Method and system for detecting reliability of automobile connector and medium

Abstract

The invention discloses a method and system for detecting the reliability of an automobile connector and a medium, and the method comprises the steps: obtaining the internal microscopic state parameters and external working condition parameters of the automobile connector in real time; preprocessing the internal microscopic state parameters and the external working condition parameters to extract health feature vectors of the automobile connector; and based on a trained reliability evaluation model, according to the health feature vector, evaluating the current health state grade of the automobile connector and predicting the remaining useful life. According to the method and the system for detecting the reliability of the automobile connector and the medium, the transient abnormality and the gradual change fault can be captured by acquiring the microscopic state and the external working condition in real time; future risks are predicted by outputting the health state level and the remaining useful life, and sudden faults are avoided.

Description

Technical Field

[0001] This invention relates to the technical field of connector testing, and more particularly to a method, system, and medium for testing the reliability of automotive connectors. Background Technology

[0002] As the neural network hub of automotive electronic systems, automotive connectors undertake the critical tasks of power transmission and signal interaction, and their reliability is directly related to the safety, functionality, and durability of the entire vehicle. With the continuous improvement of vehicle electrification and intelligence, the in-vehicle electrical system is becoming increasingly complex, and connectors need to withstand higher currents and voltages, as well as more stringent operating conditions such as mechanical vibration and temperature cycling. This poses unprecedented challenges to the reliability of connectors.

[0003] Currently, the testing and evaluation of automotive connector reliability mainly rely on the following traditional methods: (1) When the connector is replaced after a functional failure (such as signal interruption or severe overheating leading to ablation), or when routine inspections and replacements are carried out based on fixed time / mileage intervals, there is a serious lag, and sudden failures cannot be predicted, which may lead to safety accidents such as vehicle breakdowns while driving; at the same time, excessive maintenance based on fixed cycles also results in a waste of resources.

[0004] (2) When the connector is removed from the vehicle in the laboratory or repair shop and its contact resistance is statically measured using a special instrument (such as a micro-ohmmeter), or its insertion and extraction force, wear and oxidation are judged by visual inspection, it is non-real-time (cannot reflect the state of the connector under real and dynamic working conditions), non-continuous (can only obtain a snapshot at the time of detection, and cannot capture the gradual degradation process of its performance) and destructive (the disassembly process itself may change the contact state of the connector and introduce human error).

[0005] (3) Some advanced solutions attempt to indirectly infer the health status of connectors by monitoring external operating parameters (such as current and ambient temperature). However, connector failure (such as fretting wear and contact interface oxidation) is a direct result of the deterioration of its internal microstate (contact resistance, contact stress, and internal temperature field), and it is difficult to accurately and timely reflect the internal degradation mechanism by relying solely on external parameters. In addition, such methods are usually difficult to effectively capture and diagnose transient performance fluctuations caused by instantaneous overload, vibration, and shock.

[0006] Therefore, this application proposes a method, system, and medium for testing the reliability of automotive connectors. Summary of the Invention

[0007] The purpose of this invention is to provide a method, system, and medium for detecting the reliability of automotive connectors. By acquiring microscopic states and external operating conditions in real time, it can capture transient anomalies and gradual failures; by outputting health status levels and remaining useful life, it can predict future risks and avoid sudden failures.

[0008] The objective of this invention is achieved through the following technical solution: In a first aspect, the present invention provides a method for testing the reliability of automotive connectors, comprising the following steps: Real-time acquisition of the internal microscopic state parameters and external operating condition parameters of the automotive connector; The internal micro-state parameters and external operating condition parameters are preprocessed to extract the health feature vector of the automotive connector; Based on the trained reliability assessment model, the current health status level of the automotive connector is evaluated and the remaining useful life is predicted according to the health feature vector.

[0009] The beneficial effects of the above solution are as follows: Traditional detection methods are mostly offline and static measurements, while this invention achieves dynamic and continuous monitoring of the degradation process of automotive connectors under real-time microscopic conditions and external operating conditions, enabling the capture of transient anomalies and gradual failures. Furthermore, this invention outputs health status levels and remaining useful life, predicting future risks and providing direct, quantitative decision-making basis for implementing predictive maintenance, thus avoiding sudden failures. In addition, this invention automatically performs analysis and diagnosis through health feature vectors and reliability assessment models, greatly reducing the reliance on manual experience and frequent offline testing in traditional methods, and improving detection efficiency and objectivity.

