight 0.6, handling weight 0.3) and is used to guide performance trade-offs in the iterative optimization process.

[0043] In some embodiments, the historical parameter set refers to a set of suspension parameters extracted from the enterprise's historical vehicle database that are similar to or the same as the target vehicle. This set includes reasonable past values ​​of key parameters such as lower control arm length, spring stiffness, and damping coefficient, providing a reference for initial parameter generation and iterative optimization.

[0044] In some embodiments, the initial suspension parameter set refers to the initial parameter set generated based on historical parameter sets, competitor parameter ranges, and constraints. It is generated through methods such as Latin hypercube sampling to uniformly cover a reasonable design space and serve as the initial input for simulation and optimization.

[0045] In some embodiments, simulation results refer to the original output data obtained by simulating the simulation conditions after inputting the initial suspension parameter combination into a multibody dynamics simulation tool (such as ADAMS), which includes key information such as the vehicle motion state, suspension forces, and original values ​​of performance indicators.

[0046] In some embodiments, performance index constraint verification refers to the process of verifying the compliance of the core performance indicators extracted from the simulation results against the constraint conditions, and determining whether the initial suspension parameter combination meets the geometric, process, regulatory and other restrictions.

[0047] In some embodiments, a structured data table refers to a standardized data file generated after performance index constraint verification. It contains information such as the specific values ​​of the initial suspension parameter combination, the corresponding core performance index values, and the constraint satisfaction status (e.g., whether it violates the rules), providing structured input for subsequent iterative optimization.

[0048] In some embodiments, iterative optimization refers to the process of repeatedly calculating based on a weight allocation table, a structured data table, and historical parameter combinations using an improved multi-objective optimization algorithm (such as the constraint-enhanced NSGA-II), continuously filtering and optimizing parameter combinations to approximate optimal performance, and finally obtaining a feasible Pareto solution set.

[0049] In some embodiments, the parameter combination optimization result refers to the optimal suspension parameter set obtained after iterative optimization that meets the performance priority and constraint requirements. It includes the specific values ​​of key parameters such as lower control arm length, spring stiffness, damping coefficient, and kingpin angle, which can be directly used for subsequent vehicle development and design.

[0050] In the above embodiments, the method can first determine user needs and application scenarios, and obtain optimization base samples based on these, thereby obtaining high-quality basic information to ensure the pertinence of subsequent optimizations; then, it obtains a structured data table that is easy to read and optimize through simulation and constraint verification; and finally, it obtains the optimal suspension parameter combination that meets user needs through optimization processing, so that users can directly put it into use.

[0051] In some embodiments, simulation conditions, core performance indicators, constraints, and weight allocation tables are obtained based on user-input chassis suspension requirement data, including: Keyword extraction is performed on the chassis suspension requirement data input by the user to obtain a set of requirement keywords; Perform intent recognition on the keyword set to obtain the demand intent; Entity recognition is performed on the demand intent to obtain the demand entity; The simulation conditions, core performance indicators, constraints, and weight allocation tables are matched according to the required entities.

[0052] For example, engineers can input their requirements using natural language descriptions or checkboxes. For instance, they could input: "Family SUV, comfort first, handling at least 95% of the benchmark model."

[0053] At this point, the system will extract keywords from the input content (such as "family SUV", "comfort", "handling", "benchmark 95%)" and generate a set of required keywords.

[0054] In the above embodiments, the method can accurately map user needs into simulation conditions, performance indicators, constraints and weight allocation through keyword extraction, intent and entity recognition, so that subsequent design is directly aligned with core requirements, thereby improving the method's pertinence and application efficiency.

[0055] In some embodiments, historical parameter combinations and initial suspension parameter combinations are obtained based on simulation conditions and core performance indicators, including: Obtain target models that match the simulation conditions and core performance indicators from a pre-set historical vehicle database and a pre-set competitor parameter database; Obtain the actual value range of the core suspension parameters of the target vehicle model; Determine the range of core suspension parameters based on the actual value range; Determine the combination of historical parameters based on the pre-stored historical optimal parameters; Based on the preset geometric design specifications, parameter exclusion zones, suspension core parameter ranges, and historical parameter combinations, multiple initial suspension parameter combinations are generated. The initial suspension parameter combination is within the range of core suspension parameters and does not exceed the prohibited parameters specified in the geometric design specifications.

