Wiring harness connector waterproof performance attenuation intelligent early warning method and system

Abstract

According to the intelligent early warning method and system for the waterproof performance attenuation of the wire harness connector provided by the invention, the waterproof performance attenuation model is established by acquiring the deformation rate data of the sealing ring, analyzing the microstructure degradation and measuring the moisture absorption rate change. The method comprises the following steps: firstly, collecting deformation values under different disconnection durations, calculating a dynamic deformation rate sequence, analyzing a material surface microstructure degradation condition in combination with a scanning electron microscope, determining degradation distribution of a microstructure along with deformation rate change, testing a moisture absorption rate change trend, and calculating a coupling relationship with microstructure degradation; based on the data, a long-short-term memory network algorithm is adopted to train a dynamic change rule prediction model, the contribution rate of each parameter to waterproof performance attenuation is quantified through principal component analysis, and finally, real-time monitoring of related parameters, combined calculation of a current waterproof performance attenuation value and output of a dynamic early warning signal are realized. And the accuracy and timeliness of the attenuation early warning of the waterproof performance of the wire harness connector are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent manufacturing, and particularly relates to an intelligent early warning method and system for the attenuation of the waterproof performance of a wire harness connector. Background Art

[0002] The research on waterproof performance, as an important cross - field of electrical engineering and materials science, is directly related to the reliability and safety of wire harness connectors in extreme environments. Its importance is self - evident. Especially in industries such as automotive, aerospace, and industrial automation, the attenuation of waterproof performance may lead to system failures or even catastrophic consequences. Currently, many solutions mainly rely on static tests or single - parameter analysis. For example, only focusing on the initial performance of the sealing ring or the macroscopic characteristics of the material. Although this method is simple and easy to implement, it is difficult to comprehensively capture the dynamic changes caused by the disconnection state of the connector during actual use, resulting in insufficient prediction accuracy and late warning.

[0003] The limitations of existing methods have prompted researchers to re - examine the essential challenges of waterproof performance attenuation. The core problem is that the wire harness connector is affected by multiple factors during the disconnection state, especially the three key technical factors: the deformation rate of the sealing ring, the change of the microscopic structure of the material surface, and the moisture absorption rate. These factors do not exist in isolation but evolve dynamically with the disconnection duration, and their interaction makes the attenuation of waterproof performance show non - linear characteristics.

[0004] However, there is currently a lack of effective means to quantify the coupling effect of these factors in the time dimension, nor can an accurate correlation model be established between them and the performance attenuation. This directly leads to the inefficiency and unreliability of the early warning mechanism, especially in application scenarios with long - term use or complex environments, this technical problem is particularly prominent.

[0005] Therefore, how to monitor the changes of key parameters of the wire harness connector during the disconnection state, including the deformation rate of the sealing ring, the change of the microscopic structure of the material surface, and the moisture absorption rate, and combine the disconnection duration data to establish an intelligent early warning model that can dynamically reflect the law of waterproof performance attenuation has become the key problem that this research urgently needs to overcome. The solution to this problem will provide a theoretical and practical basis for improving the reliability of the connector in actual applications. Summary of the Invention

[0006] To solve the problems raised in the above - mentioned background art, in the first aspect of the present invention, an intelligent early warning method for the attenuation of the waterproof performance of a wire harness connector is provided, including: S1, obtaining the deformation rate data of the sealing ring of the wire harness connector in the disconnection state, collecting the deformation values at different disconnection durations by using a displacement sensor, and calculating the change trend of the deformation rate in combination with the evolution characteristics in the time dimension to obtain a dynamic deformation rate sequence; Step S2, extract the deformation rate values corresponding to the key time points from the dynamic deformation rate sequence, analyze the degradation distribution of the microscopic structure on the material surface at the same time points using an electron microscope, and determine the microscopic structure degradation parameters; Step S3, obtain the moisture absorption rate data of the connector at the key time points, record the moisture penetration amount through a constant humidity environment test, and calculate the coupling relationship between the moisture absorption rate and the microscopic structure degradation in combination with the disconnection duration to obtain the moisture absorption rate dynamic curve; Step S4, if the interaction between the dynamic deformation rate sequence and the moisture absorption rate dynamic curve exceeds the interaction threshold, then use the polynomial regression algorithm to fit the non-linear attenuation function and output the attenuation function parameters; Step S5, use the long short-term memory network algorithm to process the time dimension data of the dynamic deformation rate sequence, the microscopic structure degradation parameters, and the moisture absorption rate dynamic curve, train the performance degradation prediction model, and obtain the time series prediction result; Step S6, extract the parameter coupling characteristics from the time series prediction result. If the prediction error is less than the error threshold, then use the principal component analysis algorithm to quantify the contribution rate of each parameter to the performance degradation and output the coupling effect weight distribution; Step S7, adjust the prediction model according to the coupling effect weight distribution, combine the deformation rate, microscopic degradation parameters, and moisture absorption rate data monitored in real time, and calculate the current waterproof performance degradation value through the attenuation function and the prediction model to obtain the optimized warning model accuracy; Step S8, if the current waterproof performance degradation value exceeds the waterproof threshold, then use the logistic regression algorithm to judge the urgency of the degradation trend and output the dynamic warning signal sequence.

[0007] Optionally, in step S1, obtain the seal ring deformation rate data of the wire harness connector in the disconnected state, collect the deformation values at different disconnection durations using a displacement sensor, and calculate the deformation rate change trend in combination with the time dimension evolution characteristics to obtain the dynamic deformation rate sequence, including: Step S11, collect the seal ring deformation values of the wire harness connector in the disconnected state using a displacement sensor to obtain the initial deformation data set; Step S12, use the time dimension to divide the initial deformation data set to obtain the deformation value sequences corresponding to different disconnection durations; Step S13, calculate the deformation rate at each disconnection duration for the deformation value sequence to obtain the deformation rate data set; Step S14, fit the deformation rate data set by the least squares method to determine the deformation rate change trend; Step S15, extract the evolution characteristics according to the change trend to obtain the feature vector set; Step S16, use the K-means clustering algorithm to process the feature vector set and judge the classification boundary of the dynamic sequence; Step S17, divide the dynamic sequence through the classification boundary to obtain the dynamic deformation rate sequence of the seal ring deformation.

