A harness connection performance evolution analysis method and system

Abstract

The present application relates to the technical field of performance evolution analysis, in particular to a kind of wire harness connection performance evolution analysis method and system.The method comprises the following steps: obtaining various types of operating data corresponding to wire harness connection node, and generating multidimensional time series data set based on data acquisition timing;Multi-dimensional time series data set is calibrated in time sequence, and wire harness connection parameter correlation matrix is constructed;Wire harness performance correlation coefficient set is generated based on wire harness connection parameter correlation matrix;Wire harness connection performance evolution model is established based on wire harness performance correlation coefficient set, and historical performance data is input to iteratively train wire harness connection performance evolution model, and generate model training parameters;Evolution trend is predicted based on model training parameters to standardized time series data set, and wire harness performance evolution trend curve is generated;Evolution anomaly evaluation analysis is carried out based on wire harness performance evolution trend curve, and wire harness connection performance anomaly early warning information is generated.The present application can comprehensively improve wire harness connection state monitoring efficiency.

Description

Technical Field

[0001] This invention relates to the field of performance evolution analysis technology, and specifically to a method and system for analyzing the performance evolution of wire harness connections. Background Technology

[0002] Currently, most common wire harness connection performance analysis methods adopt a single-parameter monitoring mode. By collecting basic operational data of wire harness connection nodes, a fixed threshold is used to determine whether the connection performance is abnormal. However, due to the lack of a performance evolution model, existing methods cannot predict the evolution trend of wire harness connection performance. They can only detect problems after obvious performance abnormalities occur, and cannot generate early warnings of abnormalities, thus reducing the accuracy of wire harness connection performance analysis. Summary of the Invention

[0003] Therefore, it is necessary for the present invention to provide a method and system for analyzing the evolution of wire harness connection performance in order to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a method for analyzing the evolution of wire harness connection performance includes the following steps: Step S1: Obtain various types of operational data corresponding to the wire harness connection nodes, and generate a multi-dimensional time series dataset based on the data acquisition timing; perform timing calibration on the multi-dimensional time series dataset to generate a standardized time series dataset; Step S2: Extract the wire harness connection performance feature parameters based on the standardized time series dataset, and construct the wire harness connection parameter correlation matrix; generate a set of wire harness performance correlation coefficients based on the wire harness connection parameter correlation matrix; Step S3: Establish a wire harness connection performance evolution model based on the wire harness performance correlation coefficient set, input historical performance data to iteratively train the wire harness connection performance evolution model, and generate model training parameters; based on the model training parameters, predict the evolution trend of the standardized time series dataset to generate a wire harness performance evolution trend curve. Step S4: Based on the wire harness performance evolution trend curve, perform evolution anomaly assessment and analysis to generate wire harness connection performance anomaly early warning information.

[0005] Furthermore, step S1 includes the following steps: Step S11: Collect various operating data of the wire harness connection nodes under different operating conditions, including connection contact resistance, node temperature, vibration frequency, voltage transmission stability and mechanical stress data; Step S12: Sort various types of operational data according to the data acquisition time sequence, divide the data into segments according to the preset acquisition cycle, and generate a multi-dimensional time series dataset, where each data segment corresponds to a time series unit, and each time series unit contains various types of operational data under the corresponding working conditions; Step S13: Extract the time stamp of each time series unit in the multi-dimensional time series dataset, align and calibrate the time stamps of different time series units through the dynamic time warping algorithm to eliminate time series deviations, calculate the time series consistency error of various types of running data in each time series unit, correct the data segments based on the time series consistency error, and generate a time series unified intermediate dataset. Step S14: Normalize the various types of running data in the intermediate dataset to eliminate dimensional differences and generate a standardized time series dataset, which includes calibrated time series labels and corresponding normalized types of running data.

[0006] Furthermore, step S2 includes the following steps: Step S21: Extract the wire harness connection performance characteristic parameters from the standardized time series dataset, including the contact resistance change rate, temperature fluctuation amplitude, peak vibration frequency, voltage transmission attenuation, and cumulative mechanical stress. Step S22: Based on the performance characteristic parameters of the harness connection, determine the time correlation dimension of each characteristic parameter, and construct the harness connection parameter correlation matrix, where each matrix element is the corresponding value of any two characteristic parameters at different times; Step S23: Calculate the correlation matrix of the wire harness connection parameters, determine the degree of correlation between each feature parameter, and generate an initial set of correlation coefficients; Step S24: Normalize the initial correlation coefficient set to eliminate coefficient bias and obtain a standardized correlation coefficient set; based on the standardized correlation coefficient set, select feature parameter pairs with a correlation degree higher than the preset correlation threshold, and integrate all feature parameter pairs with the corresponding standardized correlation coefficients to generate a performance correlation coefficient set.

[0007] Furthermore, step S23 includes the following steps: Extract the feature parameter data of each column in the wire harness connection parameter correlation matrix and use it as the reference sequence, while the feature parameter data of the remaining columns are used as the comparison sequence. Calculate the absolute difference between each comparison sequence and the reference sequence, and determine the maximum and minimum absolute differences at each time point; The correlation coefficient between each comparison sequence and the reference sequence at each time step is calculated based on the maximum and minimum absolute differences. The mean of the correlation coefficients between each comparison sequence and the reference sequence at each time step is calculated to obtain the initial correlation coefficient between each comparison sequence and the reference sequence. Following the steps described above, each feature parameter is used as a reference sequence to complete the correlation coefficient calculation for all feature parameter pairs. All calculation results are then integrated to generate an initial correlation coefficient set. After obtaining the initial correlation coefficient between each comparison sequence and the reference sequence, the process further includes: Extract the feature parameter pairs and time series information corresponding to the initial correlation coefficient, calculate the mean correlation coefficient of each feature parameter pair in different time series stages, and generate the time series stage correlation mean; based on the time series stage correlation mean, analyze the trend of the correlation strength of each feature parameter pair over time, and generate the correlation trend curve. Based on the correlation trend curve, the abrupt change points of correlation strength are extracted, and the time-series nodes and correlation coefficient changes corresponding to the abrupt change points are determined. Based on the correlation coefficient changes corresponding to the abrupt change points, the initial correlation coefficients are corrected to obtain the corrected initial correlation coefficients. The corrected initial correlation coefficients are then classified and integrated according to the feature parameters to update the initial correlation coefficient set.

[0008] Furthermore, the calculation of the correlation coefficient between each comparison sequence and the reference sequence at each time step based on the maximum and minimum absolute differences includes the following steps: Based on the maximum and minimum absolute difference values ​​at each time point, the difference in absolute difference between the reference sequence and the comparison sequence at the same time node is analyzed, and a time difference dataset is constructed. Based on the time-series difference dataset, all time-series nodes are traversed to determine the distribution of extreme differences between each comparison sequence and the corresponding reference sequence at each time point, and global difference characterization data is extracted based on the distribution of extreme differences. Based on the global difference representation data, the single-node difference data in the time-series difference dataset is normalized and mapped to generate single-node difference normalized data. Based on single-node difference normalized data, continuous fitting is performed on all time-series nodes to generate time-series continuous difference characterization data; based on the time-series continuous difference characterization data, the degree of correlation approximation between each comparison sequence and the reference sequence at each time point is inverted, and the correlation coefficient between each comparison sequence and the reference sequence at each time point is generated.

[0009] Furthermore, the step of inverting the correlation approximation degree between each comparison sequence and the reference sequence at each time step based on the temporal continuous difference characterization data includes the following steps: Extract the curve morphology features corresponding to the time-series continuous difference characterization data, and obtain the difference curvature data at each time point based on the curve morphology features; The dynamic fit between each comparison sequence and the reference sequence at each time step is calculated based on the difference curvature data, and the time-series fit data is generated. Based on the temporal variation characteristics of the reference sequence matched with the time-fit data, a fit correction factor is generated. The fitting data at different times is dynamically calibrated based on the fitting correction factor to generate calibrated fitting data; the degree of correlation between each comparison sequence and the reference sequence at each time point is inverted based on the calibrated fitting data.

[0010] Furthermore, step S3, which involves establishing a wire harness connection performance evolution model based on the set of wire harness performance correlation coefficients and iteratively training the model using historical performance data, includes the following steps: Based on the set of correlation coefficients for wire harness performance, the influencing factors of wire harness connection performance evolution are determined, the weight allocation rules of each influencing factor are clarified, and the basic framework corresponding to the wire harness connection performance evolution model is constructed. Historical full lifecycle operation data and corresponding performance evolution results of the wire harness connection nodes are selected as the model training sample set, and the model training sample set is divided into training set and validation set according to a preset ratio; The performance correlation coefficient set and historical performance data in the training set are input into the wire harness to connect the performance evolution model, and the model parameters are iteratively optimized by the gradient descent algorithm to calculate the error between the model prediction value and the actual performance evolution result. Adjust the weight parameters and correlation factors corresponding to the wire harness connection performance evolution model based on the error value, and repeat the training process iteratively until the model prediction error is less than the preset error threshold. Use the validation set to validate the trained wire harness connection performance evolution model, calculate the model validation accuracy, and stop model training when the model validation accuracy reaches the preset accuracy threshold. Extract the current weight parameters, correlation factors and iteration number of the model to generate model training parameters.