[0056] In some embodiments, the logical formula for achieving "demand intent understanding" in the intelligent optimization system for chassis suspension parameters is as follows: Intent=LLM_Understand(Input_Text,DomainKnowledge); This logical formula describes how the Large Language Model (LLM) transforms the natural language requirements input by engineers, combined with chassis domain knowledge, into understandable requirement intentions. In this context, Intent represents the understood intent or demand. LLM_Understand is the large language model understanding function; Input_Text is the text that the engineer can input; Domain_Knowledge is a knowledge base for chassis design.

[0057] In some embodiments, the core logic formula for "entity recognition and extraction" in the intelligent optimization system for chassis suspension parameters is as follows: Entities =LLM Extract_Entities(Intent, Entity_Types); This logical formula describes the process by which a large language model, based on the understood intent, accurately extracts key entity information related to chassis suspension design according to preset entity types. Among them, Entities is the extracted set of core entities for chassis design; LLM_Extract_Entities is a dedicated function for entity extraction in large language models, used to filter and extract key information of preset types from understood demand intents; Intent is the core requirement intent that has been accurately parsed and output by the preceding steps; Entity_Types is a system-preset "list of chassis suspension design entity types", which is the "filtering rule" for extracting entities.

[0058] In some embodiments, the core logic formula for "intelligent mapping rule generation" in the intelligent chassis suspension parameter optimization system is as follows: MappingRules=LLM_Generate_Rules(Intent,Entities,Knowledge_Base); This logical formula describes the mapping rule that automatically generates "fuzzy requirements → quantitative engineering indicators" by combining the understood demand intent, the extracted core entities, and the chassis domain knowledge base with the large language model. Among them, MappingRules is the set of generated structured mapping rules; LLM_Generate_Rules is a dedicated function for generating rules for large language models, used to transform abstract intents / entities into system-executable mapping rules; Intent is the core design requirement output during the requirement intent understanding stage; Entities is the core set of entities extracted during the entity recognition process; Knowledge_Base is a comprehensive knowledge base for the chassis design field.

[0059] For example, this method can invoke an intelligent mapping engine based on a large language model to complete three types of mappings: (1) Working condition mapping: Matching commonly used simulated working conditions for family SUVs, such as ISOB / C road spectrum (60 / 90km / h), serpentine test, double lane change test; (2) Index mapping: Determine the core performance indicators, such as comfort corresponding to "body vertical acceleration", and handling corresponding to "yaw rate gain" and "roll angle"; (3) Constraint and weight mapping: Extract the constraints (suspension travel ≥70mm, geometric non-interference, and compliance with relevant regulations), and set the weights based on the priority of demand (comfort 0.6, handling 0.3, stability 0.1).

[0060] In the above embodiments, the method can generate an appropriate initial parameter combination for the simulation conditions and core performance requirements, thereby providing high-quality and highly targeted basic samples for subsequent simulation verification and iterative optimization, which is conducive to improving the overall parameter optimization accuracy.

[0061] In some embodiments, the initial suspension parameter combination is simulated according to the simulation conditions to obtain simulation results, including: The simulation task generates each initial suspension parameter group based on the simulation conditions; Simulations were performed based on the simulation task to obtain simulation results for each initial suspension parameter group.

[0062] For example, this method can retrieve matching models of "family SUVs with similar performance requirements" from historical vehicle databases and competitor parameter databases, and extract the actual value range of their core suspension parameters. Simultaneously, it combines geometric design specifications to mark parameter exclusion zones (such as control arm length < 350mm causing interference), and determines the initial optimization seed based on historically optimal parameters.

[0063] In the above embodiments, the method can generate a dedicated simulation task for each initial suspension parameter combination for the simulation conditions, thereby providing accurate simulation data support for subsequent performance constraint verification.

[0064] In some embodiments, simulation is performed according to the simulation task to obtain simulation results for each initial suspension parameter group, including: If the simulation time of a target simulation task exceeds a preset time threshold during the simulation process, the target simulation task will be simplified, and the simulation will be performed again based on the simplified target simulation task to obtain the simulation results.

[0065] For example, this method can employ Latin hypercube sampling to generate 40-60 parameter combinations within the parameter range that do not exceed the restricted area, ensuring that the sample uniformly covers the design space.

[0066] In the above embodiments, the method can simplify the model for simulation tasks that exceed a preset time threshold, thereby ensuring the rapid acquisition of effective simulation results and improving the efficiency of the overall simulation process.

