Electronic control module health state evaluation method based on parameter incidence matrix

Abstract

The invention relates to the technical field of circuit device monitoring, and discloses an electronic control module health state evaluation method based on a parameter incidence matrix, which comprises the following steps: acquiring a bus current, a bus voltage and an energy change sequence in a charging stage, and calculating a derivative response track of transient power relative to voltage square; the derivative response track is mapped to a multi-dimensional parameter space containing power, voltage and charging and discharging time sequence parameters, and a parameter incidence matrix reflecting the dynamic impedance evolution process in the energy storage unit is constructed; characteristic defects are determined by calculating curvature characteristic differences of an actually measured track and a standard healthy track, the method is based on normalization processing of an energy conversion dimension, interference of power supply ripples on impedance characteristic extraction is effectively isolated, hidden faults induced by packaging microcracks or welding spot cavities can be accurately recognized, and the method is suitable for large-scale popularization and application. And a digital health evaluation benchmark covering the whole life cycle of the module is established.

Description

Technical Field

[0001] This invention relates to a method for assessing the health status of electronic control modules based on parameter correlation matrices, belonging to the field of circuit device monitoring technology. Background Technology

[0002] Current industrial power supply and distribution systems typically employ automated testing equipment to discretely monitor static electrical parameters such as voltage, current, and resistance of electronic control modules. This ensures that the electrical parameters of each functional unit meet the factory preset range and maintains consistency across production batches. This steady-state response-based testing method plays a crucial supporting role in improving production line turnover efficiency and ensuring the reliability of basic functions.

[0003] As the requirements for power control precision and transient response stability in special applications such as industrial blasting continue to evolve, the charging and discharging behavior of the energy storage unit inside the controlled module under extremely high voltage gradients has become a core factor determining the overall safety of the system. However, in large-scale manufacturing processes, limited by hardware sampling bandwidth and interference from environmental electromagnetic ripples, existing quasi-static measurement methods struggle to capture microsecond-level transient disturbances induced by physical defects on the surface of components. For example, Chinese invention patent CN115486210A discloses an electronic control module that uses a cylindrical bracket between two substrates to form a closed space, reducing metal contamination during production and preventing short circuits induced by external impurities. Schemes that focus on physical structural protection are passive defenses against external environmental factors and cannot address the intrinsic evolution of energy storage units within the module under dynamic electric field stress. For microsecond-level transient disturbances induced by physical defects such as microcracks in component packaging and voids in solder joints, structural protection schemes struggle to capture the surface characteristics of charge throughput. If attempts are made to improve the defect detection rate simply by increasing the sampling frequency or reducing the parameter tolerance range, it will not only increase the data processing load and hardware manufacturing costs, but also fail to distinguish between normal bus power supply fluctuations and nonlinear impedance jumps caused by interlayer breakdown risks in energy storage capacitors. This can lead to metastable modules with potential failure risks flowing into subsequent processes and threatening the operational safety of the power system.

[0004] Specifically, existing technologies suffer from the following shortcomings: 1. The detection dimension is limited to static threshold judgment, thus failing to identify dynamic coupling failure characteristics in the energy conversion circuit; 2. The high degree of aliasing between environmental electromagnetic noise and intrinsic defect signals of components in the time domain restricts the accuracy of health status assessment; 3. The lack of in-depth extraction of the consistency of energy rheological trajectory makes it difficult to establish a digital quality archive with physical logic depth. Therefore, how to adopt an electronic control module health status assessment method based on parameter correlation matrix, while maintaining the existing hardware architecture, to perform correlation analysis on multi-dimensional parameters in the energy accumulation process and identify discontinuous shifts in the energy response trajectory, becomes the technical problem to be solved by this invention. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A method for assessing the health status of an electronic control module based on a parameter correlation matrix, comprising the following steps: Step S1: During the charging phase of the electronic control module, simultaneously acquire bus current data, bus voltage data, ambient temperature of the electronic control module, and energy change sequence of the energy storage unit within a preset time window. Step S2: Calculate the transient power change based on the bus current data and bus voltage data, and calculate the voltage square change corresponding to the bus voltage data. Generate the response trajectory by calculating the derivative of the transient power change with respect to the voltage square change, so as to extract the charge throughput characteristics induced by the electronic control module under the dynamic electric field gradient. Step S3: Map the response trajectory to a parameter space that includes the power change rate, the squared voltage change rate, and the charge / discharge timing deviation, and construct a parameter correlation matrix that reflects the internal impedance evolution process of the energy storage unit. Step S4: Extract the curvature feature difference between the measured response trajectory and the stored standard health trajectory in the parameter correlation matrix, calculate the normal deviation value of the measured response trajectory relative to the standard health trajectory in the parameter space, and count the time proportion of sampling points whose normal deviation value exceeds the preset safety threshold in the response trajectory. If the time proportion exceeds the preset risk threshold, it is determined that the electronic control module has a characteristic defect and the evaluation result is output.