[0010] Furthermore, the internal microstate parameters include one or more combinations of contact resistance, contact interface mechanical stress, and internal temperature distribution. The external operating parameters include one or more combinations of current load, ambient temperature, and ambient vibration.

[0011] The beneficial effects of the above scheme are as follows: This invention clarifies the specific physical quantities to be monitored. Contact resistance reflects electrical contact performance; mechanical stress at the contact interface is directly related to fretting wear; and internal temperature distribution directly reflects the electrothermal coupling effect. Furthermore, combining the internal microstate with external operating conditions enables multi-dimensional monitoring of the main failure mechanisms of automotive connectors (such as oxidation, wear, and overheating). In addition, by simultaneously monitoring the internal microstate and external operating conditions, the subsequent evaluation model can distinguish between the degradation of the automotive connector itself and the performance fluctuations caused by harsh external operating conditions, thereby making the evaluation and prediction results closer to reality.

[0012] Furthermore, obtaining the internal microscopic state parameters of the automotive connector includes: The contact resistance between the pins and sockets of the automotive connector is dynamically measured using a four-wire micro-resistance measurement method. Mechanical stress at the contact interface is monitored by a piezoelectric thin film sensor or fiber optic grating sensor mounted on the automotive connector. By using an array of multiple temperature sensors arranged inside the automotive connector, temperature data from multiple internal points is acquired to form an internal temperature distribution.

[0013] The beneficial effects of the above solution are as follows: This invention eliminates the influence of lead resistance through a four-wire micro-resistance measurement method, enabling the capture of minute changes in contact resistance to identify oxidation and fretting wear. This invention utilizes the high sensitivity of piezoelectric thin film or fiber optic grating sensors to mechanical stress and vibration, enabling early diagnosis of mechanical failures. This invention uses a temperature sensor array to map the temperature field inside automotive connectors, helping to locate localized hot spots and detect abnormal heating caused by increased contact resistance.

[0014] Furthermore, the method for extracting the health feature vector includes: Time-domain and frequency-domain features are extracted from the contact resistance, including the resistance mean, resistance variance, and resistance trend slope. Stress features are extracted from the mechanical stress at the contact interface, and the stress features include stress frequency and stress energy. Temperature features are extracted from the internal temperature distribution, including temperature peaks, temperature gradients, and temperature rise rates. The features of the extracted internal micro-state parameters are fused with the external operating condition parameters to form a multi-dimensional health feature vector.

[0015] Furthermore, the training method for the reliability assessment model includes: Obtain a historical dataset, which includes the historical health feature vector of the vehicle connector throughout its entire life cycle, as well as the historical health status level label and historical remaining useful life label corresponding to each historical health feature vector; The classification model is trained by using the historical health feature vector as input and the corresponding historical health status level label as the expected output. The regression model is trained by using the historical health feature vector as input and the corresponding historical remaining useful life label as the expected output. The classification model and the regression model are integrated to obtain the reliability assessment model.

[0016] The beneficial effects of the above solution are as follows: This invention uses labeled, full-lifecycle historical data for training, enabling the model to learn the complete degradation trajectory from health to failure, thus providing a data-driven basis for assessments and predictions. The classification model quickly and intuitively provides state levels (such as healthy, attentive, and warning), facilitating immediate decision-making; the regression model provides accurate remaining lifetime values, facilitating the development of long-term maintenance plans. By combining the two, the reliability assessment model becomes more powerful and practical.

[0017] Furthermore, the optimization method for the reliability assessment model includes: Collect real-time data, evaluation results, prediction results, and actual maintenance records of the automotive connector to form an incremental dataset; Using the incremental dataset, the parameters of the classification model and / or regression model in the reliability assessment model are optimized; The current health status level and remaining useful life of the automotive connector are updated using the optimized reliability assessment model.

[0018] The beneficial effects of the above solution are as follows: By collecting on-site data, this invention continuously optimizes the model, enabling the reliability assessment model to adapt to differences caused by different vehicle models, usage habits, and environments. Its prediction accuracy increases over time and with data accumulation. Furthermore, feeding actual maintenance records back to the reliability assessment model is equivalent to providing the model with a standard answer, allowing it to verify and correct its prediction logic, thus forming a continuously improving positive feedback loop.