[0067] In some embodiments, the performance index constraint verification of the initial suspension parameter combination is performed based on simulation results and constraints to obtain a structured data table, including: Based on the simulation results, parameter combinations that failed in the simulation are marked as invalid samples, and an invalid sample labeling table is generated based on the labeling of invalid samples; Based on the invalid sample labeling table and simulation results, determine the optimal parameter combinations that were successfully simulated. Based on the simulation results of the parameter combinations to be optimized, obtain the specific values ​​of the performance indicators corresponding to each parameter combination to be optimized; Based on the constraints, the specific values ​​of the performance indicators are checked to obtain the constraint check results for each combination of parameters to be optimized. Based on the constraint verification results, the combination of parameters to be optimized that does not meet the constraint conditions in terms of specific performance index values ​​are marked as constraint violations, and a constraint violation mark table is generated based on the constraint violation marks. A structured data table is generated based on the constraint violation flag table and the specific values ​​of the performance indicators of the combination of parameters to be optimized.

[0068] For example, this method can automatically generate ADAMS runtime scripts based on input, submit multi-condition simulation tasks in batches, and monitor task status in real time. If a single simulation times out (e.g., >120 seconds), it will automatically switch to a simplified model (e.g., reduce the precision of road surface excitation details) and rerun. If the simulation fails (e.g., the model reports an error), it will automatically retry once. If the retry fails, the parameter combination will be marked as an "invalid sample" and skipped.

[0069] In the above embodiments, the method can complete the screening of invalid samples and the verification of performance index constraints based on simulation results and constraints, thereby providing structured data support that has been verified by compliance for subsequent iterative optimization.

[0070] In some embodiments, iterative optimization is performed based on a weight allocation table, a structured data table, and historical parameter combinations to obtain parameter combination optimization results, including: The basic population is determined based on combinations of historical parameters; The initial population is obtained by initializing the basic population and the combination of parameters to be optimized in the structured data table; The initial population is identified as the population to be optimized. The population to be optimized is screened according to the weight allocation table to obtain the target population after screening. The target population is sorted to obtain a population sequence; Population genetics processing is performed on the population sequence to obtain new candidate populations; If the iteration termination condition is met, the parameter combination optimization result is generated based on the final candidate population.

[0071] In some embodiments, if the optimization termination condition is not met, the candidate population is identified as the population to be optimized, and the population to be optimized is screened according to the weight allocation table to obtain the screened target population.

[0072] In some embodiments, the parameter combination optimization results include at least the parameter combinations corresponding to individual solutions in the final candidate population, the performance index values ​​corresponding to each parameter combination, the constraint satisfaction state of the individual solutions, and the Pareto rank and distribution relationship of the individual solutions in the target space.

[0073] For example, this method can extract specific values ​​of performance indicators from simulation results (such as the root mean square value of the vehicle's vertical acceleration of 0.3g and the peak roll angle of 5°), and at the same time check whether each parameter combination meets the constraints (such as the actual suspension travel of 75mm, which meets the requirement of ≥70mm), and mark parameter combinations that do not meet the constraints as "constraint violation".

[0074] In the above embodiments, the method can perform systematic iterative optimization based on the weight allocation table, structured data table and historical parameter combinations, thereby obtaining a gradually optimized candidate population and a clear optimization path, which is conducive to improving the adaptability and reliability of the final parameter combination optimization results.

[0075] In some embodiments, the population to be optimized is screened according to a weight allocation table to obtain a screened target population, including: The constraint violation degree of each individual solution in the population to be optimized is calculated according to the weight allocation table; whereby the constraint violation degree is obtained by weighted calculation based on the weight allocation table and the constraint conditions. Individual solutions in the population to be optimized that exceed the first preset threshold for constraint violation are eliminated to obtain the filtered target population.

[0076] In some embodiments, the method can employ an improved NSGA-II algorithm for optimization. Its core lies in incorporating constraint violation degrees into non-dominated sorting and crowding calculations, achieving constraint-aware multi-objective optimization.

[0077] In some embodiments, the mathematical model for the multi-objective optimization problem is as follows: minF(x)=[f1(x),f2(x),…,f n (x)] T ; st gj (x)≤0, j=1,2, …,m; x i L ≤x i ≤x i U , i=1,2, …,p; Among them, "st" (an abbreviation for subject to) is the standard mathematical symbol for representing constraints in optimization problems; F(x) is the objective function vector; f k (x) is the k-th objective function; g j (x) is the j-th constraint function; x i L x i U These are the upper and lower limits of the design variable, respectively; p represents the dimension of the design variables; m represents the number of constraints.

[0078] In some embodiments, the formula for calculating the core constraint violation degree is: ; Wherein, CV(x) represents the degree of violation of the core constraint; w j Let be the weight coefficient of the j-th constraint; g j (x) is the j-th constraint function.

