A new energy vehicle charging cable aging performance testing method and system

Abstract

The present invention discloses a testing method and system for the aging performance of a charging cable for a new energy vehicle, and relates to the technical field of cable aging detection. When the system is running, the comprehensive use of an environment acquisition module, a data processing module, a performance test module, and an evaluation module realizes accurate evaluation and prediction of the cable aging performance. The environment acquisition module collects cable test data under temperature control, humidity control, and water immersion conditions in real time, and the data processing module cleans, standardizes, and extracts features from these data to form a high-quality feature data set X. The performance test module uses random forest, gradient boosting tree, and deep neural network models for multi-model prediction, and fits through a meta-model to obtain the prediction results of the Stacking model, which are matched with a preset evaluation threshold L to generate a cable aging performance test evaluation scheme, solving the problems of slow response, low prediction accuracy, and poor adaptability of traditional methods.

Description

Technical Field

[0001] The present invention relates to the technical field of cable aging detection, and in particular to a method and system for testing the aging performance of a charging cable for a new energy vehicle. Background Art

[0002] The rapid development of new energy vehicles has promoted the construction of charging infrastructure. Among them, charging cables are key components connecting power sources and vehicles, and their performance and safety are particularly important. In order to ensure the reliability and durability of charging cables under various extreme environmental conditions, systematic aging performance testing has become a key link.

[0003] However, there are some shortcomings and deficiencies in the current aging performance testing of charging cables. Traditional testing methods mainly rely on static and single environmental conditions testing, which can hardly fully reflect the complex stresses and environmental changes of cables in actual use. For example, tests conducted only at room temperature cannot accurately predict the performance changes of cables in high or low temperature environments. Similarly, the aging behavior of cables under extreme conditions related to high humidity and salt water immersion cannot be fully evaluated through a single laboratory test. These limitations result in the inability of existing testing methods to provide sufficient data support, affecting the optimization of cable design and material selection.

[0004] These current situations and shortcomings are mainly due to the insufficient accuracy of traditional test systems in simulating environmental conditions, which cannot accurately simulate and control a variety of extreme environmental conditions. This lack of simulation accuracy has brought about a series of abnormal effects, such as large errors in test results, and the inability to accurately predict the aging behavior of cables in actual use; delayed response to environmental stress, resulting in inaccurate aging performance evaluation; and poor test repeatability, making it difficult to conduct reliable comparative analysis. The above problems not only increase the uncertainty in the design and manufacturing process, but may also lead to safety hazards in actual cable applications. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a method and system for testing the aging performance of a new energy vehicle charging cable, which solves the problems mentioned in the background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A testing system for the aging performance of a new energy vehicle charging cable, comprising an environment acquisition module, a data processing module, a performance testing module and an evaluation module;

[0007] The environment acquisition module collects environmental test data in real time through cables for the preset environmental conditioning tests of the temperature control device, the humidity control device and the immersion device to form a test data group;

[0008] The data processing module preprocesses the test data group, including data cleaning, data standardization and feature extraction, to form a feature extraction data set X;

[0009] The performance testing module builds multiple models for the data set X, including a random forest model, a gradient boosted tree model, and a deep neural network model, and performs an integrated prediction on the random forest model. Decision tree prediction results And the prediction results of deep neural network Use the meta-model to perform fitting predictions and obtain the prediction results of the Stacking model

[0010] The evaluation module evaluates the threshold L of the preset cable performance aging test and the prediction result of the Stacking model. Matching is carried out to obtain the cable aging performance test evaluation plan, and specific display and notification are made according to the content of the cable aging performance test evaluation plan.

[0011] Preferably, the environment acquisition module includes a temperature control unit, a humidity control unit and a water immersion unit;

[0012] The temperature control unit controls the temperature of the cable test environment by using high-precision temperature control equipment, including an environmental chamber and a thermal cycle system, and then simulates extreme high and low temperature environmental conditions through the equipment, and adjusts and maintains the temperature in real time through a PID control algorithm to ensure that the temperature fluctuation range is within ±0.1°C of the set value;

[0013] The humidity control unit adjusts and controls the humidity state of the test cable environment by using humidifier and dehumidifier equipment, and then simulates a special cable use environment through the equipment, including high humidity on rainy days and low humidity on hot days;

[0014] The immersion unit simulates the immersion conditions of the cable test in different liquids, including salt water conditions and acid-base water conditions, by using an immersion tank and a liquid circulation device;

[0015] The test data set is constructed by recording parameters related to cable aging, including resistance change rate, insulation resistance, temperature, humidity, current, water immersion conditions and mechanical stress information.

