Nondestructive testing method and device for performance of high-voltage automobile wire harness material

Abstract

The invention relates to the technical field of industrial nondestructive testing, in particular to a nondestructive testing method for the performance of a high-voltage automobile wire harness material. According to the invention, through fusion of electric, thermal and electromagnetic multi-mode excitation and sensing technologies, synchronous and accurate detection and positioning of various defects such as insulation aging, mechanical damage, connection degradation and the like of the high-voltage automobile wire harness are realized, and the detection comprehensiveness and reliability are significantly improved. A DSP and a deep learning model are adopted to automatically complete signal feature extraction, fusion and intelligent diagnosis, quantitative performance indexes and defect probability are output, the detection efficiency is greatly improved, and dependence on human experience is reduced. The system has high-sensitivity early fault early warning capability, can realize conversion from post maintenance to predictive maintenance, effectively ensures safe operation of the high-voltage wiring harness, and avoids major faults.

Description

Technical Field

[0001] This invention relates to the field of industrial non-destructive testing technology, specifically to a method and apparatus for non-destructive testing of the performance of high-voltage automotive wiring harness materials. Background Technology

[0002] With the continuous improvement of voltage levels in high-voltage platforms for new energy vehicles, high-voltage wiring harnesses are subjected to multiple stress coupling effects, including electrical, thermal, and mechanical vibrations, over long periods. Their insulation materials are prone to aging, cracking, partial discharge, or carbonization, and connection points are susceptible to loosening and oxidation. This poses significant risks of leakage, short circuits, and even thermal runaway, severely threatening the safety of the entire vehicle. Therefore, accurate, efficient, and non-destructive testing of high-voltage wiring harness materials has become an urgent industry requirement.

[0003] In existing high-voltage wire harness material testing processes, most testing methods require professional personnel to analyze waveforms or images, resulting in low levels of automation and intelligence, limited testing efficiency, and highly subjective results. Furthermore, it is difficult to simultaneously perform coupled analysis of multiple parameters such as electrical, thermal, and mechanical parameters, making it impossible to comprehensively assess the health status of the wire harness.

[0004] To address the aforementioned issues, it is necessary to propose a non-destructive testing method and apparatus for the performance of high-voltage automotive wiring harness materials. Summary of the Invention

[0005] The purpose of this invention is to solve the problems existing in the background art, and to propose a non-destructive testing method and device for the performance of high-voltage automotive wiring harness materials.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] In a first aspect, the present invention provides a non-destructive testing method for the performance of high-voltage automotive wiring harness materials, comprising the following steps:

[0008] Step 1: Generation and application of multimodal excitation signals;

[0009] A preset excitation signal is generated by control commands and applied to the high-voltage automotive wiring harness under test by a specific exciter, thereby stimulating it to produce a physical response containing internal state information.

[0010] The main control module sends out control command data packets to generate various types of excitation signals. Each control command corresponds to a preset excitation signal, including sinusoidal sweep voltage, pulse current, and near-infrared light.

[0011] Each preset excitation signal has preset amplitude, voltage, current, frequency range, scan rate, and wavelength;

[0012] Among them, the sinusoidal sweep voltage is used for dielectric spectrum analysis to evaluate the changes in the dielectric constant and loss tangent of the insulating material.

[0013] Among them, pulsed current is used for time-domain reflection and vector network analysis to detect impedance non-uniformity and locate physical defects.

[0014] Near-infrared light is used for optical thermal excitation and thermal wave imaging detection.

[0015] The command signal generation module generates the corresponding excitation signal based on the control command data packet and outputs it to the high-voltage automotive wiring harness under test.

[0016] Step 2: Synchronous acquisition of multi-dimensional response signals;

[0017] The physical response signals generated by the harness under various excitation signals are synchronously captured by a sensor array, and then converted into high-resolution, low-noise pre-processed digital signals by a filter.

