A flat electronic harness environmental adaptability evaluation method and system

Abstract

The application provides a flat electronic wire harness environmental adaptability evaluation method and system, comprising: S3, extracting key parameters from quantitative characteristics of failure mechanism, combining immersion time variable, recording physical property changes of material aging under different immersion times through accelerated aging test, obtaining function relationship between immersion time and material aging rate; S6, according to the change rule of dynamic performance, a multivariate evaluation system containing water pressure change, immersion time and temperature cycle is established in advance, the weight coefficients of each variable and performance attenuation are fitted through regression analysis, and a comprehensive performance quantitative scale formula is obtained; S8, the deviation of performance prediction value and measured value is extracted from the environmental adaptability evaluation result, a machine learning algorithm is used to iteratively optimize the evaluation system, the weight coefficients are adjusted to reduce the deviation, and an optimized performance quantitative scale system is obtained. The application provides an important basis for design optimization and reliability improvement of the flat electronic wire harness.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing technology, and in particular to a method and system for evaluating the environmental adaptability of flat electronic wire harnesses. Background Technology

[0002] Waterproof performance testing of flat electronic wire harnesses is a crucial research area, holding a central position in industries such as electronics manufacturing, automotive, and consumer electronics; its importance is self-evident. With the advancement of miniaturization and integration of electronic devices, flat wire harnesses are widely used due to their high space efficiency and flexible wiring. Waterproof performance directly affects the reliability and lifespan of products in harsh environments, becoming a key indicator of their quality. However, the testing and evaluation of waterproof performance is not a simple task, involving comprehensive consideration of multiple technical factors. Current research and practice are increasingly revealing shortcomings, necessitating a more systematic solution.

[0003] Existing waterproof performance testing methods mostly rely on single test conditions or general standards, such as simple immersion tests or directly applying international protection ratings (e.g., IP67). While these methods are easy to operate, they often fail to fully reflect the actual waterproof capability of the wiring harness when facing complex application scenarios, especially its performance under dynamic conditions such as water pressure changes, long-term immersion, or temperature cycling. Furthermore, the test results typically lack detailed quantitative data, making it difficult for designers to accurately match wiring harness performance with specific requirements, easily leading to performance redundancy or inadequacy.

[0004] The core challenge in this field lies in how to scientifically quantify waterproof performance and establish an evaluation system adapted to actual application environments. Focusing on the three technical factors of water pressure, immersion time, and temperature cycling, current research has not yet effectively addressed the performance evaluation biases caused by the interaction of these variables. For example, the failure mechanism of the wiring harness sealing structure under different water pressures is not clearly defined, and the impact of prolonged immersion time combined with temperature cycling on material aging lacks systematic data support. These technical challenges make it difficult to establish a precise scale for waterproof performance, thus limiting the reliable application of flat wiring harnesses in extreme environments.

[0005] Therefore, designing a quantitative calibration system with progressive testing conditions to accurately evaluate the waterproof performance of flat electronic wire harnesses under different water pressures, immersion times, and temperature cycles is a key problem that this research urgently needs to solve. Solving this problem will provide a scientific basis for wire harness design and application, avoiding the blind spots and uncertainties in performance evaluation. Summary of the Invention

[0006] To address the technical problems mentioned in the background section, a first aspect of the present invention provides a method for evaluating the environmental adaptability of flat electronic wire harnesses, the method comprising:

[0007] S1, acquire the sealing structure response data of the flat electronic wire bundle under different water pressure changes, and record the deformation of the sealing structure and the location of the leakage point in real time when the water pressure changes from low to high through the sensor, so as to obtain the correspondence between water pressure change and sealing structure stability.

[0008] S2. Based on the correspondence between water pressure changes and the stability of the sealing structure, the failure mechanism is analyzed for the location of the leakage point. Finite element simulation technology is used to simulate the stress distribution and damage threshold of the sealing structure under water pressure, and the quantitative characteristics of the failure mechanism are determined.

[0009] S3. Key parameters are extracted from the quantitative characteristics of the failure mechanism. Combined with the immersion time variable, the changes in the physical properties of the material under different immersion times are recorded through accelerated aging tests to obtain the functional relationship between immersion time and material aging rate.

[0010] S4, based on the functional relationship between immersion time and material aging rate, superimposed temperature cycling conditions, and through environmental simulation equipment, cyclically tested wire harness samples within a preset temperature range to obtain data on the superimposed influence of temperature cycling on material aging and dynamic performance.

[0011] S5 uses the combined effect data of temperature cycling on material aging and dynamic performance to calculate the performance degradation trend under various test conditions using statistical analysis methods, and judges the change law of dynamic performance under the interaction of water pressure change, immersion time and temperature cycling.

[0012] S6. Based on the dynamic performance change pattern, a multivariate evaluation system including water pressure change, soaking time and temperature cycle is established in advance. By fitting the weight coefficients of each variable and performance decay through regression analysis, a comprehensive performance quantification formula is obtained.

[0013] S7. After obtaining the comprehensive performance quantification formula, input the water pressure, immersion time and temperature cycle parameters of the actual application scenario for specific environmental adaptation requirements, calculate the performance prediction value of the harness in the scenario through the formula, and determine the environmental adaptability assessment result.

[0014] S8 extracts the deviation between the predicted and measured values ​​of performance from the environmental adaptability assessment results, uses machine learning algorithms to iteratively optimize the assessment system, adjusts the weight coefficients to reduce the deviation, and obtains an optimized performance quantification scale system.

[0015] Optionally, step S4, based on the functional relationship between immersion time and material aging rate, superimposes temperature cycling conditions and uses an environmental simulation device to cyclically test the wire harness sample within a preset temperature range to obtain data on the superimposed influence of temperature cycling on material aging and dynamic performance, including:

[0016] Step S41: Perform a temperature cycling test on the wire harness sample using an environmental simulation device, and record the immersion time and material hardness reduction percentage of the sample in each cycle as raw data of the aging rate.

[0017] Step S42: Extract the corresponding values ​​of soaking time and hardness reduction percentage from the original data, and use the least squares method to fit the linear function relationship between the two to obtain a preliminary model of aging rate.

[0018] Step S43: Introduce temperature cycling parameters into the preliminary model, and use the difference between the highest and lowest temperatures in each cycle as an additional variable to generate the adjusted dataset.

[0019] Step S44: Use multiple linear regression to model the adjusted data and fit the equation;

[0020] Step S45: Determine the change in aging rate when the difference between the highest and lowest temperatures increases by 1 unit based on the regression coefficient.

[0021] Step S46: Calculate the dynamic performance offset of each sample based on the regression results, and construct a two-dimensional feature vector by combining the difference between the highest and lowest temperatures and the dynamic performance offset. Use K-means clustering to divide the samples into three groups of aging sensitivity categories: high, medium, and low.

[0022] Step S47: Statistically analyze the distribution range of the percentage decrease in hardness for each type of sample. If the average percentage decrease in hardness for the high-sensitivity group exceeds the threshold for percentage decrease in hardness, it is determined that the sample in this group has a synergistic accelerated aging effect of temperature cycling and immersion time.

[0023] Optionally, step S42, which involves extracting the corresponding values ​​of soaking time and hardness reduction percentage from the original data, fitting the linear function relationship between the two using the least squares method to obtain a preliminary model of the aging rate, further includes:

[0024] Step S421, calculate the percentage decrease in hardness using the following formula:

[0025] Q=aT1+b

[0026] Where Q is the percentage decrease in hardness, T1 is the soaking time, a is the first soaking time coefficient, and b is the first temperature difference coefficient.

[0027] Optionally, step S44, modeling and fitting an equation using multiple linear regression on the adjusted data, further includes:

[0028] Step S441, the fitting equation is:

[0029] Q = a'T1 + b'ΔT + c

[0030] Where a' is the second soaking time coefficient, b' is the second temperature difference coefficient, ΔT is the difference between the highest and lowest temperatures, and c is a constant.

[0031] Optionally, step S46, which calculates the dynamic performance offset for each sample based on the regression results, constructs a two-dimensional feature vector by combining the difference between the highest and lowest temperatures with the dynamic performance offset, and uses K-means clustering to classify the samples into three aging sensitivity categories: high, medium, and low, further includes:

[0032] Step S461, calculate the dynamic performance offset according to the following formula:

[0033] δ=|Q-(a'T1+b'ΔT+c)|

[0034] Where δ is the dynamic performance offset.

[0035] Optionally, in step S47, the distribution range of the percentage decrease in hardness for each type of sample is statistically analyzed. If the average percentage decrease in hardness for the high-sensitivity group exceeds the threshold for percentage decrease in hardness, it is determined that the sample in this group has a synergistic accelerated aging effect of temperature cycling and immersion time, including setting the threshold for percentage decrease in hardness to 15%.