[0016] Preferably, the data processing module includes a preprocessing unit and an extraction unit;

[0017] The preprocessing unit preprocesses the data information of the test data group, including data cleaning and data standardization, including temperature, humidity, water immersion conditions, resistance and insulation resistance related test data;

[0018] Among them, data cleaning includes identifying and deleting outliers, noise, and filling missing values; data standardization includes converting data of different dimensions to the same dimension to ensure data consistency;

[0019] The extraction unit selects characteristic information of relevant parameters according to the cable aging performance related parameters for extraction, including resistance change rate, insulation resistance, temperature, humidity, current, water immersion condition and mechanical stress, and simultaneously calculates the extracted characteristic information to form a characteristic extraction data set X;

[0020] The feature extraction data set X is specifically X{R t , R ins , T, H, I, W, F, ΔR t , △R ins}, where R t Represents the resistance change rate, R ins Indicates insulation resistance value, T indicates temperature value, H indicates humidity value, I indicates current value, W indicates water immersion condition sequence value, F indicates mechanical stress value, ΔR t Indicates the fluctuation value of the resistance change rate, △R ins Indicates the fluctuation value of insulation resistance.

[0021] Preferably, the performance testing module includes a modeling unit and a model integration unit;

[0022] The modeling unit establishes a random forest model for the feature extraction data set X, and constructs a decision tree model by splitting the Gini impurity feature of the feature extraction data set X with the smallest Gini impurity. At the same time, the random forest integrates the prediction results of multiple decision trees to improve the generalization ability and prediction accuracy of the model and obtain the integrated prediction result of the random forest. Use the feature extraction data set X to build a gradient boosting tree model, optimize the residual by iteratively training multiple decision trees, and obtain the prediction results of the decision tree after updating the residual. Use the feature extraction data set X deep neural network model, train it through forward propagation and back propagation algorithms to gradually update the model parameters, and obtain the prediction results of the deep neural network after minimizing the loss function

[0023] The model integration unit uses the stacking method to integrate the random forest prediction results Decision tree prediction results And the prediction results of deep neural network Perform integration, perform fitting prediction through the meta-model, and obtain the prediction results of the Stacking model

[0024] Preferably, the ensemble prediction result of the random forest is Obtained through the following calculation formula:

[0025]

[0026] In the formula, represents the ensemble prediction result of random forest, T RF represents the number of decision trees, Represents the prediction result of the tth decision tree;

[0027] The prediction result of the tth decision tree Obtained through the following calculation formula:

[0028]

[0029] In the formula, f t,RF represents the prediction function of the t-th decision tree, and X represents the feature extraction data set;

[0030] The Gini impurity is split by The formula is calculated, where G ini (X) represents the Gini impurity of the feature extraction dataset X, K represents the number of sample categories, P k Indicates the proportion of samples belonging to the kth class.

[0031] Preferably, the residual update is optimized by the following calculation formula:

[0032] R i,m =Y i -F m-1 *(X i );

[0033] In the formula, R i,m represents the residual of the i-th sample at the m-th iteration, Y i represents the true value of the i-th sample, F m-1 *(X i ) represents the predicted value of the feature extraction data set X of the model input for the first m-1 iterations, and the prediction error of the current model on the i-th sample is represented by the residual update calculation formula, and the prediction error of the decision tree model is corrected by fitting the residual;

[0034] The prediction results of the decision tree Obtained through the following calculation formula:

[0035]

[0036] In the formula, Represents the prediction result of the mth decision tree h mrepresents the prediction function of the mth decision tree, and X represents the feature extraction dataset.

[0037] Preferably, the prediction result of the deep neural network Obtained through the following calculation formula:

[0038]

[0039] In the formula, represents the prediction result of the deep neural network, f DNN represents the prediction function of the deep neural network, and X represents the feature extraction dataset.