[0011] Furthermore, the evolutionary trend prediction of the standardized time-series dataset based on model training parameters includes the following steps: Extract the weight parameters and correlation factors from the model training parameters, import them into the trained harness connection performance evolution model, and complete the model parameter configuration; The standardized time-series dataset is divided into multiple prediction units according to time order. Each prediction unit contains continuous standardized time-series data and corresponding performance characteristic parameters. The standardized time-series data of each prediction unit is input into the configured harness connection performance evolution model to predict the predicted value of the harness connection performance parameter corresponding to that unit. Based on the predicted values ​​of the wire harness connection performance parameters of each prediction unit, and combined with the corresponding time stamps, a time series sequence of performance parameter predictions is generated. The time series sequence of performance parameter predictions is smoothed to eliminate prediction fluctuations. Based on the smoothed time series sequence, a wire harness connection performance evolution trend curve is plotted. This curve contains the evolution trend of each performance characteristic parameter and its time correspondence.

[0012] Furthermore, step S4 includes the following steps: Extract the evolution trajectory of each performance characteristic parameter in the trend curve of wire harness connection performance, calculate the rate of change of each characteristic parameter in different time stages, and generate a rate of change sequence; Based on the rate of change sequence, the isolated forest algorithm is used to detect anomalies in the performance evolution trend, identify anomalies in the evolution trajectory, and determine the time nodes and performance parameter values ​​corresponding to the anomalies. The performance parameter values ​​corresponding to the anomalies are extracted, and combined with the performance correlation coefficient set, the associated influencing factors of the anomalies are analyzed to determine the anomaly influencing factors and their degree of influence. Based on the abnormal impact factors and their degree of impact, anomaly level assessment results are generated, which include different warning priorities corresponding to different anomaly levels; based on the anomaly level assessment results and abnormal impact factors, evolution trend adjustment strategies are formulated, including parameter adjustment schemes and time series monitoring schemes; integrating the anomaly level assessment results, abnormal impact factors, and evolution trend adjustment strategies, anomaly warning information for harness connection performance is generated, which includes the anomaly time, anomaly type, scope of impact, and adjustment suggestions.

[0013] Furthermore, the present invention also provides a wire harness connection performance evolution analysis system, including a processor, a memory, and a computer program stored in the memory and executable on the processor, for performing the wire harness connection performance evolution analysis method as described above.

[0014] The beneficial effects of this invention are: The wire harness connection performance evolution analysis method proposed in this invention, compared with the prior art, has the following advantages: by acquiring various operational data such as voltage, current, temperature, contact resistance, and vibration response corresponding to the wire harness connection nodes, a multi-dimensional time-series dataset is formed according to the actual acquisition time sequence. This overcomes the limitations of traditional methods that only collect a small amount of basic operational data and cannot fully reflect the true state of the wire harness, providing sufficient and comprehensive raw information for performance analysis. Based on this, the data undergoes time-series calibration to unify the time base, correct sampling biases, and eliminate data distortion caused by timestamp misalignment, generating a standardized time-series dataset. This allows data from different dimensions and at different times to be arranged in an orderly manner and reliably compared under a unified time axis. Secondly, by automatically extracting wire harness connection performance characteristic parameters such as contact resistance, temperature rise rate, fluctuation amplitude, and stability indicators based on the standardized time-series dataset, scattered data is transformed into key features that can be directly used for evaluation. Then, by constructing a wire harness connection parameter correlation matrix, the mutual influence, synergistic changes, and constraint relationships between various characteristic parameters are described, truly restoring the multi-physical quantity coupling state of the wire harness connection nodes during operation. Based on the matrix, a set of correlation coefficients for harness performance is further calculated, quantifying the correlation strength and variation patterns between different parameters. This upgrades performance analysis from "single-point judgment" to "global correlation analysis." This step comprehensively captures the implicit correlations between parameters, avoiding abnormal omissions and misjudgments caused by ignoring coupling relationships, significantly improving the depth and rationality of performance analysis, and providing crucial quantitative support for subsequent evolutionary model building and trend prediction. Then, by constructing a harness connection performance evolution model based on the set of correlation coefficients, a complete description of the dynamic change process of the harness from normal operation to gradual deterioration can be achieved, overcoming the limitations of traditional analysis that only focuses on the current state and lacks historical extrapolation and future prediction capabilities. Historical performance data is used to iteratively train the model, continuously optimizing its internal parameters to make the model more closely match the actual evolution patterns of harness aging, wear, and loosening, possessing stronger generalization and trend expression capabilities. By using the trained model to predict trends in standardized time-series datasets, the system generates harness performance evolution trend curves, visually displaying the direction of change, degradation rate, and key inflection points of connection performance over a future period, thus shifting from "post-event review" to "pre-event prediction." Finally, through continuous tracking and anomaly assessment of the harness performance evolution trend curves, the system can identify risk characteristics such as abrupt changes, accelerated degradation, and abnormal fluctuations that deviate from the normal evolutionary trajectory. Potential risks can be identified before performance indicators exceed traditional fixed thresholds, enabling early detection and early warning.The generated abnormal wiring harness connection performance early warning information includes the abnormal location, abnormal type, development trend and risk level, which can provide clear and actionable early warning prompts for operation and maintenance personnel. This can prevent equipment downtime, short circuits or even safety accidents caused by problems such as poor wiring harness contact, overheating, or looseness. It can effectively improve the lead time and reliability of abnormal identification, reduce the probability of failure and maintenance costs, and comprehensively improve the practical value of wiring harness connection status monitoring and performance analysis. Attached Figure Description

[0015] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram of the steps in the wire harness connection performance evolution analysis method of the present invention; Figure 2 for Figure 1 A detailed flowchart of step S2. Detailed Implementation

[0016] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0017] To achieve the above objectives, please refer to Figures 1 to 2 This invention provides a method for analyzing the evolution of wire harness connection performance. In the embodiments of this invention, please refer to... Figure 1 The diagram shown is a flowchart illustrating the steps of the wire harness connection performance evolution analysis method of the present invention. In this example, the wire harness connection performance evolution analysis method includes the following steps: Step S1: Obtain various types of operational data corresponding to the wire harness connection nodes, and generate a multi-dimensional time series dataset based on the data acquisition timing; perform timing calibration on the multi-dimensional time series dataset to generate a standardized time series dataset; In this embodiment of the invention, a data acquisition device is used to collect data on the wiring harness connection nodes under all operating conditions. The acquisition process strictly follows the acquisition specifications, covering different operating conditions such as idling, acceleration, constant speed, deceleration, and extreme environments. The collected operating data includes connection contact resistance, node temperature, vibration frequency, voltage transmission stability, and mechanical stress data. The connection contact resistance is recorded in specified units, the node temperature is collected at fixed intervals, the vibration frequency is recorded as the number of vibrations per unit time, the voltage transmission stability is recorded as the voltage fluctuation difference, and the mechanical stress data is recorded as the stress magnitude borne by the node. After the acquisition is completed, the various operating data are arranged in an orderly manner based on the data acquisition time sequence. The data is divided into segments according to the set acquisition period, which is a fixed duration. Each data segment corresponds to a time sequence unit, and each time sequence unit completely contains the connection contact resistance, node temperature, vibration frequency, voltage transmission stability, and mechanical stress data under the corresponding operating condition. All time sequence units are organized in chronological order of acquisition time to form a multi-dimensional time sequence dataset. The time stamps of each time series unit in the multi-dimensional time series dataset are extracted. A dynamic time warping algorithm is then activated to align and calibrate the time stamps of different time series units, eliminating time misalignment caused by acquisition time deviations under different operating conditions. After calibration, the time series consistency error of various types of operational data within each time series unit is calculated. The calculation method is the sum of the differences between the acquisition time of various types of data within the same time series unit and the standard time stamp. Data segments are corrected based on the time series consistency error. The correction process gradually adjusts the data correspondence according to the magnitude of the error, ensuring that the time of various types of data within the same time series unit is synchronized, generating a time-uniform intermediate dataset. The various types of operational data in the intermediate dataset are normalized. Using a normalization formula, operational data with different dimensions are converted into the same numerical range. Specifically, the data of connection contact resistance, node temperature, vibration frequency, voltage transmission stability, and mechanical stress are all normalized to the 0-1 range to eliminate the analytical bias caused by differences in dimensions. Finally, a standardized time series dataset is generated. This dataset contains the calibrated time stamps and the corresponding normalized operational data, ensuring that the data format is uniform and logically coherent, laying the foundation for subsequent feature extraction.

