Automobile circuit discharge monitoring system

Abstract

The invention discloses an automobile circuit discharge monitoring system, and belongs to the technical field of circuit monitoring. The technical problems that in an existing scheme, tiny electric leakage is difficult to recognize, and a high-voltage system monitoring blind area exists are solved. The high-pressure system realizes sensitive monitoring through curvature correlation, and the low-pressure system realizes stable monitoring through flexible tracking, so that the problem of poor adaptability of the traditional symmetric threshold to the high-pressure and low-pressure systems can be effectively solved; insulation resistance constraint is combined with cross validation, so that the false alarm rate can be effectively reduced; model parameters are iteratively updated based on newest sampling data, the dynamic adaptability of threshold adjustment is ensured, and the environment change of the whole life cycle of the automobile can be covered; through feature extraction, weight optimization, error compensation, early warning generation and blind area enhancement, a complete parameter adaptive update closed loop is formed, and the accuracy of the model in the whole life cycle can be ensured. Through cooperative operation of grading, positioning and blind area strengthening, reliability and accuracy of existing fault alarm are improved.

Description

Technical Field

[0001] This invention relates to the field of circuit monitoring technology, and more specifically to an automotive circuit discharge monitoring system. Background Technology

[0002] Automotive circuit discharge monitoring refers to the technical means of detecting and analyzing the current consumption of a vehicle's electrical system in its dormant or operating state in order to identify abnormal discharge or leakage phenomena. Its core objectives include static discharge monitoring, dynamic discharge monitoring, and partial discharge detection.

[0003] Existing automotive circuit discharge monitoring solutions suffer from limitations. Traditional multimeters cannot accurately detect static currents of less than 50mA, especially for hidden leakage caused by relay sticking or aging wiring. Partial discharge phenomena in the high-voltage system of new energy vehicles require a sensitivity of 0.1mV to be detected, which existing PD testers cannot meet, resulting in difficulties in identifying minute leakage. In addition, the BMS has a delay in monitoring the voltage and temperature of individual battery cells, and the real-time performance of high-frequency signal acquisition is insufficient, leading to delayed insulation fault warnings and blind spots in high-voltage system monitoring. Summary of the Invention

[0004] The purpose of this invention is to provide an automotive circuit discharge monitoring system to solve the technical problems of difficulty in identifying minute leakage currents and blind spots in high-voltage system monitoring in existing solutions.

[0005] The objective of this invention can be achieved through the following technical solutions: A vehicle circuit discharge monitoring system, comprising: Leakage current characteristic parameter processing and identification module: Real-time electrical parameters of the vehicle circuit are collected by multi-source sensors, and leakage current characteristic parameters are identified online based on the improved recursive least squares method. The leakage current characteristic parameters include baseline drift current and dynamic impedance change rate. Model construction multi-layer monitoring and processing module: Construct an asymmetric dynamic monitoring model, adaptively adjust the monitoring thresholds of the high-voltage system and the low-voltage system based on leakage current characteristic parameters. The high-voltage system adopts a curvature-correlated sensitivity field, while the low-voltage system maintains flexible threshold tracking. An insulation resistance constraint mechanism is designed to prevent false overload alarms. The parameter adaptive update processing module dynamically adjusts the weight matrix and compensation terms of the monitoring model based on leakage current characteristic parameters, and generates differentiated early warning signals. The early warning signals include the location coordinates of minor leakage current and the fault level of high-voltage blind zone.

[0006] Preferably, the baseline drift current is estimated based on the improved recursive least squares method, involving the following expression: ; ;in, The baseline drift current is the estimated value at time t; It is an adaptive covariance matrix; Forgetting factor; This is the current-temperature coupling matrix. ; Coupling matrix The transpose of the matrix; This is the real-time current after preprocessing.

[0007] Preferably, based on the baseline drift current residual and the insulation resistance gradient, the sensitivity to small impedance changes is enhanced by a fractional derivative term, involving the following expression: ; ;in, t represents the rate of change of dynamic impedance at time t; A is the gain coefficient; N is the sliding window size; These are fractional weighting coefficients; The impedance residual at time t; is the nominal voltage of the circuit; k is the summation index variable, k=1, 2, ..., N; It represents the insulation resistance.

[0008] Preferably, when constructing the asymmetric dynamic monitoring model, the input and output variables of the asymmetric monitoring framework, as well as the baseline parameters for initializing the high-pressure and low-pressure systems, are defined, and the relevant expressions are as follows: ;in, , These are the rates of change of dynamic impedance for high-voltage and low-voltage systems, respectively. These are the real-time monitoring thresholds for high-voltage systems and low-voltage systems, respectively.