[0008] Optionally, in step S2, extract the deformation rate values corresponding to the key time points from the dynamic deformation rate sequence, analyze the degradation distribution of the material surface microstructure at the same time points using an electron microscope, and determine the microstructure degradation parameters, including: Step S21, extract the deformation rate values of the key time points from the dynamic deformation rate sequence, use the deformation rate threshold to judge the time nodes of significant changes, and obtain the set of key time points; Step S22, obtain the microstructure images of the material surface at the set of key time points through an electron microscope, and determine the preliminary characteristics of the degradation distribution; Step S23, perform image processing on the preliminary characteristics, use a convolutional neural network to extract the degraded areas of the microstructure, and obtain the degradation distribution data; Step S24, calculate the distribution characteristics of the degradation parameters from the degradation distribution data, and determine the spatial change trend of the parameters; Step S25, according to the spatial change trend, if the change exceeds the preset range, classify the degradation degree through a support vector machine to obtain the classification result; Step S26, obtain the mapping relationship between the classification result and time, and determine the time evolution characteristics of the degradation parameters; Step S27, through the time evolution characteristics, use linear regression to fit the change trend of the degradation parameters to obtain a long-term degradation prediction model.

[0009] Optionally, in step S3, obtain the moisture absorption rate data of the connector at the key time points, record the moisture penetration amount through a constant humidity environment test, and calculate the coupling relationship between the moisture absorption rate and the microstructure degradation in combination with the disconnection duration to obtain the moisture absorption rate dynamic curve, including: Step S31, through a constant humidity environment test, obtain the moisture penetration amount of the connector from the key time points to obtain the initial moisture absorption rate data; Step S32, for the initial moisture absorption rate data, calculate the degradation relationship between moisture penetration and microstructure in combination with the disconnection duration to obtain a set of coupling parameters; Step S33, obtain the set of coupling parameters, use the support vector machine algorithm to judge the change trend of the degradation relationship, and obtain the degradation feature vector; Step S34, according to the degradation feature vector, determine the dynamic mapping between the moisture absorption rate and the microstructure degradation through time series analysis to obtain the dynamic change curve; Step S35, for the dynamic change curve, obtain the curve slope and inflection point information, judge the evolution stage of the moisture absorption rate, and obtain the stage division result; Step S36: Based on the phase division results, use a clustering algorithm to determine the degradation modes within each phase, and obtain a pattern classification set. Step S37: According to the pattern classification set, use data fitting technology to generate a prediction model for the dynamic moisture absorption rate curve, and obtain the long-term evolution trend.

[0010] Optionally, in step S4, if the interaction between the dynamic deformation rate sequence and the dynamic moisture absorption rate curve exceeds the interaction threshold, use a polynomial regression algorithm to fit a non-linear attenuation function and output the attenuation function parameters, including: Step S41: Obtain sequence data from the dynamic deformation rate sequence, and determine the interaction strength through numerical analysis. Step S42: Compare the interaction strength with the interaction threshold. If the interaction strength exceeds the interaction threshold, trigger the fitting process. Step S43: Use a polynomial regression algorithm to process the sequence data and the dynamic moisture absorption rate curve to obtain a non-linear attenuation function. Step S44: Extract the attenuation function parameters from the non-linear attenuation function to determine the parameter set. Step S45: Calculate the deviation value of the fitting curve for the parameter set. If the deviation value exceeds the preset range, optimize the function parameters by adjusting the highest degree or the amplitude of the exponential term of the polynomial regression algorithm and output.

[0011] Optionally, in step S5, use a long short-term memory network algorithm to process the time dimension data of the dynamic deformation rate sequence, the microstructure degradation parameters, and the dynamic moisture absorption rate curve, train a performance degradation prediction model, and obtain a time series prediction result, including: Step S51: Use a long short-term memory network to process the dynamic deformation rate sequence to obtain a preliminary time series trend. Step S52: Extract the change features from the preliminary time series trend, combine them with the microstructure degradation parameters, and generate a comprehensive degradation sequence. Step S53: Obtain the time point data of the dynamic moisture absorption rate curve and calculate the change trend of the moisture absorption rate. Step S54: Compare the comprehensive degradation sequence with the change trend of the moisture absorption rate. If it exceeds the trend threshold, adjust the prediction model parameters to obtain an updated time series prediction result. Step S55: For the updated time series prediction result, use a support vector machine to verify its trend consistency. Step S56: According to the trend consistency verification result, combined with the degradation parameters, judge the dynamic change law of performance degradation.

[0012] Optionally, in step S6, extract the parameter coupling features from the time series prediction results. If the prediction error is less than the error threshold, use the principal component analysis algorithm to quantify the contribution rate of each parameter to performance attenuation, and output the coupling effect weight distribution, including: Step S61, obtain the prediction results through the time series prediction algorithm, extract the parameter coupling features from the prediction results, and obtain a feature set; Step S62, calculate the error between the prediction results and the actual values. If the error is less than the error threshold, use the principal component analysis algorithm for the feature set, calculate the contribution rate of each parameter to performance attenuation, and obtain the contribution rate data; Step S63, calculate the coupling effect of each parameter according to the contribution rate data, and obtain the effect intensity ranking; Step S64, extract the weight distribution from the effect intensity ranking, and use the min-max normalization method to process the weight distribution to obtain the normalized weight distribution; Step S65, for the normalized weight distribution, use the K-means clustering algorithm for clustering analysis, obtain the association grouping between parameters, and obtain the grouping results; Step S66, calculate the correlation coefficient between parameters through the grouping results, judge the influence path of parameter coupling on performance attenuation, and obtain the influence path set; Step S67, according to the influence path set, use the matplotlib library to generate the distribution map of parameter coupling features, and obtain the final distribution structure.

[0013] Optionally, in step S7, adjust the prediction model according to the coupling effect weight distribution, combine the deformation rate, micro degradation parameters and moisture absorption rate data monitored in real time, and calculate the current waterproof performance attenuation value through the attenuation function and the prediction model to obtain the optimized warning model accuracy, including: Step S71, collect the deformation rate, micro degradation parameters and moisture absorption rate data through sensors, and store the data in the database; Step S72, obtain the real-time monitoring results from the database, use the K-means clustering algorithm to group the data, and determine the correlation between the deformation rate, micro degradation parameters and moisture absorption rate; Step S73, calculate the weight distribution of each parameter according to the clustering results; Step S74, based on the weight distribution, adjust the parameters of the prediction model, and use the support vector machine algorithm to calculate the influence values of the deformation rate and micro degradation on the waterproof performance; Step S75, after obtaining the influence values, use the exponential decay function to process the micro degradation parameters and moisture absorption rate data to obtain the current waterproof performance attenuation value; Step S76, if the attenuation value exceeds the attenuation threshold, update the parameters of the prediction model, recalculate the waterproof performance attenuation value, and judge the performance attenuation trend; Step S77: Adjust the parameters of the exponential decay function according to the performance decay trend to optimize the output result of the prediction model; Step S78: Determine the long-term change trend of the waterproof performance through the optimized prediction model.

