Structural mechanical parameter optimization method and system based on artificial intelligence

Abstract

The invention relates to the technical field of artificial intelligence, in particular to a structural mechanical parameter optimization method and system based on artificial intelligence, and the method comprises the following steps: calling an initial model and performance constraints, screening a design variable set, extracting mechanical response under a working condition, recognizing a weak region, and simulating and optimizing a structural layout based on a structural design demand. The robustness and reliability are verified, key indexes are extracted, performance fluctuation is analyzed, a scheme is adjusted, and a structural mechanical performance optimization trend analysis conclusion is obtained. According to the method, design variables are screened through a deep learning algorithm, design optimization can be achieved, the low-efficiency process depending on artificial experience and repeated trial calculation in a traditional method is avoided, structural layout and mechanical properties can be dynamically adjusted, the stability and reliability of the structure are improved, potential weak areas are effectively recognized and eliminated, and the method is suitable for large-scale popularization and application. Long-term tracking and verification ensure that the scheme has higher robustness and reliability, and ensure that the optimized structural design has wide adaptability and longer service life.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a method and system for optimizing structural mechanics parameters based on artificial intelligence. Background Technology

[0002] The field of artificial intelligence technology involves using computer systems to simulate human intelligent behavior. Its core aspects include machine learning, deep learning, neural networks, natural language processing, computer vision, and intelligent control. Based on large-scale data processing, this technology builds and trains models to achieve automatic perception, reasoning, and decision-making capabilities in complex environments. It is widely applied in various industries such as healthcare, transportation, finance, and industrial manufacturing, aiming to enhance technological autonomy and intelligence. Traditional structural mechanics parameter optimization methods refer to the process of obtaining a combination of structural parameters that meets target performance requirements during structural design and analysis. This involves iterative calculations based on empirical formulas or manual adjustments based on finite element analysis results. These methods rely on manually setting initial parameter ranges and comparing results through gradual experiments to determine the optimal parameter configuration.

[0003] Current technologies rely on manually setting initial parameter ranges and adjusting finite element calculation results during structural design. This makes the optimization process overly dependent on human experience, requiring each optimization to start from scratch, resulting in low efficiency and the potential to overlook critical weak areas. Especially under complex operating conditions, it lacks the ability to respond in real time and make dynamic adjustments. Because traditional methods lack intelligent decision support, the design scheme fails to meet the actual needs under different operating conditions to the greatest extent, ultimately affecting the stability and long-term reliability of structural performance. Especially in variable operating environments, this manual adjustment method is not only time-consuming but may also miss the best optimization opportunities. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide a method and system for optimizing structural mechanics parameters based on artificial intelligence. The technical solution is as follows: An artificial intelligence-based method for optimizing structural mechanics parameters is provided, the method comprising: S1: Based on the structural design target requirements, the initial structural model and its performance constraints are invoked to identify multiple combinations of candidate design variables. A deep learning algorithm is then used for preliminary screening to obtain a set of design variables that meet the constraints. S2: Based on the set of design variables that meet the constraints, extract the mechanical response characteristics of the model under different working conditions, analyze the structural stability and failure risk, identify key weak areas, and obtain optimized potential area data. S3: Based on the optimized potential region data, the mechanical response data is identified by finite element simulation, a mechanical behavior prediction network is constructed, and the parameter distribution is dynamically adjusted through reinforcement learning strategy to optimize the structural layout and improve mechanical performance, thereby obtaining the optimized structural design scheme. S4: Based on the optimized structural design scheme, call the structural performance verification records in the original archive, detect the robustness of the design scheme, analyze its reliability in long-term use, and obtain the structural performance optimization results.

[0005] As a further aspect of the present invention, the set of design variables satisfying the constraints includes geometric morphology data, material property configuration, and stress distribution information; the optimized potential region data includes weak region location, stress concentration distribution, and failure mode prediction; the optimized structural design scheme includes improved geometry, material distribution adjustment scheme, and mechanical performance improvement index; and the structural performance optimization results include robustness assessment conclusions and long-term reliability analysis data.

[0006] As a further aspect of the present invention, the step of setting the design variable set that satisfies the constraints specifically includes: S101: Based on the structural design target requirements, extract the geometric shape, material properties and boundary conditions of the initial structural model, construct multiple sets of candidate design variable combinations, mark the constraints of each set of variables, remove combinations that do not meet the basic requirements, store them in the candidate variable database, and generate a candidate variable sequence. S102: Based on the candidate variable sequence, call the deep learning algorithm to evaluate the performance of each group of variables, select the design variable combination that meets the constraints, store it in the design variable database, update the selection status flag, and obtain the design variable set that meets the constraints.