[0009] Preferably, the monitoring threshold is dynamically adjusted based on the nonlinear characteristics of the leakage current signal to enhance the sensitivity to minute leakage currents. The relevant expression is as follows: ; ;in, This serves as the reference threshold for high-voltage systems. This is the curvature sensitivity coefficient; Let be the curvature of the leakage signal at time t; This represents the increment of the rate of change of impedance in the high-voltage system; This is the sampling interval.

[0010] Preferably, based on the slow variation characteristics of the baseline drift current, threshold fluctuations are smoothed to meet the low-noise requirements of low-voltage systems, and the relevant expression is: ;in, For smoothing coefficients; This is the current threshold conversion coefficient; This is the minimum threshold for low-voltage systems.

[0011] Preferably, when designing an insulation resistance constraint mechanism to prevent false overload alarms, the insulation resistance is measured in real time. As a hard constraint to prevent threshold misadjustment due to load fluctuations, the relevant expression is: when hour: ;in, This is the minimum permissible value for insulation resistance; For different constraint coefficients.

[0012] Preferably, based on the preprocessed feature vector The fault identification sensitivity is optimized by dynamically adjusting the weight matrix through recursive least squares method, including initialization, prediction error calculation, gain vector update, weight matrix update and covariance matrix update.

[0013] Preferably, a temperature and insulation aging compensation term is introduced to correct system errors not covered by the weight matrix. The calculation expression for the compensation term is as follows: ;in, For compensation amount; These are the aging compensation coefficient and the temperature compensation coefficient, respectively. The initial insulation resistance; The reference temperature is T(t); the ambient temperature is T(t). compensation amount With weight matrix Superimpose to obtain the final threshold : .

[0014] Preferably, when generating differentiated early warning signals, a dynamic threshold is used. Based on fault characteristics, a warning signal containing fault level and location coordinates is generated; the reliability of the fault level is improved by weighting the signal fluctuation.

[0015] Compared to existing solutions, the beneficial effects achieved by this invention are: This invention achieves real-time updating of baseline drift current by using a forgetting factor and a temperature coupling matrix, adapting to the time-varying characteristics of automotive circuits and improving dynamic tracking capability; the design of fractional-order differential weights can effectively improve the response sensitivity of dynamic impedance change rate to weak leakage current, solving the problem of delayed identification of slowly deteriorating faults in traditional methods.

[0016] The high-voltage system in this invention achieves sensitive monitoring through curvature correlation, while the low-voltage system achieves stable monitoring through flexible tracking. This effectively solves the problem of poor adaptability of traditional symmetrical thresholds to high and low voltage systems. Combining insulation resistance constraints with cross-validation can effectively reduce the false alarm rate. Iteratively updating model parameters based on the latest sampled data ensures the dynamic adaptability of threshold adjustment, which can cover environmental changes throughout the entire life cycle of a vehicle.

[0017] This invention forms a complete parameter adaptive update closed loop through feature extraction, weight optimization, error compensation, early warning generation, and blind zone reinforcement, which can ensure the accuracy of the model throughout its entire life cycle; through the coordinated operation of level, location, and blind zone reinforcement, the reliability and accuracy of existing fault alarms are improved. Attached Figure Description

[0018] The invention will now be further described with reference to the accompanying drawings.

[0019] Figure 1 This is a flowchart illustrating the operation of an automotive circuit discharge monitoring system according to the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] like Figure 1 As shown, the present invention is an automotive circuit discharge monitoring system, comprising: Leakage current characteristic parameter processing and identification module: This module collects real-time electrical parameters of the vehicle's circuitry using multi-source sensors and identifies leakage current characteristic parameters online based on an improved recursive least squares method. These leakage current characteristic parameters include baseline drift current and dynamic impedance change rate. Specific steps include: Multi-dimensional electrical parameters of automotive circuits, including low-voltage circuits and high-voltage systems, are collected through a distributed sensor network. When acquiring data from the low-voltage circuit, high-precision Hall current sensors are deployed at the ECU power supply terminal and wiring harness connectors to collect real-time current signals. ; When acquiring data from the high-voltage system, insulation resistance sensors and temperature sensors are installed at the battery pack, motor controller, and charging interface to collect insulation resistance data. With respect to ambient temperature T(t); The collected data is preprocessed to obtain a normalized data matrix of synchronization timestamps. ; Preprocessing, including but not limited to denoising, error elimination, and time alignment, such as using wavelet thresholding to eliminate electromagnetic interference, correcting temperature drift error through Kalman filtering, and using PTP protocol or linear interpolation to achieve time alignment, are all existing conventional technical solutions. The specific implementation steps will not be elaborated here. Baseline drift current is estimated using an improved recursive least squares method, involving the following expression: ; ;in, The baseline drift current is the estimated value at time t; It is an adaptive covariance matrix; This is the forgetting factor, with a value range of 0.96-0.98. The default value can be the minimum value. This is the current-temperature coupling matrix. ; Coupling matrix The transpose of the matrix; This is the real-time current after preprocessing; Based on the baseline drift current residual and insulation resistance gradient, the sensitivity to small impedance changes is enhanced by a fractional derivative term, involving the following expression: ; ;in, t represents the rate of change of dynamic impedance at time t; A is the gain coefficient, ranging from 1.2 to 1.4, with the default value being the minimum; N is the sliding window size. This is the fractional weighting coefficient, with a value range of 0.6-0.8. The default value can be the minimum value. The impedance residual at time t; is the nominal voltage of the circuit; k is the summation index variable, k=1,2,…,N.