[0036] Optionally, step S5 involves using statistical analysis to calculate the performance degradation trend under various test conditions based on the combined effects of temperature cycling on material aging and dynamic performance, and determining the variation law of dynamic performance under the interaction of water pressure change, immersion time, and temperature cycling, including:

[0037] Step S51: Obtain the material aging feature matrix through temperature cycling and immersion time data, and use principal component analysis to extract the first three principal components as aging feature vectors.

[0038] Step S52: Input the aging feature vector into the ridge regression model and output two dynamic performance parameters: elastic modulus and elongation at break.

[0039] Step S53: Based on the time series data of dynamic performance parameters, calculate the performance degradation rate using the exponential smoothing method to obtain the degradation trend curve;

[0040] Step S54: Construct a two-factor experimental matrix for water pressure change and soaking time data, and detect the significance of interaction by variance expansion factor;

[0041] Step S55: Partial least squares regression is used to calculate the interaction weights between water pressure change and soaking time, and the weight coefficients of water pressure change and soaking time are obtained.

[0042] Step S56: Combine the interaction weight coefficients with the aging feature vector, input them into the ridge regression model, and establish the performance degradation rate prediction equation.

[0043] Step S57: When the performance degradation rate exceeds the performance degradation rate threshold, the standard deviation of the temperature cycling data is introduced as a correction factor, and the interaction weights are recalculated using the weighted average method.

[0044] Step S58: The corrected weight coefficients are superimposed on the original aging feature vector, and the decay distribution surface of dynamic performance under multi-factor conditions is generated by linear interpolation.

[0045] Step S59: Perform Z-score normalization on the attenuation distribution surface, input it into the hierarchical clustering algorithm, and determine the optimal number of clusters as 3 based on the silhouette coefficient;

[0046] Step S510 outputs three types of dynamic performance degradation modes: rapid degradation region, linear degradation region, and stable degradation region, completing the performance classification feature extraction under multiple conditions.

[0047] Optionally, in step S55, partial least squares regression is used to calculate the interaction weight between water pressure change and soaking time to obtain the water pressure change weight coefficient and the soaking time weight coefficient, including: the water pressure change weight coefficient is 0.6 and the soaking time weight coefficient is 0.4.

[0048] Optionally, in step S57, when the performance degradation rate exceeds the performance degradation rate threshold, the standard deviation of the temperature cycling data is introduced as a correction factor, and the interaction weight is recalculated using a weighted average method, including setting the performance degradation rate threshold to 0.5.

[0049] A second aspect of the present invention provides a system for evaluating the environmental adaptability of flat electronic wire harnesses, which uses the method described above to evaluate the environmental adaptability of flat electronic wire harnesses. The system includes:

[0050] The data acquisition module is used to acquire the response data of the sealing structure of the flat electronic wire bundle under different water pressure changes. The sensor records the deformation of the sealing structure and the location of the leakage point in real time when the water pressure changes from low to high, and obtains the correspondence between water pressure change and the stability of the sealing structure.

[0051] The failure analysis module is used to analyze the failure mechanism based on the correspondence between water pressure changes and the stability of the sealing structure, targeting the leakage point. It uses finite element simulation technology to simulate the stress distribution and damage threshold of the sealing structure under water pressure, and determines the quantitative characteristics of the failure mechanism.

[0052] The aging rate modeling module is used to extract key parameters from the quantitative characteristics of failure mechanisms. Combined with the immersion time variable, it records the changes in the physical properties of the material under different immersion times through accelerated aging tests, and obtains the functional relationship between immersion time and material aging rate.

[0053] The temperature superposition test module is used to superimpose temperature cycling conditions on the functional relationship between immersion time and material aging rate. The wire harness sample is cyclically tested within a preset temperature range using an environmental simulation device to obtain data on the superposition effect of temperature cycling on material aging and dynamic performance.

[0054] The performance degradation analysis module is used to calculate the performance degradation trend under various test conditions by using statistical analysis methods to analyze the superimposed effect data of material aging and dynamic performance through temperature cycling, and to determine the change law of dynamic performance under the interaction of water pressure change, immersion time and temperature cycling.

[0055] The multivariate evaluation module is used to pre-establish a multivariate evaluation system that includes water pressure changes, soaking time and temperature cycles based on the dynamic performance change pattern. By fitting the weight coefficients of each variable and performance decay through regression analysis, a comprehensive performance quantification formula is obtained.

[0056] The environmental adaptability prediction module is used to obtain the comprehensive performance quantification formula, input the water pressure, immersion time and temperature cycle parameters of the actual application scenario for specific environmental adaptability requirements, calculate the performance prediction value of the harness in that scenario through the formula, and determine the environmental adaptability assessment result.

[0057] The evaluation and optimization module is used to extract the deviation between the predicted and measured performance values ​​from the environmental adaptability evaluation results. Machine learning algorithms are used to iteratively optimize the evaluation system, adjusting weight coefficients to reduce the deviation, resulting in an optimized performance quantification scale system. The technical solution provided by this invention has the following beneficial effects:

[0058] This invention provides a method and system for evaluating the environmental adaptability of flat electronic wire harnesses. By acquiring the sealing structure response data of the wire harness under different water pressures, the failure mechanism of the leakage point is analyzed. Accelerated aging tests are conducted in conjunction with immersion time and temperature cycling conditions to obtain the changes in material aging and dynamic performance. Based on this, a multivariate evaluation system including water pressure, immersion time, and temperature cycling is established, and a comprehensive performance quantification formula is obtained through regression analysis.

[0059] This invention can predict wire harness performance and evaluate environmental adaptability by inputting actual application scenario parameters for specific environments. The evaluation system is iteratively optimized through machine learning algorithms, thereby improving prediction accuracy.

[0060] This invention enables quantitative evaluation of the performance of flat electronic wire harnesses in complex environments, providing an important basis for their design optimization and reliability improvement. Attached Figure Description

[0061] Figure 1 This is a flowchart of a method for evaluating the environmental adaptability of flat electronic wire harnesses according to the present invention.

[0062] Figure 2 This is a schematic diagram of the structure of a flat electronic wire harness environmental adaptability assessment system according to the present invention. Detailed Implementation

[0063] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0064] like Figure 1 As shown in this embodiment, a method for evaluating the environmental adaptability of flat electronic wire harnesses is provided. The method includes:

[0065] S1, acquire the sealing structure response data of the flat electronic wire bundle under different water pressure changes, and record the deformation of the sealing structure and the location of the leakage point in real time when the water pressure changes from low to high through the sensor, so as to obtain the correspondence between water pressure change and sealing structure stability.

[0066] Optionally, this step also includes:

[0067] Step S11: Acquire the sealing structure response data of the flat electronic wire bundle under water pressure changes using sensors, record the deformation and leakage point locations, and obtain an initial dataset containing the coordinate information of the deformation and leakage points. Step S12: Perform wavelet transform on the initial dataset to denoise it, obtaining smoothed deformation and leakage point data. The smoothed data retains the main features of the deformation and leakage points. Step S13: Use the K-means clustering algorithm to extract the distribution features of the deformation and leakage points from the smoothed data, determining a set of key regions. This set of key regions contains high-density areas of deformation and leakage points. Step S14: If the number of leakage points in the key region set exceeds a leakage point threshold, use linear interpolation to supplement the missing data, obtaining a complete change dataset containing complete information on all deformation and leakage points. Step S15: Based on the complete change dataset, the least squares method is used to calculate the mapping relationship between water pressure change and deformation location to obtain stability trend characteristics. These characteristics describe the linear relationship between water pressure change and deformation location. Step S16: Linear regression analysis is used to extract the correspondence between water pressure change and the stability of the sealing structure from the stability trend characteristics, resulting in a quantitative model. This model can predict the impact of water pressure change on the stability of the sealing structure. Step S17: The output of the quantitative model is obtained. The consistency of the correspondence during the change process is verified by calculating the mean square error between the predicted and actual values, determining the final response law. This final response law describes the quantitative relationship between water pressure change and the stability of the sealing structure.

[0068] Specifically, acquiring the response data of the sealing structure of the flat electronic wire bundle under water pressure changes through sensors is the basis for analyzing its performance.

[0069] For example, on a production line for automotive flat electronic wiring harnesses, sensors are placed in sealed areas to record deformation and leakage data every 0.5 seconds. The water pressure gradually increases from 0.1 MPa to 0.5 MPa. The initial dataset might show that deformation is concentrated at the joints, and leakage points mostly appear at the edge coordinates (x: 3.2, y: 1.5). This coordinate recording of data provides a spatial basis for subsequent analysis.