[0040] Preferably, the prediction result of the Stacking model is Obtained through the following calculation formula:

[0041]

[0042] In the formula, represents the prediction result of the Stacking model, g represents the meta-model, including linear regression and DNN, represents the ensemble prediction result of random forest, Represents the prediction result of the decision tree, Represents the prediction result of the deep neural network.

[0043] Preferably, the evaluation module includes a matching unit and an execution unit;

[0044] The matching unit matches the required comparison value through the preset relevant information, including the preset cable performance aging test evaluation threshold L and the prediction result of the Stacking model. Match and obtain the cable aging performance test and evaluation plan;

[0045] The cable aging performance test evaluation scheme is obtained by the following matching method:

[0046] Prediction results of the Stacking model Less than the cable performance aging test assessment threshold L, the cable performance aging test assessment is qualified;

[0047] Prediction results of the Stacking model If it is greater than or equal to the cable performance aging test evaluation threshold L, the cable performance aging test evaluation fails;

[0048] The execution unit performs specific display and notification according to the content of the cable aging performance test evaluation plan. The display includes dashboard display, test report generation and pop-up reminder, and the notification includes test report email notification, SMS notification and mobile application notification.

[0049] A method for testing the aging performance of a new energy vehicle charging cable comprises the following steps:

[0050] Step 1: The environment collection module collects environmental test data in real time through the cables of the preset environmental conditioning tests of the temperature control equipment, the humidity control equipment and the immersion equipment to form a test data group;

[0051] Step 2: The data processing module preprocesses the test data group, including data cleaning, data standardization and feature extraction, to form a feature extraction data set X;

[0052] Step 3: The performance testing module builds multiple models for data set X, including random forest model, gradient boosting tree model and deep neural network model, and integrates the prediction results of random forest. Decision tree prediction results And the prediction results of deep neural network Use the meta-model to perform fitting predictions and obtain the prediction results of the Stacking model

[0053] Step 4: The evaluation module evaluates the threshold L and the prediction results of the Stacking model through the preset cable performance aging test Matching is carried out to obtain the cable aging performance test evaluation plan, and specific display and notification are made according to the content of the cable aging performance test evaluation plan.

[0054] The present invention provides a method and system for testing the aging performance of a new energy vehicle charging cable, which has the following beneficial effects:

[0055] (1) When the system is running, the comprehensive use of the environment acquisition module, data processing module, performance test module and evaluation module realizes the accurate evaluation and prediction of the cable aging performance. The environment acquisition module collects the cable test data under temperature control, humidity control and water immersion conditions in real time. The data processing module cleans, standardizes and extracts features from these data to form a high-quality feature data set X. The performance test module uses random forest, gradient boosting tree and deep neural network models for multi-model prediction, and fits the meta-model to obtain the prediction results of the Stacking model. It is matched with the preset evaluation threshold L to generate a cable aging performance test evaluation plan, which is communicated through a variety of display and notification methods. It solves the problems of slow response, low prediction accuracy and poor adaptability of traditional methods, provides higher accuracy and reliability, and significantly improves the scientificity and effectiveness of cable aging performance testing. It provides strong support for cable maintenance and management, and ensures the safety and stability of the system.

[0056] (2) By using the feature extraction data set X, a random forest model, a gradient boosting tree model, and a deep neural network model were constructed and trained respectively. Then, the ensemble prediction results of the random forest were combined by using the stacking method. Decision tree prediction results And the prediction results of deep neural network Perform integration, use the meta-model for fitting prediction, and obtain the prediction results of the Stacking model It improves the generalization ability and prediction accuracy of the model, and effectively solves the prediction error problem that may occur in a single model under extreme environmental conditions.