[0019] Furthermore, the early degradation judgment method of the reliability assessment model includes: The external operating parameters and the contact resistance are input as boundary conditions into the digital twin model to calculate the theoretical temperature field and the theoretical stress field. The temperature deviation between the theoretical temperature field and the internal temperature distribution is calculated. If the temperature deviation continues to exceed a preset temperature threshold, the automotive connector is determined to have experienced early degradation and a first warning is triggered. The stress deviation between the theoretical stress field and the mechanical stress at the contact interface is calculated. If the stress deviation continues to exceed a preset stress threshold, the automotive connector is determined to have experienced early degradation, and a second warning is triggered.

[0020] The beneficial effects of the above solution are as follows: This invention calculates theoretical values ​​based on first-principles calculations (electrothermal-mechanical laws) using a digital twin model. When a small but persistent deviation occurs between the measured and theoretical values, it indicates that the physical properties have changed, but this occurs earlier than any absolute parameter exceeding the limit. It can issue early warnings at the earliest stage of a fault, achieving prevention before it occurs. Furthermore, the early degradation judgment of this invention does not rely on a large amount of historical fault data, but is based on a physical model, thus remaining effective even in the absence of specific fault samples. Simultaneously, different warning signals can indicate whether the abnormality lies in thermal or mechanical properties, making the diagnostic results more interpretable and facilitating engineers in locating the root cause.

[0021] Furthermore, the step of evaluating the current health status level of the automotive connector and predicting its remaining useful life based on the trained reliability assessment model and the health feature vector includes: The health feature vector is input into the reliability assessment model to generate the current health status level, predicted remaining life, and confidence interval of the automotive connector. The predicted value of the remaining lifetime and its confidence interval are compared with the reference value of the remaining lifetime; In response to the reference value of the remaining lifetime being within the confidence interval, the predicted value of the remaining useful lifetime is taken as the remaining effective lifetime. In response to the reference value of the remaining lifetime being outside the confidence interval, the remaining effective lifetime is determined by weighting the predicted value of the remaining lifetime with the reference value of the remaining lifetime.

[0022] The beneficial effects of the above solution are: the confidence interval of this invention can indicate the reliability of the prediction results, enabling decision-makers not only to know how much longer the product can be used, but also how accurate the prediction is. Furthermore, by comparing the predicted value with the reference value, the final determined remaining effective lifespan is made more reliable and practical.

[0023] Secondly, the present invention provides a system for testing the reliability of automotive connectors, comprising: The acquisition module is used to acquire the internal microscopic state parameters and external operating condition parameters of the automotive connector in real time. The extraction module is used to preprocess the internal microscopic state parameters and external operating condition parameters to extract the health feature vector of the automotive connector. The detection module is used to assess the current health status level of the automotive connector and predict its remaining useful life based on the health feature vector, using a trained reliability assessment model.

[0024] Thirdly, the present invention provides a computer-readable storage medium storing a computer program that, when executed by at least one processor, implements the steps of the above-described detection method.

[0025] Compared with the prior art, the beneficial effects of the present invention include at least the following: Traditional detection methods are mostly offline and static measurements, while this invention achieves dynamic and continuous monitoring of the degradation process of automotive connectors under real-time microscopic conditions and external operating conditions, enabling the capture of transient anomalies and gradual failures. Furthermore, this invention outputs health status levels and remaining useful life, predicting future risks and providing direct, quantitative decision-making basis for predictive maintenance, thus avoiding sudden failures. In addition, this invention automatically performs analysis and diagnosis through health feature vectors and reliability assessment models, greatly reducing the reliance on manual experience and frequent offline testing in traditional methods, improving detection efficiency and objectivity. Attached Figure Description

[0026] Figure 1 This is a schematic flowchart of a method for detecting the reliability of automotive connectors according to an embodiment of the present invention. Detailed Implementation

[0027] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the embodiments set forth herein; rather, they are provided to make the invention more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art. The same reference numerals in the drawings denote the same or similar structures, and therefore repeated descriptions of them will be omitted.

[0028] The terms used to express position and direction in this invention are illustrated with reference to the accompanying drawings, but changes can be made as needed, and all such changes are included within the scope of protection of this invention.