[0006] Preferably, step S2 includes: aligning the acquired bus current data and bus voltage data in time; using a cyclic data buffer to calculate the transient power change and voltage square change between adjacent sampling points; calculating the in-phase correlation between the transient power change and voltage square change to eliminate the interference of external power supply ripple on the extraction of internal impedance characteristics of the energy storage unit, and generating a response trajectory after eliminating ripple background noise.

[0007] Preferably, step S3 includes: establishing a feature coordinate system with the power change rate on the horizontal axis and the voltage square change rate on the vertical axis; projecting the response trajectory after eliminating ripple background noise onto the feature coordinate system to generate a scatter plot describing the charging consistency of the energy storage unit; extracting the statistical features of the scatter plot in the feature coordinate system, and using the statistical features as feature inputs to the parameter correlation matrix.

[0008] Preferably, in step S2, the calculation rule followed by the response trajectory is: Where λ is the dynamic response coefficient and ΔP is the transient power increment between adjacent sampling points. This represents the squared voltage increment between the corresponding sampling points.

[0009] Preferably, the calculation of the normal deviation value in step S4 includes: calculating the geometric deviation distance of each set of feature evaluation vectors in the parameter correlation matrix relative to the standard health trajectory; counting the number of abnormal sampling points whose geometric deviation distance exceeds the preset safety threshold during the full charging process; and determining the ratio of the number of abnormal sampling points to the total number of full sampling points as the time proportion.

[0010] Preferably, after step S4, the method further includes: step S5, monitoring the voltage holding slope of the electronic control module at the end of charging; and step S6, generating a health evaluation benchmark covering the life cycle of the electronic control module based on the nonlinear correlation weight between the voltage holding slope and the impedance evolution characteristics in the parameter correlation matrix.

[0011] Preferably, in step S1, acquiring bus current data and bus voltage data includes: using a multi-channel synchronous sampling interface to acquire data from multiple electronic control modules in parallel; and during the dynamic process of the energy storage unit charging to a preset voltage range, acquiring current and voltage signals at a sampling frequency of not less than 500kHz to capture microsecond-level transient leakage waveforms.

[0012] Preferably, after step S6, the method further includes: step S7, associating the evaluation results with the production batch information of the electronic control module; and step S8, uploading the associated health data to the production management system to perform quality risk traceability of the electronic control module.

[0013] Preferably, the output of the evaluation results includes: outputting a qualified instruction indicating that the electronic control module is in a healthy state; outputting a warning signal indicating that the electronic control module has a characteristic defect risk, and sending an instruction to interrupt the communication function of the electronic control module.

[0014] Preferably, step S1 further includes: performing temperature compensation on the energy change sequence based on the acquired ambient temperature; and adjusting the dynamic range of the current acquisition channel based on the gradient change of the bus current data to maintain constant measurement accuracy under different current gradients during the charging phase.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. In the health status assessment of electronic control modules, the characteristic consistency of energy rheological trajectories is utilized to accurately capture hidden physical damage inside the controlled modules. During the charging process of the energy storage unit, by synchronously acquiring bus current and voltage signals and calculating the derivative response of power change with voltage square change in real time, the system can identify transient charge throughput anomalies induced by microcracks in components or voids in solder joints. This dynamic monitoring mechanism based on physical mechanisms enables weak modules that perform normally under static electrical performance testing to be effectively identified due to the discontinuous deviation of their energy response trajectory under electric field gradient excitation, thereby eliminating potential safety risks caused by energy storage instability during blasting operations at the source.

[0016] 2. By dynamically correlating the power dimension and charge distribution dimension, this invention decouples test environment interference and improves the physical reliability of judgment conclusions. It utilizes the derivative characteristics of power change with respect to voltage square change to normalize the first-order fluctuation of the charging circuit in the energy conversion dimension, thereby eliminating the interference of external power supply ripple on the extraction of internal impedance characteristics. This processing method allows the assessment of health status to be directly locked into the dynamic impedance evolution process inside the module, ensuring the cleanliness of detection in complex industrial production environments and avoiding misjudgment or omission caused by small fluctuations in bus voltage.