[0070] In one possible implementation, using wavelet transform to denoise the initial dataset is particularly important. The original data may be interfered with by sensor jitter or environmental noise. Wavelet transform decomposes high-frequency noise and preserves low-frequency trend signals to obtain smoothed deformation and leakage point data.

[0071] For example, the deformation peak smoothed from a noisy 2.3 mm to 2.0 mm, while the location of the leak point remained clear, preserving key features. This step improved the reliability of the data and facilitated subsequent feature extraction.

[0072] Specifically, when using the K-means clustering algorithm to extract distribution features, smoothed data can be divided into three clusters, representing high, medium, and low density regions, respectively. In a wire harness sample, high-density regions may be concentrated in the coordinate set where deformation is greater than 1.8 mm and the number of leakage points exceeds two, such as (x: 3.0-3.5, y: 1.2-1.7). These key region sets intuitively reflect the weak points of the sealing structure, providing a basis for optimization design.

[0073] It should be noted that if the number of leakage points exceeds the preset leakage point threshold, for example, 3 points per square centimeter, while the leakage point threshold is set to 2 points per square centimeter, then linear interpolation can supplement the missing data.

[0074] In one embodiment, a certain deformation data segment was missing at a water pressure of 0.4 MPa. This was reconstructed by interpolating the data at adjacent points (1.7 mm and 1.9 mm) to obtain a 1.8 mm segment, thus forming a complete deformation dataset. This method ensures data continuity and avoids analytical bias.

[0075] Preferably, when using the least squares method to calculate the mapping between water pressure and deformation, a linear relationship can be assumed.

[0076] For example, when the water pressure increases from 0.2 MPa to 0.4 MPa, the deformation increases from 1.5 mm to 1.9 mm. A trend line is fitted, indicating that for every 0.1 MPa increase in water pressure, the deformation increases by approximately 0.2 mm. This stability trend characteristic quantifies the response law and is easy to apply in engineering.

[0077] In one embodiment, linear regression analysis further extracts the relationship between water pressure and seal stability from trend characteristics.

[0078] For example, a stability score dropping from 90% to 70% corresponds to an increase in water pressure to 0.5 MPa, and the quantitative model predicts that stability drops to 50% at a critical water pressure of 0.6 MPa. This provides a predictive tool for designing pressure-resistant wiring harnesses.

[0079] Understandably, when validating a quantification model, calculating the mean square error between the predicted and actual values ​​is crucial.

[0080] For example, the predicted deformation was 2.1 mm, while the actual deformation was 2.0 mm, with an error of only 0.1 mm, indicating high model consistency. The final response pattern reveals that for every 0.1 MPa increase in water pressure, stability decreases by approximately 10%, providing clear technical guidance for optimizing the sealing structure. This method not only improves prediction accuracy but also enhances the reliability of the wiring harness under extreme environments.

[0081] S2. Based on the correspondence between water pressure changes and the stability of the sealing structure, the failure mechanism is analyzed for the location of the leakage point. Finite element simulation technology is used to simulate the stress distribution and damage threshold of the sealing structure under water pressure, and the quantitative characteristics of the failure mechanism are determined.

[0082] Optionally, this step also includes:

[0083] Step S21: Acquire water pressure change data through sensors and combine it with sealing structure parameters to generate an initial dataset. Step S22: Extract leakage point location information from the initial dataset and determine the distribution characteristics of the leakage points using a density estimation method. Step S23: Establish a finite element model based on the distribution characteristics and simulate the stress distribution under water pressure using ANSYS software. Step S24: Acquire stress distribution data and calculate the failure value of the sealing structure based on the stress-strain relationship to obtain the failure value range. Step S25: If the failure value range exceeds the failure threshold, it is determined that stability has decreased, and the preliminary characteristics of the failure mechanism are identified. Step S26: Extract key parameters from the preliminary characteristics and classify the failure mechanism type using the support vector machine algorithm in the scikit-learn library. Step S27: Adjust the finite element model parameters based on the classification results, repeat the simulation, and determine the final quantitative characteristics.

[0084] Specifically, acquiring water pressure change data through sensors and combining it with sealing structure parameters to generate an initial dataset is an important foundation for analyzing sealing performance.

[0085] For example, in testing a flat electronic wiring harness for automobiles, sensors were placed in key areas to record pressure changes as water pressure increased from 0.2 MPa to 0.6 MPa. Combined with parameters such as the thickness and hardness of the sealing material, a multidimensional dataset incorporating time, pressure, and structural properties was created. This method ensures the comprehensiveness and dynamism of the data through real-time monitoring, laying the foundation for subsequent analysis. When extracting leak location information from the initial dataset, density estimation methods can effectively identify distribution patterns.

[0086] Specifically, kernel density estimation can be used to generate a probability density map based on the coordinates of the leak points in the dataset.

[0087] For example, a test showed that leaks were concentrated at coordinates (x: 2.8-3.3, y: 1.4-1.8), and the density peaks indicated that these areas had a higher probability of leakage than other parts. Extracting this distribution characteristic directly reflects the weak points in the sealing structure. When establishing a finite element model based on these distribution characteristics and using ANSYS software to simulate the stress distribution under water pressure, the high-density areas of leaks can be the focus of the modeling.

[0088] In one possible implementation, the model mesh is refined in these regions. After applying hydraulic loads, the simulation results may show stress concentration at the joints, with peak stress reaching 150 MPa. This simulation method digitally reproduces the actual working conditions. After acquiring the stress distribution data, the failure value is calculated based on the stress-strain relationship.

[0089] For example, assuming the material yield strength is 200 MPa, if the stress in a certain region exceeds 180 MPa in the simulation, which is close to the critical value, it can be inferred that the failure value range is between 180 and 200 MPa.

[0090] Preferably, the water pressure load is adjusted through multiple sets of simulations to refine the range of failure values ​​and improve prediction accuracy. If the range of failure values ​​exceeds a preset failure threshold, for example, if the failure threshold is set to 170 MPa and the failure value is 180 MPa, then the stability is judged to have decreased, and the preliminary characteristics of the failure mechanism are determined.

[0091] For example, failure may manifest as material fatigue or microcrack propagation.

[0092] It should be noted that this feature extraction relies on the analysis of outliers in the stress distribution, which intuitively reveals potential risks. After preliminary feature extraction of key parameters, when using the support vector machine algorithm in the scikit-learn library to classify failure mechanism types, stress peaks, leakage density, etc., can be used as input features.

[0093] In one embodiment, the training data contains 100 samples, and the classification results may categorize failures into two types: "loose seal" and "material fracture." This classification method improves recognition efficiency through machine learning. The finite element model parameters are adjusted based on the classification results, and the simulation is repeated to determine the final quantitative characteristics.

[0094] For example, if the classification points to "material fracture", the material stiffness parameter can be increased, and the simulation shows that the stress peak drops to 160 MPa, which is below the failure threshold.

[0095] Understandably, this iterative optimization method effectively verifies the directionality of parameter adjustments, providing a basis for design improvement.

[0096] S3 extracts key parameters from the quantitative characteristics of the failure mechanism, combines them with the immersion time variable, and records the changes in the physical properties of the material under different immersion times through accelerated aging tests to obtain the functional relationship between immersion time and material aging rate.

[0097] Optionally, this step also includes:

[0098] Step S31: Extract quantitative features from the failure mechanism and determine key parameters through feature analysis. Step S32: Obtain accelerated aging test data for the key parameters and immersion time, recording changes in physical properties. Step S33: Determine the relationship between aging rate and immersion time based on changes in physical properties, obtaining a preliminary functional expression. Step S34: Perform regression analysis using the least squares method to determine the functional relationship of aging rate from the preliminary functional expression. Step S35: If the functional relationship is inconsistent with the test records, adjust the time variable based on changes in physical properties to obtain a corrected functional expression. Step S36: Determine the quantitative correlation between material aging and immersion time based on the corrected functional expression, obtaining the final functional relationship. Step S37: Determine the trend of aging rate with immersion time using the final functional relationship.

[0099] Specifically, when extracting quantitative features from failure mechanisms, key parameters can be determined by analyzing the performance of the sealing structure under different operating conditions.

[0100] For example, under continuous water pressure, the stress concentration area of ​​the seal may be one of the key parameters.

[0101] In one possible approach, the deformation of the seal under different pressures can be experimentally measured, and its displacement value recorded (e.g., 0.5 mm or 1.2 mm). This, combined with material properties, allows for a preliminary assessment of which parameters have the greatest impact on failure. The advantage of this method is that it can quickly identify key areas of concern, providing data support for subsequent analysis. When obtaining accelerated aging test data regarding the relationship between key parameters and immersion time, the long-term use of the seal in a high-humidity environment can be simulated.