[0007] As a further aspect of the present invention, the step of optimizing the potential region data specifically includes: S201: Based on the set of design variables that meet the constraints, extract the mechanical response characteristics of the model under different working conditions, including stress distribution, deformation and fatigue life, analyze its stability and failure risk, identify key weak areas, and obtain a distribution map of weak areas. S202: Based on the weak area distribution map, extract the location and size of stress concentration points, combine with the failure mode prediction model, calculate the potential failure probability value, record key mechanical behavior characteristics, and obtain optimized potential area data.

[0008] As a further aspect of the present invention, the potential failure probability value is expressed by the formula: ; in, Represents the potential failure probability value. This represents the total number of stress concentration points. Representing the Stress values ​​at stress concentration points This represents the average stress value at the stress concentration point. Representing the The distance from each stress concentration point to the boundary of the weak area. The maximum stress value at the stress concentration point. This represents the average distance from the stress concentration point to the boundary. Representing the Index of stress concentration points.

[0009] As a further aspect of the present invention, the steps of the optimized structural design scheme are as follows: S301: Based on the optimized potential region data, extract the mechanical behavior features of the weak region, construct a mechanical behavior prediction network, use reinforcement learning strategy to dynamically adjust the parameter distribution, optimize the structural layout, improve the overall mechanical performance, and obtain a preliminary optimization scheme. S302: Based on the preliminary optimization scheme, analyze the improved geometry and material distribution, compare the mechanical performance indicators before and after optimization, statistically analyze the optimization effect, and obtain the optimized structural design scheme.

[0010] As a further aspect of the present invention, the steps for achieving the structural performance optimization result are specifically as follows: S401: Based on the optimized structural design scheme, extract the structural performance verification records from the original archive, including failure cases and reliability test data during long-term use, detect the robustness of the design scheme, analyze its adaptability under different environmental conditions, and obtain a robustness assessment conclusion. S402: Based on the robustness assessment conclusions and combined with the long-term reliability analysis model, predict the performance change trend of the structural design scheme within its service life, statistically analyze the fluctuation range of key performance indicators, and obtain the structural performance optimization results.

[0011] As a further aspect of the present invention, the method further includes step S5: S5: Based on the structural performance optimization results, extract the key performance indicators of the optimization scheme, analyze their performance under different working conditions, screen the performance fluctuation working conditions, combine the mechanical behavior prediction network, adjust the design scheme, and obtain the structural mechanical performance optimization trend analysis conclusion. The conclusions of the structural mechanical performance optimization trend analysis include the changing trends of key performance indicators, working condition adaptability analysis, and optimization potential assessment.

[0012] As a further aspect of the present invention, the steps for drawing the structural mechanical performance optimization trend analysis conclusion are as follows: S501: Based on the structural performance optimization results, extract key performance index data under different working conditions, remove outliers, match working condition numbers, analyze performance fluctuation range, and obtain a set of performance indicators. S502: Based on the performance index set, classify and arrange them according to the working condition type, identify the working conditions with large performance fluctuations, analyze their mechanical behavior characteristics, combine with the optimization potential assessment model, identify the optimization space, and obtain the structural mechanical performance optimization trend analysis conclusion.

[0013] The electric vehicle status monitoring system is used to execute the above-mentioned electric vehicle status monitoring method, and the system includes: The design variable selection module calls the design parameter set and performance constraints in the initial structural model according to the design target requirements, groups the design parameter set, checks whether each group of parameters meets the performance constraints, and selects the design variable combinations that meet the requirements to obtain the candidate design variable set. The response analysis module extracts the mechanical response data of the model under different working conditions based on the candidate design variable set, and analyzes the nodal stress distribution, strain behavior and mechanical stability under each set of design variables. It compares and analyzes the mechanical response of each combination, marks high-risk areas, and obtains the stress response set of key areas. The optimization scheme generation module is based on the stress response set of key areas. It adjusts and optimizes the structural geometric parameters, material properties and boundary conditions in high-risk areas. By analyzing stress distribution and local deformation, it dynamically adjusts parameters, optimizes the structural layout, and obtains the optimized structural parameter set. The stability verification module, based on the optimized set of structural parameters, calls the original performance verification data, combines long-term usage data and environmental change parameters, analyzes the stability and durability of the structure under different usage conditions, identifies potential failure points, and obtains structural durability assessment results. Based on the structural durability assessment results and combined with environmental change parameters, the adaptive adjustment module analyzes the structural adaptability and performance fluctuations, adjusts design variables that are unstable under target working conditions, optimizes the design scheme, optimizes structural adaptability, and obtains the conclusion of structural mechanical performance optimization trend analysis.