[0079] In some embodiments, the improved non-dominated sorting strategy employs a three-level sorting mechanism: First-level sorting: Sorting by constraint violation rate; Second-level sorting: When two solutions have the same degree of constraint violation, the Pareto dominance relation is used for sorting. Third-level sorting: Crowding distance sorting: When two solutions belong to the same Pareto level, they are sorted by crowding distance, where the crowding distance is calculated by the following formula: ; Among them, CD i It is the crowding distance of the i-th solution; α1 is a constraint influence factor (0 < α1 < 1); CV i Let be the constraint violation degree of the i-th solution; f k (i+1), f k (i-1) represent the objective function values ​​of adjacent solutions; f k max f k min These are the maximum and minimum values ​​of the objective function, respectively.

[0080] In some embodiments, the characteristic function based on cooperative game theory is as follows (using energy saving and comfort as examples only): v(S)=α2·ΔE(S)+β·ΔC(S); Where v(S) is the characteristic function value of the alliance S; where the alliance S refers to multiple core performance objectives (such as energy efficiency and comfort) that need to be coordinated and balanced. ΔE(S) represents the energy-saving benefit; ΔC(S) represents the benefit of improved comfort. α2 and β are weighting factors, respectively.

[0081] For example, the method can use the initial sample obtained in the above steps as the basic population and incorporate the parameter combination that satisfies the constraints during the calculation of indicators and constraints. Then, during the iteration process, solutions with constraint violation degrees greater than the threshold are automatically eliminated, and "nearly feasible solutions" (CV(x)≤ε) are given a higher selection probability; where ε is the preset constraint violation degree threshold, that is, the maximum allowable constraint violation degree critical value; Finally, based on the improved non-dominated sorting and crowding calculation, after 50-100 iterations, a set of uniformly distributed feasible Pareto solutions (containing multiple sets of performance-parameter correspondences) is obtained.

[0082] In the above embodiments, the method can calculate the constraint violation degree of individual solutions based on the weight allocation table and eliminate individuals that exceed the standard, thereby obtaining a target population that meets the constraint requirements, which can improve the pertinence and efficiency of subsequent iterative optimization.

[0083] In some embodiments, the sorting process for sorting the target population first sorts the individual solutions in the target population according to the constraint violation degree; for individual solutions with equal or all lower than a second preset threshold constraint violation degree, sorting is performed according to the Pareto dominance relationship of the individual solutions; for individual solutions belonging to the same non-dominance level, sorting is finally performed according to the crowding distance of the individual solutions.

[0084] For example, this method can display the solution set as a scatter plot (horizontal axis for comfort indicators and vertical axis for handling indicators) or a radar chart (comparison of multiple performance dimensions), allowing engineers to intuitively view the performance trade-offs of each solution.

[0085] For example, engineers can adjust requirements based on experience (such as increasing the comfort weight from 0.6 to 0.7) or specify preferences such as "prioritizing the reduction of peak acceleration" so that the system can perform secondary optimization using the following strategies after receiving adjustment instructions: Weight adjustment and further optimization; Local search strategy: Narrowing the search range near the existing optimal solution; Fast iteration: Reduce the number of iterations (e.g., reduce by 10-20) to quickly generate improved solutions.

[0086] In the above embodiments, the method can obtain an ordered and high-quality population sequence through hierarchical sorting logic based on constraint violation degree, Pareto dominance relationship, and crowding distance, thereby improving the targeting and accuracy of subsequent population genetic processing and optimization direction.

[0087] In some embodiments, after outputting the optimized result of the parameter combination, the method further includes: If an adjustment instruction is received for the parameter combination optimization result, the weight allocation table is adjusted according to the adjustment instruction to obtain the adjusted target weight allocation table. Then, the optimization is performed again based on the target weight allocation table, the structured data table, and the historical parameter combinations to obtain the optimized optimal parameter combination result. An optimization report is generated and output based on the optimal parameter combination results.

[0088] In some embodiments, the method can save all input and output data (requirement text, simulation script, optimization log, intermediate results, etc.) in all the above steps, and store them in the database along with the vehicle model number.

[0089] For example, this method can export optimal parameters (hard point coordinates, spring stiffness, damper settings, etc.) into a standardized Excel spreadsheet and automatically generate an optimization report using a large language model. Report=LLM_Generate_Report(OptimizationData,Template); Among them, Report is an automatically generated optimization report; LLM_Generate_Report is the function for generating reports for large language models; Optimization_Data represents the optimization process data; Template is the report template.

[0090] In the above embodiments, the method can respond to the adjustment instructions of the parameter combination optimization results, carry out iterative optimization again by adjusting the weight allocation table, obtain the optimized optimal parameter combination results and generate an optimization report, thereby improving the flexibility and adaptability of parameter optimization.