[0102] Specifically, samples can be placed in a test chamber at 80°C and 90% humidity for 100 hours, 200 hours, and 500 hours respectively, and their hardness changes can be recorded, such as from an initial 70 Shore A to 65 Shore A. This data acquisition method can intuitively reflect the impact of time on material properties, laying the foundation for aging rate analysis.

[0103] When determining the relationship between aging rate and immersion time based on changes in physical properties, it is preferable to observe the degree of surface cracking or the decreasing trend of tensile strength of the seal.

[0104] In one embodiment, if the tensile strength decreases from 15 MPa to 12 MPa after immersion for 300 hours and to 9 MPa after 600 hours, it can be preliminarily inferred that the aging rate exhibits a non-linear accelerating trend over time. The advantage of this analysis is that it provides a basis for predicting material lifespan.

[0105] When using the least squares method for regression analysis, it can be understood that a curve is fitted using existing hardness or strength data points.

[0106] For example, assuming the aging rate and soaking time have a quadratic function relationship, parameters can be adjusted based on experimental data to obtain a preliminary expression. The advantage of this method is that it can transform discrete data into a continuous trend description, facilitating engineering applications. If the functional relationship is inconsistent with the experimental records, the time variable can be adjusted by changes in physical properties.

[0107] For example, if the hardness decreases more than expected after 400 hours, the time period can be refined to 350 hours and 450 hours for retesting to observe the details of the changes and ultimately correct the function. This adjustment improves the accuracy of the model and ensures it matches the actual aging process. The quantitative correlation between material aging and immersion time is then determined based on the corrected function expression.

[0108] One possible implementation approach is to incorporate temperature factor analysis.

[0109] For example, by fitting the function at 60°C and 80°C respectively, it was concluded that the aging rate increases by 20% at high temperatures. This multi-condition verification enhances the applicability of the function.

[0110] When determining the trend of aging rate through the final functional relationship, for example, a curve can be plotted to visually show the performance degradation of the seal during immersion in 0-1000 hours.

[0111] For example, aging is slow in the early stages, but accelerates after 500 hours. This trend analysis helps in developing maintenance plans in advance and avoiding the risk of failure.

[0112] It should be noted that the practicality of this method lies in its ability to provide quantitative references for the design optimization of seals, while also extending their service life.

[0113] S4, based on the functional relationship between immersion time and material aging rate, superimposed temperature cycling conditions, and using environmental simulation equipment to cycle test wire harness samples within a preset temperature range, obtain data on the superimposed influence of temperature cycling on material aging and dynamic performance.

[0114] Optionally, this step also includes:

[0115] Step S41: Perform temperature cycling tests on the wire harness samples using an environmental simulation device, recording the immersion time and the percentage decrease in material hardness within each cycle as raw data for the aging rate. Step S42: Extract the corresponding values ​​of immersion time and percentage decrease in hardness from the raw data, and fit a linear function relationship between the two using the least squares method to obtain a preliminary aging rate model. Step S43: Introduce temperature cycling parameters into the preliminary model, using the difference ΔT between the highest temperature T_max and the lowest temperature T_min in each cycle as an additional variable to generate an adjusted dataset (T1, ΔT, Q). Step S44: Model and fit the equation using multiple linear regression on the adjusted data. Step S45: Determine the change in aging rate for each unit increase in the difference ΔT between the highest and lowest temperatures based on the regression coefficients. Step S46: Calculate the dynamic performance offset for each sample based on the regression results, and construct a two-dimensional feature vector by combining the difference ΔT between the highest and lowest temperatures and the dynamic performance offset δ. Use K-means clustering to classify the samples into three aging sensitivity categories: high, medium, and low. Step S47: Statistically analyze the distribution range of the percentage decrease in hardness for each type of sample. If the average percentage decrease in hardness for the high-sensitivity group exceeds the hardness decrease threshold of 15%, it is determined that the sample in this group has a synergistic accelerated aging effect of temperature cycling and immersion time.

[0116] In one possible implementation, temperature cycling tests are performed on the wire harness sample using an environmental simulation device to simulate the alternating hot and cold environment in actual use.

[0117] For example, samples are placed in a cycle ranging from -40°C to 120°C, with each cycle lasting 4 hours, including 2 hours of immersion at both high and low temperatures. When recording the percentage decrease in hardness, a hardness tester can be used to measure the change from the initial value to the value after each cycle. Assuming an initial hardness of 100 units, decreasing to 85 units after 5 cycles, the percentage decrease in hardness is 15%. This method visually reflects the aging trend of materials under temperature stress.

[0118] Optionally, step S42, which involves extracting the corresponding values ​​of soaking time and hardness reduction percentage from the original data, fitting the linear function relationship between the two using the least squares method to obtain a preliminary model of the aging rate, further includes:

[0119] Step S421, calculate the percentage decrease in hardness using the following formula:

[0120] Q = aT1 + b,

[0121] Where Q is the percentage decrease in hardness, T1 is the soaking time, a is the first soaking time coefficient, and b is the first temperature difference coefficient.

[0122] Specifically, when extracting the corresponding values ​​of soaking time and percentage decrease in hardness from the raw data, the data can be organized into pairs, such as 2 hours corresponding to 3%, 4 hours to 7%, and 6 hours to 12%. When fitting a linear function relationship using the least squares method, assuming the data shows a linear trend, the preliminary model may take the form of y=2x+1. This provides a basic framework for subsequent analysis and helps to quickly determine the approximate correlation between aging rate and soaking time.

[0123] It should be noted that introducing the temperature cycling parameter ΔT into the preliminary model can more comprehensively reflect the environmental impact.

[0124] Optionally, step S44, modeling and fitting an equation using multiple linear regression on the adjusted data, further includes:

[0125] Step S441, the fitting equation is:

[0126] Q = a'T1 + b'ΔT + c

[0127] Where a' is the second soaking time coefficient, b' is the second temperature difference coefficient, and c is a constant.

[0128] For example, when the difference between the highest and lowest temperatures, ΔT, is 160°C (120°C minus -40°C), the decrease in hardness may be accelerated, while it is more gradual when ΔT is 80°C. The adjusted dataset can be recorded as (6, 160, 12), (4, 80, 5), etc. After fitting with multiple linear regression, the coefficients a' and b' quantify the independent contributions of immersion time and temperature difference, respectively. For example, a'=1.5 indicates an increase of 1.5% in hardness decrease per hour of immersion, and b'=0.02 indicates an additional 0.02% decrease for every 1°C increase in ΔT. This method improves the model's adaptability to complex environments.

[0129] Optionally, step S46, which calculates the dynamic performance offset for each sample based on the regression results, constructs a two-dimensional feature vector by combining the difference between the highest and lowest temperatures with the dynamic performance offset, and uses K-means clustering to classify the samples into three aging sensitivity categories: high, medium, and low, further includes:

[0130] Step S461, calculate the dynamic performance offset for each sample according to the following formula:

[0131] δ=|Q-(a'T1+b'ΔT+c)|,

[0132] Where δ is the dynamic performance offset.

[0133] For example, when judging the impact of ΔT based on the regression coefficient, if b'=0.02, then an increase in ΔT from 80°C to 160°C may increase the aging rate by 1.6%. This suggests that an increase in temperature range has a significant amplifying effect on material damage. When calculating the dynamic performance offset δ based on the regression results, assuming a sample's measured Q=10% and predicted value is 9%, then δ=1%. After constructing a two-dimensional feature vector by combining δ and ΔT, K-means clustering can divide the samples into three groups. For example, the δ of the high-sensitivity group is concentrated above 1.5%, the medium-sensitivity group is 0.5%-1.5%, and the low-sensitivity group is below 0.5%. This provides data support for sample classification, facilitating targeted optimization design.

[0134] In one embodiment, when statistically analyzing the distribution range of the percentage decrease in hardness, the high-sensitivity group may have a range of 15%-20%, the medium-sensitivity group 8%-14%, and the low-sensitivity group 3%-7%. If the average value of the high-sensitivity group reaches 16%, exceeding the hardness decrease percentage threshold of 15%, it indicates a synergistic effect between temperature cycling and immersion time.

[0135] For example, after a certain wire harness was immersed in high temperature for a long time, the molecular chain breakage accelerated, resulting in a greater-than-expected decrease in hardness. This kind of analysis helps to identify weak points and guide material selection or process improvement.