[0018] Access the sensor array to acquire the physical response signals generated by the high-voltage automotive wiring harness under test under various excitation signals, including leakage current signals, surface temperature change signals, and reflected electromagnetic wave signals.

[0019] The leakage current signal refers to the current signal generated when a sinusoidal sweep voltage or pulsed current excitation is applied, causing a small current path to form on the surface or in the bulk of the insulating medium due to defects or aging of the insulating material. It reflects the conductivity characteristics of the insulating material and its degree of degradation. Its mathematical model is as follows:

[0020]

[0021] in The leakage current signal to be measured, wherein... The voltage of the applied excitation signal, where represents the complex conductivity of the insulating material, which is related to the dielectric constant and loss factor; where j is the imaginary unit. The phase offset angle of the leakage current signal to be measured. ω is the angular frequency.

[0022] The surface temperature change signal refers to the dynamic temperature distribution signal formed by the instantaneous temperature rise on the surface of the wire harness due to light energy absorption under near-infrared photothermal excitation, through processes such as heat conduction, convection, and radiation. Its mathematical model is as follows:

[0023]

[0024] Among them Let be the temperature at location r as a function of time t, where r is the location number; The preset thermal diffusivity; Let be the heat source power density at position r as a function of time t, where ρ is the material density. This represents the specific heat capacity of the material.

[0025] The temperature at each location r changes with time t. A temperature field is generated based on the coordinates of its own position r, and then arranged in order according to time t to obtain the surface temperature change signal.

[0026] Reflected electromagnetic wave signals refer to signals formed by partial reflection of electromagnetic waves due to impedance discontinuities caused by damage, joints, or deformation under pulsed current or high-frequency voltage excitation. This signal can be received by a high-frequency current probe or antenna, and its time-domain waveform contains reflection coefficient information. Its mathematical model is as follows:

[0027]

[0028] in, The reflected electromagnetic wave signal to be measured is related to the defect type and location; among which, For local impedance, The preset characteristic impedance is 50Ω.

[0029] The sensor array is non-contactly aligned with the wire harness under test to capture the weak response signals of leakage current, surface temperature change, and reflected electromagnetic wave in real time. These signals are then input into a preamplifier and an anti-aliasing filter for preliminary signal conditioning to obtain pre-processed leakage current, surface temperature change, and reflected electromagnetic wave signals, thereby enhancing the signal-to-noise ratio and preventing sampling distortion.

[0030] The preamplifier amplifies the weak response signals captured by the sensor array, including leakage current, surface temperature change, and reflected electromagnetic wave signals, increasing the signal amplitude and suppressing interference during transmission. Its mathematical model for operation is as follows:

[0031]

[0032] Among them and These are the output and input signals of the preamplifier, including the output and input signals of the leakage current signal, surface temperature change signal, and reflected electromagnetic wave signal; where F is the preset amplification gain. This is the preset offset voltage.

[0033] The anti-aliasing filter is a low-pass filter used to limit the signal bandwidth to below the Nyquist frequency, preventing spectral aliasing in the high-frequency components during signal sampling. Its transfer function is:

[0034]

[0035] Among them The sampling frequency for the leakage current signal, surface temperature change signal, and reflected electromagnetic wave signal is denoted as . The preset cutoff frequency is used, and the following conditions are met: ; n is the filter order, where Let be the transfer function of the anti-aliasing filter, representing how the anti-aliasing filter affects the frequency components of the signal. In leakage current signals, surface temperature change signals, and reflected electromagnetic wave signals, frequencies higher than the transfer function will be blocked, while frequencies lower than the transfer function will be output.

[0036] In a preferred embodiment of the present invention, the preprocessed leakage current signal, surface temperature change signal, and reflected electromagnetic wave signal are sent as preprocessed digital signals to step three.

[0037] Step 3: Signal feature extraction and fusion processing;

[0038] The digital signal is reduced in dimension and refined. The digital signal is then input into the DSP digital signal processing module for feature extraction. Key feature information that is strongly correlated with material performance degradation is extracted from the noise and then fused.