[0018] Step S2: Extract the wire harness connection performance feature parameters based on the standardized time series dataset, and construct the wire harness connection parameter correlation matrix; generate a set of wire harness performance correlation coefficients based on the wire harness connection parameter correlation matrix; In this embodiment of the invention, the performance characteristic parameters of the wire harness connection are extracted one by one from the standardized time-series dataset. The extraction process strictly corresponds to various types of operating data in the standardized time-series dataset. The extracted performance characteristic parameters include contact resistance change rate, temperature fluctuation amplitude, vibration frequency peak value, voltage transmission attenuation, and mechanical stress accumulation value. The contact resistance change rate is calculated as the ratio of the difference in contact resistance between adjacent time-series units to the contact resistance of the previous time-series unit. The temperature fluctuation amplitude is the difference between the maximum and minimum node temperatures within the same time-series unit. The vibration frequency peak value is the maximum vibration frequency within the same time-series unit. The voltage transmission attenuation is the difference between the standard voltage and the actual transmission voltage. The mechanical stress accumulation value is the sum of the mechanical stress data within the same time-series unit. All characteristic parameters are organized and recorded according to the time-series unit. Based on the extracted harness connection performance characteristic parameters, the temporal correlation dimension of each characteristic parameter is determined. This temporal correlation dimension is divided according to time-series units, with each dimension corresponding to the characteristic parameter value of a time-series unit. Using matrix construction techniques, a harness connection parameter correlation matrix is ​​built. The rows of the matrix correspond to the time markers of different time-series units, the columns correspond to each performance characteristic parameter, and each matrix element represents the corresponding value of any two characteristic parameters at a given time, fully presenting the numerical correlation of each characteristic parameter at different times. The harness connection parameter correlation matrix is ​​then processed. Matrix operations are used to calculate the degree of correlation between any two characteristic parameters at different times. The degree of correlation is calculated based on the consistency of the numerical change trends of the two characteristic parameters; the more consistent the numerical change trends, the higher the degree of correlation. The results of the correlation degree calculations for all characteristic parameter pairs are compiled to form an initial correlation coefficient set. This initial correlation coefficient set is then normalized using the same normalization formula as in the previous steps, converting the correlation coefficients to the 0-1 interval to eliminate deviations in the correlation coefficients for different characteristic parameter pairs, resulting in a standardized correlation coefficient set. A correlation threshold is set, which is a fixed value. Based on the standardized correlation coefficient set, feature parameter pairs with a correlation degree higher than the set correlation threshold are selected, and feature parameter pairs with a correlation degree lower than the set threshold are removed. All selected feature parameter pairs and their corresponding standardized correlation coefficients are integrated and archived according to the feature parameter pair type to form a performance correlation coefficient set, which provides correlation data support for subsequent wire harness connection performance evolution analysis.

[0019] Step S3: Establish a wire harness connection performance evolution model based on the wire harness performance correlation coefficient set, input historical performance data to iteratively train the wire harness connection performance evolution model, and generate model training parameters; based on the model training parameters, predict the evolution trend of the standardized time series dataset to generate a wire harness performance evolution trend curve. In this embodiment of the invention, by analyzing the correlation strength of each feature parameter pair in the set of wire harness performance correlation coefficients, feature parameter pairs that have a significant impact on the evolution of wire harness connection performance are selected. These feature parameter pairs are determined as influencing factors of wire harness connection performance evolution. The weight allocation rules for each influencing factor are clarified. The weight allocation is set according to the correlation strength of the influencing factors. The higher the correlation strength, the larger the weight ratio. The specific weight value of each influencing factor is calculated using weight calculation technology. Based on the influencing factors and weight allocation rules, the basic framework corresponding to the wire harness connection performance evolution model is built. The framework includes an input layer, a hidden layer, and an output layer. The input layer contains influencing factors and related performance data. The output layer contains the predicted results of wire harness connection performance evolution. The hidden layer is responsible for data calculation and feature mapping. Historical full lifecycle operation data and corresponding performance evolution results of wire harness connection nodes are selected as the model training sample set. The historical full lifecycle operation data includes various operation data such as historical contact resistance, node temperature, and vibration frequency. The performance evolution results include the performance level and evolution trend at different time stages. The model training sample set is divided into a training set and a validation set according to a set ratio. The set ratio is a fixed value. The training set is used for iterative model training, and the validation set is used for accuracy verification after model training. Input the performance correlation coefficient set and historical performance data from the training set into the harness connection performance evolution model, start the gradient descent algorithm, and iteratively optimize the model parameters through the algorithm. During the iteration process, calculate the error value between the model's predicted value and the actual performance evolution result in the training set. The error value is calculated as the sum of the absolute values ​​of the differences between the model's predicted value and the actual performance evolution result. Adjust the weight parameters and correlation factors corresponding to the harness connection performance evolution model according to the error value. The adjustment method is to adjust the weight parameters and correlation factors in reverse according to the error value. The larger the error value, the larger the adjustment. Repeat the above iterative training process until the model prediction error is less than the set error threshold. The set error threshold is a fixed value to ensure that the model prediction accuracy meets the requirements. The trained wire harness connection performance evolution model is validated using a validation set. Historical performance data from the validation set is input into the model to obtain the model prediction results. The model validation accuracy is calculated by dividing the number of samples whose prediction results match the actual performance evolution results in the validation set by the total number of samples in the validation set. When the model validation accuracy reaches a set accuracy threshold (which is a fixed value), model training is stopped. The current weight parameters, correlation factors, and iteration count of the model are extracted, organized and archived according to the model parameter type, forming the model training parameters.From the model training parameters, weight parameters and correlation factors are extracted one by one. The extraction process strictly follows the classification and archiving format of the model training parameters to ensure that the extracted weight parameters and correlation factors are complete and without omissions. The weight parameters correspond to the weight ratio of each influencing factor, and the correlation factors correspond to the correlation relationship of each feature parameter pair. The extracted weight parameters and correlation factors are imported into the trained harness connection performance evolution model. The model parameters are configured according to the model parameter configuration rules. The configuration process strictly follows the parameter setting logic during model training to ensure that the model can perform its predictive function normally and meet the prediction requirements of harness connection performance evolution. The standardized time series dataset is divided into multiple prediction units according to the time series labels in the standardized time series dataset. Each prediction unit contains continuous standardized time series data and corresponding performance feature parameters. The number of continuous time series data is a fixed value. The performance feature parameters include all extracted feature parameters such as contact resistance change rate and temperature fluctuation amplitude. Each prediction unit is numbered in chronological order to ensure the continuity and integrity of the prediction units. The standardized time-series data of each prediction unit are sequentially input into the configured harness connection performance evolution model. The model's built-in computation unit performs prediction calculations, incorporating configured weight parameters and correlation factors. Based on various operational data within the standardized time-series data, the model predicts the harness connection performance parameters corresponding to each unit. These predicted performance parameter values ​​include the predicted values ​​of each characteristic parameter. The results are organized and recorded by prediction unit to form the performance parameter prediction results for each unit. Based on the predicted harness connection performance parameters of each prediction unit, and combined with the corresponding time stamps, the predicted values ​​are organized in chronological order to form a performance parameter prediction time-series sequence. This sequence fully presents the predicted changes in performance parameters at each time point. The performance parameter prediction time-series sequence is then smoothed to eliminate outliers and compensate for the discreteness of the prediction data, resulting in a continuous and stable trend and eliminating analytical biases caused by prediction fluctuations. Based on the smoothed time series, the evolution trend curve of the wire harness connection performance is plotted. The horizontal axis of the curve is the time mark, and the vertical axis is the predicted value of each performance characteristic parameter. Each curve corresponds to a performance characteristic parameter, which fully includes the evolution trend and time correspondence of each performance characteristic parameter. It clearly presents the evolution law of wire harness connection performance over time, and provides intuitive trend support for subsequent anomaly detection and early warning.

[0020] Step S4: Based on the wire harness performance evolution trend curve, perform evolution anomaly assessment and analysis to generate wire harness connection performance anomaly early warning information.