[0014] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: By applying deep learning algorithms, the optimal combination of structural performance requirements can be selected from a large number of design variables, thereby achieving design optimization and avoiding the inefficient process of relying on manual experience and repeated trial and error in traditional methods. The introduction of reinforcement learning strategies enables dynamic adjustment and optimization of structural layout and mechanical properties, not only improving structural stability and reliability but also effectively identifying and eliminating potential weak areas, avoiding the risk of overlooking details during manual adjustments. Through long-term tracking and verification of structural performance, the final solution can be ensured to have higher robustness and reliability, exhibiting greater stability under varying operating conditions, thus achieving broad adaptability and a longer service life for the optimized structural design in different environments. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the workflow of the present invention; Figure 2 This is a detailed flowchart of S1 of the present invention; Figure 3 This is a detailed flowchart of the S2 process of the present invention; Figure 4 This is a detailed flowchart of the S3 process of the present invention; Figure 5 This is a detailed flowchart of the S4 process of the present invention; Figure 6 This is a detailed flowchart of S5 of the present invention; Figure 7 This is a system flowchart of the present invention. Detailed Implementation

[0016] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0017] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0018] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0019] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0020] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0021] Please see Figure 1 This invention provides a method for optimizing structural mechanics parameters based on artificial intelligence. The processing flow of this method may include the following steps: S1: Based on the structural design target requirements, the initial structural model and its performance constraints are invoked to identify multiple combinations of candidate design variables. A deep learning algorithm is then used for preliminary screening to obtain a set of design variables that meet the constraints. S2: Based on the set of design variables that meet the constraints, extract the mechanical response characteristics of the model under different working conditions, analyze the structural stability and failure risk, identify key weak areas, and obtain optimized potential area data. S3: Based on the optimized potential region data, the mechanical response data is identified by finite element simulation, a mechanical behavior prediction network is constructed, and the parameter distribution is dynamically adjusted through reinforcement learning strategy to optimize the structural layout and improve mechanical performance, thus obtaining the optimized structural design scheme. S4: Based on the optimized structural design scheme, call the structural performance verification records in the original archive, test the robustness of the design scheme, analyze its reliability in long-term use, and obtain the structural performance optimization results. S5: Based on the results of structural performance optimization, extract the key performance indicators of the optimization scheme, analyze their performance under different working conditions, screen the working conditions with performance fluctuations, combine the mechanical behavior prediction network, adjust the design scheme, and obtain the conclusion of structural mechanical performance optimization trend analysis. The set of design variables that satisfy the constraints includes geometric data, material property configuration, and stress distribution information. The optimized potential area data includes the location of weak areas, stress concentration distribution, and failure mode prediction. The optimized structural design scheme includes the improved geometry, material distribution adjustment scheme, and mechanical performance improvement indicators. The structural performance optimization results include robustness assessment conclusions and long-term reliability analysis data. The structural mechanical performance optimization trend analysis conclusions include the changing trends of key performance indicators, working condition adaptability analysis, and optimization potential assessment.

[0022] Specifically, such as Figure 2 As shown, the specific steps for setting a design variable set that satisfies the constraints are as follows: S101: Based on the structural design target requirements, extract the geometric shape, material properties and boundary conditions of the initial structural model, construct multiple sets of candidate design variable combinations, mark the constraints of each set of variables, remove combinations that do not meet the basic requirements, store them in the candidate variable database, and generate a candidate variable sequence. The initial structural model is the existing design or benchmark design of the object to be optimized. It is created by computer-aided design (CAD) software or obtained from existing engineering databases and historical design documents. It defines in detail all the basic physical information of the structure, such as its geometry, material properties, boundary conditions and external loads. Based on the structural design objectives, the initial structural model is assumed to be an aero-engine compressor blade. Its geometry includes parameters such as blade chord length, thickness distribution, twist angle, and curvature. The material properties are set as TC4 titanium alloy with an elastic modulus of 110 GPa, Poisson's ratio of 0.34, and yield strength of 880 MPa. The boundary conditions are set as fixed at the blade root and free at the blade tip, with the blade surface subjected to aerodynamic loads. Multiple sets of candidate design variable combinations are constructed, where the blade chord length is adjusted in 1 mm increments between 200 mm and 220 mm, and the blade thickness is adjusted in 0.1 mm increments between 4 mm and 6 mm. The length is adjusted, the blade twist angle is adjusted in 0.5° increments between 15° and 20°, and the curvature is adjusted in 0.1° increments between 10° and 12°. The constraints for each set of variables include blade mass less than 0.5 kg, blade modal vibration frequency must avoid the engine's operating speed range, and maximum blade stress less than 90% of the material's yield strength, i.e., maximum stress less than 792 MPa. Combinations that do not meet the basic requirements are removed, such as combinations with blade mass exceeding 0.5 kg or combinations with maximum blade stress exceeding 792 MPa. The results are stored in the candidate variable database to generate a candidate variable sequence.