[0039] As a preferred embodiment of the present invention, statistical features are extracted, and the effective value, peak value, skewness, kurtosis and pulse count of the preprocessed leakage current signal, surface temperature change signal and reflected electromagnetic wave signal are calculated respectively.

[0040] In a preferred embodiment of the present invention, spectral features are extracted, and fast Fourier transforms are performed on the preprocessed leakage current signal, surface temperature change signal, and reflected electromagnetic wave signal to obtain their frequency domain signals. Harmonic component amplitudes and spectral centroids are then extracted from the frequency domain signals.

[0041] Feature vectors are generated based on the statistical and spectral characteristics of the preprocessed leakage current signal, surface temperature change signal, and reflected electromagnetic wave signal to obtain key feature information.

[0042] The feature vector contains several elements, which correspond to the statistical and spectral characteristics of each preprocessed leakage current signal, surface temperature change signal, and reflected electromagnetic wave signal.

[0043] Step 4: Performance degradation assessment and diagnosis based on deep learning models;

[0044] Performance degradation is assessed and diagnosed based on key feature information, and diagnostic results are obtained. Judgment signals are then derived from the diagnostic results.

[0045] The elements in the feature vector are processed by a multi-branch deep neural network, and the output layer yields an output value that includes a regression branch: performance degradation coefficient and defect type probability.

[0046] The performance degradation coefficient ranges from 0 to 1, where 0 represents an undamaged state and 1 represents a complete failure state; the defect type probability ranges from 0% to 100%, including:

[0047] The probability of defect types for insulation aging, partial discharge, insulation carbonization, moisture contamination, mechanical damage, and connector degradation.

[0048] The multi-branch deep neural network includes an input layer, a convolutional layer, a fully connected layer, and an output layer. The input layer contains several nodes, which are used to receive each element of the feature vector. The convolutional layer extracts abstract features from each element of the feature vector through convolutional kernels and outputs them to the fully connected layer.

[0049] The fully connected layer calculates the degradation coefficient and the probability of each defect type using the sigmoid activation function.

[0050] If the degradation coefficient of the output is found to be greater than the preset threshold of 0.8, the probability of the defect type is numerically judged. If the probability of a certain defect type is found to be greater than the preset confidence level of 70%, the corresponding judgment signal is output, including insulation aging judgment signal, partial discharge judgment signal, insulation carbonization judgment signal, moisture contamination judgment signal, mechanical damage judgment signal, and connector degradation judgment signal.

[0051] The generated decision signal is sent to the display output device.

[0052] Secondly, the present invention provides a non-destructive testing device for the performance of high-voltage automotive wiring harness materials, comprising a main control module, a signal generation module, a DSP digital signal processing module, and a diagnostic module.

[0053] The main control module stores several control instruction data packets, each corresponding to a preset excitation signal, including sinusoidal sweep voltage, pulse current, and near-infrared light.

[0054] The signal generation module includes a high-voltage probe, an antenna array, and an infrared laser transmitter. After receiving a control command data packet, it parses the excitation signal type, amplitude, frequency range, scan rate, and wavelength contained therein, and outputs the corresponding excitation signal.

[0055] The DSP digital signal processing module acquires leakage current signals, surface temperature change signals, and reflected electromagnetic wave signals generated under the excitation signal environment in real time, and preprocesses them. Feature extraction is performed on the preprocessed leakage current signals, surface temperature change signals, and reflected electromagnetic wave signals to obtain statistical features and spectral features. Finally, the statistical features and spectral features are fused to obtain a feature vector.

[0056] The diagnostic module inputs the feature vector into a multi-branch deep neural network to process the elements of the feature vector. The output layer then outputs values ​​containing regression branches: performance degradation coefficient and defect type probability. A judgment signal is output based on the performance degradation coefficient and defect type probability.