[0057] (3) Evaluation of the threshold L and the prediction results of the Stacking model through the preset cable performance aging test Matching can be carried out to accurately determine whether the cable aging performance is qualified, and the cable aging performance test evaluation plan can be obtained simultaneously, and a detailed cable aging performance test evaluation plan can be generated. When the evaluation result is unqualified, the execution unit will display and notify in a variety of ways, including dashboard display, test report generation, pop-up reminders, and notifications through emails, text messages and mobile applications. This multi-channel and instant feedback mechanism ensures that relevant personnel can quickly obtain evaluation results and suggestions and take corresponding measures in a timely manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 This is a schematic diagram of a flow chart of a testing system for aging performance of a new energy vehicle charging cable according to the present invention;

[0059] Figure 2 The present invention is a schematic diagram of the steps of a method for testing the aging performance of a new energy vehicle charging cable. DETAILED DESCRIPTION

[0060] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0061] The rapid development of new energy vehicles has promoted the construction of charging infrastructure. Among them, charging cables are key components connecting power sources and vehicles, and their performance and safety are particularly important. In order to ensure the reliability and durability of charging cables under various extreme environmental conditions, systematic aging performance testing has become a key link.

[0062] However, there are some shortcomings and deficiencies in the current aging performance testing of charging cables. Traditional testing methods mainly rely on static and single environmental conditions testing, which can hardly fully reflect the complex stresses and environmental changes of cables in actual use. For example, tests conducted only at room temperature cannot accurately predict the performance changes of cables in high or low temperature environments. Similarly, the aging behavior of cables under extreme conditions related to high humidity and salt water immersion cannot be fully evaluated through a single laboratory test. These limitations result in the inability of existing testing methods to provide sufficient data support, affecting the optimization of cable design and material selection.

[0063] These current situations and shortcomings are mainly due to the insufficient accuracy of traditional test systems in simulating environmental conditions, which cannot accurately simulate and control a variety of extreme environmental conditions. This lack of simulation accuracy has brought about a series of abnormal effects, such as large errors in test results, and the inability to accurately predict the aging behavior of cables in actual use; delayed response to environmental stress, resulting in inaccurate aging performance evaluation; and poor test repeatability, making it difficult to conduct reliable comparative analysis. The above problems not only increase the uncertainty in the design and manufacturing process, but may also lead to safety hazards in actual cable applications.

[0064] Example 1

[0065] The present invention provides a testing system for the aging performance of charging cables for new energy vehicles. Figure 1 , including environment acquisition module, data processing module, performance testing module and evaluation module;

[0066] The environment acquisition module collects environmental test data in real time through cables for the preset environmental conditioning tests of the temperature control device, the humidity control device and the immersion device to form a test data group;

[0067] The data processing module preprocesses the test data group, including data cleaning, data standardization and feature extraction, to form a feature extraction data set X;

[0068] The performance testing module builds multiple models for the data set X, including a random forest model, a gradient boosted tree model, and a deep neural network model, and performs an integrated prediction on the random forest model. Decision tree prediction results And the prediction results of deep neural network Use the meta-model to perform fitting predictions and obtain the prediction results of the Stacking model

[0069] The evaluation module evaluates the threshold L of the preset cable performance aging test and the prediction result of the Stacking model. Matching is carried out to obtain the cable aging performance test evaluation plan, and specific display and notification are made according to the content of the cable aging performance test evaluation plan.

[0070] In this embodiment, the comprehensive use of the environment acquisition module, data processing module, performance test module and evaluation module realizes the accurate evaluation and prediction of the cable aging performance. The environment acquisition module collects the cable test data under temperature control, humidity control and water immersion conditions in real time, and the data processing module cleans, standardizes and extracts features from these data to form a high-quality feature data set X. The performance test module uses random forest, gradient boosting tree and deep neural network models for multi-model prediction, and fits them through the meta-model to obtain the prediction results of the Stacking model. By matching with the preset evaluation threshold L, a cable aging performance test evaluation plan is generated and communicated through a variety of display and notification methods. The system solves the problems of slow response, low prediction accuracy and poor adaptability of traditional methods, provides higher accuracy and reliability, and significantly improves the scientificity and effectiveness of cable aging performance testing. It provides strong support for cable maintenance and management, and ensures the safety and stability of the system.