[0023] S102: Based on the candidate variable sequence, call the deep learning algorithm to evaluate the performance of each group of variables, select the design variable combination that meets the constraints, store it in the design variable database, update the selection status flag, and obtain the design variable set that meets the constraints; Based on the candidate variable sequence, a deep learning algorithm is invoked. This algorithm serves as a proxy model for finite element simulation analysis. It has been pre-trained to learn the complex mapping relationship from candidate variable combinations to structural performance evaluation results. Using candidate variable combinations as input and structural performance evaluation results as output, the algorithm evaluates the performance of each set of variables. This evaluation process is simulated by performing finite element simulation analysis on the blade. The finite element simulation analysis is used to generate the dataset required for training the deep learning model. After the deep learning algorithm predicts the results, it verifies some key design schemes. In the rapid evaluation phase, the deep learning algorithm can quickly predict the structural performance evaluation results of each set of variables based on its learned knowledge, significantly reducing the time-consuming finite element calculations. For each candidate variable combination, static analysis is performed to obtain stress distribution, modal analysis is performed to obtain vibration frequency, and fatigue analysis is performed to obtain fatigue life. The combination of design variables that meets the constraints is then selected. The selection process is carried out by comparing simulation results with constraints, such as determining whether the maximum stress of the blade is less than 792 MPa and whether the first-order vibration frequency of the blade is outside the engine operating speed (e.g., 3000-5000 Hz). The combination that meets all constraints is stored in the design variable database, and the selection status is updated to "selected", thus obtaining the set of design variables that meet the constraints.

[0024] Specifically, such as Figure 3 As shown, the specific steps for optimizing the potential region data are as follows: S201: Based on the set of design variables that meet the constraints, extract the mechanical response characteristics of the model under different working conditions, including stress distribution, deformation and fatigue life, analyze its stability and failure risk, identify key weak areas, and obtain a distribution map of weak areas. Based on a set of design variables satisfying the constraints, a set of blade design schemes satisfying the constraints is extracted from the design variable database. These schemes have a blade chord length of 210 mm, a thickness of 5.2 mm, a twist angle of 18°, and a curvature of 11.5°. The mechanical response characteristics of this model under different operating conditions are extracted, including rated speed, overspeed, vibration, and shock. Under rated speed conditions (e.g., 6000 rpm), static simulations are performed on the blade, and stress distribution data on the blade surface is extracted. The maximum stress value is obtained as 650 MPa, and the deformation at the blade tip is extracted. The deformation was 1.2 mm. Under vibration and impact conditions, an impact load of 200 g and 4000 Hz was applied to perform transient dynamic simulation. The fatigue life data of the blade under impact was extracted, and its stability and failure risk were analyzed. Key weak areas were identified. Specifically, by analyzing the stress distribution map, areas with stress values ​​exceeding 600 MPa were marked as high stress areas, areas with deformation exceeding 1 mm were marked as large deformation areas, and areas with fatigue life less than 500,000 cycles were marked as fatigue weak areas, thus obtaining a weak area distribution map.

[0025] S202: Based on the distribution map of weak areas, extract the location and size of stress concentration points, combine with the failure mode prediction model, calculate the potential failure probability value, record key mechanical behavior characteristics, and obtain optimized potential area data. The potential failure probability value is calculated using the formula: ; in, Represents the potential failure probability value. This represents the total number of stress concentration points. Representing the Stress values ​​at stress concentration points This represents the average stress value at the stress concentration point. Representing the The distance from each stress concentration point to the boundary of the weak area. The maximum stress value at the stress concentration point. This represents the average distance from the stress concentration point to the boundary. Representing the Index of stress concentration points; The "potential failure probability value" is a key indicator used to assess the risk of failure in certain stress concentration areas of a structure under specific working conditions. This probability value is calculated by combining the stress magnitude at the stress concentration point with its spatial distribution. The core idea is that the higher the stress and the farther away from the boundary of the weak area, the greater the potential failure probability. Specifically, multiple stress concentration points are first identified, and their stress values ​​and distances from the boundary are extracted. Then, a specific formula is used to weight and summarize the failure risks of each point to obtain the overall potential failure probability value. This indicator not only reflects local safety hazards in the structure but also provides a quantitative basis for subsequent structural optimization, improving the overall stability and reliability of the structure. This method allows for the scientific identification of high-risk areas, avoiding the design deficiencies caused by neglecting stress distribution characteristics in traditional methods. Based on the distribution map of weak areas, multiple stress concentration points are identified, such as at the transition between the leaf root and leaf blade, and at the leading edge of the leaf tip, and their stress values ​​are obtained. MPa MPa MPa, etc., combined with the failure mode prediction model, which takes the stress value at the stress concentration point, stress gradient, and distance from the stress concentration point to the boundary of the weak area as input, calculates the potential failure probability value, records key mechanical behavior characteristics, such as the stress peak value, stress gradient, and average distance from the boundary at the stress concentration point, and obtains the optimized potential area data.