[0021] In this embodiment of the invention, the evolution trajectory of each performance characteristic parameter is extracted one by one from the plotted wire harness connection performance evolution trend curve. The extraction process strictly corresponds to the time nodes and parameter values ​​in the curve to ensure that the evolution trajectory fully reflects the change process of each characteristic parameter over time. The extracted evolution trajectories of each performance characteristic parameter are organized according to time stages to clarify the trajectory change characteristics of each time stage. The rate of change of each characteristic parameter in different time stages is calculated. The rate of change is calculated by dividing the difference between the maximum and minimum values ​​of the characteristic parameter in that time stage by the time length of that time stage. The rate of change of all time stages is organized in chronological order to form a rate of change sequence, which fully presents the temporal change of the rate of change of each characteristic parameter. The Isolation Forest algorithm is activated to detect anomalies in performance evolution trends. By constructing multiple isolated trees, the algorithm analyzes data points in the rate of change sequence in isolation, identifying data points deviating from the normal range as anomalies. All anomalies in the evolution trajectory are identified, and the corresponding time point and performance parameter value for each anomaly are determined. The specific location and time-series features of the anomalies are labeled, ensuring comprehensive and thorough anomaly identification. The performance parameter values ​​corresponding to the anomalies are extracted, and combined with a performance correlation coefficient set, the influencing factors of each anomaly are analyzed. The analysis focuses on the correlation between the performance parameters corresponding to the anomalies and other feature parameters. Based on the correlation strength in the performance correlation coefficient set, the anomaly influencing factors and their degree of influence are determined. The degree of influence is categorized according to the strength of the correlation; the higher the correlation strength, the greater the influence. The anomaly influencing factors and their specific degree of influence for each anomaly are labeled. Based on the anomaly's impact factors and severity, anomaly level assessment results are generated, divided into three fixed levels: Level 1 anomalies correspond to situations with high impact and numerous impact factors; Level 2 anomalies correspond to situations with moderate impact and fewer impact factors; and Level 3 anomalies correspond to situations with low impact and a single impact factor. Different anomaly levels correspond to different warning priorities, with Level 1 having the highest priority, followed by Level 2, and Level 3 the lowest. Based on the anomaly level assessment results and anomaly impact factors, evolution trend adjustment strategies are formulated. These strategies include parameter adjustment schemes and time-series monitoring schemes. The parameter adjustment schemes specify specific parameter adjustment values ​​and timely adjustment methods for the anomaly's impact factors. The time-series monitoring schemes set fixed monitoring cycles and indicators, clearly defining monitoring priorities and time-series nodes. The anomaly level assessment results, anomaly impact factors, and evolution trend adjustment strategies are integrated, organized in a fixed format, and supplemented with anomaly time, anomaly type, impact scope, and adjustment suggestions to form a harness connection performance anomaly warning information. This information clearly presents the anomaly details and response plans, providing guidance for harness connection performance maintenance and optimization, and supporting the stable operation of harness connection performance.

[0022] Furthermore, step S1 includes the following steps: Step S11: Collect various operating data of the wire harness connection nodes under different operating conditions, including connection contact resistance, node temperature, vibration frequency, voltage transmission stability and mechanical stress data; Step S12: Sort various types of operational data according to the data acquisition time sequence, divide the data into segments according to the preset acquisition cycle, and generate a multi-dimensional time series dataset, where each data segment corresponds to a time series unit, and each time series unit contains various types of operational data under the corresponding working conditions; Step S13: Extract the time stamp of each time series unit in the multi-dimensional time series dataset, align and calibrate the time stamps of different time series units through the dynamic time warping algorithm to eliminate time series deviations, calculate the time series consistency error of various types of running data in each time series unit, correct the data segments based on the time series consistency error, and generate a time series unified intermediate dataset. Step S14: Normalize the various types of running data in the intermediate dataset to eliminate dimensional differences and generate a standardized time series dataset, which includes calibrated time series labels and corresponding normalized types of running data.

[0023] In this embodiment of the invention, full-condition data acquisition is carried out on the wiring harness connection nodes. The acquisition process strictly follows fixed acquisition specifications, covering different operating conditions such as idling, acceleration, constant speed, deceleration, and extreme environments. The acquired operating data includes connection contact resistance, node temperature, vibration frequency, voltage transmission stability, and mechanical stress data. The connection contact resistance is recorded in fixed units, the node temperature is collected at fixed intervals, the vibration frequency is recorded as the number of vibrations per unit time, the voltage transmission stability is recorded as the voltage fluctuation difference, and the mechanical stress data is recorded as the stress magnitude borne by the node. After the acquisition is completed, the various operating data are arranged in an orderly manner based on the data acquisition time sequence. The data is divided into segments according to a preset acquisition period, which is set to a fixed duration. Each data segment corresponds to a time sequence unit, and each time sequence unit completely contains the connection contact resistance, node temperature, vibration frequency, voltage transmission stability, and mechanical stress data under the corresponding operating condition. All time sequence units are arranged in chronological order of acquisition time to generate a multi-dimensional time sequence dataset. The time stamps of each time series unit in the multi-dimensional time series dataset are extracted. A dynamic time warping algorithm is then activated to align and calibrate the time stamps of different time series units, eliminating time misalignment caused by acquisition time deviations under different operating conditions. After calibration, the time series consistency error of various types of operational data within each time series unit is calculated. The calculation method is the sum of the differences between the acquisition time of various types of data within the same time series unit and the standard time stamp. Based on the time series consistency error, data segments are corrected. The correction process gradually adjusts the data correspondence according to the magnitude of the error, ensuring that the time of various types of data within the same time series unit is synchronized, generating a time-uniform intermediate dataset. The various types of operational data in the intermediate dataset are normalized using a fixed normalization formula to convert operational data with different dimensions into the same numerical range. Specifically, the data of connection contact resistance, node temperature, vibration frequency, voltage transmission stability, and mechanical stress are all normalized to the 0-1 range to eliminate the analytical bias caused by differences in dimensions. Finally, a standardized time series dataset is generated. This dataset contains the calibrated time stamps and the corresponding normalized operational data, ensuring that the data format is unified and logically consistent, laying the foundation for subsequent feature extraction.

[0024] Furthermore, as an embodiment of the present invention, reference is made to... Figure 2 As shown, Figure 1 A detailed flowchart of step S2 is shown below. In this embodiment, step S2 includes the following steps: Step S21: Extract the wire harness connection performance characteristic parameters from the standardized time series dataset, including the contact resistance change rate, temperature fluctuation amplitude, peak vibration frequency, voltage transmission attenuation, and cumulative mechanical stress. Step S22: Based on the performance characteristic parameters of the harness connection, determine the time correlation dimension of each characteristic parameter, and construct the harness connection parameter correlation matrix, where each matrix element is the corresponding value of any two characteristic parameters at different times; Step S23: Calculate the correlation matrix of the wire harness connection parameters, determine the degree of correlation between each feature parameter, and generate an initial set of correlation coefficients; Step S24: Normalize the initial correlation coefficient set to eliminate coefficient bias and obtain a standardized correlation coefficient set; based on the standardized correlation coefficient set, select feature parameter pairs with a correlation degree higher than the preset correlation threshold, and integrate all feature parameter pairs with the corresponding standardized correlation coefficients to generate a performance correlation coefficient set.

[0025] In this embodiment of the invention, the performance characteristic parameters of the wire harness connection are extracted one by one from the standardized time-series dataset using a fixed extraction method. The extraction process strictly corresponds to various types of operating data in the standardized time-series dataset. The extracted performance characteristic parameters include contact resistance change rate, temperature fluctuation amplitude, vibration frequency peak value, voltage transmission attenuation, and mechanical stress accumulation value. The contact resistance change rate is calculated as the ratio of the difference in contact resistance between adjacent time-series units to the contact resistance of the previous time-series unit. The temperature fluctuation amplitude is the difference between the maximum and minimum node temperatures within the same time-series unit. The vibration frequency peak value is the maximum vibration frequency within the same time-series unit. The voltage transmission attenuation is the difference between the standard voltage and the actual transmission voltage. The mechanical stress accumulation value is the sum of the mechanical stress data within the same time-series unit. All characteristic parameters are organized and recorded according to the time-series unit. Based on the extracted harness connection performance characteristic parameters, the temporal correlation dimension of each characteristic parameter is determined. This temporal correlation dimension is divided according to time-series units, with each dimension corresponding to the characteristic parameter value of a time-series unit. A matrix construction method is used to construct a harness connection parameter correlation matrix. The rows of the matrix correspond to the time markers of different time-series units, and the columns correspond to each performance characteristic parameter. Each matrix element represents the corresponding value of any two characteristic parameters at a given time, fully presenting the numerical correlation of each characteristic parameter at different times. The harness connection parameter correlation matrix is ​​then subjected to fixed calculations. Using matrix operations, the degree of correlation between any two characteristic parameters at different times is calculated. The degree of correlation is calculated based on the consistency of the numerical change trends of the two characteristic parameters; the more consistent the numerical change trends, the higher the degree of correlation. The results of the correlation degree calculations for all characteristic parameter pairs are compiled to generate an initial correlation coefficient set. This initial correlation coefficient set is then normalized using the same normalization formula as in the previous steps, converting the correlation coefficients to the 0-1 interval to eliminate deviations in the correlation coefficients for different characteristic parameter pairs, resulting in a standardized correlation coefficient set. A fixed correlation threshold is preset, and the correlation threshold is set to a fixed value. Based on the standardized correlation coefficient set, feature parameter pairs with a correlation degree higher than the preset correlation threshold are selected, and feature parameter pairs with a correlation degree lower than the preset threshold are removed. All selected feature parameter pairs and their corresponding standardized correlation coefficients are integrated, and the pairs are organized and archived according to the feature parameter pair type to generate a performance correlation coefficient set, which provides correlation data support for subsequent wire harness connection performance evolution analysis.