[0071] Example 2

[0072] This embodiment is explained in Example 1, please refer to Figure 1 ,Specifically: the environment acquisition module includes a temperature control unit, a humidity control unit and a water immersion unit;

[0073] The temperature control unit controls the temperature of the cable test environment by using high-precision temperature control equipment, including an environmental chamber and a thermal cycle system, and then simulates extreme high and low temperature environmental conditions through the equipment, and adjusts and maintains the temperature in real time through a PID control algorithm to ensure that the temperature fluctuation range is within ±0.1°C of the set value;

[0074] The humidity control unit adjusts and controls the humidity state of the test cable environment by using humidifier and dehumidifier equipment, and then simulates a special cable use environment through the equipment, including high humidity on rainy days and low humidity on hot days;

[0075] The immersion unit simulates the immersion conditions of the cable test in different liquids, including salt water conditions and acid-base water conditions, by using an immersion tank and a liquid circulation device;

[0076] The test data set is constructed by recording parameters related to cable aging, including resistance change rate, insulation resistance, temperature, humidity, current, water immersion conditions and mechanical stress information.

[0077] The data processing module includes a preprocessing unit and an extraction unit;

[0078] The preprocessing unit preprocesses the data information of the test data group, including data cleaning and data standardization, including temperature, humidity, water immersion conditions, resistance and insulation resistance related test data;

[0079] Among them, data cleaning includes identifying and deleting outliers, noise, and filling missing values; data standardization includes converting data of different dimensions to the same dimension to ensure data consistency;

[0080] The extraction unit selects characteristic information of relevant parameters according to the cable aging performance related parameters for extraction, including resistance change rate, insulation resistance, temperature, humidity, current, water immersion condition and mechanical stress, and simultaneously calculates the extracted characteristic information to form a characteristic extraction data set X;

[0081] The feature extraction data set X is specifically X{R t , R ins , T, H, I, W, F, ΔR t , △R ins}, where R t Represents the resistance change rate, R ins Indicates insulation resistance value, T indicates temperature value, H indicates humidity value, I indicates current value, W indicates water immersion condition sequence value, F indicates mechanical stress value, ΔR t Indicates the fluctuation value of the resistance change rate, △R ins Indicates the fluctuation value of insulation resistance.

[0082] Example 3

[0083] This embodiment is explained in Example 1, please refer to Figure 1 ,Specifically: the performance testing module includes a modeling unit and a model integration unit;

[0084] The modeling unit establishes a random forest model for the feature extraction data set X, and constructs a decision tree model by splitting the Gini impurity feature of the feature extraction data set X with the smallest Gini impurity. At the same time, the random forest integrates the prediction results of multiple decision trees to improve the generalization ability and prediction accuracy of the model and obtain the integrated prediction result of the random forest. Use the feature extraction data set X to build a gradient boosting tree model, optimize the residual by iteratively training multiple decision trees, and obtain the prediction results of the decision tree after updating the residual. Use the feature extraction data set X deep neural network model, train it through forward propagation and back propagation algorithms to gradually update the model parameters, and obtain the prediction results of the deep neural network after minimizing the loss function

[0085] The model integration unit uses the stacking method to integrate the random forest prediction results Decision tree prediction results And the prediction results of deep neural network Perform integration, perform fitting prediction through the meta-model, and obtain the prediction results of the Stacking model

[0086] The ensemble prediction results of the random forest Obtained through the following calculation formula:

[0087]

[0088] In the formula, represents the ensemble prediction result of random forest, T RF represents the number of decision trees, Represents the prediction result of the tth decision tree;

[0089] The prediction result of the tth decision tree Obtained through the following calculation formula:

[0090]

[0091] In the formula, f t,RF represents the prediction function of the t-th decision tree, and X represents the feature extraction data set;

[0092] The Gini impurity is split by The formula is calculated, where G ini (X) represents the Gini impurity of the feature extraction dataset X, K represents the number of sample categories, P k Indicates the proportion of samples belonging to the kth class.

[0093] The residual update is optimized by the following calculation formula:

[0094] R i,m =Y i -F m-1 *(X i );

[0095] In the formula, R i,m represents the residual of the i-th sample at the m-th iteration, Y i represents the true value of the i-th sample, F m-1 *(X i ) represents the predicted value of the feature extraction data set X of the model input for the first m-1 iterations, and the prediction error of the current model on the i-th sample is represented by the residual update calculation formula, and the prediction error of the decision tree model is corrected by fitting the residual;

[0096] The prediction results of the decision tree Obtained through the following calculation formula:

[0097]

[0098] In the formula, Represents the prediction result of the mth decision tree h m represents the prediction function of the mth decision tree, and X represents the feature extraction dataset.