[0057] Compared with the prior art, the beneficial effects of the present invention are:

[0058] 1. This invention achieves multi-dimensional, high-precision defect detection and accurate positioning, significantly improving the comprehensiveness and reliability of detection;

[0059] 2. This invention realizes intelligent automatic diagnosis from "signal" to "decision", which greatly improves detection efficiency and reduces reliance on operator experience;

[0060] 3. This invention has the ability to provide early fault warning, realizing the transformation from "reactive maintenance" to "predictive maintenance", which has significant economic benefits and safety value. Attached Figure Description

[0061] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings:

[0062] Figure 1 This is a flowchart of a non-destructive testing method for the performance of high-voltage automotive wiring harness materials proposed in the embodiments of the present invention;

[0063] Figure 2 This is a schematic diagram of a module for a non-destructive testing device for the performance of high-voltage automotive wiring harness materials, as proposed in an embodiment of the present invention. Detailed Implementation

[0064] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. 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.

[0065] Please see Figure 1 As shown, a non-destructive testing method for the properties of high-voltage automotive wiring harness materials includes the following steps:

[0066] Step 1: Generation and application of multimodal excitation signals;

[0067] A preset excitation signal is generated by control commands and applied to the high-voltage automotive wiring harness under test by a specific exciter, thereby stimulating it to produce a physical response containing internal state information.

[0068] The main control module sends out control command data packets to generate various types of excitation signals. Each control command corresponds to a preset excitation signal, including sinusoidal sweep voltage, pulse current, and near-infrared light.

[0069] Each preset excitation signal has preset amplitude, voltage, current, frequency range, scan rate, and wavelength;

[0070] Among them, the sinusoidal sweep voltage is used for dielectric spectrum analysis to evaluate the changes in the dielectric constant and loss tangent of the insulating material.

[0071] Among them, pulsed current is used for time-domain reflection and vector network analysis to detect impedance non-uniformity and locate physical defects.

[0072] Near-infrared light is used for optical thermal excitation and thermal wave imaging detection.

[0073] The command signal generation module generates the corresponding excitation signal based on the control command data packet and outputs it to the high-voltage automotive wiring harness under test.

[0074] Step 2: Synchronous acquisition of multi-dimensional response signals;

[0075] The physical response signals generated by the harness under various excitation signals are synchronously captured by a sensor array, and then converted into high-resolution, low-noise pre-processed digital signals by a filter.

[0076] Access the sensor array to acquire the physical response signals generated by the high-voltage automotive wiring harness under test under various excitation signals, including leakage current signals, surface temperature change signals, and reflected electromagnetic wave signals.

[0077] The leakage current signal refers to the current signal generated when a sinusoidal sweep voltage or pulsed current excitation is applied, causing a small current path to form on the surface or in the bulk of the insulating medium due to defects or aging of the insulating material. It reflects the conductivity characteristics of the insulating material and its degree of degradation. Its mathematical model is as follows:

[0078]

[0079] in The leakage current signal to be measured, wherein... The voltage of the applied excitation signal, where represents the complex conductivity of the insulating material, which is related to the dielectric constant and loss factor; where j is the imaginary unit. The phase offset angle of the leakage current signal to be measured. ω is the angular frequency.

[0080] The surface temperature change signal refers to the dynamic temperature distribution signal formed by the instantaneous temperature rise on the surface of the wire harness due to light energy absorption under near-infrared photothermal excitation, through processes such as heat conduction, convection, and radiation. Its mathematical model is as follows:

[0081]

[0082] Among them Let be the temperature at location r as a function of time t, where r is the location number; The preset thermal diffusivity; Let be the heat source power density at position r as a function of time t, where ρ is the material density. This represents the specific heat capacity of the material.

[0083] The temperature at each location r changes with time t. A temperature field is generated based on the coordinates of its own position r, and then arranged in order according to time t to obtain the surface temperature change signal.