[0026] Furthermore, step S23 includes the following steps: Extract the feature parameter data of each column in the wire harness connection parameter correlation matrix and use it as the reference sequence, while the feature parameter data of the remaining columns are used as the comparison sequence. Calculate the absolute difference between each comparison sequence and the reference sequence, and determine the maximum and minimum absolute differences at each time point; The correlation coefficient between each comparison sequence and the reference sequence at each time step is calculated based on the maximum and minimum absolute differences. The mean of the correlation coefficients between each comparison sequence and the reference sequence at each time step is calculated to obtain the initial correlation coefficient between each comparison sequence and the reference sequence. Following the steps described above, each feature parameter is used as a reference sequence to complete the correlation coefficient calculation for all feature parameter pairs. All calculation results are then integrated to generate an initial correlation coefficient set. After obtaining the initial correlation coefficient between each comparison sequence and the reference sequence, the process further includes: Extract the feature parameter pairs and time series information corresponding to the initial correlation coefficient, calculate the mean correlation coefficient of each feature parameter pair in different time series stages, and generate the time series stage correlation mean; based on the time series stage correlation mean, analyze the trend of the correlation strength of each feature parameter pair over time, and generate the correlation trend curve. Based on the correlation trend curve, the abrupt change points of correlation strength are extracted, and the time-series nodes and correlation coefficient changes corresponding to the abrupt change points are determined. Based on the correlation coefficient changes corresponding to the abrupt change points, the initial correlation coefficients are corrected to obtain the corrected initial correlation coefficients. The corrected initial correlation coefficients are then classified and integrated according to the feature parameters to update the initial correlation coefficient set.

[0027] In this embodiment of the invention, by extracting the feature parameter data of each column in the correlation matrix of the harness connection parameters, a single column of feature parameter data is used as a reference sequence, and the feature parameter data of the remaining columns are used as comparison sequences. Correlation analysis is performed one by one to ensure that each feature parameter can be used as a reference sequence to compare with all other feature parameters. For each set of reference and comparison sequences, a fixed calculation method is used to calculate the absolute difference between each comparison sequence and the reference sequence at the same time. The absolute difference is calculated as the absolute value of the difference between the comparison sequence value and the reference sequence value at the same time. By traversing all time points, the maximum and minimum absolute differences of the reference and comparison sequences at each time point are determined. The maximum absolute difference is the largest value among all absolute differences at all time points, and the minimum absolute difference is the smallest value among all absolute differences at all time points. Based on the calculated maximum and minimum absolute differences, a fixed correlation coefficient calculation formula is used to calculate the correlation coefficient between each comparison sequence and the reference sequence at each time point. The correlation coefficient calculation combines the absolute difference, the maximum absolute difference, and the minimum absolute difference; the larger the value, the stronger the correlation between the two feature parameters at that time point. The mean correlation coefficient between each comparison sequence and the reference sequence at each time step is calculated. The mean is calculated by dividing the sum of the correlation coefficients at all time steps by the number of time steps, resulting in the initial correlation coefficient between each comparison sequence and the reference sequence. Following the same steps, each feature parameter is used as a reference sequence, and the comparison calculation is repeated until the correlation coefficients for all feature parameter pairs are calculated. All calculation results are then integrated to generate an initial correlation coefficient set. After obtaining the initial correlation coefficients between each comparison sequence and the reference sequence, the feature parameter pairs and their time series information corresponding to the initial correlation coefficients are extracted. According to a fixed time series stage division rule, all time series units are divided into multiple time series stages. The mean correlation coefficient of each feature parameter pair in different time series stages is calculated. The mean is calculated by dividing the sum of the correlation coefficients at all time steps within that time series stage by the number of time steps in that stage, generating the time series stage correlation mean. Based on the time series stage correlation mean, a curve plotting method is used to plot the trend of the correlation strength of each feature parameter pair over time, generating a correlation trend curve. The horizontal axis of the curve represents the time series stage, and the vertical axis represents the mean correlation coefficient, clearly presenting the temporal variation pattern of the correlation strength. Based on the correlation trend curve, a mutation point identification method is used to extract mutation points in correlation strength. The criterion for determining a mutation point is that the difference between the mean correlation coefficients of adjacent time series stages is greater than a fixed difference. The time series node corresponding to the mutation point and the change in correlation coefficient are determined. The change in correlation coefficient is the difference between the mean correlation coefficients of the time series stages before and after the mutation point. Based on the change in correlation coefficient corresponding to the mutation point, the initial correlation coefficient is corrected. The correction method is to add a correction coefficient corresponding to the change in correlation coefficient to the initial correlation coefficient. The correction coefficient is set according to the magnitude of the change in correlation coefficient; the larger the change, the larger the correction coefficient. This yields the corrected initial correlation coefficient.The corrected initial correlation coefficients are classified and integrated according to the feature parameters, and the corresponding values ​​in the original initial correlation coefficient set are replaced to update the initial correlation coefficient set. This ensures that the initial correlation coefficients can accurately reflect the temporal changes of the feature parameters on the correlation strength, and provide accurate correlation data support for subsequent performance evolution analysis.

[0028] Furthermore, the calculation of the correlation coefficient between each comparison sequence and the reference sequence at each time step based on the maximum and minimum absolute differences includes the following steps: Based on the maximum and minimum absolute difference values ​​at each time point, the difference in absolute difference between the reference sequence and the comparison sequence at the same time node is analyzed, and a time difference dataset is constructed. Based on the time-series difference dataset, all time-series nodes are traversed to determine the distribution of extreme differences between each comparison sequence and the corresponding reference sequence at each time point, and global difference characterization data is extracted based on the distribution of extreme differences. Based on the global difference representation data, the single-node difference data in the time-series difference dataset is normalized and mapped to generate single-node difference normalized data. Based on single-node difference normalized data, continuous fitting is performed on all time-series nodes to generate time-series continuous difference characterization data; based on the time-series continuous difference characterization data, the degree of correlation approximation between each comparison sequence and the reference sequence at each time point is inverted, and the correlation coefficient between each comparison sequence and the reference sequence at each time point is generated.

[0029] In this embodiment of the invention, based on the calculated maximum and minimum absolute differences at each time point, a fixed difference analysis method is used to analyze the absolute difference differences between the reference sequence and the comparison sequence at the same time node. The difference analysis focuses on the same time node, examining the degree of deviation between the absolute difference of the comparison sequence and the reference sequence from the maximum and minimum absolute differences, clarifying the magnitude and variation characteristics of the difference at each time node. The absolute difference differences, maximum absolute differences, and minimum absolute differences of all time nodes are organized in chronological order to construct a time-series difference dataset. This dataset fully presents the difference details of each time node, providing basic data for subsequent difference analysis and correlation coefficient calculation. Based on the time-series difference dataset, a fixed traversal method is used to traverse all time-series nodes, recording the absolute difference value of each node. Through extreme value analysis, the distribution of extreme differences between each comparison sequence and its corresponding reference sequence at each time point is determined. This distribution includes the maximum, minimum, and fluctuation range of the difference within each time interval. Based on this distribution, a feature extraction method is used to extract global difference characterization data. This data includes the overall difference mean, fluctuation amplitude, and distribution density, comprehensively reflecting the overall difference characteristics between the reference and comparison sequences. Using this global difference characterization data, a fixed normalization mapping method is employed to normalize the single-node difference data in the time-series difference dataset. The mapping process uses the difference mean and fluctuation amplitude in the global difference characterization data as a benchmark, transforming the single-node difference data to the 0-1 interval to eliminate numerical biases in the difference data from different time-series nodes, generating normalized single-node difference data to ensure the comparability of the difference data across nodes. Based on single-node difference normalized data, a continuous fitting method is employed to perform fitting operations across all time-series nodes. The fitting process follows the chronological order of the time-series nodes. The fitting algorithm compensates for the discreteness of single-node difference data, generating continuous time-series difference characterization data. This data is presented as a continuous curve, fully reflecting the continuous change of differences over time. Based on this continuous time-series difference characterization data, a correlation inversion method is used to invert the degree of correlation approximation between each comparison sequence and the reference sequence at each time point. The inversion logic is that the smaller the value of the continuous time-series difference characterization data, the higher the degree of correlation approximation. By using a fixed inversion formula, the continuous time-series difference characterization data is converted into correlation coefficients, with the correlation coefficient values ​​set to a range of 0-1. Finally, the correlation coefficients between each comparison sequence and the reference sequence at each time point are generated, providing accurate data support for subsequent initial correlation coefficient calculations and performance evolution analysis.