[0029] This invention introduces a method, system, and medium for testing the reliability of automotive connectors. refer to Figure 1 The method for detecting the reliability of automotive connectors of the present invention includes steps SS1 to SS3.

[0030] Step SS1: Real-time acquisition of the internal microscopic state parameters and external operating condition parameters of the automotive connector.

[0031] In application, the external operating parameters of this invention include one or more combinations of current load, ambient temperature, and environmental vibration. Specifically, the external operating parameters of the automotive connector are obtained via the vehicle's CAN bus.

[0032] In practical applications, the internal microscopic state parameters of this invention include one or more combinations of contact resistance, contact interface mechanical stress, and internal temperature distribution.

[0033] Furthermore, obtaining the internal microscopic state parameters of the automotive connector includes steps SS11 to SS13.

[0034] Step SS11: Use the four-wire micro-resistance measurement method to dynamically measure the contact resistance between the pins and sockets of the automotive connector.

[0035] Specifically, a measurement circuit with four-wire measurement capability includes a pair of drive lines and a pair of sensing lines.

[0036] A pair of drive leads connect to a constant current source to apply a stable test current to the automotive connector. This test current is significantly lower than the connector's operating current to avoid the measurement itself affecting the connector's condition (the test current is typically less than 100mA). A pair of sensing leads connect directly to the measurement points on the connector's pins and sockets (e.g., the base of the pin and the corresponding spring base of the socket) to measure the voltage drop between these two points. The contact resistance is determined using Ohm's law: Contact Resistance = Voltage Drop / Test Current.

[0037] Step SS12: Monitor the mechanical stress at the contact interface using a piezoelectric thin film sensor or fiber optic grating sensor mounted on the automotive connector.

[0038] In some embodiments, to monitor in real time changes in contact pressure caused by fretting wear due to vibration and thermal cycling, a miniature, flexible piezoelectric thin-film sensor (such as PVDF) is attached inside the automotive connector housing, close to the non-contact area on the back of the socket spring. When the mechanical stress at the contact interface fluctuates (mainly caused by external vibration), the pin inserts and applies positive pressure to the spring. The minute deformation of the spring is transmitted to the piezoelectric thin-film sensor, causing it to generate an electric charge.

[0039] At this point, the amplitude of the electrical signal generated by the piezoelectric thin-film sensor is related to the amount of stress change, and the frequency characteristics are related to the vibration source. In some embodiments, the fiber optic sensor is glued to the substrate of the socket spring, allowing it to deform in tandem with the spring. When the mechanical stress at the contact interface fluctuates (mainly caused by external vibration), the spring deforms under stress, causing a change in the grating pitch of the fiber optic sensor, thereby altering the center wavelength of its reflected light. The drift Δλ of the center wavelength of the fiber optic sensor is positively correlated with the fluctuation of the mechanical stress at the contact interface.

[0040] Step SS13: Acquire temperature data at multiple internal points using an array of multiple temperature sensors arranged inside the automotive connector to form an internal temperature distribution.

[0041] Specifically, multiple temperature sensor arrays can be positioned near high-current-carrying pins, behind the automotive connector mating interface, and in the center and edges of the automotive connector housing.

[0042] Step SS2: Preprocess the internal micro-state parameters and external operating condition parameters to extract the health feature vector of the automotive connector.

[0043] When applied, the method for extracting the health feature vector includes steps SS21 to SS24.

[0044] Step SS21: Extract time-domain and frequency-domain features from the contact resistance, including the resistance mean, resistance variance, and resistance trend slope.

[0045] Step SS22: Extract stress features from the mechanical stress at the contact interface, the stress features including stress frequency and stress energy.

[0046] Step SS23: Extract temperature features from the internal temperature distribution, including temperature peaks, temperature gradients, and temperature rise rates.

[0047] Step SS24: The features of the extracted internal microstate parameters are fused with the external operating condition parameters to form a multidimensional health feature vector.

[0048] Specifically, the average resistance, resistance variance, resistance trend slope, stress frequency, stress energy, temperature peak, temperature gradient, and temperature rise rate are all stamped with a unified timestamp, and features at different scales are converted to the same scale.

[0049] Step SS3: Based on the trained reliability assessment model and the health feature vector, assess the current health status level of the automotive connector and predict its remaining useful life.

[0050] When applied, the training method of the reliability assessment model of the present invention includes steps SS311 to SS314.