[0091] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below. In some embodiments, through... Figure 2 A system flowchart illustrating the chassis suspension parameter optimization method is presented. This system flowchart relies on the coordinated operation of nine functional modules to achieve intelligent optimization of chassis suspension parameters from input requirements to output results. The functions and core roles of each module are as follows: 1. Requirements Input Module: As the interaction point between engineers and the system, it supports two input methods: natural language description (such as "family SUV, comfort first") or preset option selection (such as selecting vehicle type and core performance priority from the drop-down menu). Engineers do not need to master professional algorithms or simulation operations, thus reducing the threshold for use.

[0092] 2. Target parsing module: The core is an intelligent mapping engine based on a large language model, with a built-in automotive chassis performance knowledge base (including association rules of working conditions, indicators, and constraints). It can automatically transform the fuzzy requirements input by engineers into calculable simulation working conditions, performance indicators, constraints, and weights, so as to realize the quantitative implementation of requirements.

[0093] 3. Benchmarking and Range Setting Module: Connects to the enterprise's historical vehicle database and competitor parameter library. Based on the performance requirements output by the target analysis module, it automatically extracts the suspension parameter range of similar models (such as lower control arm length 350-420mm). At the same time, it marks "parameter forbidden zones" such as geometric interference and manufacturing process, providing reasonable search boundaries for subsequent optimization and avoiding invalid calculations.

[0094] 4. Initial Sample Generation Module: Based on the parameter range, parameter restricted area, and historical best parameters output by the benchmarking and range setting module, multiple initial suspension parameter combinations covering a reasonable design space are generated to ensure that the samples are evenly distributed and do not exceed the parameter restricted area, providing specific and calculable initial input samples for the subsequent simulation scheduling module.

[0095] 5. Simulation Scheduling Module: As the "intelligent manager" of simulation tasks, it can automatically generate running scripts for multibody dynamics simulation tools (such as ADAMS) and submit simulation tasks in batches; at the same time, it has a built-in exception handling mechanism to automatically execute strategies such as downgrading and retrying in case of simulation timeouts or failures, ensuring the continuity of the process.

[0096] 6. Index and Constraint Calculation Module: Automatically extracts key performance indicators (such as the root mean square value of the vehicle's vertical acceleration) from the simulation results, and verifies them against the constraint conditions (such as suspension travel ≥70mm). It outputs structured data of "index calculation results + constraint satisfaction status" to provide input for the optimization module.

[0097] 7. Multi-objective optimization module: It adopts an improved multi-objective optimization algorithm, the core of which is the "constraint awareness enhancement strategy". It automatically eliminates or reduces the selection probability of solutions that do not meet the constraints, while retaining "near feasible solutions" to increase the number of final feasible solutions, and iteratively generates a Pareto solution set that meets the performance and constraint requirements.

[0098] 8. Interaction and Re-optimization Module: Display optimization results with visual charts (scatter plots, radar charts) to allow engineers to intuitively weigh performance; when engineers adjust weights (such as increasing the comfort weight by 10%) or add new preferences, the system can quickly perform a local search near the existing solution set to generate an improved solution that meets the new requirements.

[0099] 9. Export and Archive Module: Exports the final suspension parameters (hard point coordinates, spring stiffness, etc.) into a standardized table and automatically generates an optimization report containing "requirements-process-results"; at the same time, archives all process data (input requirements, simulation scripts, optimization logs, etc.) to ensure that the design results are traceable and reproducible.

[0100] In the above embodiments, the method can combine a pre-trained large language model with knowledge of chassis design domain to achieve a precise understanding of the engineer's fuzzy needs, thereby improving the accuracy of intent recognition and realizing "natural language interaction".

[0101] In the above embodiments, the method can also utilize the knowledge reasoning capabilities of large language models to automatically map natural language requirements into specific working conditions, indicators, constraints, and weights, without requiring manual configuration by engineers, thus significantly reducing learning costs.

[0102] In the above embodiments, the method can also automatically output parameter ranges and restricted areas based on historical and competitor data, avoiding the optimization algorithm from searching in invalid space, improving the efficiency of initial sample responses, and significantly shortening the optimization cycle.

[0103] In the above embodiments, the method can also solve common problems in the simulation process without manual intervention by using a preset strategy of "timeout downgrade and failure retry", ensuring the continuity of the entire process and greatly improving the simulation task completion rate.

[0104] In the above embodiments, the method can also incorporate "core constraint violation degree" into non-dominated sorting and congestion calculation to ensure that the final optimization result not only meets the performance requirements, but also meets the feasibility requirements such as geometry and process, thereby improving the feasible solution output rate and reducing the workload of engineers in subsequent adjustments.