[0014] Optionally, in step S8, if the current waterproof performance decay value exceeds the waterproof threshold, the logistic regression algorithm is used to judge the urgency of the decay trend, and a dynamic warning signal sequence is output, including: Step S81: Obtain the real-time data of the waterproof performance through the sensor and calculate the current decay value; Step S82: If the current decay value exceeds the waterproof threshold, the decay value is processed using the logistic regression algorithm in the Scikit-learn library to determine the urgency of the decay trend; Step S83: Generate a dynamic warning signal sequence according to the urgency of the decay trend; Step S84: Process the dynamic warning signal sequence through the time series analysis method to judge the persistence of the trend change; Step S85: If the persistence of the trend change exceeds the normal range, the support vector machine algorithm in the Scikit-learn library is used to classify the urgency to obtain the classification result; Step S86: Update the dynamic warning signal sequence using the classification result and output the adjusted signal sequence; Step S87: Determine the real-time state of the waterproof performance through the adjusted signal sequence.

[0015] In the second aspect of the present invention, an intelligent early warning system for the attenuation of the waterproof performance of a wire harness connector is provided. The method described above is used to perform intelligent early warning on the attenuation of the waterproof performance of the wire harness connector. The system includes: A deformation rate data acquisition module, which is used to acquire the deformation rate data of the sealing ring under the disconnected state of the wire harness connector, collect the deformation values at different disconnection durations using a displacement sensor, and calculate the change trend of the deformation rate in combination with the evolution characteristics of the time dimension to obtain a dynamic deformation rate sequence; A microstructure analysis module, which is used to extract the deformation rate values corresponding to the key time points from the dynamic deformation rate sequence, analyze the degradation distribution of the material surface microstructure at the same time point using an electron microscope, and determine the microstructure degradation parameters; A moisture absorption rate data acquisition module, which is used to acquire the moisture absorption rate data of the connector at the key time point, record the moisture penetration amount through a constant humidity environment test, and calculate the coupling relationship between the moisture absorption rate and the microstructure degradation in combination with the disconnection duration to obtain a moisture absorption rate dynamic curve; An attenuation function fitting module, which is used to fit a non-linear attenuation function by using a polynomial regression algorithm and output attenuation function parameters if the interaction between the dynamic deformation rate sequence and the moisture absorption rate dynamic curve exceeds an interaction threshold; A performance attenuation prediction module, which is used to process the time dimension data of the dynamic deformation rate sequence, the microstructure degradation parameters and the moisture absorption rate dynamic curve by using a long short-term memory network algorithm, train a performance attenuation prediction model, and obtain a time series prediction result; A coupling effect analysis module, which is used to extract parameter coupling features from the time series prediction result. If the prediction error is less than an error threshold, a principal component analysis algorithm is used to quantify the contribution rate of each parameter to performance attenuation, and a coupling effect weight distribution is output; An early warning model optimization module, which is used to adjust the prediction model according to the coupling effect weight distribution, combine the deformation rate, microstructure degradation parameters and moisture absorption rate data monitored in real time, calculate the current waterproof performance attenuation value through the attenuation function and the prediction model, and obtain the accuracy of the optimized early warning model; A dynamic early warning module, which is used to judge the urgency of the attenuation trend by using a logistic regression algorithm and output a dynamic early warning signal sequence if the current waterproof performance attenuation value exceeds a waterproof threshold.

[0016] An intelligent early warning method and system for waterproof performance attenuation of a wire harness connector provided by the present invention establish a waterproof performance attenuation model by acquiring seal ring deformation rate data, analyzing microstructure degradation, and measuring moisture absorption rate changes. First, deformation values under different disconnection durations are collected, a dynamic deformation rate sequence is calculated, the microstructure degradation of the material surface is analyzed in combination with a scanning electron microscope, the degradation distribution of the microstructure with the change of the deformation rate is determined, the change trend of the moisture absorption rate is tested at the same time, and the coupling relationship with the microstructure degradation is calculated. Based on these data, a dynamic change law prediction model is trained by using a long short-term memory network algorithm, and the contribution rate of each parameter to the waterproof performance attenuation is quantified by principal component analysis. Finally, the present invention monitors relevant parameters in real time, jointly calculates the current waterproof performance attenuation value, and outputs a dynamic early warning signal, effectively improving the accuracy and timeliness of the early warning of the waterproof performance attenuation of the wire harness connector. Description of the Drawings

[0017] Figure 1 It is a flowchart of an intelligent early warning method for waterproof performance attenuation of a wire harness connector according to the present invention.

[0018] Figure 2 It is a schematic structural diagram of an intelligent early warning system for waterproof performance attenuation of a wire harness connector according to the present invention. Detailed Embodiments

[0019] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0020] As Figure 1 , in the first aspect of the present invention, an intelligent early warning method for the attenuation of the waterproof performance of a wire harness connector is provided, including: S1, Obtain the seal ring deformation rate data in the disconnected state of the wire harness connector, collect the deformation values at different disconnection durations using a displacement sensor, calculate the change trend of the deformation rate in combination with the evolution characteristics in the time dimension, and obtain a dynamic deformation rate sequence.

[0021] Optionally, this step further includes: Step S11, Collect the seal ring deformation values of the wire harness connector in the disconnected state through a displacement sensor to obtain an initial deformation data set.

[0022] Step S12, Divide the initial deformation data set using the time dimension to obtain the deformation value sequences corresponding to different disconnection durations.

[0023] Step S13, Calculate the deformation rate at each disconnection duration for the deformation value sequences to obtain a deformation rate data set.

[0024] Step S14, Fit the deformation rate data set by the least squares method to determine the change trend of the deformation rate.

[0025] Step S15, Extract the evolution characteristics according to the change trend to obtain a set of feature vectors.

[0026] Step S16, Process the set of feature vectors using the K-means clustering algorithm to judge the classification boundary of the dynamic sequence.

[0027] Step S17, Divide the dynamic sequence through the classification boundary to obtain the dynamic deformation rate sequence of the seal ring deformation.

[0028] Exemplarily, when obtaining the seal ring deformation rate data in the disconnected state of the wire harness connector, first use a high-precision displacement sensor to collect the deformation values of the seal ring at different disconnection durations.