[0026] The calculation process for the potential failure probability value P is as follows: First, the parameters in this formula... This represents the total number of stress concentration points, such as those identified in this study. One stress concentration point; Representing the The stress value at each stress concentration point is expressed in MPa. The average stress value at stress concentration points is calculated by adding the stress values ​​at all stress concentration points and then dividing by the number of points. get; Representing the The distance from each stress concentration point to the boundary of the weak area is expressed in mm. The maximum stress value at the stress concentration point is expressed in MPa. This represents the average distance from each stress concentration point to the boundary, calculated by summing the distances from all stress concentration points to the boundary and then dividing by the number of points. get; The formula's operational logic lies in comprehensively assessing the failure risk of each stress concentration point by using the difference between the stress value at the stress concentration point and the average stress value, as well as the distance from the stress concentration point to the boundary of the weak region. (The product of the numerators...) This reflects that the greater the deviation of stress from the average value and the farther the point is from the boundary, the greater its contribution to the failure probability. (Denominator) As a normalization factor, its unit is MPa·mm, consistent with the numerator unit, thus ensuring the dimensionality of the formula and keeping the probability value within a reasonable range. Finally, by summing and averaging the contributions of all stress concentration points, the overall potential failure probability is obtained. ; Now, substituting the parameters into the formula for calculation, assuming five stress concentration points are identified, their stress values ​​and distances from the boundary are as follows: MPa mm MPa mm MPa mm MPa mm MPa mm; First, calculate the average stress value at the stress concentration point. : MPa; Then calculate the average distance from the stress concentration point to the boundary. : mm; Maximum stress value at stress concentration point It is 695 MPa; Substitute these values ​​into the formula: ; The result indicates that the potential failure probability of this structure under current operating conditions is [value missing]. This value represents a relatively low failure probability, meaning that the failure risk in this area is within a controllable range. This potential failure probability value, as part of the optimized potential area data, will be used for subsequent structural layout optimization. The advantage of the formula lies in introducing the distance from the stress concentration point to the boundary of the weak region. As a weighting factor, it comprehensively considers the stress magnitude and the spatial distribution of stress concentration areas, enabling a more comprehensive assessment of failure risk. When the stress value deviates significantly from the average value and is far from the boundary, its contribution to the failure probability is greater. This helps to identify stress concentration points that are high in stress but located in non-critical positions, and to focus on critical points that are high in stress and close to the failure boundary.

[0027] Specifically, such as Figure 4 As shown, the specific steps of the optimized structural design scheme are as follows: S301: Based on the optimized potential region data, extract the mechanical behavior features of the weak region, construct a mechanical behavior prediction network, use reinforcement learning strategy to dynamically adjust the parameter distribution, optimize the structural layout, improve the overall mechanical performance, and obtain a preliminary optimization scheme; Based on optimized potential region data, including stress values, stress gradients, and potential failure probabilities at stress concentration points, a mechanical behavior prediction network is constructed. This network typically employs a deep neural network structure (such as a multilayer perceptron or convolutional neural network) and undergoes supervised learning training using a large amount of historical finite element simulation data (containing the mechanical behavior characteristics of weak regions under different structural parameters and their corresponding mechanical responses, such as stress distribution and deformation). This allows the network to learn and capture the complex nonlinear mapping relationship between input features and structural mechanical responses. After training, the network can quickly and accurately predict the optimized stress distribution using the extracted mechanical behavior features as input. The specific adjustment process involves: first, setting a reward function for reinforcement learning, which aims to quantify the contribution of each parameter adjustment to the improvement of structural performance. It typically combines the improvements of multiple performance indicators through a weighted summation. Specifically, the reward function can be defined as: Reward (R) = w1 * (reduction in maximum stress value) + w2 * (reduction in deformation) + w3 * (reduction in potential failure probability value), where w1, w2, and w3 are weighting coefficients used to balance the importance of different performance indicators, and their values ​​can be adjusted according to design requirements and expert experience. For example, if the maximum stress is the most critical indicator in the current optimization, w1 can be set relatively high. These "reductions" or "decreases" are performance improvements relative to the previous iteration or initial state, i.e., (previous iteration value - current iteration value). The reduction in maximum stress, deformation, and potential failure probability are used as rewards. In each iteration, the agent selects an action based on the current structural parameters and mechanical behavior characteristics, such as adjusting the material thickness of weak areas, changing the size of stiffeners, or adjusting the internal geometry of the blade. The environment executes the action and returns a new structural model and mechanical response data. The agent calculates the reward value based on the new data and updates its strategy. For example, if the agent chooses to increase the thickness at the blade root by 0.5 mm, and finite element simulation shows that the maximum stress value decreases by 20 MPa, the reward value increases, and the agent will be more inclined to choose this action. Through repeated iterations, the agent can learn the optimal parameter adjustment strategy, optimize the structural layout, and improve the overall mechanical performance. For example, reducing the maximum stress value at the blade root from 650 MPa to 600 MPa yields a preliminary optimization scheme.