[0030] Furthermore, the step of inverting the correlation approximation degree between each comparison sequence and the reference sequence at each time step based on the temporal continuous difference characterization data includes the following steps: Extract the curve morphology features corresponding to the time-series continuous difference characterization data, and obtain the difference curvature data at each time point based on the curve morphology features; The dynamic fit between each comparison sequence and the reference sequence at each time step is calculated based on the difference curvature data, and the time-series fit data is generated. Based on the temporal variation characteristics of the reference sequence matched with the time-fit data, a fit correction factor is generated. The fitting data at different times is dynamically calibrated based on the fitting correction factor to generate calibrated fitting data; the degree of correlation between each comparison sequence and the reference sequence at each time point is inverted based on the calibrated fitting data.

[0031] In this embodiment of the invention, curve morphological features corresponding to the temporal continuous difference characterization data are extracted. A fixed feature extraction method is used to extract morphological features such as slope changes, curvature, and extreme point distribution of the curves one by one. These features directly reflect the trend of difference changes over time. Based on the extracted curve morphological features, a curvature calculation method is used to calculate the difference curvature data at each time point. The curvature calculation method is to calculate the curvature value of the curve at that time point. The larger the curvature value, the more obvious the curvature of the curve at that time point, and the more drastic the difference change. The difference curvature data at all times are organized in temporal order to form a complete difference curvature dataset. Based on the difference curvature data, a fixed fit calculation method is used to calculate the dynamic fit between each comparison sequence and the reference sequence at each time point. The dynamic fit calculation combines the difference curvature data and single-node difference normalized data. The smaller the difference curvature and the smaller the single-node difference normalized data, the higher the dynamic fit. The fit value range is set to 0-1, generating time-based fit data. This data fully presents the fit between the reference sequence and the comparison sequence at each time point. Based on the time-fit data, a fixed matching method is used to match the temporal variation characteristics of the reference sequence. These characteristics include the trend, rate of change, and frequency of fluctuation. The time-fit data is compared one by one with the temporal variation characteristics of the reference sequence. A fit correction factor is generated based on the comparison results. The correction factor value is set according to the degree of matching between the fit and the reference sequence's variation characteristics; the higher the matching degree, the closer the correction factor is to 1, and vice versa. This corrects for deviations in the time-fit data. Based on the fit correction factor, a dynamic calibration method is used to dynamically calibrate the time-fit data time-by-time. The calibration method involves multiplying the time-fit data by the corresponding fit correction factor to eliminate deviations caused by the mismatch between the fit data and the reference sequence's variation characteristics, generating calibrated fit data. Based on the calibrated fit data, the correlation inversion method is used to invert the degree of correlation approximation between each comparison sequence and the reference sequence at each time point. The inversion logic is that the larger the value of the calibrated fit data, the higher the degree of correlation approximation. The calibrated fit data is directly mapped to the value of correlation approximation, with the value range set to 0-1. This fully presents the degree of correlation between the reference sequence and the comparison sequence at each time point, providing support for subsequent correlation coefficient generation and performance evolution analysis.

[0032] Furthermore, step S3, which involves establishing a wire harness connection performance evolution model based on the set of wire harness performance correlation coefficients and iteratively training the model using historical performance data, includes the following steps: Based on the set of correlation coefficients for wire harness performance, the influencing factors of wire harness connection performance evolution are determined, the weight allocation rules of each influencing factor are clarified, and the basic framework corresponding to the wire harness connection performance evolution model is constructed. Historical full lifecycle operation data and corresponding performance evolution results of the wire harness connection nodes are selected as the model training sample set, and the model training sample set is divided into training set and validation set according to a preset ratio; The performance correlation coefficient set and historical performance data in the training set are input into the wire harness to connect the performance evolution model, and the model parameters are iteratively optimized by the gradient descent algorithm to calculate the error between the model prediction value and the actual performance evolution result. Adjust the weight parameters and correlation factors corresponding to the wire harness connection performance evolution model based on the error value, and repeat the training process iteratively until the model prediction error is less than the preset error threshold. Use the validation set to validate the trained wire harness connection performance evolution model, calculate the model validation accuracy, and stop model training when the model validation accuracy reaches the preset accuracy threshold. Extract the current weight parameters, correlation factors and iteration number of the model to generate model training parameters.

[0033] In this embodiment of the invention, based on the set of correlation coefficients for wire harness performance, the correlation strength of each feature parameter pair in the set of performance correlation coefficients is analyzed one by one. Feature parameter pairs that have a significant impact on the evolution of wire harness connection performance are screened out, and these feature parameter pairs are determined as the influencing factors of wire harness connection performance evolution. The weight allocation rules for each influencing factor are clarified. The weight allocation is set according to the correlation strength of the influencing factors. The higher the correlation strength, the larger the weight ratio. A fixed weight calculation method is used to calculate the specific weight value of each influencing factor. Based on the influencing factors and the weight allocation rules, the basic framework corresponding to the wire harness connection performance evolution model is constructed. The framework includes an input layer, a hidden layer, and an output layer. The input layer is the influencing factors and related performance data, the output layer is the prediction result of wire harness connection performance evolution, and the hidden layer is responsible for data operation and feature mapping. Historical full lifecycle operation data and corresponding performance evolution results of wire harness connection nodes are selected as the model training sample set. The historical full lifecycle operation data includes various operation data such as historical contact resistance, node temperature, and vibration frequency. The performance evolution results include the performance level and evolution trend at different time stages. A fixed partitioning ratio is adopted to divide the model training sample set into a training set and a validation set according to a preset ratio. The preset partitioning ratio is set to a fixed value. The training set is used for iterative model training, and the validation set is used for accuracy verification after model training. Input the performance correlation coefficient set and historical performance data from the training set into the harness connection performance evolution model, start the gradient descent algorithm, and iteratively optimize the model parameters through the algorithm. During the iteration process, calculate the error value between the model's predicted value and the actual performance evolution result in the training set. The error value is calculated as the sum of the absolute values ​​of the differences between the model's predicted value and the actual performance evolution result. Based on the error value, adjust the weight parameters and correlation factors corresponding to the harness connection performance evolution model. The adjustment method is to adjust the weight parameters and correlation factors in reverse according to the magnitude of the error value. The larger the error value, the larger the adjustment magnitude. Repeat the above iterative training process until the model prediction error is less than the preset error threshold. The preset error threshold is set to a fixed value to ensure that the model prediction accuracy meets the requirements. The trained harness connection performance evolution model is validated using a validation set. Historical performance data from the validation set is input into the model to obtain prediction results. The model validation accuracy is calculated by dividing the number of samples whose prediction results match the actual performance evolution results in the validation set by the total number of samples in the validation set. When the model validation accuracy reaches a preset accuracy threshold (which is then set to a fixed value), model training is stopped. The current weight parameters, correlation factors, and iteration count of the model are extracted, organized and archived according to model parameter types, and model training parameters are generated to provide model support for subsequent harness connection performance evolution prediction and analysis.

[0034] Furthermore, the evolutionary trend prediction of the standardized time-series dataset based on model training parameters includes the following steps: Extract the weight parameters and correlation factors from the model training parameters, import them into the trained harness connection performance evolution model, and complete the model parameter configuration; The standardized time-series dataset is divided into multiple prediction units according to time order. Each prediction unit contains continuous standardized time-series data and corresponding performance characteristic parameters. The standardized time-series data of each prediction unit is input into the configured harness connection performance evolution model to predict the predicted value of the harness connection performance parameter corresponding to that unit. Based on the predicted values ​​of the wire harness connection performance parameters of each prediction unit, and combined with the corresponding time stamps, a time series sequence of performance parameter predictions is generated. The time series sequence of performance parameter predictions is smoothed to eliminate prediction fluctuations. Based on the smoothed time series sequence, a wire harness connection performance evolution trend curve is plotted. This curve contains the evolution trend of each performance characteristic parameter and its time correspondence.