[0029] S2, Extract the deformation rate values corresponding to the key time points from the dynamic deformation rate sequence, analyze the degradation distribution of the microscopic structure of the material surface at the same time points using an electron microscope, and determine the microscopic structure degradation parameters.

[0030] Optionally, this step further includes: Step S21: Extract the deformation rate values at key time points from the dynamic deformation rate sequence, and use a preset deformation rate threshold to judge the time nodes with significant changes, obtaining a set of key time points.

[0031] Step S22: Obtain the microscopic structure images of the material surface at the key time points in the set through an electron microscope, and determine the preliminary characteristics of the degradation distribution.

[0032] Step S23: Perform image processing on the preliminary characteristics, and use a convolutional neural network to extract the degraded areas of the microscopic structure, obtaining degradation distribution data.

[0033] Step S24: Calculate the distribution characteristics of the degradation parameters from the degradation distribution data, and determine the spatial change trend of the parameters.

[0034] Step S25: According to the spatial change trend, if the change exceeds the preset range, classify the degradation degree through a support vector machine to obtain a classification result.

[0035] Step S26: Obtain the mapping relationship between the classification result and time, and determine the time evolution characteristics of the degradation parameters.

[0036] Step S27: Through the time evolution characteristics, use linear regression to fit the change trend of the degradation parameters to obtain a long-term degradation prediction model.

[0037] Exemplarily, extract the deformation rate values corresponding to the key time points from the dynamic deformation rate sequence. First, identify the moments when the deformation rate changes significantly through time series analysis methods.

[0038] For example, in a time series, by calculating the difference in the deformation rate between adjacent time points and setting the deformation rate threshold to 05, when the difference exceeds this threshold, the time point is considered a critical time point. Suppose three critical time points are identified in the time series, namely t1 = 10 minutes, t2 = 30 minutes, and t3 = 60 minutes, and the corresponding deformation rate values are ε1 = 12, ε2 = 25, and ε3 = 38 respectively. Next, an electron microscope is used to analyze the degradation distribution of the material surface microstructure at the same time points. At time t1, microcracks are observed on the material surface through the electron microscope, with a crack density of 5 per square millimeter and an average crack length of 10 micrometers. At time t2, the crack density increases to 15 per square millimeter, and the average crack length increases to 20 micrometers. At time t3, the crack density further increases to 25 per square millimeter, and the average crack length increases to 30 micrometers. Based on these observation data, microstructure degradation parameters such as crack density and crack length are determined as quantitative indicators of the material degradation degree. Through linear regression analysis, a relationship model between crack density and time is established, and the model is ρ(t)=2t + 3, where ρ(t) is the crack density and t is the time. This model can predict the crack density at future time points and provide a basis for material life assessment.

[0039] S3. Obtain the moisture absorption rate data of the connector at the critical time point, record the moisture penetration amount through a constant humidity environment test, and calculate the coupling relationship between the moisture absorption rate and the microstructure degradation in combination with the disconnection duration to obtain the moisture absorption rate dynamic curve.

[0040] Optionally, this step further includes: Step S31. Through a constant humidity environment test, obtain the moisture penetration amount of the connector from the critical time point to obtain the initial moisture absorption rate data.

[0041] Step S32. For the initial moisture absorption rate data, calculate the degradation relationship between moisture penetration and microstructure in combination with the disconnection duration to obtain a set of coupling parameters.

[0042] Step S33. Obtain the set of coupling parameters, use the support vector machine algorithm to judge the change trend of the degradation relationship, and obtain the degradation feature vector.

[0043] Step S34. According to the degradation feature vector, determine the dynamic mapping between the moisture absorption rate and the microstructure degradation through time series analysis to obtain the dynamic change curve.

[0044] Step S35. For the dynamic change curve, obtain the curve slope and inflection point information, judge the evolution stage of the moisture absorption rate, and obtain the stage division result.

[0045] Step S36. Through the stage division result, use the clustering algorithm to determine the degradation mode within each stage to obtain the mode classification set.

[0046] Step S37: According to the pattern classification set, generate a prediction model of the moisture absorption rate dynamic curve through data fitting technology to obtain the long-term evolution trend.

[0047] Exemplarily, when obtaining the moisture absorption rate data of the connector at key time points, first place the connector in an environmental test chamber with a constant humidity of 60%, and continuously record its moisture penetration amount. Through an infrared spectroscopy analyzer, measure the moisture content on the surface of the connector every 2 hours, record the data, and plot the curve of the moisture penetration amount changing with time. In terms of the disconnection duration, simulate the disconnection times of the connector in different usage scenarios, which are 1 hour, 3 hours, and 5 hours respectively, and record the moisture absorption rate data for each disconnection duration. In order to calculate the coupling relationship between the moisture absorption rate and the microstructure degradation, use finite element analysis software to model the microstructure of the connector, and combine the experimental data to analyze the influence of the moisture absorption rate on the internal porosity and grain boundary strength of the material. By fitting the experimental data, the mathematical expression of the moisture absorption rate dynamic curve is obtained as Y = 0.05X + 0.002X², where Y is the moisture absorption rate and X is the time (hour).

[0048] Further analysis shows that when the moisture absorption rate reaches 5%, obvious degradation phenomena begin to appear in the microstructure of the connector, especially microcracks appear at the grain boundaries, resulting in a decrease in the mechanical properties of the material. Through this analysis, the service life of the connector in different humidity environments can be predicted, and a theoretical basis can be provided for material improvement.

[0049] S4: If the interaction between the dynamic deformation rate sequence and the moisture absorption rate dynamic curve exceeds the interaction threshold, then use the polynomial regression algorithm to fit the non-linear attenuation function and output the attenuation function parameters.

[0050] Optionally, this step further includes: Step S41: Obtain sequence data from the dynamic deformation rate sequence and determine the interaction strength through numerical analysis.

[0051] Step S42: Compare the interaction strength with the interaction threshold. If the interaction strength exceeds the interaction threshold, trigger the fitting process.

[0052] Step S43: Use the polynomial regression algorithm to process the sequence data and the moisture absorption rate dynamic curve to obtain the non-linear attenuation function.

[0053] Step S44: Extract the attenuation function parameters from the non-linear attenuation function to determine the parameter set.

[0054] Step S45: Calculate the deviation value of the fitting curve for the parameter set. If the deviation value exceeds the preset range, optimize the function parameters by adjusting the highest degree or exponential term amplitude of the polynomial regression algorithm and output.