[0017] 3. Relying on the synergistic effect of multidimensional parameter correlation matrix and trajectory curvature verification, a digital health evaluation benchmark covering the entire life cycle is established. This invention maps the current and voltage sequences obtained by synchronous sampling into a parameter correlation matrix and performs consistency comparison by combining the curvature characteristics of the ideal model trajectory. This constructs a multidimensional evaluation system from single parameter qualification to consistent dynamic response. Under the action of this multi-mechanism coupling, the system can not only monitor the current electrical performance indicators, but also discover metastable structural risks inside the module through the dispersion assessment of trajectory distortion, providing data support with physical and logical depth for subsequent quality traceability and process improvement. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the health status assessment logic of the parameter correlation matrix in this invention. Figure 2 This is a schematic diagram of the hierarchical architecture of the electronic control module health assessment system of the present invention. Detailed Implementation

[0019] The method claimed in this invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The following embodiments are intended for explanation and illustration, and are not intended to limit the scope of protection of this invention.

[0020] A method for assessing the health status of electronic control modules (ECMs) based on parameter correlation matrices is proposed. During the charging phase of the ECMs, a multi-channel synchronous sampling interface is used to acquire bus current data, bus voltage data, ambient temperature of the ECMs, and energy change sequences of energy storage units within a preset time window in parallel from multiple ECMs. Specifically, during the dynamic charging process of the energy storage units to a preset voltage range, the system acquires current and voltage signals at a sampling frequency of 500kHz, capturing microsecond-level transient leakage waveforms. During waveform capture, the core evaluation processor calls internal memory to define a sliding observation window with a length of 256 sampling points, setting the step length between adjacent windows to 64 sampling points, thereby achieving a 75% data overlap rate. This ensures that extremely narrow current pulse transitions lasting only 5 to 10 microseconds can be captured by at least three consecutive windows, avoiding the loss of transient characteristics due to sampling window breaks. Furthermore, based on… The acquired ambient temperature is used to compensate for the energy change sequence, and the dynamic range of the current acquisition channel is adjusted based on the gradient change of the bus current data to maintain constant measurement accuracy under different current gradients during the charging phase. The transient power change is calculated based on the bus current and bus voltage data, and the corresponding square voltage change is calculated. A response trajectory is generated by calculating the derivative of the transient power change with respect to the square voltage change. The charge throughput characteristics induced by the electronic control module under the dynamic electric field gradient are extracted. The acquired bus current and bus voltage data are time-aligned, and the transient power change and square voltage change between adjacent sampling points are calculated using a cyclic data buffer. The in-phase correlation between the transient power change and the square voltage change is calculated to eliminate the interference of external power supply ripple on the extraction of the internal impedance characteristics of the energy storage unit. A response trajectory after eliminating ripple background noise is generated, and the calculation rules followed by the response trajectory are as follows: Where λ is the dynamic response coefficient and ΔP is the transient power increment between adjacent sampling points. This represents the squared voltage increment between the corresponding sampling points.

[0021] In step S2, when generating the response trajectory, the processor retrieves the bus voltage and bus current data of N adjacent sampling points stored in the circular data buffer, and uses the sliding window algorithm to calculate the transient power change ΔP and the voltage square change. The cross-correlation coefficient ρ between sequences is used to determine if the energy accumulation process is affected by external power supply ripple when ρ exceeds the preset ripple correlation threshold of 0.85. The processor then calls a preset second-order Butterworth low-pass filter to reduce noise in the original sampled sequence. The cutoff frequency is set to 5 to 10 times the charging characteristic frequency of the electronic control module. The charging characteristic frequency is defined as the reciprocal of the product of the equivalent capacitance of the energy storage unit and the total resistance of the charging circuit, and is set to 50Hz in this hardware system. Accordingly, the cutoff frequency of the second-order Butterworth low-pass filter is fixed at 400Hz to retain the high-frequency components of the microsecond-level impedance transition while forcibly filtering out the 100Hz ripple noise generated by the bus rectifier circuit and the power frequency interference generated by spatial electromagnetic induction. It also filters out in-phase superimposed background noise and extracts transient transition characteristics induced by internal physical defects in the energy storage unit. The response trajectory follows the calculation formula. Where λ is the dynamic response coefficient and ΔP is the transient power increment between adjacent sampling points. The dynamic response coefficient λ represents the squared voltage increment between corresponding sampling points, and it characterizes the energy throughput rate of the electronic control module under a unit electric field gradient change. The numerical discontinuous fluctuation is used to identify internal conductive paths or surface damage of the dielectric layer.