[0035] In this embodiment of the invention, weight parameters and correlation factors are extracted one by one from the model training parameters. The extraction process strictly follows the classification and archiving format of the model training parameters to ensure that the extracted weight parameters and correlation factors are complete and without omissions. The weight parameters correspond to the weight ratio of each influencing factor, and the correlation factors correspond to the correlation relationship of each feature parameter pair. The extracted weight parameters and correlation factors are imported into the trained wire harness connection performance evolution model. According to the model parameter configuration rules, the model parameters are accurately configured. The configuration process strictly follows the parameter setting logic during model training to ensure that the model can perform its prediction function normally and meet the prediction requirements of wire harness connection performance evolution. A fixed partitioning rule is adopted to divide the generated standardized time series dataset into multiple prediction units according to the time sequence. The partitioning is based on the time series labels in the standardized time series dataset. Each prediction unit contains continuous standardized time series data and corresponding performance feature parameters. The number of continuous time series data is set to a fixed value. The performance feature parameters include all extracted feature parameters such as contact resistance change rate and temperature fluctuation amplitude. Each prediction unit is numbered in time sequence to ensure the continuity and integrity of the prediction units. The standardized time-series data of each prediction unit are sequentially input into the configured harness connection performance evolution model. The model's built-in computation unit performs prediction calculations, incorporating configured weight parameters and correlation factors. Based on various operational data within the standardized time-series data, the model predicts the harness connection performance parameters corresponding to each unit. These predicted performance parameter values ​​include the predicted values ​​of each characteristic parameter. The results are organized and recorded by prediction unit to generate the performance parameter prediction results for each unit. Based on the predicted harness connection performance parameters of each prediction unit, and combined with the corresponding time stamps, a time-series arrangement method is used to organize the predicted values ​​according to the order of the time stamps, generating a performance parameter prediction time-series sequence. This sequence fully presents the predicted changes in performance parameters at each time point. A fixed smoothing method is used to smooth the performance parameter prediction time-series sequence. This process eliminates outliers in the sequence, compensates for the discreteness of the prediction data, and ensures that the time-series sequence exhibits a continuous and stable trend, eliminating analytical biases caused by prediction fluctuations. Based on the smoothed time series, a curve plotting method is used to plot the evolution trend curve of the wire harness connection performance. The horizontal axis of the curve is the time mark, and the vertical axis is the predicted value of each performance characteristic parameter. Each curve corresponds to a performance characteristic parameter, which fully includes the evolution trend and time correspondence of each performance characteristic parameter. It clearly presents the evolution law of wire harness connection performance over time, and provides intuitive trend support for subsequent anomaly detection and early warning.

[0036] Furthermore, step S4 includes the following steps: Extract the evolution trajectory of each performance characteristic parameter in the trend curve of wire harness connection performance, calculate the rate of change of each characteristic parameter in different time stages, and generate a rate of change sequence; Based on the rate of change sequence, the isolated forest algorithm is used to detect anomalies in the performance evolution trend, identify anomalies in the evolution trajectory, and determine the time nodes and performance parameter values ​​corresponding to the anomalies. The performance parameter values ​​corresponding to the anomalies are extracted, and combined with the performance correlation coefficient set, the associated influencing factors of the anomalies are analyzed to determine the anomaly influencing factors and their degree of influence. Based on the abnormal impact factors and their degree of impact, anomaly level assessment results are generated, which include different warning priorities corresponding to different anomaly levels; based on the anomaly level assessment results and abnormal impact factors, evolution trend adjustment strategies are formulated, including parameter adjustment schemes and time series monitoring schemes; integrating the anomaly level assessment results, abnormal impact factors, and evolution trend adjustment strategies, anomaly warning information for harness connection performance is generated, which includes the anomaly time, anomaly type, scope of impact, and adjustment suggestions.

[0037] In this embodiment of the invention, the evolution trajectory of each performance characteristic parameter is extracted one by one from the plotted wire harness connection performance evolution trend curve. The extraction process strictly corresponds to the time nodes and parameter values ​​in the curve to ensure that the evolution trajectory fully reflects the change process of each characteristic parameter over time. The extracted evolution trajectories of each performance characteristic parameter are organized according to time stages to clarify the trajectory change characteristics of each time stage. A fixed calculation method is used to calculate the change rate of each characteristic parameter in different time stages. The change rate is calculated by dividing the difference between the maximum and minimum values ​​of the characteristic parameter in that time stage by the time length of that time stage. The change rates of all time stages are organized in chronological order to generate a change rate sequence, which fully presents the temporal change of the change rate of each characteristic parameter. Based on the rate of change sequence, the Isolation Forest algorithm is initiated to detect anomalies in performance evolution trends. The algorithm constructs multiple isolated trees to analyze data points in the rate of change sequence in isolation, identifying data points deviating from the normal range as anomalies. All anomalies in the evolution trajectory are identified, and the corresponding time node and performance parameter value for each anomaly are determined. The specific location and time-series features of the anomaly are labeled to ensure comprehensive and thorough anomaly identification. The performance parameter values ​​corresponding to the anomalies are extracted, and combined with a performance correlation coefficient set, anomaly analysis methods are used to analyze the associated influencing factors of each anomaly. The analysis focuses on the correlation between the performance parameters corresponding to the anomaly and other feature parameters. Based on the correlation strength in the performance correlation coefficient set, the anomaly influencing factors and their degree of influence are determined. The degree of influence is categorized according to the correlation strength; the higher the correlation strength, the greater the influence. The anomaly influencing factors and their specific degree of influence for each anomaly are labeled. Based on the anomaly impact factors and their severity, a fixed-level assessment rule is adopted to generate anomaly level assessment results. Anomalies are divided into three fixed levels: Level 1 anomalies correspond to situations with high impact and many impact factors; Level 2 anomalies correspond to situations with medium impact and few impact factors; and Level 3 anomalies correspond to situations with low impact and a single impact factor. Different anomaly levels correspond to different warning priorities, with Level 1 having the highest priority, followed by Level 2, and Level 3 the lowest. Based on the anomaly level assessment results and anomaly impact factors, a strategy formulation method is used to develop evolution trend adjustment strategies. These strategies include parameter adjustment schemes and time-series monitoring schemes. The parameter adjustment schemes specify specific parameter adjustment values ​​and timely adjustment methods for the anomaly impact factors. The time-series monitoring schemes set fixed monitoring cycles and indicators, clearly defining monitoring priorities and time-series nodes. The anomaly level assessment results, anomaly impact factors, and evolution trend adjustment strategies are integrated, organized in a fixed format, and supplemented with anomaly time, anomaly type, impact scope, and adjustment suggestions to generate anomaly warning information for harness connection performance. This information clearly presents the anomaly details and response solutions, providing precise guidance for harness connection performance maintenance and optimization, and supporting the stable operation of harness connection performance.

[0038] Furthermore, the present invention also provides a wire harness connection performance evolution analysis system, including a processor, a memory, and a computer program stored in the memory and executable on the processor, for performing the wire harness connection performance evolution analysis method as described above.