[0051] Step SS311: Obtain historical dataset, which includes historical health feature vectors of the vehicle connector throughout its entire life cycle, as well as historical health status level labels and historical remaining useful life labels corresponding to each historical health feature vector.

[0052] The term "full life cycle" refers to the entire time from when a brand-new, intact automotive connector is first installed and put into use until it is replaced or scrapped due to performance degradation and failure.

[0053] Step SS312: Train the classification model using the historical health feature vector as input and the corresponding historical health status level label as the expected output.

[0054] Step SS313: Train the regression model using the historical health feature vector as input and the corresponding historical remaining useful life label as the expected output.

[0055] Step SS314: Integrate the classification model and the regression model to obtain the reliability assessment model.

[0056] In practical applications, the optimization method of the reliability assessment model of the present invention includes steps SS321 to SS323.

[0057] Step SS321: Collect real-time data, evaluation results, prediction results, and actual maintenance record data of the automotive connector to form an incremental dataset.

[0058] The real-time data includes internal micro-state parameters, external operating condition parameters, and health feature vectors; the assessment results include the current health status level; the prediction results include the remaining useful life; the actual maintenance record data includes: the final failure mode, i.e., the actual problems found during maintenance (e.g., socket ablation, spring loosening, insulation aging); the actual remaining useful life, i.e., the actual time elapsed from a model prediction to the actual maintenance (e.g., if the model predicts at time t1, and maintenance is forced due to a fault 30 days later, then the actual remaining useful life is 30 days); and the actual health status, i.e., the technician's final judgment on the condition of the automotive connector during maintenance (e.g., severe oxidation, poor contact).

[0059] Step SS322: Optimize the parameters of the classification model and / or regression model in the reliability assessment model using the incremental dataset.

[0060] In some embodiments, the parameters of the classification model and / or regression model in the reliability assessment model are periodically optimized using the incremental dataset.

[0061] Specifically, a scheduled task can be set up (e.g., early morning of the first week of each month) to automatically check whether the incremental dataset accumulated over the past month has reached a preset minimum number of samples (e.g., 100 new samples). If it has, the retraining process for the classification and regression models can be automatically started. For example, the original training set can be merged with the new incremental dataset to perform incremental learning on the classification and regression models.

[0062] In some embodiments, the parameters of the classification model and / or regression model in the reliability assessment model are triggered for optimization using the incremental dataset.

[0063] Specifically, the predictive performance of the reliability assessment model is monitored in real time. Optimization is triggered immediately when one of the following conditions is detected: (1) The average error between the model-predicted remaining useful life of any batch of automotive connectors and the actual remaining useful life continues (e.g., for 10 consecutive seconds or within a 1-minute time window) to exceed an error threshold. The error threshold can be determined based on the mean and standard deviation of historical prediction errors, or based on the current remaining useful life (e.g., 5% to 30% of the current remaining useful life), or periodically updated based on prediction error data within the most recent time window.

[0064] (2) Compare the current health feature vector with the historical health feature vector. When a significant change in the data distribution is found, it indicates that the model may no longer be applicable to the current working conditions.

[0065] (3) Within a short period of time (e.g., one week), a large number of model warning signals or inaccurate prediction feedback manually marked by technicians are generated. Step SS323: Update the current health status level and remaining useful life of the automotive connector using the optimized reliability assessment model. In actual implementation, the early degradation judgment method of the reliability assessment model includes steps SS331 to SS333.

[0066] Step SS331: Input the external operating parameters and the contact resistance as boundary conditions into the digital twin model to calculate the theoretical temperature field and theoretical stress field.

[0067] Specifically, the heat generation rate Q inside the automotive connector is calculated based on the current load and contact resistance. Using the heat generation rate Q as the heat source and the ambient temperature as the external boundary condition, the theoretical temperature field inside the entire automotive connector is calculated by solving the heat conduction equation. Using the theoretical temperature field as input, and simultaneously using the spectrum and acceleration of environmental vibration as mechanical boundary conditions, the dynamic response is calculated to obtain the theoretical stress field.

[0068] Step SS332: Calculate the temperature deviation between the theoretical temperature field and the internal temperature distribution. If the temperature deviation is continuously greater than a preset temperature threshold for 10 consecutive seconds or within a 1-minute time window, determine that the automotive connector has experienced early degradation and trigger an early warning.