[0022] The response trajectory is mapped to a parameter space containing the power change rate, voltage squared change rate, and charge / discharge timing deviation, constructing a parameter correlation matrix reflecting the impedance evolution process within the energy storage unit. This matrix is ​​built as a 3-column, 1024-row digital data array. The first column stores the normalized power change rate values ​​in sampling time order, the second column stores the voltage squared change rate values, and the third column stores the charge / discharge timing deviation values ​​in microseconds for each sampling point. The row indices of the matrix directly correspond to 2-microsecond equally spaced sampling periods starting from the charging start point. A fixed-dimensional matrix structure performs logical alignment and correlation analysis on a unified time axis for heterogeneous electrical parameters. A feature coordinate system is established with the power change rate on the horizontal axis and the voltage square change rate on the vertical axis. The response trajectory after eliminating ripple background noise is projected onto this feature coordinate system, forming a scatter plot describing the charging consistency of the energy storage units. Statistical features of the scatter plot in the feature coordinate system are extracted and used as feature inputs to the parameter correlation matrix M. Specifically, when determining the charging / discharging timing deviation, the processor extracts the energy storage unit voltage rise to a preset segmented voltage point. Actual measured time and compared it with the standard health trajectory Reference time for the corresponding voltage point Perform differential calculations to determine the charging / discharging timing deviation. The calculation formula is as follows: ,in, This refers to the charge / discharge timing deviation. This is the actual measured time. Using the reference time, this difference value serves as the time-domain component of the parameter correlation matrix M, and is used to characterize the abnormal energy accumulation rate caused by the increase in leakage current.

[0023] When constructing the parameter correlation matrix M in step S3 and calculating the normal deviation value in step S4, the processor projects the response trajectory onto the power change rate, the voltage squared change rate, and the charge / discharge timing deviation. The constructed three-dimensional Euclidean space is used for each sampling point on the measured trajectory. In the pre-set standard health trajectory Search for the reference point with the shortest Euclidean distance to this point. The normal deviation value is taken as the magnitude of the two vectors. The formula for calculating this deviation value is as follows: ,in, This is the normal deviation value. , and These are the coordinate components of the measured sampling points, , and These are the coordinate components of the reference point corresponding to the standard health trajectory, obtained by statistically analyzing the entire charging process. Exceeding the preset safety threshold The sampling point ratio is used to quantitatively assess the microsecond-level impedance evolution anomalies and their persistence; the curvature feature difference between the measured response trajectory and the stored standard healthy trajectory in the parameter correlation matrix is ​​extracted, the normal deviation value of the measured response trajectory relative to the standard healthy trajectory in the parameter space is calculated, and the time proportion of sampling points exceeding the preset safety threshold in the response trajectory is counted. If the time proportion exceeds the preset risk threshold, the electronic control module is determined to have a characteristic defect and the evaluation result is output. The calculation of the normal deviation value includes: calculating the geometric deviation distance of each set of feature evaluation vectors in the parameter correlation matrix relative to the standard healthy trajectory, counting the number of abnormal sampling points whose geometric deviation distance exceeds the preset safety threshold in the full charging process, and determining the time proportion as the ratio of the number of abnormal sampling points to the total number of full sampling points.