[0105] In the above embodiments, the method can also support engineers to dynamically adjust requirements based on experience. The system can quickly perform local searches near the existing solution set without re-executing the full process optimization, thereby effectively shortening the response time and taking into account both "algorithm automation" and "human experience".

[0106] In the above embodiments, the method can also utilize the text generation capabilities of large language models to automatically generate professional and detailed optimization reports, including requirements descriptions, optimization processes, performance comparisons, parameter details, etc., thereby reducing the report generation time.

[0107] Figure 3 A schematic diagram of a chassis suspension parameter optimization device is shown. It should be understood that this device is related to... Figure 1 The method executed in the middle corresponds to the steps involved in the aforementioned method. The specific functions and effects of the device can be found in the description above. To avoid repetition, detailed descriptions are omitted here.

[0108] The chassis suspension parameter optimization device includes: The first acquisition unit 210 is used to acquire simulation conditions, core performance indicators, constraints and weight allocation tables based on the chassis suspension requirement data input by the user. The second acquisition unit 220 is used to acquire historical parameter combinations and initial suspension parameter combinations based on simulation conditions and core performance indicators. Simulation unit 230 is used to simulate the initial suspension parameter combination according to the simulation conditions and obtain simulation results; The constraint verification unit 240 is used to perform performance index constraint verification on the initial suspension parameter combination based on simulation results and constraint conditions, and obtain a structured data table. Iterative optimization unit 250 is used to perform iterative optimization based on the weight allocation table, structured data table and historical parameter combinations to obtain the parameter combination optimization result; Output unit 260 is used to output the result of parameter combination optimization.

[0109] In some embodiments, the first acquisition unit 210 includes: Extraction subunit 211 is used to extract keywords from the chassis suspension requirement data input by the user to obtain a set of requirement keywords; The identification subunit 212 is used to perform intent recognition on the keyword set to obtain the demand intent; The identification subunit 212 is also used to perform entity recognition on the demand intent to obtain the demand entity; Matching subunit 213 is used to match simulation conditions, core performance indicators, constraints and weight allocation tables according to the required entities.

[0110] In some embodiments, the second acquisition unit 220 includes: The first acquisition subunit 221 is used to acquire target models that match the simulation conditions and core performance indicators from a preset historical vehicle database and a preset competitor parameter database. The first acquisition subunit 221 is also used to acquire the actual value range of the core parameters of the target vehicle's suspension. The first determining subunit 222 is used to determine the range of core suspension parameters based on the actual value range. The first determining subunit 222 is also used to determine the combination of historical parameters based on the pre-stored historical optimal parameters; The first generation subunit 223 is used to generate multiple initial suspension parameter combinations based on preset geometric design specifications, marked parameter restricted areas, suspension core parameter ranges, and historical parameter combinations. The initial suspension parameter combination is within the range of core suspension parameters and does not exceed the prohibited parameters specified in the geometric design specifications.

[0111] In some embodiments, the simulation unit 230 includes: The second generation subunit 231 is used to generate simulation tasks for each initial suspension parameter group according to the simulation conditions. Simulation subunit 232 is used to perform simulations according to the simulation task and obtain simulation results for each initial suspension parameter group.

[0112] In some embodiments, the simulation subunit 232 is specifically used to simplify the target simulation task if the simulation time of the target simulation task exceeds a preset time threshold during the simulation process according to the simulation task, and then re-simulate the target simulation task according to the simplified target simulation task to obtain the simulation result.

[0113] In some embodiments, the constraint verification unit 240 includes: The marking subunit 241 is used to mark the parameter combinations that failed in the simulation as invalid samples based on the simulation results, and to generate an invalid sample marking table based on the marking of the invalid samples. The second determining subunit 242 is used to determine the combination of parameters to be optimized that has been successfully simulated, based on the invalid sample labeling table and the simulation results. The second acquisition subunit 243 is used to acquire the specific values ​​of the performance index corresponding to each combination of parameters to be optimized based on the simulation results of the combination of parameters to be optimized. Verification subunit 244 is used to perform constraint verification on the specific values ​​of performance indicators according to the constraint conditions, and obtain the constraint verification results corresponding to each combination of parameters to be optimized. The annotation sub-unit 245 is used to mark the combination of parameters to be optimized that does not meet the constraint conditions according to the constraint verification results as constraint violation, and to generate a constraint violation mark table based on the constraint violation mark; The third generation subunit 246 is used to generate a structured data table based on the constraint violation flag table and the specific values ​​of the performance indicators of the combination of parameters to be optimized.