[0099] Prediction results of the deep neural network Obtained through the following calculation formula:

[0100]

[0101] In the formula, represents the prediction result of the deep neural network, f DNN represents the prediction function of the deep neural network, and X represents the feature extraction dataset.

[0102] Prediction results of the Stacking model Obtained through the following calculation formula:

[0103]

[0104] In the formula, represents the prediction result of the Stacking model, g represents the meta-model, including linear regression and DNN, represents the ensemble prediction result of random forest, Represents the prediction result of the decision tree, Represents the prediction result of the deep neural network.

[0105] In this embodiment, a random forest model, a gradient boosting tree model and a deep neural network model are constructed by using the feature extraction data set X, and then the integrated prediction results of the random forest are combined by using the stacking method. Decision tree prediction results And the prediction results of deep neural network Perform integration, use the meta-model for fitting prediction, and obtain the prediction results of the Stacking model It improves the generalization ability and prediction accuracy of the model, and effectively solves the prediction error problem that may occur in a single model under extreme environmental conditions. Through these technical means, the system has demonstrated higher accuracy and reliability in dealing with cable aging performance tests, providing a scientific basis and guidance for cable maintenance and management, and significantly improving the safety and stability of the system.

[0106] Example 4

[0107] This embodiment is explained in Example 1, please refer to Figure 1 ,Specifically: the evaluation module includes a matching unit and an execution unit;

[0108] The matching unit matches the required comparison value through the preset relevant information, including the preset cable performance aging test evaluation threshold L and the prediction result of the Stacking model. Match and obtain the cable aging performance test and evaluation plan;

[0109] The cable aging performance test evaluation scheme is obtained by the following matching method:

[0110] Prediction results of the Stacking model Less than the cable performance aging test assessment threshold L, the cable performance aging test assessment is qualified;

[0111] Prediction results of the Stacking model If it is greater than or equal to the cable performance aging test evaluation threshold L, the cable performance aging test evaluation fails;

[0112] The execution unit performs specific display and notification according to the content of the cable aging performance test evaluation plan. The display includes dashboard display, test report generation and pop-up reminder, and the notification includes test report email notification, SMS notification and mobile application notification.

[0113] In this embodiment, the preset cable performance aging test evaluation threshold L and the prediction result of the Stacking model are used. Matching is carried out to accurately determine whether the cable aging performance is qualified, and the cable aging performance test evaluation plan is obtained simultaneously, and a detailed cable aging performance test evaluation plan is generated. When the evaluation result is unqualified, the execution unit will display and notify in a variety of ways, including dashboard display, test report generation, pop-up reminders, and notifications through emails, text messages and mobile applications. This multi-channel and instant feedback mechanism ensures that relevant personnel can quickly obtain evaluation results and suggestions, and take corresponding measures in time to solve the potential risks brought by cable aging. This not only improves the safety and reliability of the cable system, but also optimizes the maintenance and management processes, and solves the problems of slow response, untimely information transmission, and incomplete treatment measures of traditional aging detection methods.

[0114] Example 5

[0115] A test method for the aging performance of new energy vehicle charging cables, please refer to Figure 2 , specifically: including the following steps:

[0116] Step 1: The environment collection module collects environmental test data in real time through the cables of the preset environmental conditioning tests of the temperature control equipment, the humidity control equipment and the immersion equipment to form a test data group;

[0117] Step 2: The data processing module preprocesses the test data group, including data cleaning, data standardization and feature extraction, to form a feature extraction data set X;

[0118] Step 3: The performance testing module builds multiple models for data set X, including random forest model, gradient boosting tree model and deep neural network model, and integrates the prediction results of random forest. Decision tree prediction results And the prediction results of deep neural network Use the meta-model to perform fitting predictions and obtain the prediction results of the Stacking model

[0119] Step 4: The evaluation module evaluates the threshold L and the prediction results of the Stacking model through the preset cable performance aging test Matching is carried out to obtain the cable aging performance test evaluation plan, and specific display and notification are made according to the content of the cable aging performance test evaluation plan.