[0024] After determining that the electronic control module has a characteristic defect and outputting the evaluation result, the voltage holding slope of the electronic control module at the end of charging is monitored. Based on the nonlinear correlation weight between the voltage holding slope and the impedance evolution characteristics in the parameter correlation matrix, a digital health evaluation benchmark covering the life cycle of the electronic control module is generated. The nonlinear correlation weight is updated by calculating the offset rate of the eigenvalues ​​of the parameter correlation matrix M with the number of charge and discharge cycles. The digital health evaluation benchmark H is determined by a weighted summation model, and its calculation formula is as follows: In step S6, when determining the digital health assessment benchmark H, the processor retrieves historical data from pre-stored sample sets of different defect degrees and uses a sensitivity analysis algorithm to calculate the feature offset. Maintain slope with voltage The ratio of the sensitivity of the two parameters to the partial derivatives with respect to the preset failure modes is used as a weighting coefficient. and The initial value is adjusted as the number of charge / discharge cycles of the electronic control module increases. The processor monitors the slope of the eigenvalues ​​of the parameter correlation matrix M as a function of the number of cycles, and adjusts the weighting coefficients accordingly. and Online gradient descent correction is used to offset baseline drift caused by natural material aging. The formula for calculating the digital health assessment baseline H is shown below: Where H represents the digital health assessment benchmark. For feature offset, To maintain the voltage slope at the end of charging. and As the weighting coefficient, the benchmark H provides a digital measurement basis for the traceability of quality risks of electronic control modules, and the evaluation results are associated with the production batch information of electronic control modules. The associated health data is uploaded to the production management system to perform the traceability of quality risks of electronic control modules. The output of the evaluation results includes a qualified instruction indicating that the electronic control module is in a healthy state, or an early warning signal indicating that the electronic control module has a characteristic defect risk, and a command to interrupt the communication function of the electronic control module is sent.

[0025] Example 1: In an automated assembly scenario for electronic control modules with high-frequency electromagnetic noise and a single detection cycle of less than 500ms, the voltage ripple of the bus power supply system overlaps with the transient leakage waveform generated by the energy storage unit during the charging phase. This makes it difficult for the quasi-static sampling mode to capture the impedance discontinuous jump caused by micro-cracks in the capacitor package, resulting in modules with potential physical structural damage passing conventional electrical performance tests. The processor uses the acquired bus current data and bus voltage data to construct a parameter correlation matrix M, and generates a response trajectory by performing derivative operations in the power conversion dimension, using the formula... Calculate the dynamic response coefficient λ, where λ is the dynamic response coefficient and ΔP is the transient power increment between adjacent sampling points. To correspond to the squared voltage increment between sampling points, this calculation process utilizes the derivative relationship between the power change and the squared voltage change to decouple the charge throughput characteristics inside the electronic control module from the external voltage fluctuations, thereby enabling the eigenvectors in the parameter space to directly characterize the dynamic impedance evolution law inside the electronic control module.

[0026] After feature extraction from the parameter correlation matrix M, the system identified that the measured statistical feature quantity had a trajectory offset in the feature coordinate system. The measured normal deviation value exceeded the preset safety threshold within a specific sampling time period, and the time proportion of abnormal sampling points met the judgment condition of the preset risk threshold. Based on this, the system determined that the electronic control module was a defective part and executed the instruction blocking of the communication function to prevent the electronic control module from entering the subsequent assembly stage.

[0027] Example 2: To verify the detection stability of the electronic control module health status assessment method based on parameter correlation matrix under strong electromagnetic interference environment, the experiment was established on a multi-channel synchronous sampling physical experimental platform with a 500V range and 0.05% measurement accuracy. The experimental data came from the measured current and voltage sequences of the simulated electronic detonator control module under fast charging cycle. In order to control the processor's computational load while ensuring the capture of transient impedance jumps, the sampling frequency was set to 500kHz based on the trade-off between signal bandwidth and real-time requirements, i.e., the sampling interval was 2μs. Gaussian white noise with a signal-to-noise ratio of 20dB was actively superimposed in the experimental signal source to simulate the real electromagnetic environment background. 200 modules with different degrees of encapsulation microcracks were selected as the test group, and another 200 structurally intact modules were selected as the control group to form a verification sample set covering the physical defect gradient.

[0028] In the stage of obtaining the original physical quantities, for the electronic control module with 5μm-level microcracks in the experimental group, the system extracted its data during charging cycles. The bus current data I(t) and bus voltage data V(t) within the system were observed to be affected by the superposition of power supply ripple and Gaussian white noise. The amplitude fluctuation characteristics were mixed with the 50Hz power frequency interference harmonics in the time domain, resulting in the quasi-static mean judgment method showing that the characteristic parameters of this sample group were within the qualified range within a 100ms window, with a detection rate of 12.3%. The system used a cyclic data buffer to calculate the transient power change ΔP and voltage square change between adjacent sampling points. The response trajectory is generated through derivative operations, specifically using the formula... Calculate the dynamic response coefficient λ, where λ is the dynamic response coefficient and ΔP is the transient power increment between adjacent sampling points. To correspond to the squared voltage increment between sampling points, the data shows that when the charging voltage climbs to 25V, the response trajectory exhibits discontinuous phase jumps in the parameter space. Through full-process evaluation of sample groups with different defect gradients, the experimental group achieved a comprehensive detection rate of 98.6% in samples containing microcracks ranging from 5μm to 20μm. However, the detection rate of the control sample group, which removed the dynamic response coefficient λ calculation step and retained only power change monitoring, dropped to 45.8%, and a false alarm rate of 15.4% was generated due to the influence of bus voltage ripple. When the sampling frequency was reduced from 500kHz to below 100kHz, the system's accuracy in extracting the normal deviation value decreased due to the inability to effectively capture the differential characteristics of charge throughput, resulting in a detection rate drop to 32.4%. This confirms the necessity of the limited sampling frequency and derivative response algorithm in identifying metastable defects. The experimental data demonstrates that through feature extraction of the parameter correlation matrix M, the system can identify latent physical damage induced by encapsulation microcracks, resolving the contradiction between testing efficiency and detection reliability.