[0136] Preferably, when determining the synergistic accelerated aging effect, real-world cases can be considered. For example, a batch of wire harnesses, after being immersed at ΔT=160°C for 6 hours, showed a 18% decrease in hardness, far exceeding the prediction value of a single factor. This phenomenon indicates the nonlinear effect of multiple factors superimposed. Through clustering and statistics, the root cause of the problem can be more accurately located, providing a basis for subsequent durability improvement. The advantage of this method is that it quantifies the aging pattern and is easy to apply in engineering.

[0137] Understandably, the above examples, from data collection to model adjustment, and then to classification and judgment, proceed step by step, ensuring the rigor of the analysis.

[0138] For example, initial hardness measurement, data fitting, and the introduction of temperature variables support each other, jointly revealing the changing trend of the aging rate. This clear logical chain effectively assists R&D personnel in optimizing the anti-aging performance of wiring harnesses.

[0139] S5 uses statistical analysis to calculate the performance degradation trend under various test conditions by using the superimposed effect data of temperature cycling on material aging and dynamic performance, and judges the change law of dynamic performance under the interaction of water pressure change, immersion time and temperature cycling.

[0140] Optionally, this step also includes:

[0141] Step S51: Obtain the material aging feature matrix using temperature cycling and immersion time data, and extract the first three principal components as aging feature vectors using principal component analysis. Step S52: Input the aging feature vectors into a ridge regression model, outputting two dynamic performance parameters: elastic modulus and elongation at break. Step S53: Calculate the performance degradation rate using exponential smoothing based on the time series data of the dynamic performance parameters, obtaining the degradation trend curve. Step S54: Construct a two-factor experimental matrix for water pressure change and immersion time data, and detect the significance of the interaction using the variance expansion factor. Step S55: Calculate the interaction weights of water pressure change and immersion time using partial least squares regression, obtaining a water pressure change weight coefficient of 0.6 and an immersion time weight coefficient of 0.4. Step S56: Combine the interaction weight coefficients with the aging feature vectors, input them into the ridge regression model, and establish a performance degradation rate prediction equation. Step S57: When the performance degradation rate exceeds the performance degradation rate threshold of 0.5, introduce the standard deviation of the temperature cycling data as a correction factor, and recalculate the interaction weights using a weighted average method. Step S58: The corrected weight coefficients are superimposed on the original aging feature vector, and a dynamic performance decay distribution surface under multi-factor conditions is generated through linear interpolation. Step S59: The decay distribution surface is Z-score standardized and input into a hierarchical clustering algorithm. The optimal number of clusters is determined to be 3 based on the silhouette coefficient. Step S510: Three types of dynamic performance decay modes are output: rapid decay region, linear decay region, and stable decay region, completing the performance classification feature extraction under multiple conditions.

[0142] Specifically, when obtaining the material aging characteristic matrix through temperature cycling and immersion time data, it can be understood as collecting multi-dimensional data from experiments.

[0143] For example, in wire harness sample testing, the immersion time, number of temperature cycles, and corresponding hardness change values ​​of each sample are recorded to form a multi-column matrix.

[0144] For example, suppose there are 10 samples with soaking times ranging from 50 to 500 hours and temperature cycling ranges from -40°C to 85°C. The hardness decrease data is then output as a matrix. This method can comprehensively capture aging characteristics. When principal component analysis is used to extract the first three principal components as the aging feature vector...

[0145] Preferably, the high-dimensional data in the matrix can be reduced in dimensionality.

[0146] Specifically, assuming the original matrix contains five variables, including soaking time, temperature difference, and hardness decrease, principal component analysis may reveal that the first three principal components explain 90% of the variance. These principal components can be considered as a "condensed expression" of aging characteristics, facilitating subsequent modeling.

[0147] When inputting aging feature vectors into a ridge regression model to predict elastic modulus and elongation at break, one possible implementation is to train the model using experimental data.

[0148] For example, the aging feature vector of a sample is [0.8, -0.3, 0.5], corresponding to an elastic modulus decreasing from an initial 200 MPa to 180 MPa, and an elongation at break changing from 15% to 12%. Ridge regression, through regularization, can effectively avoid overfitting and improve prediction stability. When calculating the performance degradation rate using exponential smoothing based on time-series data of dynamic performance parameters, it can be understood as smoothing the fluctuating data.

[0149] For example, if the elastic modulus data for 10 consecutive periods are 200, 198, 195, 190 MPa, etc., after exponential smoothing, the decay rate can be obtained as 0.02 / period. Plotting the trend curve can help predict long-term performance.

[0150] When constructing a two-factor experimental matrix for water pressure variation and soaking time, an example can be designed: a combination of experiments with water pressure ranging from 0.1 MPa to 0.5 MPa and soaking time ranging from 100 hours to 300 hours. Using the variance inflation factor (VOP), a value of 1.2 indicates that the interaction is not significant; if it exceeds 10, the model needs adjustment. This method can clearly separate the main effect from the interaction effect.

[0151] When using partial least squares regression to calculate the interaction weights, in one embodiment, the weight of water pressure change is 0.6 and the weight of soaking time is 0.4, indicating that water pressure has a greater impact on aging.

[0152] For example, a sample showed a 10% decrease in hardness under 0.3 MPa water pressure, but only a 6% decrease under 0.1 MPa; the weighting coefficient quantifies this difference.

[0153] When combining interaction weights with aging feature vectors as input to support vector regression, specifically, the weights can be used as additional features.

[0154] For example, expanding the feature vector from [0.8, -0.3, 0.5] to [0.8, -0.3, 0.5, 0.6, 0.4] makes the model's predicted decay rate closer to reality. This improves prediction accuracy under multi-factor conditions. When the performance decay rate exceeds the performance decay rate threshold of 0.5, the temperature cycle standard deviation is introduced as a correction factor, and the weights are recalculated using the entropy weight method.

[0155] For example, with a standard deviation of 20°C for temperature cycling data, the entropy weighting method might adjust the water pressure weight to 0.55 and the soaking time to 0.45. This dynamic adjustment can better adapt to complex environments.

[0156] When generating a decay distribution surface using Kriging interpolation, preferably, a smooth surface can be generated based on discrete data points from 10 samples.

[0157] For example, a certain region shows an attenuation rate that increases from 0.1 to 0.8, revealing spatial distribution patterns and facilitating visual analysis.

[0158] When the attenuation distribution surface is normalized by Z-score and then input into K-means clustering, in one embodiment, the best results are achieved when the silhouette coefficient calculation shows a cluster number of 3.

[0159] For example, the decay rate of samples in the rapid decay region is concentrated in the range of 0.6-0.8, linear decay is 0.3-0.5, and steady decay is 0.1-0.2. This classification can accurately distinguish aging modes and provide a basis for material optimization.

[0160] S6. Based on the dynamic performance change pattern, a multivariate evaluation system including water pressure change, soaking time and temperature cycle is established in advance. By fitting the weight coefficients of each variable and performance decay through regression analysis, a comprehensive performance quantification formula is obtained.

[0161] Optionally, this step also includes:

[0162] Step S61: Acquire real-time data on water pressure changes, soaking time, and temperature cycling using sensors, and store the data in a database. Step S62: Extract the multivariate raw dataset from the database, and use linear regression to fit the relationship between water pressure changes, soaking time, temperature cycling, and the percentage decrease in material strength, obtaining the weight coefficients for each variable. Step S63: After normalizing the weight coefficients, calculate the contribution ratio of each variable using the formula: the weight coefficient of each variable divided by the sum of all weight coefficients. Step S64: Based on the contribution ratios and a pre-established weighted summation formula, generate a quantitative scale value for the comprehensive performance. The formula is: the contribution ratio of water pressure changes multiplied by the water pressure change value, the contribution ratio of soaking time multiplied by the soaking time value, and the contribution ratio of temperature cycling multiplied by the temperature cycling value. Step S65: If the quantitative scale value exceeds a preset quantitative scale value threshold, output a warning signal. Step S66: Substitute the final scale value into the moving average method to calculate the sliding window mean of historical data, and output the predicted result of the material strength decline trend.

[0163] Specifically, real-time data on water pressure changes, soaking time, and temperature cycles are acquired through sensors and stored in a database. This process forms the basis for material performance monitoring.

[0164] In one possible implementation, a high-precision pressure sensor can be used to record water pressure fluctuations every second, the immersion time can be measured to the minute using a timer, and the temperature cycle can be captured hourly by a thermocouple to capture temperature changes. This data is stored as timestamps for easy subsequent analysis.

[0165] It should be noted that sensor selection must consider corrosion resistance and long-term stability to ensure data reliability. A multivariate raw dataset was extracted from the database, and linear regression was used to fit the relationship between water pressure changes, immersion time, temperature cycling, and the percentage decrease in material strength, obtaining the weight coefficients for each variable.