[0114] In some embodiments, the iterative optimization unit 250 includes: The third determining subunit 251 is used to determine the basic population based on the combination of historical parameters; Initialization subunit 252 is used to initialize the population based on the combination of the basic population and the parameters to be optimized in the structured data table, so as to obtain the initial population. The third determining subunit 251 is also used to determine the initial population as the population to be optimized. The filtering subunit 253 is used to filter the population to be optimized according to the weight allocation table to obtain the filtered target population. The sorting subunit 254 is used to sort the target population to obtain a population sequence; Processing subunit 255 is used to perform population genetic processing on the population sequence to obtain new candidate populations; The fourth generation subunit 256 is used to generate the parameter combination optimization result based on the final candidate population when the iteration termination condition is reached.

[0115] In some embodiments, the filtering subunit 253 is specifically used to calculate the constraint violation degree of each individual solution in the population to be optimized according to the weight allocation table; wherein, the constraint violation degree is obtained by weighted calculation based on the weight allocation table and the constraint conditions; Individual solutions in the population to be optimized that exceed the first preset threshold for constraint violation are eliminated to obtain the filtered target population.

[0116] In some embodiments, the sorting process for sorting the target population first sorts the individual solutions in the target population according to the constraint violation degree; for individual solutions with equal or all lower than a second preset threshold constraint violation degree, sorting is performed according to the Pareto dominance relationship of the individual solutions; for individual solutions belonging to the same non-dominance level, sorting is finally performed according to the crowding distance of the individual solutions.

[0117] In some embodiments, the chassis suspension parameter optimization device further includes: The adjustment unit 270 is used to adjust the weight allocation table according to the adjustment instruction after the output unit 260 outputs the parameter combination optimization result, if it receives an adjustment instruction for the parameter combination optimization result, to obtain the adjusted target weight allocation table, and to perform iterative optimization again according to the target weight allocation table, the structured data table and the historical parameter combination to obtain the optimized optimal parameter combination result. Output unit 260 is also used to generate and output an optimization report based on the optimal parameter combination result.

[0118] like Figure 4As shown, this application provides an electronic device 300, which includes a processor 301 and a memory 302. The processor 301 and the memory 302 are interconnected and communicate with each other through a communication bus 303 and / or other forms of connection mechanism (not shown). The memory 302 stores a computer program that can be executed by the processor 301. When the computing device is running, the processor 301 executes the computer program to perform the method in any of the aforementioned optional implementations.

[0119] This application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the method in any of the aforementioned optional implementations.

[0120] The computer-readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0121] This application provides a computer program product, which includes a computer program that, when run by a processor, executes the method in any of the aforementioned optional implementations.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and not to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application, and they should all be covered within the scope of the claims and specification of this application. In particular, as long as there is no conflict, the various technical features mentioned in the embodiments can be combined in any way. This application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for optimizing chassis suspension parameters, characterized in that, include: Based on the chassis suspension requirements data input by the user, obtain the simulation conditions, core performance indicators, constraints, and weight allocation table; Based on the simulation conditions and the core performance indicators, obtain the historical parameter combinations and the initial suspension parameter combinations; The initial suspension parameter combination was simulated based on the simulation conditions to obtain simulation results. Based on the simulation results and the constraints, the performance index of the initial suspension parameter combination is checked to obtain a structured data table; Iterative optimization is performed based on the weight allocation table, the structured data table, and the historical parameter combinations to obtain the parameter combination optimization result, and the parameter combination optimization result is output.

2. The chassis suspension parameter optimization method according to claim 1, characterized in that, The process of obtaining simulation conditions, core performance indicators, constraints, and weight allocation tables based on user-input chassis suspension requirement data includes: Keyword extraction is performed on the chassis suspension requirement data input by the user to obtain a set of requirement keywords; The intent is determined by performing intent recognition on the keyword set. Entity recognition is performed on the stated demand intent to obtain the demand entity; The simulation conditions, core performance indicators, constraints, and weight allocation tables are matched according to the required entities.

3. The chassis suspension parameter optimization method according to claim 1, characterized in that, The step of obtaining historical parameter combinations and initial suspension parameter combinations based on the simulation conditions and the core performance indicators includes: Obtain target models that match the simulation conditions and core performance indicators from a preset historical vehicle database and a preset competitor parameter database; Obtain the actual value range of the core suspension parameters of the target vehicle model; The range of core suspension parameters is determined based on the actual value range. Determine the combination of historical parameters based on the pre-stored historical optimal parameters; Based on the preset geometric design specifications, the parameter restriction areas, the core parameter ranges of the suspension, and the historical parameter combinations, multiple initial suspension parameter combinations are generated. The initial suspension parameter combination is within the range of the core suspension parameters and does not exceed the restricted areas specified in the geometric design specifications.