[0029] Example 3: This example combines Figures 1 to 2 The method for assessing the health status of electronic control modules based on parameter correlation matrices is explained, such as... Figure 1 As shown, in step S1, during the charging phase of the electronic control module, bus current data, bus voltage data, ambient temperature, and energy change sequence of the energy storage unit within a preset time window are acquired simultaneously. In step S2, the transient power change and voltage square change are calculated based on the bus current and voltage data. The response trajectory is generated through derivative calculation to extract the charge throughput characteristics of the electronic control module under dynamic electric field gradient. Then, in step S3, the response trajectory is mapped to a parameter space containing the power change rate, voltage square change rate, and charge / discharge timing deviation. A parameter correlation matrix reflecting the internal impedance evolution process of the energy storage unit is constructed. Finally, step S4 is executed to calculate the normal deviation value of the measured trajectory relative to the standard healthy trajectory. The percentage of sampling points with deviations exceeding the threshold is statistically analyzed. If the percentage exceeds the risk threshold, a characteristic defect is determined and the evaluation result is output.

[0030] like Figure 2 As shown, the electronic control module in the physical testing terminal collects data through a multi-channel synchronous sampling interface, transmitting bus current, voltage, and ambient temperature signals to the core evaluation processor. Internally, it integrates a data preprocessing layer to perform cyclic data buffer operations and time alignment, a response trajectory generation layer to perform transient power and voltage square derivative calculations, a parameter correlation matrix construction layer to perform multi-dimensional parameter space mapping, and a difference judgment layer to call the standard health trajectory to calculate the normal deviation value and compare curvature features. Finally, the system outputs signals to the control execution end, transmitting the evaluation results and quality traceability data to the production management system on the one hand, and sending early warning signals and interrupt commands to the control bus on the other.

[0031] Example 4: In a batch reliability assessment of special electronic control modules, due to process inconsistencies in the dielectric loss tangent of different batches of energy storage units, the system's pre-stored fixed reference trajectory cannot adapt to the fluctuations of the current production waveform, leading to false alarms in the health status assessment. To eliminate the mismatch between the reference trajectory and the physical properties of the current batch, the system collects data from 100 electronic control modules whose structural integrity has been confirmed by X-ray non-destructive testing. During their charging to rated voltage, the corresponding dynamic response coefficient λ sequence is obtained. The processor calculates the arithmetic mean of the 100 sets of original response trajectories according to the sampling time index to determine the standard health trajectory. The processor calculates the original response trajectory of each group relative to the standard health trajectory. The residual distribution is calculated, and the root mean square value σ of the residual sequence is determined, with a preset safety threshold. Follow the formula ,in, To preset a safety threshold, where σ is the root mean square value of the residual sequence, this procedure utilizes the three-standard-deviation principle to quantify and isolate normal process fluctuations from physical defect disturbances, thus ensuring the preset safety threshold is met. The settings are dynamically anchored to the physical consistency level of the current batch.

[0032] For the feature extraction step in the parameter correlation matrix M, the processor divides the response trajectory projected onto the feature coordinate system into N equally timed local observation windows. It calculates the centroid position offset and covariance features of the scattered point coordinates within each local observation window, extracts the trace of the local covariance matrix over the entire charging cycle, and determines it as a statistical feature quantity. This is used to characterize the smoothness of impedance evolution during energy accumulation in the energy storage unit. Through this step-by-step processing, trajectory consistency assessment is transformed into a quantitative judgment of the covariance matrix trace. The processing results are directly input into the parameter space, realizing a physical mapping from the original energy fluctuation to the causal chain of structural damage. After executing the above reconstruction and calibration procedures, the system improves the accuracy of statistical feature extraction by 15.6% compared to the fixed benchmark method when processing aging batch modules. Furthermore, under interference conditions with an ambient temperature change of 10℃, the background noise of the normal deviation value remains within a preset safety threshold. Within this scope, automated control of the risk of false alarms in the assessment system has been achieved.