[0069] Specifically, T threshold =T+k·T0, where T threshold Here, T is the temperature threshold, T is the base temperature, T0 is the theoretical temperature rise, and k is an adjustment coefficient. Further adjustments to k can be made dynamically based on model uncertainty and measurement noise.

[0070] Step SS333: Calculate the stress deviation between the theoretical stress field and the mechanical stress at the contact interface. If the stress deviation is greater than a preset stress threshold for an extended period (e.g., for 10 consecutive seconds or within a 1-minute time window), it is determined that the automotive connector has experienced early degradation and an early warning is triggered.

[0071] Specifically, the stress threshold can be dynamically set and can be determined based on the theoretical stress field and / or the mechanical stress at the contact interface.

[0072] In some embodiments, step SS3 of the present invention includes steps SS341 to SS344.

[0073] Step SS341: Input the health feature vector into the reliability assessment model to generate the current health status level, predicted remaining lifespan, and confidence interval of the automotive connector. Specifically, firstly, the health feature vector is simultaneously input into the first layer of the reliability assessment model; the first layer includes parallel temporal feature extraction branches and statistical feature extraction branches, used to extract dynamic temporal patterns and static statistical features from the health feature vector, respectively, and output the corresponding first and second hidden vectors; through an attention fusion layer, the first and second hidden vectors are dynamically weighted and summed to generate a fused feature vector.

[0074] Next, the fused feature vector is input into the second layer network of the reliability assessment model; the classifier in the second layer network outputs the current health status level of the vehicle connector based on the fused feature vector. The quantile regressor in the second layer network outputs multiple quantile values ​​of the remaining useful life of the automotive connector based on the fused feature vector. The 50th quantile value is used as the predicted value of the remaining useful life, and the preset upper and lower quantile values ​​constitute the confidence interval of the predicted value.

[0075] Step SS3412: Compare the predicted value of the remaining lifetime and its confidence interval with the reference value of the remaining lifetime.

[0076] Specifically, based on the current load and contact resistance values ​​obtained from the health feature vector, a physical-based remaining lifetime reference value is calculated using a preset empirical model; the predicted remaining lifetime value is then compared with the physical-based remaining lifetime reference value.

[0077] Step SS343: In response to the reference value of the remaining lifetime being within the confidence interval, the predicted value of the remaining useful lifetime is taken as the remaining effective lifetime.

[0078] Step SS344: In response to the reference value of the remaining lifetime being outside the confidence interval, the remaining effective lifetime is determined by weighting the predicted value of the remaining lifetime with the reference value of the remaining lifetime.

[0079] Specifically, the remaining effective lifespan includes the following formula: W1∝1 / (C high -C low ); W2 = 1 - W1; RUL = (W1·RUL1 + W2·RUL2); In the formula, RUL is the remaining effective lifetime; W1 is the weight of the reference value of the remaining lifetime; W2 is the weight of the predicted value of the remaining lifetime; RUL1 is the reference value of the remaining lifetime; and RUL2 is the predicted value of the remaining lifetime.

[0080] Secondly, the automotive connector reliability testing system of the present invention includes: an acquisition module, an extraction module, and a testing module.

[0081] The acquisition module acquires the internal micro-state parameters and external operating condition parameters of the automotive connector in real time. The extraction module preprocesses the internal micro-state parameters and external operating condition parameters to extract the health feature vector of the automotive connector. The detection module, based on a trained reliability assessment model and the health feature vector, evaluates the current health status level of the automotive connector and predicts its remaining useful life.

[0082] The specific implementation of the above functional modules refers to the specific steps of the above detection method.

[0083] Thirdly, the computer-readable storage medium of the present invention stores a computer program that, when executed by at least one processor, implements the steps of the above-described detection method.

[0084] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the invention without departing from the principles and spirit of the invention, and all such changes should fall within the protection scope of the claims of the present invention.

Claims

1. A method for testing the reliability of automotive connectors, characterized in that, Includes the following steps: Real-time acquisition of the internal microscopic state parameters and external operating condition parameters of the automotive connector; The internal micro-state parameters and external operating condition parameters are preprocessed to extract the health feature vector of the automotive connector; Based on the trained reliability assessment model, the current health status level of the automotive connector is evaluated and the remaining useful life is predicted according to the health feature vector.