[0028] S302: Based on the preliminary optimization scheme, analyze the improved geometry and material distribution, compare the mechanical performance indicators before and after optimization, statistically analyze the optimization effect, and obtain the optimized structural design scheme. Based on the preliminary optimization scheme, the improved geometry and material distribution are analyzed. For example, the geometric changes such as the increase of 0.5 mm in the thickness at the blade root and the increase of 0.2 mm in the chamfer radius at the blade tip are analyzed, as well as the changes in material distribution due to the increase in material density in certain areas. The mechanical performance indicators before and after optimization are compared, specifically the maximum stress, maximum deformation, and first-order vibration frequency. For example, the maximum stress before optimization was 650 MPa, and after optimization it was 600 MPa. The maximum deformation decreased from 1.2 mm to 1.0 mm, and the first-order vibration frequency was adjusted from 4500 Hz to 4800 Hz. The optimization effect is statistically analyzed, and the optimization effect is quantified by calculating the percentage improvement of performance indicators. For example, the percentage reduction in maximum stress is (650-600) / 650%=7.69%, resulting in the optimized structural design scheme.

[0029] Specifically, such as Figure 5 As shown, the specific steps for optimizing structural performance are as follows: S401: Based on the optimized structural design scheme, extract the structural performance verification records from the original archive, including failure cases and reliability test data during long-term use, detect the robustness of the design scheme, analyze its adaptability under different environmental conditions, and obtain robustness assessment conclusions. Based on the optimized structural design scheme, for example, service data of the structural model in the past five years can be retrieved from the archive, including 12 failure cases caused by fatigue cracks, 5 failure cases caused by excessive deformation, and reliability test data under vibration table, high temperature and high pressure and other environments, to test the robustness of the design scheme. The specific testing method is as follows: under uncertain environment, such as aerodynamic load fluctuation amplitude of % and material performance deviation of %, Monte Carlo simulation is performed on the optimized design scheme to analyze its performance under uncertainty and its adaptability under different environmental conditions. For example, thermal stress simulation is performed in the temperature range of -40℃ to 100℃ to evaluate its thermal stability and obtain robustness assessment conclusions.

[0030] S402: Based on the robustness assessment conclusions and combined with the long-term reliability analysis model, predict the performance change trend of the structural design scheme within its service life, statistically analyze the fluctuation range of key performance indicators, and obtain the structural performance optimization results. Based on the robustness assessment conclusions and combined with the long-term reliability analysis model, this model takes the robustness assessment conclusions, failure case data, and reliability test data as inputs to predict the performance change trend of the structural design scheme during its service life. For example, it predicts the maximum stress value change curve of the blade during its 10,000-hour service life and the fatigue life decay trend, and statistically analyzes the fluctuation range of key performance indicators. For example, the fluctuation range of the maximum stress value is 600MPa, and the fluctuation range of the first-order vibration frequency is 4800Hz, thus obtaining the structural performance optimization results.

[0031] Specifically, such as Figure 6 As shown, the specific steps for analyzing the trend of structural mechanical performance optimization are as follows: S501: Based on the results of structural performance optimization, extract key performance index data under different working conditions, remove outliers, match working condition numbers, analyze the performance fluctuation range, and obtain a set of performance indicators. Based on the results of structural performance optimization, key performance index data under different operating conditions are extracted, including maximum stress values ​​and deformation data under rated speed, overspeed, vibration and shock, and high and low temperature conditions. For example, the maximum stress is 600 MPa under rated speed condition, 720 MPa under overspeed condition, and 1.0 mm under vibration and shock condition. Outliers are removed. For example, stress data exceeding the material yield strength (880 MPa) are marked as outliers and removed. Operating condition numbers are matched, and a corresponding operating condition number is assigned to each data point. For example, the maximum stress of 600 MPa is matched as operating condition 1 (rated speed). The performance fluctuation range is analyzed. For example, the fluctuation range of the maximum stress value under all operating conditions is statistically analyzed, from 600 MPa to 720 MPa, to obtain a set of performance indexes.