[0039] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for analyzing the evolution of wire harness connection performance, characterized in that, Includes the following steps: Step S1: Obtain various types of operational data corresponding to the wire harness connection nodes, and generate a multi-dimensional time series dataset based on the data acquisition time sequence; Perform time series calibration on multi-dimensional time series datasets to generate standardized time series datasets; Step S2: Extract the performance characteristic parameters of the harness connection based on the standardized time series dataset, and construct the correlation matrix of the harness connection parameters; Generate a set of wire harness performance correlation coefficients based on the correlation matrix of wire harness connection parameters; wherein, step S2 includes the following steps: Step S21: Extract the wire harness connection performance characteristic parameters from the standardized time series dataset, including the contact resistance change rate, temperature fluctuation amplitude, vibration frequency peak value, voltage transmission attenuation, and mechanical stress accumulation value. Step S22: Based on the performance characteristic parameters of the harness connection, determine the time correlation dimension of each characteristic parameter, and construct the harness connection parameter correlation matrix, where each matrix element is the corresponding value of any two characteristic parameters at different times; Step S23: Calculate the correlation matrix of the wire harness connection parameters to determine the degree of correlation between each feature parameter and generate an initial correlation coefficient set; wherein, step S23 includes the following steps: Extract the feature parameter data of each column in the wire harness connection parameter correlation matrix and use it as the reference sequence, while the feature parameter data of the remaining columns are used as the comparison sequence. Calculate the absolute difference between each comparison sequence and the reference sequence, and determine the maximum and minimum absolute differences at each time point; The correlation coefficient between each comparison sequence and the reference sequence at each time step is calculated based on the maximum and minimum absolute differences; this includes the following steps: Based on the maximum and minimum absolute difference values ​​at each time point, the difference in absolute difference between the reference sequence and the comparison sequence at the same time node is analyzed, and a time difference dataset is constructed. Based on the time-series difference dataset, all time-series nodes are traversed to determine the distribution of extreme differences between each comparison sequence and the corresponding reference sequence at each time point, and global difference characterization data is extracted based on the distribution of extreme differences. Based on the global difference representation data, the single-node difference data in the time-series difference dataset is normalized and mapped to generate single-node difference normalized data. Based on single-node difference normalized data, continuous fitting is performed across all time-series nodes to generate time-series continuous difference characterization data; based on the time-series continuous difference characterization data, the degree of correlation approximation between each comparison sequence and the reference sequence at each time step is retrieved, generating the correlation coefficient between each comparison sequence and the reference sequence at each time step; wherein, the step of retrieving the degree of correlation approximation between each comparison sequence and the reference sequence at each time step based on the time-series continuous difference characterization data includes the following steps: Extract the curve morphology features corresponding to the continuous time-series difference representation data, and obtain the difference curvature data at each time point based on the curve morphology features; in particular, the feature extraction method is used to extract the slope change, curvature degree, and extreme point distribution morphology features of the curve one by one; based on the extracted curve morphology features, the curvature calculation method is used to calculate the difference curvature data at each time point, and the curvature calculation method is the curvature degree value of the curve at that time point. The dynamic fit between each comparison sequence and the reference sequence at each time step is calculated based on the difference curvature data, generating time-based fit data. The fit calculation method is used to calculate the dynamic fit between each comparison sequence and the reference sequence at each time step. The dynamic fit calculation combines the difference curvature data and the single-node difference normalized data. The smaller the difference curvature and the smaller the single-node difference normalized data, the higher the dynamic fit. The fit value range is set to 0-1. A fitting correction factor is generated by matching the temporal variation characteristics of the reference sequence with the time-fit data. Specifically, a matching method is used to match the temporal variation characteristics of the reference sequence, which include the trend, rate of change, and frequency of fluctuation. The time-fit data is compared with the temporal variation characteristics of the reference sequence one by one, and the fitting correction factor is generated based on the comparison results. The value of the correction factor is set according to the degree of matching between the fitting and the variation characteristics of the reference sequence. Dynamic calibration of the time-bound fit data is performed based on a fit correction factor to generate calibrated fit data. The correlation approximation between each comparison sequence and the reference sequence at each time step is then retrieved based on the calibrated fit data. Specifically, a dynamic calibration method is used to dynamically calibrate the time-bound fit data moment-by-moment. The calibration method involves multiplying the time-bound fit data by the corresponding fit correction factor to eliminate the bias caused by the mismatch between the fit data and the reference sequence's variation characteristics, thus generating calibrated fit data. Based on the calibrated fit data, a correlation inversion method is used to retrieve the correlation approximation between each comparison sequence and the reference sequence at each time step. The inversion logic is that the larger the value of the calibrated fit data, the higher the correlation approximation. The mean of the correlation coefficients between each comparison sequence and the reference sequence at each time step is calculated to obtain the initial correlation coefficient between each comparison sequence and the reference sequence. Following the steps described above, each feature parameter is used as a reference sequence to complete the correlation coefficient calculation for all feature parameter pairs. All calculation results are then integrated to generate an initial correlation coefficient set. After obtaining the initial correlation coefficient between each comparison sequence and the reference sequence, the process further includes: Extract the feature parameter pairs and time series information corresponding to the initial correlation coefficient, calculate the mean correlation coefficient of each feature parameter pair in different time series stages, and generate the time series stage correlation mean; based on the time series stage correlation mean, analyze the trend of the correlation strength of each feature parameter pair over time, and generate the correlation trend curve. Based on the correlation trend curve, extract the abrupt change points of correlation strength, determine the time-series nodes and the change in correlation coefficients corresponding to the abrupt change points; based on the change in correlation coefficients corresponding to the abrupt change points, correct the initial correlation coefficients to obtain the corrected initial correlation coefficients; classify and integrate the corrected initial correlation coefficients according to the feature parameters, and update the initial correlation coefficient set. Step S24: Normalize the initial correlation coefficient set to eliminate coefficient bias and obtain a standardized correlation coefficient set; based on the standardized correlation coefficient set, select feature parameter pairs with a correlation degree higher than the preset correlation threshold, and integrate all feature parameter pairs with the corresponding standardized correlation coefficients to generate a performance correlation coefficient set; Step S3: Establish a wire harness connection performance evolution model based on the wire harness performance correlation coefficient set, input historical performance data to iteratively train the wire harness connection performance evolution model, and generate model training parameters; based on the model training parameters, predict the evolution trend of the standardized time series dataset to generate a wire harness performance evolution trend curve. Step S4: Based on the wire harness performance evolution trend curve, perform evolution anomaly assessment and analysis to generate wire harness connection performance anomaly early warning information.

2. The method for analyzing the evolution of wire harness connection performance according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Collect various operating data of the wire harness connection nodes under different operating conditions, including connection contact resistance, node temperature, vibration frequency, voltage transmission stability and mechanical stress data; Step S12: Sort various types of operational data according to the data acquisition time sequence, divide the data into segments according to the preset acquisition cycle, and generate a multi-dimensional time series dataset, where each data segment corresponds to a time series unit, and each time series unit contains various types of operational data under the corresponding working conditions; Step S13: Extract the time stamp of each time series unit in the multi-dimensional time series dataset, align and calibrate the time stamps of different time series units through the dynamic time warping algorithm to eliminate time series deviations, calculate the time series consistency error of various types of running data in each time series unit, correct the data segments based on the time series consistency error, and generate a time series unified intermediate dataset. Step S14: Normalize the various types of running data in the intermediate dataset to eliminate dimensional differences and generate a standardized time series dataset, which includes calibrated time series labels and corresponding normalized types of running data.

3. The method for analyzing the evolution of wire harness connection performance according to claim 1, characterized in that, Step S3, which involves establishing a wire harness connection performance evolution model based on a set of wire harness performance correlation coefficients and iteratively training the model using historical performance data, includes the following steps: Based on the set of correlation coefficients for wire harness performance, the influencing factors of wire harness connection performance evolution are determined, the weight allocation rules of each influencing factor are clarified, and the basic framework corresponding to the wire harness connection performance evolution model is constructed. Historical full lifecycle operation data and corresponding performance evolution results of the wire harness connection nodes are selected as the model training sample set, and the model training sample set is divided into training set and validation set according to a preset ratio; The performance correlation coefficient set and historical performance data in the training set are input into the wire harness to connect the performance evolution model, and the model parameters are iteratively optimized by the gradient descent algorithm to calculate the error between the model prediction value and the actual performance evolution result. Adjust the weight parameters and correlation factors corresponding to the wire harness connection performance evolution model based on the error value, and repeat the training process iteratively until the model prediction error is less than the preset error threshold. Use the validation set to validate the trained wire harness connection performance evolution model, calculate the model validation accuracy, and stop model training when the model validation accuracy reaches the preset accuracy threshold. Extract the current weight parameters, correlation factors and iteration number of the model to generate model training parameters.

4. The method for analyzing the evolution of wire harness connection performance according to claim 3, characterized in that, The method of predicting the evolutionary trend of a standardized time-series dataset based on model training parameters includes the following steps: Extract the weight parameters and correlation factors from the model training parameters, import them into the trained harness connection performance evolution model, and complete the model parameter configuration; The standardized time-series dataset is divided into multiple prediction units according to time order. Each prediction unit contains continuous standardized time-series data and corresponding performance characteristic parameters. The standardized time-series data of each prediction unit is input into the configured harness connection performance evolution model to predict the predicted value of the harness connection performance parameter corresponding to that unit. Based on the predicted values ​​of the wire harness connection performance parameters of each prediction unit, and combined with the corresponding time stamps, a time series sequence of performance parameter predictions is generated. The time series sequence of performance parameter predictions is smoothed to eliminate prediction fluctuations. Based on the smoothed time series sequence, a wire harness connection performance evolution trend curve is plotted. This curve contains the evolution trend of each performance characteristic parameter and its time correspondence.

5. The method for analyzing the evolution of wire harness connection performance according to claim 1, characterized in that, Step S4 includes the following steps: Extract the evolution trajectory of each performance characteristic parameter in the trend curve of wire harness connection performance, calculate the rate of change of each characteristic parameter in different time stages, and generate a rate of change sequence; Based on the rate of change sequence, the isolated forest algorithm is used to detect anomalies in the performance evolution trend, identify anomalies in the evolution trajectory, and determine the time nodes and performance parameter values ​​corresponding to the anomalies. The performance parameter values ​​corresponding to the anomalies are extracted, and combined with the performance correlation coefficient set, the associated influencing factors of the anomalies are analyzed to determine the anomaly influencing factors and their degree of influence. Based on the abnormal impact factors and their degree of impact, anomaly level assessment results are generated, which include different warning priorities corresponding to different anomaly levels; based on the anomaly level assessment results and abnormal impact factors, evolution trend adjustment strategies are formulated, including parameter adjustment schemes and time series monitoring schemes; integrating the anomaly level assessment results, abnormal impact factors, and evolution trend adjustment strategies, anomaly warning information for harness connection performance is generated, which includes the anomaly time, anomaly type, scope of impact, and adjustment suggestions.

6. A wire harness connection performance evolution analysis system, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor, for performing the harness connection performance evolution analysis method as described in any one of claims 1-5.

Images