[0022] In this embodiment of the invention, by collecting current, insulation resistance, and temperature in a coordinated manner, the blind spots of single-parameter monitoring can be eliminated, and the integrity of leakage current characteristics can be improved. By using the forgetting factor and temperature coupling matrix, the baseline drift current can be updated in real time, adapting to the time-varying characteristics of automotive circuits, such as start-stop conditions and temperature fluctuations, thus improving dynamic tracking capabilities. The design of fractional-order differential weights can effectively improve the response sensitivity of dynamic impedance change rate to weak leakage current, solving the problem of delayed identification of slowly deteriorating faults in traditional methods.

[0023] The model constructs a multi-layered monitoring and processing module: An asymmetric dynamic monitoring model is built, adaptively adjusting the monitoring thresholds for the high-voltage and low-voltage systems based on leakage current characteristic parameters. The high-voltage system employs a curvature-correlated sensitivity field, while the low-voltage system maintains flexible threshold tracking. An insulation resistance constraint mechanism is designed to prevent false overload alarms. Specific steps include: When constructing the asymmetric dynamic monitoring model, the input and output variables of the asymmetric monitoring framework are defined, as well as the baseline parameters for initializing the high-pressure and low-pressure systems. The relevant expressions are as follows: ;in, , These are the rates of change of dynamic impedance for high-voltage and low-voltage systems, respectively. These are the real-time monitoring thresholds for high-voltage systems and low-voltage systems, respectively. When adaptively adjusting the monitoring thresholds of high-voltage and low-voltage systems based on leakage current characteristic parameters, the monitoring thresholds are dynamically adjusted based on the nonlinear characteristics of the leakage current signal. The nonlinear characteristic is curvature, which enhances the sensitivity to minute leakage currents. The relevant expression is: ; ;in, This serves as the reference threshold for high-voltage systems. This is the curvature sensitivity coefficient, with a value range of 0.6-0.8. The default value can be the minimum value. Let be the curvature of the leakage signal at time t; This represents the increment of the rate of change of impedance in the high-voltage system; This is the sampling interval time; It should be noted that, through curvature By capturing the accelerated change characteristics of leakage current signals, the threshold can be dynamically reduced. To trigger early warnings.

[0024] Based on the slow variation characteristics of the baseline drift current, a time window weighted algorithm is used to smooth the threshold fluctuation, adapting to the low noise requirements of low-voltage systems. The relevant expression is: ;in, This is a smoothing coefficient, with a value range of 0.8-0.95. The default value can be the minimum value. This is the current threshold conversion coefficient; This is the minimum threshold for low-voltage systems; It should be noted that, unlike high-voltage systems which prioritize sensitivity, low-voltage systems utilize a smoothing coefficient. and minimum threshold Suppressing noise interference while coordinating baseline drift current This enables adaptive tracking.

[0025] When designing an insulation resistance constraint mechanism to prevent false overload alarms, the insulation resistance is measured in real time. As a hard constraint to prevent threshold misadjustment due to load fluctuations, the relevant expression is: when hour: ;in, This is the minimum permissible value for insulation resistance; The default values ​​for the different constraint coefficients are 1.5 and 1.2, respectively. It should be noted that, through Increase threshold This can prevent normal load fluctuations from being misjudged as leakage.

[0026] Cross-validation is performed using thresholds from both high-voltage and low-voltage systems to eliminate false alarm triggering conditions from a single system. The warning triggering condition is as follows: ; where ∩ is the logical AND operator; It should be noted that by combining the results of high voltage threshold, low voltage threshold, and insulation resistance constraints, multi-condition cross-validation can effectively reduce the false alarm rate.