[0166] Specifically, features can be extracted from 1000 sets of data collected over a week, revealing that water pressure change has a weighting coefficient of 0.5, soaking time has a weighting coefficient of 0.3, and temperature cycle has a weighting coefficient of 0.2. This method intuitively reflects the degree of influence of each factor on the intensity reduction, helping to quickly identify key variables. After normalizing the weighting coefficients, the contribution ratio of each variable is calculated using the formula: the weighting coefficient of each variable divided by the sum of all weighting coefficients.

[0167] For example, if the sum of 0.5, 0.3, and 0.2 is 1, then the contribution ratio of water pressure change is 50%, the contribution ratio of soaking time is 30%, and the contribution ratio of temperature cycle is 20%.

[0168] In one embodiment, if the water pressure suddenly increases by 10 units, the soaking time is extended by 5 hours, and the temperature cycle is increased by 3 times, the overall performance quantification value can be calculated using a weighted summation formula. This quantification method simplifies multi-factor analysis and facilitates real-time monitoring. The overall performance quantification value is generated based on the contribution ratio and the pre-established weighted summation formula.

[0169] Preferably, if the water pressure change value is 10, contributing 50%, the soaking time value is 5, contributing 30%, and the temperature cycle value is 3, contributing 20%, then the scale value is 10×0.5+5×0.3+3×0.2=7.1.

[0170] Understandably, when the quantization scale threshold is set to 6, if 7.1 exceeds the threshold, the system will output a warning signal. This mechanism can promptly alert to potential risks and improve response efficiency. If the quantization scale value exceeds the preset threshold, a warning signal will be output.

[0171] For example, when the scale value increases from 5.8 to 7.1, the system can notify the operator via an audible and visual alarm. This early warning design effectively prevents material failure and extends its service life.

[0172] In one possible implementation, the warning signal can also trigger equipment to automatically adjust water pressure or temperature, reducing manual intervention. The final scale value is then substituted into a moving average method to calculate the sliding window mean of historical data, outputting a prediction of the material strength decline trend.

[0173] Specifically, the window is set to 5 days, and the scale values ​​for the past 5 days are 5.0, 5.5, 6.0, 6.5, and 7.1, with an average of 6.02.

[0174] For example, if the trend shows a continuous increase in the mean, it can be inferred that the material strength is declining at an accelerated pace. This predictive method smooths out short-term fluctuations, is suitable for judging long-term trends, and helps in planning maintenance in advance.

[0175] S7. After obtaining the comprehensive performance quantification formula, input the water pressure, immersion time and temperature cycle parameters of the actual application scenario for specific environmental adaptation requirements, calculate the performance prediction value of the harness in the scenario through the formula, and determine the environmental adaptability assessment result.

[0176] Optionally, this step also includes:

[0177] Step S71: Obtain the pre-established comprehensive performance quantification formula. Step S72: Input parameters such as water pressure P, soaking time T1, and temperature cycle T2 from the actual scenario to calculate the initial performance prediction value R. Step S73: Input R into the SVC classifier of scikit-learn to determine if R exceeds the performance threshold and output the environmental adaptability classification result. Step S74: Extract the time series of R from historical data, use linear regression to fit its trend, and generate the trend slope k. Step S75: Input k into the RandomForestClassifier of scikit-learn to determine if the slope stability meets the scenario requirements and output the adjusted performance prediction value R'. Step S76: If R' is lower than the performance threshold, substitute R' into the comprehensive performance quantification formula to calculate the optimized performance prediction value R''. Step S77: Based on the difference between R'' and the performance prediction value R, determine the final conclusion regarding the harness's environmental adaptability.

[0178] Optionally, step S71, obtaining the pre-established comprehensive performance quantification formula, further includes:

[0179] Step S711, the comprehensive performance quantification formula is as follows:

[0180] R = w1P + w2T1 + w3T2

[0181] Where R is the initial performance prediction value, P is the water pressure parameter, T1 is the soaking time, T2 is the temperature cycle parameter, and w1, w2, and w3 are preset weighting coefficients.

[0182] For example, the pre-established comprehensive performance quantification formula R=w1P+w2T1+w3T2 can be regarded as an environmental adaptability assessment tool based on multivariate weighted summation. Here, P represents the water pressure parameter, T1 is the immersion time, T2 is the temperature cycle parameter, and w1, w2, and w3 are weighting coefficients derived from historical data analysis. The core of this method lies in integrating multiple influencing factors into a single performance value index.

[0183] For example, in the field of wire harness manufacturing, water pressure may refer to the pressure applied by the external environment, immersion time reflects the duration of material exposure to humid conditions, and temperature cycling simulates the actual usage scenario of alternating hot and cold conditions. Assuming a scenario where P is 5 units, T1 is 10 hours, T2 is 20 cycles, and the weighting coefficients are w1=0.4, w2=0.3, and w3=0.3 respectively, the initial performance prediction value R can be calculated by substituting these values ​​into the formula. The advantage of this method is that it can quickly integrate multi-dimensional data, providing a foundation for subsequent classification.

[0184] In one possible implementation, R is input into scikit-learn's SVC classifier to determine if it exceeds a preset performance threshold. The SVC classifier maps the R value to a category of environmental adaptability using the principle of support vector machines.

[0185] Specifically, assuming a performance threshold of 15, if R = 16, the classification result might be "out of safe range"; if R = 12, it might be classified as "good environmental adaptability". The advantage of this classification is that it transforms complex numerical judgments into intuitive category outputs, facilitating quick decision-making by technical personnel.

[0186] For example, in high humidity scenarios, the value of R may rise rapidly due to the higher weight of T1, indicating that the material formulation needs to be adjusted.

[0187] It should be noted that extracting the time series of R from historical data and fitting the trend using linear regression is to capture the pattern of performance changes over time. After generating the trend slope k, the rate of increase or decrease of R can be visually reflected.

[0188] For example, if the R-value of a batch of wiring harnesses increases from 10 to 15 within 30 days with a positive slope k, it indicates that performance degradation is accelerating. Conversely, if k is close to 0, the trend is stable. The advantage of this analysis is that it can provide early warning of potential risks and prevent sudden failures.

[0189] Specifically, inputting the slope k into RandomForestClassifier to determine its stability is a step in further optimizing the prediction. Random Forest uses the ensemble of multiple decision trees to determine whether k is within an acceptable range.

[0190] For example, a k value of 0.2 might be classified as "stable," while a k value of 0.8 might indicate "excessive volatility." The adjusted performance prediction value R' is then generated. If R' falls below the performance value threshold, the optimized performance prediction value R'' needs to be recalculated. The benefit of this step is that it enhances the robustness of trend judgment through machine learning.

[0191] For example, when k fluctuates greatly, R' may reflect a risk that the performance may seem normal in the short term but is unreliable in the long term.

[0192] In one embodiment, if R' is lower than a preset performance value threshold, R' is substituted into the comprehensive performance quantification formula to calculate the optimized performance prediction value R'', and the final conclusion is determined based on the difference between R'' and R.

[0193] For example, R is 16, R'' is 13, and the difference is 3, indicating that the optimization measures are effective and the adaptability of the harness is improved.

[0194] Preferably, this iterative process can verify the reliability of the results from multiple perspectives.

[0195] For example, a decrease in R'' might stem from a reduction in the effect of T1, while difference analysis further confirms the correctness of the adjustment direction. The advantage of this method is that it clarifies the room for improvement through numerical comparison.

[0196] Understandably, the above process, from the core formula to the classifier and then to trend analysis, forms a complete evaluation chain.

[0197] For example, in high temperature and high humidity environments, the weight of T2 may be increased, and the R value will change accordingly. The combined use of SVC and random forest ensures the accuracy of classification and trend prediction.

[0198] In one embodiment, if R is high and R'' approaches a reasonable range, it indicates that the harness can maintain performance under harsh conditions. The benefit of this multi-level verification is that it ensures both real-time monitoring and long-term prediction, providing comprehensive support for environmental adaptability.

[0199] S8 extracts the deviation between the predicted and measured values ​​of performance from the environmental adaptability assessment results, uses machine learning algorithms to iteratively optimize the assessment system, adjusts the weight coefficients to reduce the deviation, and obtains an optimized performance quantification scale system.