[0084] Reflected electromagnetic wave signals refer to signals formed by partial reflection of electromagnetic waves due to impedance discontinuities caused by damage, joints, or deformation under pulsed current or high-frequency voltage excitation. This signal can be received by a high-frequency current probe or antenna, and its time-domain waveform contains reflection coefficient information. Its mathematical model is as follows:

[0085]

[0086] in, The reflected electromagnetic wave signal to be measured is related to the defect type and location; among which, For local impedance, The preset characteristic impedance is 50Ω.

[0087] The sensor array is non-contactly aligned with the wire harness under test to capture the weak response signals of leakage current, surface temperature change, and reflected electromagnetic wave in real time. These signals are then input into a preamplifier and an anti-aliasing filter for preliminary signal conditioning to obtain pre-processed leakage current, surface temperature change, and reflected electromagnetic wave signals, thereby enhancing the signal-to-noise ratio and preventing sampling distortion.

[0088] The preamplifier amplifies the weak response signals captured by the sensor array, including leakage current, surface temperature change, and reflected electromagnetic wave signals, increasing the signal amplitude and suppressing interference during transmission. Its mathematical model for operation is as follows:

[0089]

[0090] Among them and These are the output and input signals of the preamplifier, including the output and input signals of the leakage current signal, surface temperature change signal, and reflected electromagnetic wave signal; where F is the preset amplification gain. This is the preset offset voltage.

[0091] The anti-aliasing filter is a low-pass filter used to limit the signal bandwidth to below the Nyquist frequency, preventing spectral aliasing in the high-frequency components during signal sampling. Its transfer function is:

[0092]

[0093] Among them The sampling frequency for the leakage current signal, surface temperature change signal, and reflected electromagnetic wave signal is denoted as . The preset cutoff frequency is used, and the following conditions are met: ; n is the filter order, where Let be the transfer function of the anti-aliasing filter, representing how the anti-aliasing filter affects the frequency components of the signal. In leakage current signals, surface temperature change signals, and reflected electromagnetic wave signals, frequencies higher than the transfer function will be blocked, while frequencies lower than the transfer function will be output.

[0094] Furthermore, the preprocessed leakage current signal, surface temperature change signal, and reflected electromagnetic wave signal are sent as preprocessed digital signals to step three.

[0095] Step 3: Signal feature extraction and fusion processing;

[0096] The digital signal is reduced in dimension and refined. The digital signal is then input into the DSP digital signal processing module for feature extraction. Key feature information that is strongly correlated with material performance degradation is extracted from the noise and then fused.

[0097] Furthermore, statistical features are extracted, and the effective value, peak value, skewness, kurtosis, and pulse count of the preprocessed leakage current signal, surface temperature change signal, and reflected electromagnetic wave signal are calculated respectively.

[0098] Furthermore, spectral features are extracted, and fast Fourier transforms are performed on the preprocessed leakage current signal, surface temperature change signal, and reflected electromagnetic wave signal to obtain their frequency domain signals. Harmonic component amplitudes and spectral centroids are then extracted from these frequency domain signals.

[0099] Feature vectors are generated based on the statistical and spectral characteristics of the preprocessed leakage current signal, surface temperature change signal, and reflected electromagnetic wave signal to obtain key feature information.

[0100] The feature vector contains several elements, which correspond to the statistical and spectral characteristics of each preprocessed leakage current signal, surface temperature change signal, and reflected electromagnetic wave signal.

[0101] Step 4: Performance degradation assessment and diagnosis based on deep learning models;

[0102] Performance degradation is assessed and diagnosed based on key feature information, and diagnostic results are obtained. Judgment signals are then derived from the diagnostic results.

[0103] The elements in the feature vector are processed by a multi-branch deep neural network, and the output layer yields an output value that includes a regression branch: performance degradation coefficient and defect type probability.