[0027] When iteratively updating model parameters based on the latest sampled data, the high-voltage system updates every 10ms. The low-voltage system updates every 100ms. Insulation resistance constraint verification every 500ms This ensures the dynamic adaptability of threshold adjustment.

[0028] In this embodiment of the invention, the high-voltage system achieves sensitive monitoring through curvature correlation, while the low-voltage system achieves stable monitoring through flexible tracking. This effectively solves the problem of poor adaptability of traditional symmetrical thresholds to high and low voltage systems. Combining insulation resistance constraints with cross-validation can effectively reduce the false alarm rate. Iteratively updating model parameters based on the latest sampled data ensures the dynamic adaptability of threshold adjustment, which can cover environmental changes throughout the entire life cycle of the vehicle.

[0029] The parameter adaptive update processing module dynamically adjusts the weight matrix and compensation terms of the monitoring model based on leakage current characteristic parameters, generating differentiated early warning signals. These signals include the location coordinates of minor leakage currents and the fault level in high-voltage blind zones. Specific steps include: Key feature parameters of the leakage current signal are extracted and standardized from the raw signal acquired through a distributed sensor network. Standardization specifically involves normalizing the feature parameters to the [0,1] interval to eliminate the influence of dimensions. This is a conventional technique, and the specific implementation steps are not detailed here. The preprocessed feature vector is output. ;in, The feature parameter vector at time t contains the dynamic and static characteristics of the leakage current signal; Based on the preprocessed feature vectors The fault identification sensitivity is optimized by dynamically adjusting the weight matrix through recursive least squares method, including initialization, prediction error calculation, gain vector update, weight matrix update and covariance matrix update. During initialization, covariance matrix Forgetting factor ; The expression for calculating the prediction error is: ; The expression for updating the gain vector is: ; The expression for updating the weight matrix is: ; The expression for updating the covariance matrix is: ;in, Let be the weight matrix at time t, containing the weight coefficients of the four characteristic parameters: for Weight, for Weight, for Weight, for Weight; It is the covariance matrix; For prediction error, i.e., actual Deviation from model predictions; Output the optimized weight matrix This provides a dynamic weighting basis for the calculation of compensation items; A temperature and insulation aging compensation term is introduced to correct for system errors not covered by the weight matrix. The calculation expression for the compensation term is as follows: ;in, For compensation amount; These are the aging compensation coefficient and the temperature compensation coefficient, with default values ​​of 0.3 and 0.005, respectively. The initial insulation resistance; Reference temperature; compensation amount The final threshold is obtained by superimposing it with the weight matrix. : ; It should be noted that introducing temperature and insulation aging compensation terms can effectively reduce threshold errors caused by temperature drift and insulation aging, and improve the long-term stability of the model.

[0030] When generating differentiated early warning signals, based on dynamic thresholds Based on the fault characteristics, a warning signal containing the fault level and location coordinates is generated. The expression for fault level classification is as follows: Among them, levels 1, 2, and 3 represent mild, moderate, and severe, respectively. The expression involved in multi-sensor signal time difference positioning is: ;in, The two-dimensional spatial coordinates of the leakage fault point; The independent variable that minimizes the following expression; i is the sensor index, and n is the total number of sensors; Let be the rate of change of dynamic impedance measured by the i-th sensor at time t; The reference dynamic impedance change rate under fault-free conditions; For areas with weak signals in the high-voltage system, such as connectors and busbars, when improving the reliability of fault levels by weighting signal fluctuations, the blind zone locations of the high-voltage system are predefined, such as the positive and negative connectors of the battery pack and the motor controller interface, and the corresponding sensor numbers are marked. The expression involved in performing grade correction is: ;in, The standard deviation of the signal in the blind zone; To enhance the weighting of weak signal fault levels, the default value is 1; This is the baseline value for the signal standard deviation in the normal region; This is the initial fault level.

[0031] It should be noted that, based on the initial fault level and the standard deviation of the blind zone signal, the enhanced fault level is output, which can avoid the underreporting of weak signals and solve the problem of the traditional model being insensitive to weak signal areas.

[0032] In this embodiment of the invention, a complete parameter adaptive update closed loop is formed through feature extraction, weight optimization, error compensation, early warning generation, and blind zone reinforcement, which can ensure the accuracy of the model throughout its entire life cycle; through the coordinated operation of level, location, and blind zone reinforcement, the reliability and accuracy of existing fault alarms are improved.

[0033] In the several embodiments provided by this invention, it should be understood that the disclosed system can be implemented in other ways. For example, the embodiments of the invention described above are merely illustrative; for example, the division of modules is only a logical functional division, and there may be other division methods in actual implementation.