[0200] Optionally, this step also includes:

[0201] Step S81: Obtain measured values ​​of environmental parameters through environmental monitoring equipment, and compare the measured values ​​with preset environmental conditions to generate numerical evaluation results. Step S82: Process the evaluation results using the XGBoost algorithm, outputting the absolute difference between the predicted performance value and the measured value as deviation data. Step S83: Use scikit-learn's analysis of variance tool to extract key features from the deviation data, including the fluctuation range and duration of environmental parameters. Step S84: Input the features into a random forest algorithm to train a regression model, outputting feature importance scores for each environmental parameter. Step S85: Perform min-max normalization on the feature importance scores to generate initial weight coefficients. Step S86: Adjust the evaluation system parameters according to the initial weight coefficients, and use the least squares method to fit the linear relationship between the parameters and the deviation. Step S88: Calculate the mean absolute percentage error corresponding to the adjusted weight data; if the error is greater than the error threshold of 5%, retrain the random forest model. Step S89: When the mean absolute percentage error is less than the mean absolute percentage threshold of 5%, update the final weight coefficients to the evaluation system. Step S810: Process the newly collected environmental parameters using the updated system and output performance optimization prediction values. Step S811: Compare the performance optimization prediction values ​​with the measured values ​​of the K-fold cross-validation set to generate an environmental adaptability verification report.

[0202] Specifically, obtaining measured values ​​of environmental parameters through environmental monitoring equipment is a fundamental aspect of the assessment system.

[0203] For example, in the environmental adaptability test of wire harnesses, high-precision sensors can be used to collect parameters such as water pressure, temperature and humidity in real time.

[0204] For example, suppose that in a certain scenario, the monitoring device records a water pressure of 2.5 bar, a temperature that cycles from -20°C to 60°C, and a humidity of 90%. These measured values ​​reflect the actual stress on the harness under specific environmental conditions.

[0205] It should be noted that the sampling frequency of the measured values ​​needs to be adjusted according to the needs of the scenario, such as sampling once per second, in order to capture parameter fluctuations.

[0206] When comparing measured values ​​with preset environmental conditions to generate numerical evaluation results, it is understandable that the preset conditions may be the ideal range during the design, such as water pressure not exceeding 3 bar and temperature range of -15°C to 50°C.

[0207] In one possible implementation, a comparison revealed that the measured water pressure did not exceed the standard, but the temperature exceeded the upper limit by 10°C. A preliminary assessment result was generated by quantifying the difference. This comparison intuitively reflects the degree of deviation of the wiring harness in the actual scenario, providing a data foundation for subsequent analysis.

[0208] When using the XGBoost algorithm to process the evaluation results, specifically, the deviation between the measured value and the predicted value can be used as the target variable input.

[0209] Preferably, assuming a predicted performance of 90%, an actual measured performance of 85%, and an absolute difference of 5%, XGBoost analyzes the relationship between bias and environmental parameters using gradient boosting. This method effectively captures nonlinear characteristics, and the generated bias data can be used for further prediction optimization.

[0210] When using scikit-learn's analysis of variance (ANOVA) tool to extract key features, in one embodiment, water pressure fluctuation range (e.g., 0.5 bar) and temperature duration (e.g., 2 hours) can be separated from the deviation data as key factors. These features reflect the actual impact of the environment on the harness performance.

[0211] For example, frequent fluctuations in water pressure may cause seal failure more easily than a single high temperature, and extraction helps to focus on the main risk points.

[0212] When features are input into a random forest algorithm to train a regression model and the algorithm outputs feature importance scores, for example, water pressure fluctuation might score 0.45, temperature duration 0.35, and humidity 0.20. These scores quantify the degree to which each parameter affects performance.

[0213] In one embodiment, water pressure fluctuations are given priority due to their high score, indicating a need to strengthen the harness's pressure resistance design. When generating initial weight coefficients by performing min-max normalization on the feature importance scores, 0.45, 0.35, and 0.20 can be mapped to the 0-1 interval, resulting in adjusted weights such as 0.56, 0.44, and 0.25. These weights are used for subsequent adjustments to the evaluation system to ensure a closer alignment with real-world needs.

[0214] When the evaluation system parameters are adjusted based on the initial weighting coefficients and fitted using the least squares method, historical data analysis can reveal a linear relationship between the deviation and the duration of temperature. The adjusted weights can reduce prediction bias and improve evaluation accuracy.

[0215] When calculating the mean absolute percentage error corresponding to the adjusted weighted data, specifically, if the error is 4%, which is less than the preset error threshold of 5%, then there is no need to retrain the random forest model. This error control ensures the reliability of the system.

[0216] When the final weighting coefficients are updated to the evaluation system and new collected parameters are processed, such as a new scenario with a water pressure of 2.8 bar and a temperature cycle of -10°C to 55°C, the updated system can output more accurate performance predictions. This method improves the adaptability of the predictions. K-fold cross-validation is used to compare the predicted values ​​with the measured values ​​to generate a validation report.

[0217] In one possible implementation, the data is divided into five parts, and the verification results show that the average deviation between the predicted and measured values ​​is only 3%. This verification method enhances the credibility of the evaluation system and provides a reliable basis for the environmental adaptability of the harness.

[0218] like Figure 2 As shown, in a second aspect, the present invention provides a flat electronic wire harness environmental adaptability assessment system, which uses the method described above to assess the environmental adaptability of flat electronic wire harnesses. The system includes:

[0219] The data acquisition module is used to acquire the response data of the sealing structure of the flat electronic wire bundle under different water pressure changes. The sensor records the deformation of the sealing structure and the location of the leakage point in real time when the water pressure changes from low to high, and obtains the correspondence between water pressure change and the stability of the sealing structure.

[0220] The failure analysis module is used to analyze the failure mechanism based on the correspondence between water pressure changes and the stability of the sealing structure, targeting the leakage point. It uses finite element simulation technology to simulate the stress distribution and damage threshold of the sealing structure under water pressure, and determines the quantitative characteristics of the failure mechanism.

[0221] The aging rate modeling module is used to extract key parameters from the quantitative characteristics of failure mechanisms. Combined with the immersion time variable, it records the changes in the physical properties of the material under different immersion times through accelerated aging tests, and obtains the functional relationship between immersion time and material aging rate.

[0222] The temperature superposition test module is used to superimpose temperature cycling conditions on the functional relationship between immersion time and material aging rate. The wire harness sample is cyclically tested within a preset temperature range using an environmental simulation device to obtain data on the superposition effect of temperature cycling on material aging and dynamic performance.

[0223] The performance degradation analysis module is used to calculate the performance degradation trend under various test conditions by using statistical analysis methods to analyze the superimposed effect data of material aging and dynamic performance through temperature cycling, and to determine the change law of dynamic performance under the interaction of water pressure change, immersion time and temperature cycling.

[0224] The multivariate evaluation module is used to pre-establish a multivariate evaluation system that includes water pressure changes, soaking time and temperature cycles based on the dynamic performance change pattern. By fitting the weight coefficients of each variable and performance decay through regression analysis, a comprehensive performance quantification formula is obtained.

[0225] The environmental adaptability prediction module is used to obtain the comprehensive performance quantification scale formula, and then input the water pressure, immersion time and temperature cycle parameters of the actual application scenario for specific environmental adaptability requirements. The module calculates the predicted performance value of the harness in that scenario using the formula, and determines the environmental adaptability assessment result. The evaluation optimization module is used to extract the deviation between the predicted performance value and the measured value from the environmental adaptability assessment result, and uses machine learning algorithms to iteratively optimize the evaluation system, adjust the weight coefficients to reduce the deviation, and obtain an optimized performance quantification scale system.

[0226] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the concept of this application. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for evaluating the environmental adaptability of flat electronic wire harnesses, characterized in that, The method includes: S1, acquire the sealing structure response data of the flat electronic wire bundle under different water pressure changes, and record the deformation of the sealing structure and the location of the leakage point in real time when the water pressure changes from low to high through the sensor, so as to obtain the correspondence between water pressure change and sealing structure stability. S2. Based on the correspondence between water pressure changes and the stability of the sealing structure, the failure mechanism is analyzed for the location of the leakage point. Finite element simulation technology is used to simulate the stress distribution and damage threshold of the sealing structure under water pressure, and the quantitative characteristics of the failure mechanism are determined. S3. Key parameters are extracted from the quantitative characteristics of the failure mechanism. Combined with the immersion time variable, the changes in the physical properties of the material under different immersion times are recorded through accelerated aging tests to obtain the functional relationship between immersion time and material aging rate. S4, based on the functional relationship between immersion time and material aging rate, superimposed temperature cycling conditions, and through environmental simulation equipment, cyclically tested wire harness samples within a preset temperature range to obtain data on the superimposed influence of temperature cycling on material aging and dynamic performance. S5 uses the combined effect data of temperature cycling on material aging and dynamic performance to calculate the performance degradation trend under various test conditions using statistical analysis methods, and judges the change law of dynamic performance under the interaction of water pressure change, immersion time and temperature cycling. S6. Based on the dynamic performance change pattern, a multivariate evaluation system including water pressure change, soaking time and temperature cycle is established in advance. By fitting the weight coefficients of each variable and performance decay through regression analysis, a comprehensive performance quantification formula is obtained. S7. After obtaining the comprehensive performance quantification formula, input the water pressure, immersion time and temperature cycle parameters of the actual application scenario for specific environmental adaptation requirements, calculate the performance prediction value of the harness in the scenario through the formula, and determine the environmental adaptability assessment result. S8 extracts the deviation between the predicted and measured values ​​of performance from the environmental adaptability assessment results, uses machine learning algorithms to iteratively optimize the assessment system, adjusts the weight coefficients to reduce the deviation, and obtains an optimized performance quantification scale system.