[0032] S502: Based on the performance index set, the working conditions are classified and arranged according to the working condition type. The working conditions with large performance fluctuations are identified, their mechanical behavior characteristics are analyzed, and combined with the optimization potential assessment model, the optimization space is identified, and the conclusions of the structural mechanical performance optimization trend analysis are obtained. Based on a set of performance indicators, such as dividing all data into three categories: "speed conditions," "vibration conditions," and "temperature conditions," conditions with large performance fluctuations are identified. This is done by calculating the standard deviation of performance indicators under each condition. Conditions with a standard deviation exceeding 50 MPa are identified as having large fluctuations. Their mechanical behavior characteristics are then analyzed. For example, under overspeed conditions, the centrifugal force on the blade increases significantly, leading to significant stress concentration. Combined with an optimization potential assessment model, which takes the fluctuation range of the condition, the degree of stress concentration, and the failure risk as inputs, the optimization space is identified. For example, it is found that there is a huge optimization space in the root and tip regions of the blade under overspeed conditions, leading to the conclusion of structural mechanical performance optimization trend analysis.

[0033] like Figure 7 As shown, an artificial intelligence-based structural mechanics parameter optimization system includes: The design variable selection module calls the design parameter set and performance constraints in the initial structural model according to the design target requirements, groups the design parameter set, checks whether each group of parameters meets the performance constraints, and selects the design variable combinations that meet the requirements to obtain the candidate design variable set. The response analysis module extracts the mechanical response data of the model under different working conditions based on the candidate design variable set, and analyzes the nodal stress distribution, strain behavior and mechanical stability under each set of design variables. It compares and analyzes the mechanical response of each combination, marks high-risk areas, and obtains the stress response set of key areas. The optimization scheme generation module is based on the stress response set of key areas. It adjusts and optimizes the structural geometric parameters, material properties and boundary conditions in high-risk areas. By analyzing stress distribution and local deformation, it dynamically adjusts parameters, optimizes the structural layout, and obtains the optimized structural parameter set. The stability verification module, based on the optimized set of structural parameters, calls the original performance verification data, combines long-term usage data and environmental change parameters, analyzes the stability and durability of the structure under different usage conditions, identifies potential failure points, and obtains structural durability assessment results. Based on the structural durability assessment results and combined with environmental change parameters, the adaptive adjustment module analyzes the structural adaptability and performance fluctuations, adjusts design variables that are unstable under target working conditions, optimizes the design scheme, optimizes structural adaptability, and obtains the conclusion of structural mechanical performance optimization trend analysis.

[0034] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for optimizing structural mechanics parameters based on artificial intelligence, characterized in that, Includes the following steps: S1: Based on the structural design target requirements, the initial structural model and its performance constraints are invoked to identify multiple combinations of candidate design variables. A deep learning algorithm is then used for preliminary screening to obtain a set of design variables that meet the constraints. S2: Based on the set of design variables that meet the constraints, extract the mechanical response characteristics of the model under different working conditions, analyze the structural stability and failure risk, identify key weak areas, and obtain optimized potential area data. S3: Based on the optimized potential region data, mechanical response data is identified using finite element simulation, a mechanical behavior prediction network is constructed, and parameter distribution is dynamically adjusted through reinforcement learning strategy to optimize structural layout and improve mechanical performance, thereby obtaining an optimized structural design scheme. S4: Based on the optimized structural design scheme, call the structural performance verification records in the original archive, detect the robustness of the design scheme, analyze its reliability in long-term use, and obtain the structural performance optimization results. S5: Based on the structural performance optimization results, extract the key performance indicators of the optimization scheme, analyze their performance under different working conditions, screen the performance fluctuation working conditions, combine the mechanical behavior prediction network, adjust the design scheme, and obtain the structural mechanical performance optimization trend analysis conclusion. The conclusions of the structural mechanical performance optimization trend analysis include the changing trends of key performance indicators, working condition adaptability analysis, and optimization potential assessment.

2. The structural mechanics parameter optimization method based on artificial intelligence according to claim 1, characterized in that, The set of design variables that satisfy the constraints includes geometric morphology data, material property configuration, and stress distribution information. The optimized potential region data includes the location of weak regions, stress concentration distribution, and failure mode prediction. The optimized structural design scheme includes the improved geometry, material distribution adjustment scheme, and mechanical performance improvement index. The structural performance optimization results include robustness assessment conclusions and long-term reliability analysis data.

3. The structural mechanics parameter optimization method based on artificial intelligence according to claim 1, characterized in that, The specific steps for establishing the design variable set that satisfies the constraints are as follows: S101: Based on the structural design target requirements, extract the geometric shape, material properties and boundary conditions of the initial structural model, construct multiple sets of candidate design variable combinations, mark the constraints of each set of variables, remove combinations that do not meet the basic requirements, store them in the candidate variable database, and generate a candidate variable sequence. S102: Based on the candidate variable sequence, call the deep learning algorithm to evaluate the performance of each group of variables, select the design variable combination that meets the constraints, store it in the design variable database, update the selection status flag, and obtain the design variable set that meets the constraints.