[0120] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A test system for the aging performance of new energy vehicle charging cables, characterized by: It includes environment acquisition module, data processing module, performance testing module and evaluation module; The environment acquisition module collects environmental test data in real time through cables for the preset environmental conditioning tests of the temperature control device, the humidity control device and the immersion device to form a test data group; The data processing module preprocesses the test data group, including data cleaning, data standardization and feature extraction, to form a feature extraction data set X; The performance testing module builds multiple models for the data set X, including a random forest model, a gradient boosted tree model, and a deep neural network model, and performs an integrated prediction on the random forest model. Decision tree prediction results And the prediction results of deep neural network Use the meta-model to perform fitting predictions and obtain the prediction results of the Stacking model The performance test module includes a modeling unit and a model integration unit; the modeling unit establishes a random forest model for the feature extraction data set X, and constructs a decision tree model by splitting the Gini impurity feature of the feature extraction data set X with the smallest Gini impurity. At the same time, the random forest integrates the prediction results of multiple decision trees to improve the generalization ability and prediction accuracy of the model and obtain the integrated prediction results of the random forest Use the feature extraction data set X to build a gradient boosting tree model, optimize the residual by iteratively training multiple decision trees, and obtain the prediction results of the decision tree after updating the residual. Use the feature extraction data set X deep neural network model, train it through forward propagation and back propagation algorithms, gradually update the model parameters, minimize the loss function, and obtain the prediction results of the deep neural network. The model integration unit uses the stacking method to integrate the random forest prediction results Decision tree prediction results And the prediction results of deep neural network Perform integration, perform fitting prediction through the meta-model, and obtain the prediction results of the Stacking model The evaluation module evaluates the threshold L of the preset cable performance aging test and the prediction result of the Stacking model. Matching is carried out to obtain the cable aging performance test evaluation plan, and specific display and notification are made according to the content of the cable aging performance test evaluation plan.

2. A testing system for aging performance of a new energy vehicle charging cable according to claim 1, characterized in that: The environment acquisition module includes a temperature control unit, a humidity control unit and a water immersion unit; The temperature control unit controls the temperature of the cable test environment by using high-precision temperature control equipment, including an environmental chamber and a thermal cycle system, and then simulates extreme high and low temperature environmental conditions through the equipment, and adjusts and maintains the temperature in real time through a PID control algorithm to ensure that the temperature fluctuation range is within ±0.1°C of the set value; The humidity control unit adjusts and controls the humidity state of the test cable environment by using humidifier and dehumidifier equipment, and then simulates a special cable use environment through the equipment, including high humidity on rainy days and low humidity on hot days; The immersion unit simulates the immersion conditions of the cable test in different liquids, including salt water conditions and acid-base water conditions, by using an immersion tank and a liquid circulation device; The test data set is constructed by recording parameters related to cable aging, including resistance change rate, insulation resistance, temperature, humidity, current, water immersion conditions and mechanical stress information.

3. A testing system for aging performance of a new energy vehicle charging cable according to claim 1, characterized in that: The data processing module includes a preprocessing unit and an extraction unit; The preprocessing unit preprocesses the data information of the test data group, including data cleaning and data standardization, including temperature, humidity, water immersion conditions, resistance and insulation resistance related test data; Among them, data cleaning includes identifying and deleting outliers, noise, and filling missing values; data standardization includes converting data of different dimensions to the same dimension to ensure data consistency; The extraction unit selects characteristic information of relevant parameters according to the cable aging performance related parameters for extraction, including resistance change rate, insulation resistance, temperature, humidity, current, water immersion condition and mechanical stress, and simultaneously calculates the extracted characteristic information to form a characteristic extraction data set X; The feature extraction data set X is specifically X{R t , R ins , T, H, I, W, F, ΔR t , △R ins }, where R t Represents the resistance change rate, R ins Indicates insulation resistance value, T indicates temperature value, H indicates humidity value, I indicates current value, W indicates water immersion condition sequence value, F indicates mechanical stress value, ΔR t Indicates the fluctuation value of the resistance change rate, △R ins Indicates the fluctuation value of insulation resistance.