2. The method according to claim 1, characterized in that, Step S4 involves applying temperature cycling conditions to the functional relationship between immersion time and material aging rate. The wire harness sample is then tested cyclically within a preset temperature range using an environmental simulation device to obtain data on the combined effects of temperature cycling on material aging and dynamic performance. This includes: Step S41: Perform a temperature cycling test on the wire harness sample using an environmental simulation device, and record the immersion time and material hardness reduction percentage of the sample in each cycle as raw data of the aging rate. Step S42: Extract the corresponding values ​​of soaking time and hardness reduction percentage from the original data, and use the least squares method to fit the linear function relationship between the two to obtain a preliminary model of aging rate. Step S43: Introduce temperature cycling parameters into the preliminary model, and use the difference between the highest and lowest temperatures in each cycle as an additional variable to generate the adjusted dataset. Step S44: Use multiple linear regression to model the adjusted data and fit the equation; Step S45: Determine the change in aging rate when the difference between the highest and lowest temperatures increases by 1 unit based on the regression coefficient. Step S46: Calculate the dynamic performance offset of each sample based on the regression results, and construct a two-dimensional feature vector by combining the difference between the highest and lowest temperatures and the dynamic performance offset. Use K-means clustering to divide the samples into three groups of aging sensitivity categories: high, medium, and low. Step S47: Statistically analyze the distribution range of the percentage decrease in hardness for each type of sample. If the average percentage decrease in hardness for the high-sensitivity group exceeds the threshold for percentage decrease in hardness, it is determined that the sample in this group has a synergistic accelerated aging effect of temperature cycling and immersion time.

3. The method according to claim 2, characterized in that, Step S42, which involves extracting the corresponding values ​​of soaking time and hardness reduction percentage from the original data, fitting the linear function relationship between the two using the least squares method to obtain a preliminary model of the aging rate, further includes: Step S421, calculate the percentage decrease in hardness using the following formula: Q=aT1+b Where Q is the percentage decrease in hardness, T1 is the soaking time, a is the first soaking time coefficient, and b is the first temperature difference coefficient.

4. The method according to claim 3, characterized in that, Step S44, which uses multiple linear regression to model the adjusted data and fit the equation, further includes: Step S441, the fitting equation is: Q = a'T1 + b'ΔT + c Where a' is the second soaking time coefficient, b' is the second temperature difference coefficient, ΔT is the difference between the highest and lowest temperatures, and c is a constant.

5. The method according to claim 4, characterized in that, Step S46 involves calculating the dynamic performance offset for each sample based on the regression results, constructing a two-dimensional feature vector by combining the difference between the highest and lowest temperatures with the dynamic performance offset, and using K-means clustering to classify the samples into three aging sensitivity categories: high, medium, and low. The step also includes: Step S461, calculate the dynamic performance offset according to the following formula: δ=|Q-(a'T1+b'ΔT+c)| Where δ is the dynamic performance offset.

6. The method according to claim 5, characterized in that, Step S47: Statistically analyze the distribution range of the percentage decrease in hardness for each type of sample. If the average percentage decrease in hardness of the high-sensitivity group exceeds the threshold for percentage decrease in hardness, it is determined that the sample in this group has a synergistic accelerated aging effect of temperature cycling and immersion time, including: the threshold for percentage decrease in hardness is set to 15%.

7. The method according to claim 1, characterized in that, Step S5 involves using statistical analysis to calculate the performance degradation trend under various test conditions based on the combined effects of temperature cycling on material aging and dynamic performance. This process determines the dynamic performance variation under the interaction of water pressure changes, immersion time, and temperature cycling, including: Step S51: Obtain the material aging feature matrix through temperature cycling and immersion time data, and use principal component analysis to extract the first three principal components as aging feature vectors. Step S52: Input the aging feature vector into the ridge regression model and output two dynamic performance parameters: elastic modulus and elongation at break. Step S53: Based on the time series data of dynamic performance parameters, calculate the performance degradation rate using the exponential smoothing method to obtain the degradation trend curve; Step S54: Construct a two-factor experimental matrix for water pressure change and soaking time data, and detect the significance of interaction by variance expansion factor; Step S55: Partial least squares regression is used to calculate the interaction weights between water pressure change and soaking time, and the weight coefficients of water pressure change and soaking time are obtained. Step S56: Combine the interaction weight coefficients with the aging feature vector, input them into the ridge regression model, and establish the performance degradation rate prediction equation. Step S57: When the performance degradation rate exceeds the performance degradation rate threshold, the standard deviation of the temperature cycling data is introduced as a correction factor, and the interaction weights are recalculated using the weighted average method. Step S58: The corrected weight coefficients are superimposed on the original aging feature vector, and the decay distribution surface of dynamic performance under multi-factor conditions is generated by linear interpolation. Step S59: Perform Z-score normalization on the attenuation distribution surface, input it into the hierarchical clustering algorithm, and determine the optimal number of clusters as 3 based on the silhouette coefficient; Step S510 outputs three types of dynamic performance degradation modes: rapid degradation region, linear degradation region, and stable degradation region, completing the performance classification feature extraction under multiple conditions.

8. The method according to claim 7, characterized in that, In step S55, partial least squares regression is used to calculate the interaction weight between water pressure change and soaking time, and the weight coefficients of water pressure change and soaking time are obtained, including: the weight coefficient of water pressure change is 0.6 and the weight coefficient of soaking time is 0.

4.

9. The method according to claim 8, characterized in that, In step S57, when the performance degradation rate exceeds the performance degradation rate threshold, the standard deviation of the temperature cycling data is introduced as a correction factor, and the interaction weight is recalculated using a weighted average method, including setting the performance degradation rate threshold to 0.

5.

10. A flat electronic wire harness environmental adaptability assessment system, characterized in that, The system for assessing the environmental adaptability of flat electronic wire harnesses using the method described in any one of claims 1-9, the system comprising: The data acquisition module is used to acquire the response data of the sealing structure of the flat electronic wire bundle under different water pressure changes. The sensor records the deformation of the sealing structure and the location of the leakage point in real time when the water pressure changes from low to high, and obtains the correspondence between water pressure change and the stability of the sealing structure. The failure analysis module is used to analyze the failure mechanism based on the correspondence between water pressure changes and the stability of the sealing structure, targeting the leakage point. It uses finite element simulation technology to simulate the stress distribution and damage threshold of the sealing structure under water pressure, and determines the quantitative characteristics of the failure mechanism. The aging rate modeling module is used to extract key parameters from the quantitative characteristics of failure mechanisms. Combined with the immersion time variable, it records the changes in the physical properties of the material under different immersion times through accelerated aging tests, and obtains the functional relationship between immersion time and material aging rate. The temperature superposition test module is used to superimpose temperature cycling conditions on the functional relationship between immersion time and material aging rate. The wire harness sample is cyclically tested within a preset temperature range using an environmental simulation device to obtain data on the superposition effect of temperature cycling on material aging and dynamic performance. The performance degradation analysis module is used to calculate the performance degradation trend under various test conditions by using statistical analysis methods to analyze the superimposed effect data of material aging and dynamic performance through temperature cycling, and to determine the change law of dynamic performance under the interaction of water pressure change, immersion time and temperature cycling. The multivariate evaluation module is used to pre-establish a multivariate evaluation system that includes water pressure changes, soaking time and temperature cycles based on the dynamic performance change pattern. By fitting the weight coefficients of each variable and performance decay through regression analysis, a comprehensive performance quantification formula is obtained. The environmental adaptability prediction module is used to obtain the comprehensive performance quantification formula, input the water pressure, immersion time and temperature cycle parameters of the actual application scenario for specific environmental adaptability requirements, calculate the performance prediction value of the harness in that scenario through the formula, and determine the environmental adaptability assessment result. The evaluation and optimization module is used to extract the deviation between the predicted and measured values ​​of performance from the environmental adaptability evaluation results. It uses machine learning algorithms to iteratively optimize the evaluation system, adjust the weight coefficients to reduce the deviation, and obtain an optimized performance quantification scale system.

Images