[0055] S5. Use the long short - term memory network algorithm to process the time - dimension data of the dynamic deformation rate sequence, the microstructure degradation parameter, and the moisture absorption rate dynamic curve, train the performance degradation prediction model, and obtain the time - series prediction result.

[0056] Optionally, this step further includes: Step S51. Use the long short - term memory network to process the dynamic deformation rate sequence and obtain the preliminary time - series trend.

[0057] Step S52. Extract the change features from the preliminary time - series trend, combine them with the microstructure degradation parameter, and generate the comprehensive degradation sequence.

[0058] Step S53. Obtain the time - point data of the moisture absorption rate dynamic curve and calculate the change trend of the moisture absorption rate.

[0059] Step S54. Compare the comprehensive degradation sequence with the change trend of the moisture absorption rate. If it exceeds the preset trend threshold, adjust the prediction model parameters to obtain the updated time - series prediction result.

[0060] Step S55. For the updated time - series prediction result, use the support vector machine to verify its trend consistency.

[0061] Step S56. According to the trend consistency verification result, combined with the degradation parameter, judge the dynamic change law of the performance degradation.

[0062] Exemplarily, when processing the time - dimension data of the dynamic deformation rate sequence, the microstructure degradation parameter, and the moisture absorption rate dynamic curve, the long short - term memory network (LSTM) algorithm is used to model these time - series data. First, the dynamic deformation rate sequence is collected at intervals of 0.01 seconds, with a total of 1000 data points. The microstructure degradation parameter is obtained as a 10 - dimensional feature vector through scanning electron microscope image feature extraction. The moisture absorption rate dynamic curve is recorded hourly and covers 30 days of data. After uniformly standardizing these data, they are input into the LSTM network. The LSTM network contains 3 hidden layers, with the number of neurons in each layer being 64, 128, and 64 respectively. The activation function uses ReLU, the loss function uses the mean square error (MSE), the optimizer selects Adam, and the learning rate is set to 0.001. During training, 80% of the data is used as the training set, 20% as the test set, the training iteration times are 500 times, and the batch size is 32. After training, the model can accurately predict the performance degradation trend in the next 5 days. The average absolute error (MAE) between the predicted value and the actual value is 0.15, and the root mean square error (RMSE) is 0.22. By analyzing the prediction results, the accuracy of the material performance evaluation is further optimized, providing data support for subsequent material design.

[0063] S6. Extract parameter coupling features from the timing prediction results. If the prediction error is less than a preset error threshold, use the principal component analysis algorithm to quantify the contribution rate of each parameter to performance degradation, and output the coupling effect weight distribution.

[0064] Optionally, this step further includes: Step S61. Obtain the prediction results through the timing prediction algorithm, extract the parameter coupling features from the prediction results, and obtain a feature set.

[0065] Step S62. Calculate the error between the prediction results and the actual values. If the error is less than the preset error threshold, use the principal component analysis algorithm for the feature set to calculate the contribution rate of each parameter to performance degradation, and obtain the contribution rate data.

[0066] Step S63. Calculate the coupling effect of each parameter according to the contribution rate data to obtain the effect intensity ranking.

[0067] Step S64. Extract the weight distribution from the effect intensity ranking, and use the min-max normalization method to process the weight distribution to obtain the normalized weight distribution.

[0068] Step S65. For the normalized weight distribution, use the K-means clustering algorithm for clustering analysis to obtain the association grouping between parameters and obtain the grouping result.

[0069] Step S66. Calculate the correlation coefficient between parameters through the grouping result, judge the influence path of parameter coupling on performance degradation, and obtain the influence path set.

[0070] Step S67. According to the influence path set, use the matplotlib library to generate the distribution diagram of parameter coupling features to obtain the final distribution structure.

[0071] Exemplarily, when extracting parameter coupling features from the timing prediction results, first predict the system performance through a time series model, such as using a long short-term memory network (LSTM) model, input historical performance data such as CPU utilization rate, memory occupancy rate, etc., and predict the performance values for the next 10 time steps. If the prediction error is less than the preset threshold (such as the mean square error MSE is less than 0.05), then perform principal component analysis (PCA). The PCA algorithm performs dimensionality reduction processing on multiple parameters (such as temperature, voltage, load, etc.) and calculates the principal component contribution rate of each parameter.

[0072] For example, the contribution rate of the first principal component is 60%, the contribution rate of the second principal component is 25%, and the rest is 15%. Quantify the impact of each parameter on performance degradation according to the contribution rate, and output the weight distribution of the coupling effect. For example, the weight of temperature on performance degradation is 40%, voltage is 30%, load is 20%, and other factors are 10%. This process is implemented through an automated script to ensure the efficiency and accuracy of data processing and provide data support for system performance optimization.

[0073] S7. Adjust the prediction model according to the weight distribution of the coupling effect. Combine the deformation rate, microscopic degradation parameters, and moisture absorption rate data monitored in real time, and calculate the current waterproof performance degradation value through the attenuation function and the prediction model to obtain the optimized accuracy of the early warning model.

[0074] Optionally, this step further includes: Step S71. Collect the deformation rate, microscopic degradation parameters, and moisture absorption rate data through sensors and store the data in the database.

[0075] Step S72. Obtain the real-time monitoring results from the database, use the K-means clustering algorithm to group the data, and determine the correlation between the deformation rate, microscopic degradation parameters, and moisture absorption rate.

[0076] In one embodiment, the following formula is used to calculate the microscopic degradation parameters: ; where D represents the microscopic degradation parameter, m represents the number of influencing factors, represents the weight of the i-th degradation factor, represents the corresponding attenuation coefficient, and t represents time.

[0077] Step S73. Calculate the weight distribution of each parameter according to the clustering results.

[0078] Step S74. Based on the weight distribution, adjust the parameters of the prediction model, and use the support vector machine algorithm to calculate the influence values of the deformation rate and microscopic degradation on the waterproof performance.

[0079] Step S75. After obtaining the influence values, use the exponential decay function to process the microscopic degradation parameters and moisture absorption rate data to obtain the current waterproof performance degradation value.

[0080] Step S76. If the degradation value exceeds the degradation threshold, update the parameters of the prediction model, recalculate the waterproof performance degradation value, and judge the performance degradation trend.

[0081] Step S77. According to the performance degradation trend, adjust the parameters of the exponential decay function to optimize the output results of the prediction model.

[0082] Step S78. Determine the long-term change trend of the waterproof performance through the optimized prediction model.