[0033] Example 5: When the system faces the first-piece debugging of a new electronic control module hardware model deployed to the production line, the electrical characteristics of the passive components deviate due to changes in the packaging process, causing a mismatch between the pre-stored evaluation benchmark and the dynamic impedance response law of the hardware. The system selects 50 modules whose parameters have been verified by a precision bridge and which have no physical damage, sets the ambient temperature to 25°C, and performs data acquisition for the charging range. The processor performs amplitude normalization processing on the acquired bus current data and bus voltage data to obtain the original trajectory set, and calculates the dynamic response coefficient at each sampling point in the set. The arithmetic mean of the values ​​is used to generate a standard health trajectory corresponding to the physical attributes of the current batch. .

[0034] In deployment scenarios where the operating environment temperature ranges from -20℃ to 60℃, the nonlinear drift caused by temperature changes in the internal resistance of the energy storage unit can lead to a shift in the baseline value of the energy change sequence. To eliminate the interference of ambient temperature fluctuations on detection accuracy, the system executes a coefficient calibration procedure within a temperature control chamber. The energy sequence of the module is measured at temperature nodes with a 10℃ step gradient, and the energy deviation of each node relative to the nominal 25℃ environment is calculated. A temperature compensation coefficient is then generated using the least squares method, and the compensated energy value is determined accordingly. The calculation formula is as follows: ,in, The compensated energy value, Here is the original energy sample value, and k is the temperature compensation coefficient. Using the measured ambient temperature, after calibration, the background noise fluctuation range of the normal deviation value decreased from 4.5% to 0.8%, and the evaluation results remained consistent under different temperature environments.

[0035] Example 6: When the system faces a situation where the charge absorption pattern deviates due to the inconsistent dielectric polarization delay characteristics of the energy storage units within the electronic control module, the processor establishes a local observation window width adapted to the hardware characteristics by executing an initialization scan procedure. The initialization scan procedure includes acquiring the original physical quantity sequence containing 1024 consecutive sampling points during the rising phase of the charging electric field, and the processor calculates the sequence variation coefficient under different window step sizes. Sequence variation coefficient The calculation formula is as follows: ,in, The coefficient of variation is the sequence variation. The standard deviation of the sampling sequence. As the average value of the sampled sequence, the processor selects the window length that maximizes the energy of the first derivative of the response trajectory as the current equal time interval, so that the cluster center shift features of the statistical features in the multidimensional parameter space can be quantified and captured.

[0036] In deployment scenarios where humidity fluctuations exceed 15% in the production environment, the micro-leakage current on the module surface increases with rising humidity, interfering with the accuracy of extracting the internal impedance evolution characteristics of the energy storage unit. The processor eliminates this interference by initiating an environmental adaptive calibration procedure, which includes real-time monitoring of the ambient temperature of the electronic control module. and bus no-load current data Dynamically correct the starting point of the response trajectory in the voltage range during the sampling phase. voltage range start point The calculation formula is as follows: ,in, This is the starting point of the voltage range. The reference voltage is the starting point, and α is the temperature influence factor. The actual ambient temperature. For reference to ambient temperature, and with the processor executing based on bus no-load current data in the measured energy sequence. After bias subtraction and execution of the environmental adaptive calibration procedure, the measured response trajectory is compared with the standard health trajectory. The fluctuation range of the normal deviation value converged to within 0.9%, and the evaluation results remained consistent under different workshop humidity conditions.

[0037] 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 spirit or essential characteristics of the present invention.