[0104] The performance degradation coefficient ranges from 0 to 1, where 0 represents an undamaged state and 1 represents a complete failure state; the defect type probability ranges from 0% to 100%, including:

[0105] The probability of defect types for insulation aging, partial discharge, insulation carbonization, moisture contamination, mechanical damage, and connector degradation.

[0106] The multi-branch deep neural network includes an input layer, a convolutional layer, a fully connected layer, and an output layer. The input layer contains several nodes, which are used to receive each element of the feature vector. The convolutional layer extracts abstract features from each element of the feature vector through convolutional kernels and outputs them to the fully connected layer.

[0107] The fully connected layer calculates the degradation coefficient and the probability of each defect type using the sigmoid activation function.

[0108] If the degradation coefficient of the output is found to be greater than the preset threshold of 0.8, the probability of the defect type is numerically judged. If the probability of a certain defect type is found to be greater than the preset confidence level of 70%, the corresponding judgment signal is output, including insulation aging judgment signal, partial discharge judgment signal, insulation carbonization judgment signal, moisture contamination judgment signal, mechanical damage judgment signal, and connector degradation judgment signal.

[0109] The generated decision signal is sent to the display output device.

[0110] Please see Figure 2 As shown, a non-destructive testing device for the performance of high-voltage automotive wiring harness materials includes a main control module, a signal generation module, a DSP digital signal processing module, and a diagnostic module.

[0111] The main control module stores several control instruction data packets, each corresponding to a preset excitation signal, including sinusoidal sweep voltage, pulse current, and near-infrared light.

[0112] The signal generation module includes a high-voltage probe, an antenna array, and an infrared laser transmitter. After receiving a control command data packet, it parses the excitation signal type, amplitude, frequency range, scan rate, and wavelength contained therein, and outputs the corresponding excitation signal.

[0113] The DSP digital signal processing module acquires leakage current signals, surface temperature change signals, and reflected electromagnetic wave signals generated under the excitation signal environment in real time, and preprocesses them. Feature extraction is performed on the preprocessed leakage current signals, surface temperature change signals, and reflected electromagnetic wave signals to obtain statistical features and spectral features. Finally, the statistical features and spectral features are fused to obtain a feature vector.

[0114] The diagnostic module inputs the feature vector into a multi-branch deep neural network to process the elements of the feature vector. The output layer then outputs values ​​containing regression branches: performance degradation coefficient and defect type probability. A judgment signal is output based on the performance degradation coefficient and defect type probability.

[0115] It should be understood that the terms “comprising” and “including” used in this disclosure and claims indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0116] It should also be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the disclosure. As used in this disclosure and claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this disclosure and claims means any combination and all possible combinations of one or more of the associated listed items, and includes such combinations;

[0117] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A non-destructive testing method for high voltage automotive wiring harness material performance, characterized by, Comprising the following steps: Step one, multi-modal excitation signal generation and application; Generate preset excitation signals through control instructions, and apply them to the high-voltage automotive wiring harness under test through specific exciters to stimulate the generation of physical responses containing internal state information; Step two, multi-dimensional response signal synchronous acquisition; Synchronously capture the physical response signals generated by the wiring harness under various excitation signals through a sensor array, and convert them into high-resolution, low-noise preprocessed digital signals through a filter; Step three, signal feature extraction and fusion processing; According to the digital signal, dimensionality reduction and refinement are performed, the digital signal is input into the DSP digital signal processing module for feature extraction, the key feature information strongly related to material performance degradation is extracted from the noise, and feature fusion is performed; Step four, performance degradation evaluation and diagnosis based on deep learning model; According to the key feature information, performance degradation evaluation and diagnosis are performed to obtain diagnosis result information, and a determination signal is obtained according to the diagnosis result information.