[0083] S8. If the current waterproof performance attenuation value exceeds the waterproof threshold, use the logistic regression algorithm to determine the urgency of the attenuation trend and output a dynamic warning signal sequence.

[0084] Optionally, this step further includes: Step S81. Obtain the real-time data of the waterproof performance through the sensor and calculate the current attenuation value.

[0085] Step S82. If the current attenuation value exceeds the preset waterproof threshold, use the logistic regression algorithm in the Scikit-learn library to process the attenuation value and determine the urgency of the attenuation trend.

[0086] Step S83. Generate a dynamic warning signal sequence according to the urgency of the attenuation trend.

[0087] Step S84. Process the dynamic warning signal sequence through the time series analysis method to judge the persistence of the trend change.

[0088] Step S85. If the persistence of the trend change exceeds the normal range, use the support vector machine algorithm in the Scikit-learn library to classify the urgency level and obtain the classification result.

[0089] Step S86. Update the dynamic warning signal sequence with the classification result and output the adjusted signal sequence.

[0090] Step S87. Determine the real-time state of the waterproof performance through the adjusted signal sequence.

[0091] Exemplarily, in the waterproof performance monitoring system, assume that the current waterproof performance attenuation value is 85 dB, and the preset waterproof threshold is 80 dB. The system detects that this value has exceeded the waterproof threshold and immediately starts the logistic regression algorithm for attenuation trend analysis. The algorithm first constructs a model based on historical data, with inputs including time series waterproof performance data, environmental humidity, temperature and other features, and optimizes the model parameters through the gradient descent method to output the urgency of the attenuation trend.

[0092] For example, the model predicts that the waterproof performance will drop to 75 dB within the next 24 hours. The system generates a dynamic warning signal sequence according to the prediction result, such as "Urgent: The waterproof performance is dropping rapidly, please repair immediately". At the same time, the system associates the warning signal with business data such as equipment status and maintenance records, automatically generates a maintenance work order and pushes it to the relevant responsible person to ensure timely handling. The whole process is realized through an automated process without manual intervention, ensuring the efficiency and accuracy of monitoring and warning.

[0093] Please refer to Figure 2, the second aspect of the present invention provides an intelligent early warning system for the attenuation of the waterproof performance of a wire harness connector. The intelligent early warning of the attenuation of the waterproof performance of the wire harness connector is carried out by using the method described above. The system further includes: A deformation rate data acquisition module, which is used to acquire the deformation rate data of the sealing ring in the disconnected state of the wire harness connector. A displacement sensor is used to collect deformation values at different disconnection durations, and the change trend of the deformation rate is calculated by combining the evolution characteristics in the time dimension to obtain a dynamic deformation rate sequence; A microstructure analysis module, which is used to extract the deformation rate values corresponding to key time points from the dynamic deformation rate sequence, and uses an electron microscope to analyze the degradation distribution of the material surface microstructure at the same time point to determine the microstructure degradation parameters; A moisture absorption rate data acquisition module, which is used to acquire the moisture absorption rate data of the connector at the key time point. The moisture penetration amount is recorded through a constant humidity environment test, and the coupling relationship between the moisture absorption rate and the microstructure degradation is calculated in combination with the disconnection duration to obtain a moisture absorption rate dynamic curve; An attenuation function fitting module, which is used to fit a non-linear attenuation function by using a polynomial regression algorithm and output the attenuation function parameters if the interaction between the dynamic deformation rate sequence and the moisture absorption rate dynamic curve exceeds the interaction threshold; A performance attenuation prediction module, which is used to process the time dimension data of the dynamic deformation rate sequence, the microstructure degradation parameters and the moisture absorption rate dynamic curve by using a long short-term memory network algorithm, train a performance attenuation prediction model, and obtain a time series prediction result; A coupling effect analysis module, which is used to extract parameter coupling characteristics from the time series prediction result. If the prediction error is less than the error threshold, a principal component analysis algorithm is used to quantify the contribution rate of each parameter to the performance attenuation and output the coupling effect weight distribution; An early warning model optimization module, which is used to adjust the prediction model according to the coupling effect weight distribution, combine the real-time monitored deformation rate, microstructure degradation parameters and moisture absorption rate data, calculate the current waterproof performance attenuation value through the attenuation function and the prediction model, and obtain the optimized accuracy of the early warning model; A dynamic early warning module, which is used to judge the urgency of the attenuation trend by using a logistic regression algorithm and output a dynamic early warning signal sequence if the current waterproof performance attenuation value exceeds the waterproof threshold.

[0094] An intelligent early warning method and system for the attenuation of the waterproof performance of a wire harness connector provided by the present invention establish a waterproof performance attenuation model by obtaining the data of the deformation rate of the sealing ring, analyzing the degradation of the microstructure, and measuring the change in the moisture absorption rate. First, the deformation values under different disconnection durations are collected, the dynamic deformation rate sequence is calculated, the degradation of the material surface microstructure is analyzed in combination with a scanning electron microscope, the degradation distribution of the microstructure with the change of the deformation rate is determined, and at the same time, the change trend of the moisture absorption rate is tested, and the coupling relationship with the degradation of the microstructure is calculated; based on these data, the present invention uses the long short-term memory network algorithm to train the dynamic change law prediction model, and quantifies the contribution rate of each parameter to the attenuation of the waterproof performance through principal component analysis. Finally, the relevant parameters are monitored in real time, the current attenuation value of the waterproof performance is calculated jointly, and a dynamic early warning signal is output, effectively improving the accuracy and timeliness of the early warning of the attenuation of the waterproof performance of the wire harness connector.

[0095] Only some preferred embodiments of the present invention are listed above, but the present invention is not limited thereto, and many improvements and transformations can be made. As long as the improvements and transformations are made on the basis of the basic principle of the present invention, they should be regarded as falling within the protection scope of the present invention.