4. The chassis suspension parameter optimization method according to claim 1, characterized in that, The simulation of the initial suspension parameter combination based on the simulation conditions, and the resulting simulation results, include: A simulation task is generated for each of the initial suspension parameter groups based on the simulation conditions. If the simulation time of a target simulation task exceeds a preset time threshold during the simulation process, the target simulation task will be simplified, and the simulation will be performed again based on the simplified target simulation task to obtain the simulation result.

5. The chassis suspension parameter optimization method according to claim 1, characterized in that, The initial suspension parameter combination is subjected to performance index constraint verification based on the simulation results and the constraints, resulting in a structured data table, including: Based on the simulation results, parameter combinations that failed in the simulation are marked as invalid samples, and an invalid sample marking table is generated based on the marking of the invalid samples; Based on the invalid sample labeling table and the simulation results, determine the combination of parameters to be optimized that was successfully simulated; Based on the simulation results of the parameter combinations to be optimized, obtain the specific values ​​of the performance indicators corresponding to each parameter combination to be optimized; The specific values ​​of the performance indicators are constrained and checked according to the constraints, and the constraint check results corresponding to each combination of parameters to be optimized are obtained. Based on the constraint verification results, the combination of parameters to be optimized that does not meet the constraint conditions in terms of specific performance index values ​​are marked as constraint violations, and a constraint violation mark table is generated based on the constraint violation marks. A structured data table is generated based on the constraint violation flag table and the specific values ​​of the performance indicators of the combination of parameters to be optimized.

6. The chassis suspension parameter optimization method according to claim 1, characterized in that, The iterative optimization based on the weight allocation table, the structured data table, and the historical parameter combinations to obtain the parameter combination optimization result includes: The basic population is determined based on the combination of historical parameters. The initial population is obtained by initializing the basic population and the combination of parameters to be optimized in the structured data table; The initial population is identified as the population to be optimized. The population to be optimized is screened according to the weight allocation table to obtain the screened target population; The target population is sorted to obtain a population sequence; Population genetic processing is performed on the population sequence to obtain new candidate populations; If the iteration termination condition is met, the parameter combination optimization result is generated based on the final candidate population.

7. The chassis suspension parameter optimization method according to claim 6, characterized in that, The step of filtering the population to be optimized according to the weight allocation table to obtain the filtered target population includes: The constraint violation degree of each individual solution in the population to be optimized is calculated according to the weight allocation table; wherein, the constraint violation degree is obtained by weighted calculation based on the weight allocation table and the constraint conditions; Individual solutions in the population to be optimized whose constraint violation degree exceeds the first preset threshold are eliminated to obtain the filtered target population.

8. The chassis suspension parameter optimization method according to claim 6, characterized in that, The sorting process for sorting the target population first sorts the individual solutions based on the constraint violation degree; for individual solutions with equal or all lower than a second preset threshold constraint violation degree, sorting is performed based on the Pareto dominance relationship of the individual solutions; for individual solutions belonging to the same non-dominance level, sorting is finally performed based on the crowding distance of the individual solutions.

9. The chassis suspension parameter optimization method according to claim 1, characterized in that, After outputting the optimized result of the parameter combination, the method further includes: If an adjustment instruction is received for the optimization result of the parameter combination, the weight allocation table is adjusted according to the adjustment instruction to obtain the adjusted target weight allocation table, and iterative optimization is performed again according to the target weight allocation table, the structured data table and the historical parameter combinations to obtain the optimized optimal parameter combination result; An optimization report is generated and output based on the optimal parameter combination results.

10. A chassis suspension parameter optimization device, characterized in that, The chassis suspension parameter optimization device includes: The first acquisition unit is used to acquire simulation conditions, core performance indicators, constraints and weight allocation tables based on the chassis suspension requirement data input by the user. The second acquisition unit is used to acquire historical parameter combinations and initial suspension parameter combinations based on the simulation conditions and the core performance indicators. The simulation unit is used to simulate the initial suspension parameter combination according to the simulation conditions and obtain simulation results. The constraint verification unit is used to perform performance index constraint verification on the initial suspension parameter combination based on the simulation results and the constraint conditions, and obtain a structured data table. The iterative optimization unit is used to perform iterative optimization based on the weight allocation table, the structured data table, and the historical parameter combinations to obtain the parameter combination optimization result; The output unit is used to output the optimization result of the parameter combination.

Images