4. A testing system for aging performance of a new energy vehicle charging cable according to claim 1, characterized in that: The ensemble prediction results of the random forest Obtained through the following calculation formula: In the formula, represents the ensemble prediction result of random forest, T RF represents the number of decision trees, Represents the prediction result of the tth decision tree; The prediction result of the tth decision tree Obtained through the following calculation formula: In the formula, f t,RF represents the prediction function of the t-th decision tree, and X represents the feature extraction data set; The Gini impurity is split by The formula is calculated, where G ini (X) represents the Gini impurity of the feature extraction dataset X, K represents the number of sample categories, P k Indicates the proportion of samples belonging to the kth class.

5. A testing system for aging performance of a new energy vehicle charging cable according to claim 1, characterized in that: The residual update is optimized by the following calculation formula: R i,m =Y i -F m-1 *(X i ); In the formula, R i,m represents the residual of the i-th sample at the m-th iteration, Y i represents the true value of the i-th sample, F m-1 *(X i ) represents the predicted value of the feature extraction data set X of the model input for the first m-1 iterations, and the prediction error of the current model on the i-th sample is represented by the residual update calculation formula, and the prediction error of the decision tree model is corrected by fitting the residual; The prediction results of the decision tree Obtained through the following calculation formula: In the formula, Represents the prediction result of the mth decision tree h m represents the prediction function of the mth decision tree, and X represents the feature extraction dataset.

6. A new energy vehicle charging cable aging performance testing system according to claim 1, characterized in that: Prediction results of the deep neural network Obtained through the following calculation formula: In the formula, represents the prediction result of the deep neural network, f DNN represents the prediction function of the deep neural network, and X represents the feature extraction dataset.

7. A new energy vehicle charging cable aging performance testing system according to claim 1, characterized in that: Prediction results of the Stacking model Obtained through the following calculation formula: In the formula, represents the prediction result of the Stacking model, g represents the meta-model, including linear regression and DNN, represents the ensemble prediction result of random forest, Represents the prediction result of the decision tree. Represents the prediction result of the deep neural network.

8. A new energy vehicle charging cable aging performance testing system according to claim 1, characterized in that: The evaluation module includes a matching unit and an execution unit; The matching unit matches the required comparison value through the preset relevant information, including the preset cable performance aging test evaluation threshold L and the prediction result of the Stacking model. Match and obtain the cable aging performance test and evaluation plan; The cable aging performance test evaluation scheme is obtained by the following matching method: Prediction results of the Stacking model If it is less than the cable performance aging test evaluation threshold L, it is determined that the cable performance aging test evaluation is qualified; Prediction results of the Stacking model If it is greater than or equal to the cable performance aging test evaluation threshold L, the cable performance aging test evaluation fails; The execution unit performs specific display and notification according to the content of the cable aging performance test evaluation plan. The display includes dashboard display, test report generation and pop-up reminder, and the notification includes test report email notification, SMS notification and mobile application notification.

9. A method for testing the aging performance of a new energy vehicle charging cable, applied to a system for testing the aging performance of a new energy vehicle charging cable as claimed in any one of claims 1 to 8, characterized in that: The following steps are involved: Step 1: The environment collection module collects environmental test data in real time through the cables of the preset environmental conditioning tests of the temperature control equipment, the humidity control equipment and the immersion equipment to form a test data group; Step 2: The data processing module preprocesses the test data group, including data cleaning, data standardization and feature extraction, to form a feature extraction data set X; Step 3: The performance testing module builds multiple models for data set X, including random forest model, gradient boosting tree model and deep neural network model, and integrates the prediction results of random forest. Decision tree prediction results And the prediction results of deep neural network Use the meta-model to perform fitting predictions and obtain the prediction results of the Stacking model Step 4: The evaluation module evaluates the threshold L and the prediction results of the Stacking model through the preset cable performance aging test Matching is carried out to obtain the cable aging performance test evaluation plan, and specific display and notification are made according to the content of the cable aging performance test evaluation plan.

Images