2. The method for non-destructive testing of the properties of high-voltage automobile wiring harness material according to claim 1, characterized in that, The specific process of generating preset excitation signals through control instructions is as follows: The main control module sends control instruction data packets to generate various types of excitation signals, each control instruction corresponds to a preset excitation signal, including sinusoidal sweep voltage, pulse current and near-infrared light; Each preset excitation signal has a preset amplitude, voltage, current, frequency range, scan rate and wavelength; Among them, the sinusoidal sweep voltage is used for dielectric spectrum analysis to evaluate the changes of dielectric constant and loss tangent of insulating materials; Among them, the pulse current is used for time domain reflection and vector network analysis to detect impedance inhomogeneity and locate physical defects; Among them, the near-infrared light is used for optical thermal excitation for thermal wave imaging detection; The command signal generation module generates corresponding excitation signals according to the control instruction data packet and outputs them to the high-voltage automotive wiring harness under test.

3. The method for non-destructive testing of the properties of high-voltage automobile wiring harness material according to claim 1, characterized in that, The specific process of collecting physical response signals is as follows: Access the sensor array to obtain the physical response signals generated by the high-voltage automotive wiring harness under test under various excitation signals, including leakage current signal, surface temperature change signal and reflected electromagnetic wave signal; Among them, the leakage current signal refers to the current signal generated by the micro current path formed by the partial current passing through the surface or body region of the insulating material due to the existence of defects or aging of the insulating material when the sinusoidal sweep voltage or pulse current excitation is applied, which reflects the conductivity characteristics and degradation degree of the insulating material; Among them, the surface temperature change signal refers to the dynamic temperature distribution signal formed by the instantaneous temperature rise of the wiring harness surface due to the absorption of light energy under near-infrared light thermal excitation, through heat conduction, convection and radiation processes; Among them, the reflected electromagnetic wave signal refers to the signal formed by the partial reflection of electromagnetic waves due to impedance discontinuity caused by damage, joints or deformation under pulse current or high-frequency voltage excitation; This signal can be received by a high-frequency current probe or an antenna, and its time domain waveform contains reflection coefficient information; The sensor array is non-contact aligned with the measured wiring harness to real-time capture the weak response signals of the leakage current signal, surface temperature change signal and reflected electromagnetic wave signal, which are respectively input into the preamplifier and anti-aliasing filter for preliminary signal conditioning.

4. The method for non-destructive testing of the properties of high-voltage automobile wiring harness material according to claim 1, characterized in that, The specific process of preliminary signal conditioning is as follows: The weak response signals of the leakage current signal, the surface temperature change signal and the reflected electromagnetic wave signal are captured in real time by non-contact alignment of the measured wire harness through the sensor array, and are input into a preamplifier and an anti-aliasing filter for preliminary conditioning to obtain the preprocessed leakage current signal, the surface temperature change signal and the reflected electromagnetic wave signal, so as to enhance the signal-to-noise ratio and prevent sampling distortion; The preamplifier amplifies the weak response signals of the sensor array capturing the leakage current signals, the surface temperature change signals and the reflected electromagnetic wave signals, improves the signal amplitude and suppresses the interference in the transmission process, and the mathematical model of the working mode is: , Wherein and are the output signal and the input signal of the preamplifier, the output signal and the input signal including the leakage current signal, the surface temperature change signal and the reflected electromagnetic wave signal; wherein F is a preset amplification gain, and wherein is a preset offset voltage; Wherein, the anti-aliasing filter is a low-pass filter, which is used to limit the signal bandwidth to below the Nyquist frequency, to prevent the high frequency part from producing spectral aliasing when the signal is sampled, and its transfer function is: , wherein is a sampling frequency of the leakage current signal, the surface temperature change signal and the reflected electromagnetic wave signal, wherein is a preset cutoff frequency, and satisfies the condition: ; n is a filter order, wherein is a transfer function of the anti-aliasing filter, representing an influence mode of the anti-aliasing filter on a frequency component of the signal; in the leakage current signal, the surface temperature change signal and the reflected electromagnetic wave signal, a part with a frequency greater than the transfer function is blocked, and a part with a frequency less than the transfer function is output.