2. The method for testing the reliability of automotive connectors according to claim 1, characterized in that, The internal microstate parameters include one or more combinations of contact resistance, contact interface mechanical stress, and internal temperature distribution. The external operating parameters include one or more combinations of current load, ambient temperature, and ambient vibration.

3. The method for testing the reliability of automotive connectors according to claim 2, characterized in that, The process of obtaining the internal microscopic state parameters of the automotive connector includes: The contact resistance between the pins and sockets of the automotive connector is dynamically measured using a four-wire micro-resistance measurement method. Mechanical stress at the contact interface is monitored by a piezoelectric thin film sensor or fiber optic grating sensor mounted on the automotive connector. By using an array of multiple temperature sensors arranged inside the automotive connector, temperature data from multiple internal points is acquired to form an internal temperature distribution.

4. The method for testing the reliability of automotive connectors according to claim 1, characterized in that, The method for extracting the health feature vector includes: Time-domain and frequency-domain features are extracted from the contact resistance, including the resistance mean, resistance variance, and resistance trend slope. Stress features are extracted from the mechanical stress at the contact interface, and the stress features include stress frequency and stress energy. Temperature features are extracted from the internal temperature distribution, including temperature peaks, temperature gradients, and temperature rise rates. The features of the extracted internal micro-state parameters are fused with the external operating condition parameters to form a multi-dimensional health feature vector.

5. The method for testing the reliability of automotive connectors according to claim 1, characterized in that, The training method for the reliability assessment model includes: Obtain a historical dataset, which includes the historical health feature vector of the vehicle connector throughout its entire life cycle, as well as the historical health status level label and historical remaining useful life label corresponding to each historical health feature vector; The classification model is trained by using the historical health feature vector as input and the corresponding historical health status level label as the expected output. The regression model is trained by using the historical health feature vector as input and the corresponding historical remaining useful life label as the expected output. The classification model and the regression model are integrated to obtain the reliability assessment model.

6. The method for testing the reliability of automotive connectors according to claim 1, characterized in that, The optimization method for the reliability assessment model includes: Collect real-time data, evaluation results, prediction results, and actual maintenance records of the automotive connector to form an incremental dataset; Using the incremental dataset, the parameters of the classification model and / or regression model in the reliability assessment model are optimized; The current health status level and remaining useful life of the automotive connector are updated using the optimized reliability assessment model.

7. The method for testing the reliability of automotive connectors according to claim 4, characterized in that, The early degradation detection method of the reliability assessment model includes: The external operating parameters and the contact resistance are input as boundary conditions into the digital twin model to calculate the theoretical temperature field and the theoretical stress field. The temperature deviation between the theoretical temperature field and the internal temperature distribution is calculated. If the temperature deviation continues to exceed a preset temperature threshold, the automotive connector is determined to have experienced early degradation and a first warning is triggered. The stress deviation between the theoretical stress field and the mechanical stress at the contact interface is calculated. If the stress deviation continues to exceed a preset stress threshold, the automotive connector is determined to have experienced early degradation, and a second warning is triggered.

8. The method for testing the reliability of automotive connectors according to claim 4, characterized in that, The method based on the trained reliability assessment model, according to the health feature vector, assesses the current health status level of the automotive connector and predicts its remaining useful life, including: The health feature vector is input into the reliability assessment model to generate the current health status level, predicted remaining life, and confidence interval of the automotive connector. The predicted value of the remaining lifetime and its confidence interval are compared with the reference value of the remaining lifetime; In response to the reference value of the remaining lifetime being within the confidence interval, the predicted value of the remaining useful lifetime is taken as the remaining effective lifetime. In response to the reference value of the remaining lifetime being outside the confidence interval, the remaining effective lifetime is determined by weighting the predicted value of the remaining lifetime with the reference value of the remaining lifetime.

9. A reliability testing system for automotive connectors, characterized in that, include: The acquisition module is used to acquire the internal microscopic state parameters and external operating condition parameters of the automotive connector in real time. The extraction module is used to preprocess the internal microscopic state parameters and external operating condition parameters to extract the health feature vector of the automotive connector. The detection module is used to assess the current health status level of the automotive connector and predict its remaining useful life based on the health feature vector, using a trained reliability assessment model.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by at least one processor, implements the steps of the detection method according to any one of claims 1 to 8.

Images