Claims

1. An intelligent early warning method for the attenuation of the waterproof performance of a wire harness connector, characterized in that, The method includes: S1. Obtain the seal ring deformation rate data under the disconnected state of the wire harness connector. Use a displacement sensor to collect deformation values at different disconnection durations, and calculate the change trend of the deformation rate by combining the evolution characteristics in the time dimension to obtain a dynamic deformation rate sequence. S2. Extract the deformation rate values corresponding to the key time points from the dynamic deformation rate sequence. Use an electron microscope to analyze the degradation distribution of the material surface microstructure at the same time points to determine the microstructure degradation parameters. S3. Obtain the moisture absorption rate data of the connector at the key time points. Record the moisture penetration amount through a constant humidity environment test, and calculate the coupling relationship between the moisture absorption rate and the microstructure degradation by combining the disconnection duration to obtain a moisture absorption rate dynamic curve. S4. If the interaction between the dynamic deformation rate sequence and the moisture absorption rate dynamic curve exceeds the interaction threshold, then use the polynomial regression algorithm to fit the non-linear attenuation function and output the attenuation function parameters. S5. Use the long short-term memory network algorithm to process the time dimension data of the dynamic deformation rate sequence, the microstructure degradation parameters, and the moisture absorption rate dynamic curve, train the performance degradation prediction model, and obtain the time series prediction result. S6. Extract the parameter coupling characteristics from the time series prediction result. If the prediction error is less than the error threshold, then use the principal component analysis algorithm to quantify the contribution rate of each parameter to the performance degradation and output the coupling effect weight distribution. S7. Adjust the prediction model according to the coupling effect weight distribution. Combine the deformation rate, microstructure degradation parameters, and moisture absorption rate data monitored in real time, and calculate the current waterproof performance degradation value through the attenuation function and the prediction model to obtain the optimized accuracy of the early warning model. S8. If the current waterproof performance degradation value exceeds the waterproof threshold, then use the logistic regression algorithm to judge the urgency of the degradation trend and output a dynamic early warning signal sequence.

2. The method according to claim 1, wherein The step S1, obtaining the seal ring deformation rate data under the disconnected state of the wire harness connector, using a displacement sensor to collect deformation values at different disconnection durations, and calculating the change trend of the deformation rate by combining the evolution characteristics in the time dimension to obtain a dynamic deformation rate sequence, includes: Step S11. Collect the seal ring deformation values of the wire harness connector in the disconnected state through a displacement sensor to obtain an initial deformation data set. Step S12. Use the time dimension to divide the initial deformation data set to obtain the deformation value sequences corresponding to different disconnection durations. Step S13. Calculate the deformation rate at each disconnection duration for the deformation value sequence to obtain a deformation rate data set. Step S14. Fit the deformation rate data set by the least squares method to determine the change trend of the deformation rate. Step S15. Extract the evolution characteristics according to the change trend to obtain a set of feature vectors. Step S16. Use the K-means clustering algorithm to process the set of feature vectors and judge the classification boundary of the dynamic sequence. Step S17. Divide the dynamic sequence through the classification boundary to obtain the dynamic deformation rate sequence of the seal ring deformation.

3. The method according to claim 1, characterized in that, The step S2, extracting the deformation rate values corresponding to the key time points from the dynamic deformation rate sequence, using an electron microscope to analyze the degradation distribution of the material surface microstructure at the same time points to determine the microstructure degradation parameters, includes: Step S21: Extract the deformation rate values at key time points from the dynamic deformation rate sequence, and use the deformation rate threshold to judge the time nodes of significant changes to obtain the set of key time points; Step S22: Obtain the microscopic structure images of the material surface at the set of key time points through an electron microscope, and determine the preliminary characteristics of the degradation distribution; Step S23: Perform image processing on the preliminary characteristics, and use a convolutional neural network to extract the degraded areas of the microscopic structure to obtain the degradation distribution data; Step S24: Calculate the distribution characteristics of the degradation parameters from the degradation distribution data, and determine the spatial variation trend of the parameters; Step S25: According to the spatial variation trend, if the change exceeds the preset range, classify the degradation degree through a support vector machine to obtain the classification result; Step S26: Obtain the mapping relationship between the classification result and time, and determine the time evolution characteristics of the degradation parameters; Step S27: Through the time evolution characteristics, use linear regression to fit the change trend of the degradation parameters to obtain a long-term degradation prediction model.

4. The method according to claim 1, wherein In step S3, obtain the moisture absorption rate data of the connector at the key time point, record the moisture penetration amount through a constant humidity environment test, and calculate the coupling relationship between the moisture absorption rate and the microscopic structure degradation in combination with the disconnection duration to obtain the moisture absorption rate dynamic curve, including: Step S31: Through a constant humidity environment test, obtain the moisture penetration amount of the connector from the key time point to obtain the initial moisture absorption rate data; Step S32: For the initial moisture absorption rate data, calculate the degradation relationship between moisture penetration and the microscopic structure in combination with the disconnection duration to obtain a set of coupling parameters; Step S33: Obtain the set of coupling parameters, and use a support vector machine algorithm to judge the change trend of the degradation relationship to obtain a degradation feature vector; Step S34: According to the degradation feature vector, determine the dynamic mapping between the moisture absorption rate and the microscopic structure degradation through time series analysis to obtain a dynamic change curve; Step S35: For the dynamic change curve, obtain the curve slope and inflection point information, and judge the evolution stage of the moisture absorption rate to obtain the stage division result; Step S36: Through the stage division result, use a clustering algorithm to determine the degradation mode within each stage to obtain a set of mode classifications; Step S37: According to the set of mode classifications, generate a prediction model of the moisture absorption rate dynamic curve through data fitting technology to obtain the long-term evolution trend.

5. The method according to claim 1, wherein In step S4, if the interaction between the dynamic deformation rate sequence and the moisture absorption rate dynamic curve exceeds the interaction threshold, then use a polynomial regression algorithm to fit a non-linear attenuation function and output the attenuation function parameters, including: Step S41: Obtain the sequence data from the dynamic deformation rate sequence, and determine the interaction intensity through numerical analysis; Step S42: Compare the interaction intensity with the interaction threshold. If the interaction intensity exceeds the interaction threshold, trigger the fitting process; Step S43: Use a polynomial regression algorithm to process the sequence data and the moisture absorption rate dynamic curve to obtain a non-linear attenuation function; Step S44: Extract the attenuation function parameters from the non-linear attenuation function to determine the parameter set; Step S45: Calculate the deviation value of the fitting curve for the parameter set. If the deviation value exceeds the preset range, optimize the function parameters by adjusting the highest degree of the polynomial regression algorithm or the amplitude of the exponential term, and output the results.

6. The method according to claim 1, wherein In step S5, use the long short-term memory network algorithm to process the time dimension data of the dynamic deformation rate sequence, microstructure degradation parameters, and moisture absorption rate dynamic curve, and train the performance degradation prediction model to obtain the time series prediction results, including: Step S51: Use the long short-term memory network to process the dynamic deformation rate sequence to obtain the preliminary time series trend. Step S52: Extract the change features from the preliminary time series trend, combine them with the microstructure degradation parameters, and generate a comprehensive degradation sequence. Step S53: Obtain the time point data of the moisture absorption rate dynamic curve and calculate the change trend of the moisture absorption rate. Step S54: Compare the comprehensive degradation sequence with the change trend of the moisture absorption rate. If it exceeds the trend threshold, adjust the prediction model parameters to obtain the updated time series prediction results. Step S55: For the updated time series prediction results, use the support vector machine to verify their trend consistency. Step S56: According to the trend consistency verification results, combine with the degradation parameters to judge the dynamic change law of performance degradation.