[0038] 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. A method for assessing the health status of an electronic control module based on a parameter correlation matrix, characterized in that, Includes the following steps: Step S1: During the charging phase of the electronic control module, simultaneously acquire bus current data, bus voltage data, ambient temperature of the electronic control module, and energy change sequence of the energy storage unit within a preset time window. Step S2: Calculate the transient power change based on the bus current data and bus voltage data, and calculate the voltage square change corresponding to the bus voltage data. Generate the response trajectory by calculating the derivative of the transient power change with respect to the voltage square change, so as to extract the charge throughput characteristics induced by the electronic control module under the dynamic electric field gradient. Step S3: Map the response trajectory to a parameter space that includes the power change rate, the voltage square change rate, and the charge / discharge timing deviation, and construct a parameter correlation matrix that reflects the internal impedance evolution process of the energy storage unit. Step S4: Extract the curvature feature difference between the measured response trajectory and the stored standard health trajectory in the parameter correlation matrix, calculate the normal deviation value of the measured response trajectory relative to the standard health trajectory in the parameter space, and count the time proportion of sampling points whose normal deviation value exceeds the preset safety threshold in the response trajectory. If the time proportion exceeds the preset risk threshold, it is determined that the electronic control module has a characteristic defect and the evaluation result is output.

2. The method for assessing the health status of an electronic control module based on a parameter correlation matrix according to claim 1, characterized in that, Step S2 includes: aligning the acquired bus current data and bus voltage data in time; using a cyclic data buffer to calculate the transient power change and voltage square change between adjacent sampling points; calculating the in-phase correlation between the transient power change and voltage square change to eliminate the interference of external power ripple on the extraction of internal impedance characteristics of the energy storage unit, and generating a response trajectory after eliminating ripple background noise.

3. The method for assessing the health status of an electronic control module based on a parameter correlation matrix according to claim 2, characterized in that, Step S3 includes: establishing a feature coordinate system with the power change rate on the horizontal axis and the voltage square change rate on the vertical axis; projecting the response trajectory after eliminating ripple background noise onto the feature coordinate system to generate a scatter plot describing the charging consistency of the energy storage unit; extracting the statistical features of the scatter plot in the feature coordinate system, and using the statistical features as feature inputs to the parameter correlation matrix.

4. The method for assessing the health status of an electronic control module based on a parameter correlation matrix according to claim 3, characterized in that, In step S2, the calculation rules for the response trajectory are as follows: Where λ is the dynamic response coefficient and ΔP is the transient power increment between adjacent sampling points. This represents the squared voltage increment between the corresponding sampling points.

5. The method for assessing the health status of an electronic control module based on a parameter correlation matrix according to claim 1, characterized in that, Step S4 involves calculating the normal deviation value, including: calculating the geometric deviation distance of each set of feature evaluation vectors in the parameter correlation matrix relative to the standard health trajectory; counting the number of abnormal sampling points whose geometric deviation distance exceeds the preset safety threshold during the full charging process; and determining the ratio of the number of abnormal sampling points to the total number of full sampling points as the time percentage.

6. The method for assessing the health status of an electronic control module based on a parameter correlation matrix according to claim 5, characterized in that, Step S4 is followed by: Step S5, monitoring the voltage holding slope of the electronic control module at the end of charging; Step S6, generating a health evaluation benchmark covering the life cycle of the electronic control module based on the nonlinear correlation weight between the voltage holding slope and the impedance evolution characteristics in the parameter correlation matrix.

7. The method for assessing the health status of an electronic control module based on a parameter correlation matrix according to claim 1, characterized in that, Step S1 involves acquiring bus current and bus voltage data, including: using a multi-channel synchronous sampling interface to acquire data from multiple electronic control modules in parallel; and acquiring current and voltage signals at a sampling frequency of not less than 500kHz during the dynamic process of charging the energy storage unit to a preset voltage range, in order to capture microsecond-level transient leakage waveforms.

8. The method for assessing the health status of an electronic control module based on a parameter correlation matrix according to claim 6, characterized in that, Step S6 is followed by: Step S7, which associates the evaluation results with the production batch information of the electronic control module; Step S8, which uploads the associated health data to the production management system to perform quality risk traceability of the electronic control module.

9. The method for assessing the health status of an electronic control module based on a parameter correlation matrix according to claim 1, characterized in that, The evaluation results output includes: a qualified instruction indicating that the electronic control module is in a healthy state; a warning signal indicating that the electronic control module has a risk of characteristic defects, and an instruction to interrupt the communication function of the electronic control module.

10. The method for assessing the health status of an electronic control module based on a parameter correlation matrix according to claim 1, characterized in that, Step S1 also includes: performing temperature compensation on the energy change sequence based on the acquired ambient temperature; and adjusting the dynamic range of the current acquisition channel based on the gradient change of the bus current data to maintain constant measurement accuracy under different current gradients during the charging phase.

Images