4. The structural mechanics parameter optimization method based on artificial intelligence according to claim 3, characterized in that, The specific steps for optimizing the potential region data are as follows: S201: Based on the set of design variables that meet the constraints, extract the mechanical response characteristics of the model under different working conditions, including stress distribution, deformation and fatigue life, analyze its stability and failure risk, identify key weak areas, and obtain a distribution map of weak areas. S202: Based on the weak area distribution map, extract the location and size of stress concentration points, combine with the failure mode prediction model, calculate the potential failure probability value, record key mechanical behavior characteristics, and obtain optimized potential area data.

5. The structural mechanics parameter optimization method based on artificial intelligence according to claim 4, characterized in that, The potential failure probability value is calculated using the following formula: ； in, Represents the potential failure probability value. This represents the total number of stress concentration points. Representing the Stress values ​​at stress concentration points This represents the average stress value at the stress concentration point. Representing the The distance from each stress concentration point to the boundary of the weak area. The maximum stress value at the stress concentration point. This represents the average distance from the stress concentration point to the boundary. Representing the Index of stress concentration points.

6. The structural mechanics parameter optimization method based on artificial intelligence according to claim 4, characterized in that, The specific steps of the optimized structural design scheme are as follows: S301: Based on the optimized potential region data, extract the mechanical behavior features of the weak region, construct a mechanical behavior prediction network, use reinforcement learning strategy to dynamically adjust the parameter distribution, optimize the structural layout, improve the overall mechanical performance, and obtain a preliminary optimization scheme. S302: Based on the preliminary optimization scheme, analyze the improved geometry and material distribution, compare the mechanical performance indicators before and after optimization, statistically analyze the optimization effect, and obtain the optimized structural design scheme.

7. The structural mechanics parameter optimization method based on artificial intelligence according to claim 6, characterized in that, The specific steps for achieving the structural performance optimization results are as follows: S401: Based on the optimized structural design scheme, extract the structural performance verification records from the original archive, including failure cases and reliability test data during long-term use, detect the robustness of the design scheme, analyze its adaptability under different environmental conditions, and obtain a robustness assessment conclusion. S402: Based on the robustness assessment conclusions and combined with the long-term reliability analysis model, predict the performance change trend of the structural design scheme within its service life, statistically analyze the fluctuation range of key performance indicators, and obtain the structural performance optimization results.

8. The structural mechanics parameter optimization method based on artificial intelligence according to claim 1, characterized in that, The specific steps for drawing conclusions regarding the structural mechanical performance optimization trend analysis are as follows: S501: Based on the structural performance optimization results, extract key performance index data under different working conditions, remove outliers, match working condition numbers, analyze performance fluctuation range, and obtain a set of performance indicators. S502: Based on the set of performance indicators, classify and arrange them according to the type of working condition, identify the working conditions with large performance fluctuations, analyze their mechanical behavior characteristics, combine with the optimization potential assessment model, identify the optimization space, and obtain the structural mechanical performance optimization trend analysis conclusion.

9. A structural mechanics parameter optimization system based on artificial intelligence, characterized in that, The system is used to implement the artificial intelligence-based structural mechanics parameter optimization method according to any one of claims 1-8, and the system comprises: The design variable filtering module calls the design parameter set and performance constraints in the initial structural model according to the design target requirements, groups the design parameter set, checks whether each group of parameters meets the performance constraints, and filters the design variable combinations that meet the requirements to obtain the candidate design variable set. The response analysis module extracts the mechanical response data of the model under different working conditions based on the candidate design variable set, and analyzes the nodal stress distribution, strain behavior and mechanical stability under each set of design variables. It compares and analyzes the mechanical response of each combination, marks high-risk areas, and obtains the stress response set of key areas. The optimization scheme generation module is based on the stress response set of key areas. It adjusts and optimizes the structural geometric parameters, material properties and boundary conditions in high-risk areas. By analyzing stress distribution and local deformation, it dynamically adjusts parameters, optimizes the structural layout, and obtains the optimized structural parameter set. The stability verification module, based on the optimized set of structural parameters, calls the original performance verification data, combines long-term usage data and environmental change parameters, analyzes the stability and durability of the structure under different usage conditions, identifies potential failure points, and obtains the structural durability assessment results. Based on the structural durability assessment results and combined with environmental change parameters, the adaptive adjustment module analyzes the structural adaptability and performance fluctuations, adjusts design variables that are unstable under target working conditions, optimizes the design scheme, optimizes structural adaptability, and obtains the trend analysis conclusion of structural mechanical performance optimization.

Images