7. The method according to claim 1, characterized in that, In step S6, extract the parameter coupling features from the time series prediction results. If the prediction error is less than the error threshold, use the principal component analysis algorithm to quantify the contribution rate of each parameter to performance degradation, and output the coupling effect weight distribution, including: Step S61: Obtain the prediction results through the time series prediction algorithm, extract the parameter coupling features from the prediction results to obtain a feature set. Step S62: Calculate the error between the prediction results and the actual values. If the error is less than the error threshold, use the principal component analysis algorithm for the feature set to calculate the contribution rate of each parameter to performance degradation, and obtain the contribution rate data. Step S63: Calculate the coupling effect of each parameter according to the contribution rate data to obtain the effect intensity ranking. Step S64: Extract the weight distribution from the effect intensity ranking, and use the min-max normalization method to process the weight distribution to obtain the normalized weight distribution. Step S65: For the normalized weight distribution, use the K-means clustering algorithm for clustering analysis to obtain the association groups between parameters and obtain the grouping results. Step S66: Calculate the correlation coefficient between parameters through the grouping results, and judge the influence path of parameter coupling on performance degradation to obtain the influence path set. Step S67: According to the influence path set, use the matplotlib library to generate the distribution map of parameter coupling features to obtain the final distribution structure.

8. The method according to claim 1, characterized in that In step S7, adjust the prediction model according to the coupling effect weight distribution, and combine the real-time monitored deformation rate, microstructure degradation parameters, and moisture absorption rate data. Calculate the current waterproof performance degradation value through the attenuation function and the prediction model to obtain the optimized accuracy of the warning model, including: Step S71: Collect the deformation rate, microstructure degradation parameters, and moisture absorption rate data through sensors and store the data in the database. Step S72: Obtain the real-time monitoring results from the database, use the K-means clustering algorithm to group the data, and determine the correlation between the deformation rate, microstructure degradation parameters, and moisture absorption rate. Step S73: Calculate the weight distribution of each parameter according to the clustering result; Step S74: Based on the weight distribution, adjust the parameters of the prediction model, and use the support vector machine algorithm to calculate the influence values of the deformation rate and microscopic degradation on the waterproof performance; Step S75: After obtaining the influence values, use the exponential decay function to process the microscopic degradation parameters and moisture absorption rate data to obtain the current waterproof performance decay value; Step S76: If the decay value exceeds the decay threshold, update the parameters of the prediction model, recalculate the waterproof performance decay value, and judge the performance decay trend; Step S77: According to the performance decay trend, adjust the parameters of the exponential decay function to optimize the output result of the prediction model; Step S78: Determine the long-term change trend of the waterproof performance through the optimized prediction model.

9. The method according to claim 1, wherein In step S8, if the current waterproof performance decay value exceeds the waterproof threshold, use the logistic regression algorithm to judge the urgency of the decay trend and output a dynamic warning signal sequence, including: Step S81: Obtain the real-time data of the waterproof performance through the sensor and calculate the current decay value; Step S82: If the current decay value exceeds the waterproof threshold, use the logistic regression algorithm in the Scikit-learn library to process the decay value and determine the urgency of the decay trend; Step S83: Generate a dynamic warning signal sequence according to the urgency of the decay trend; Step S84: Process the dynamic warning signal sequence through the time series analysis method to judge the persistence of the trend change; Step S85: If the persistence of the trend change exceeds the normal range, use the support vector machine algorithm in the Scikit-learn library to classify the urgency level to obtain the classification result; Step S86: Update the dynamic warning signal sequence with the classification result and output the adjusted signal sequence; Step S87: Determine the real-time state of the waterproof performance through the adjusted signal sequence.

10. An intelligent early warning system for the attenuation of the waterproof performance of a wire harness connector, characterized in that, Use the method described in any one of claims 1 to 9 to perform intelligent warning on the attenuation of the waterproof performance of the wire harness connector. The system includes: A deformation rate data acquisition module, which is used to acquire the seal ring deformation rate data when the wire harness connector is in the disconnected state, collect the deformation values at different disconnection durations using a displacement sensor, and calculate the deformation rate change trend in combination with the evolution characteristics of the time dimension to obtain a dynamic deformation rate sequence; A microscopic structure analysis module, which is used to extract the deformation rate values corresponding to the key time points from the dynamic deformation rate sequence, analyze the degradation distribution of the material surface microscopic structure at the same time point using an electron microscope, and determine the microscopic structure degradation parameters; A moisture absorption rate data acquisition module, which is used to acquire the moisture absorption rate data of the connector at the key time point, record the moisture penetration amount through a constant humidity environment test, and calculate the coupling relationship between the moisture absorption rate and microscopic structure degradation in combination with the disconnection duration to obtain a moisture absorption rate dynamic curve; An attenuation function fitting module, which is used to, if the interaction between the dynamic deformation rate sequence and the moisture absorption rate dynamic curve exceeds the interaction threshold, use the polynomial regression algorithm to fit the non-linear attenuation function and output the attenuation function parameters; A performance degradation prediction module, which is used to process the time - dimension data of the dynamic deformation rate sequence, the microstructure degradation parameters, and the moisture absorption rate dynamic curve by using the long - short - term memory network algorithm, train a performance degradation prediction model, and obtain a time - series prediction result; A coupling effect analysis module, which is used to extract parameter coupling features from the time - series prediction result. If the prediction error is less than the error threshold, the principal component analysis algorithm is used to quantify the contribution rate of each parameter to performance degradation, and the coupling effect weight distribution is output; An early - warning model optimization module, which is used to adjust the prediction model according to the coupling effect weight distribution, combine the deformation rate, microstructure degradation parameters, and moisture absorption rate data monitored in real - time, calculate the current waterproof performance degradation value through the attenuation function and the prediction model, and obtain the optimized accuracy of the early - warning model; A dynamic early - warning module, which is used to, if the current waterproof performance degradation value exceeds the waterproof threshold, use the logistic regression algorithm to judge the urgency of the attenuation trend and output a dynamic early - warning signal sequence.

Images