[0034] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0035] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated module can be implemented in hardware or in the form of hardware plus software functional modules.

[0036] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the essential characteristics of the present invention.

[0037] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An automotive circuit discharge monitoring system, characterized by, Comprise: Leakage characteristic parameter processing identification module: collect real-time electrical parameters of automobile circuit through multi-source sensors, identify leakage characteristic parameters based on improved recursive least squares method online, leakage characteristic parameters include baseline drift current and dynamic impedance change rate; Model construction multi-layer supervision processing module: construct asymmetric dynamic monitoring model, adjust monitoring threshold of high-voltage system and low-voltage system based on leakage characteristic parameters, high-voltage system adopts curvature correlation sensitivity field, low-voltage system keeps flexible threshold tracking, and designs insulation resistance constraint mechanism to prevent overload false alarm; Parameter adaptive update processing module: dynamically adjust weight matrix and compensation term of monitoring model based on leakage characteristic parameters, generate differentiated early warning signal, early warning signal contains micro leakage positioning coordinates and high-voltage blind area fault grade.

2. The automotive circuit discharge monitoring system of claim 1, wherein The baseline drift current is estimated based on an improved recursive least square method, and the expression involved is: ; ; wherein is a baseline drift current estimate at time t; is an adaptive covariance matrix; is a forgetting factor; is a current-temperature coupling matrix, ; is a coupling matrix is a transpose matrix of the coupling matrix is a pre-processed real-time current.

3. The automotive circuit discharge monitoring system of claim 2, wherein Based on baseline drift current residual and insulation resistance gradient, enhance the sensitivity of micro impedance change through fractional order differential term, the expression involved is: ； ; wherein, is the dynamic impedance change rate at time t; A is the gain coefficient; N is the sliding window size; is the fractional order weight coefficient; is the impedance residual error at time t; is the circuit nominal voltage; k is the summation index variable, k = 1, 2, …, N; is the insulation resistance.

4. The automotive circuit discharge monitoring system of claim 3, wherein, When constructing the asymmetric dynamic monitoring model, define the input and output variables of the asymmetric monitoring framework, and initialize the reference parameters of the high-voltage and low-voltage systems, the expression involved is: ; wherein, , are high-voltage and low-voltage system dynamic impedance change rates, respectively; are high-voltage system real-time monitoring threshold and low-voltage system real-time monitoring threshold, respectively.

5. The automotive circuit discharge monitoring system of claim 4, wherein, Based on the nonlinear characteristics of leakage signal, dynamically adjust the monitoring threshold to enhance the sensitivity to micro leakage, the expression involved is: ； ; wherein, is a high voltage system reference threshold value; is a curvature sensitivity coefficient; is a curvature of the leakage signal at time t; is an increment of the high voltage system impedance change rate; is a sampling interval time.

6. The automotive circuit discharge monitoring system of claim 5, wherein, Based on the slow change characteristics of baseline drift current, smooth the threshold fluctuation, adapt to the low noise demand of low voltage system, the expression involved is: ; wherein, is a smoothing coefficient; is a current threshold conversion coefficient; is a minimum threshold for low voltage system.

7. The automotive circuit discharge monitoring system of claim 6, wherein, When designing the insulation resistance constraint mechanism to prevent overload false alarms, through the real-time measurement value of insulation resistance As a hard constraint, prevent threshold misadjustment due to load fluctuations, the expression involved is: When Time: ; wherein, is the minimum permissible value of the insulation resistance; are different constraint coefficients.

8. The automotive circuit discharge monitoring system of claim 7, wherein, based on the pre-processed feature vectors The weight matrix is dynamically adjusted by the recursive least square method to optimize the fault identification sensitivity, including initialization, prediction error calculation, gain vector update, weight matrix update and covariance matrix update.

9. The automotive circuit discharge monitoring system of claim 8, wherein, Introduce temperature and insulation aging compensation term to correct the system error not covered by the weight matrix, the calculation expression of the compensation term is: ; wherein is a compensation quantity; are an aging compensation coefficient and a temperature compensation coefficient, respectively; is an initial insulation resistance; is a reference temperature; and T(t) is an ambient temperature; compensate the amount with the weight matrix superimposed to obtain a final threshold : 。 10. The automotive circuit discharge monitoring system of claim 9, wherein, Generating a differentiated early warning signal based on dynamic thresholds Generating an early warning signal containing a failure level and localization coordinates based on the failure characteristics; the reliability of the failure level can be improved by weighting the signal volatility.

Images