5. The method for non-destructive testing of high voltage automotive wiring harness material properties according to claim 1, characterized in that, The specific process of feature extraction is as follows: Statistical features are extracted, and the effective value, peak value, skewness, kurtosis and pulse count of the preprocessed leakage current signal, the surface temperature change signal and the reflected electromagnetic wave signal are calculated respectively; Spectrum features are extracted, and the frequency domain signals of the preprocessed leakage current signal, the surface temperature change signal and the reflected electromagnetic wave signal are obtained by performing fast Fourier transform on the preprocessed leakage current signal, the surface temperature change signal and the reflected electromagnetic wave signal respectively; and the harmonic component amplitude and the spectral centroid are extracted from the frequency domain signals.

6. The method for non-destructive testing of high voltage automotive wiring harness material properties according to claim 1, characterized in that, The specific process of feature fusion is as follows: The feature vectors are generated according to the statistical features and the spectrum features of the preprocessed leakage current signal, the surface temperature change signal and the reflected electromagnetic wave signal, and the key feature information is obtained. The feature vectors contain several elements, which correspond to the statistical features and the spectrum features of the preprocessed leakage current signal, the surface temperature change signal and the reflected electromagnetic wave signal respectively.

7. The method for non-destructive testing of high voltage automotive wiring harness material properties according to claim 1, characterized in that, The specific process of performance degradation evaluation and diagnosis according to the key feature information is as follows: The elements in the feature vectors are processed by a multi-branch deep neural network, and the output values containing the regression branch of the performance degradation coefficient and the defect type probability are obtained through the output layer; The numerical range of the performance degradation coefficient is 0 to 1, 0 represents a lossless state, and 1 represents a complete failure state; the numerical range of the defect type probability is 0% to 100%, including: the defect type probability of insulation aging, the defect type probability of partial discharge, the defect type probability of insulation carbonization, the defect type probability of moisture pollution, the defect type probability of mechanical damage and the defect type probability of connector degradation; The multi-branch deep neural network includes an input layer, a convolution layer, a fully connected layer and an output layer, the input layer contains several nodes for receiving the elements of the feature vectors, and the convolution layer extracts abstract features from the elements of the feature vectors through convolution kernels and outputs them to the fully connected layer. The fully connected layer calculates the degradation coefficient and the defect type probability by using a sigmoid activation function.

8. The method for non-destructive testing of high voltage automotive wiring harness material properties according to claim 1, characterized in that, The specific process of obtaining the determination signal according to the diagnosis result information is as follows: If the output degradation coefficient is greater than the preset threshold value 0.8, the numerical value of the defect type probability is judged; if it is identified that there is a defect type probability greater than the preset confidence 70%, the corresponding determination signal is output, including the insulation aging determination signal, the partial discharge determination signal, the insulation carbonization determination signal, the moisture pollution determination signal, the mechanical damage determination signal and the connector degradation determination signal; The generated determination signal is sent to a display output device.

9. A device for non-destructive testing of high-voltage automobile wiring harness material performance, used to implement the method of non-destructive testing of high-voltage automobile wiring harness material performance according to any one of claims 1-8, characterized in that: The device includes a main control module, a signal generation module, a DSP digital signal processing module and a diagnosis module; The main control module stores a plurality of control instruction data packets, each of which corresponds to a preset excitation signal, including a sinusoidal sweep voltage, a pulse current and a near-infrared light. The signal generation module includes a high-voltage probe, an antenna array and an infrared laser emitter, and after receiving a control instruction data packet, parses the excitation signal type, amplitude, frequency range, scanning rate and wavelength contained therein, and outputs the excitation signal corresponding thereto; The DSP digital signal processing module obtains the leakage current signal, surface temperature change signal and reflected electromagnetic wave signal generated in the excitation signal environment in real time, and pre-processes the same; feature extraction is performed on the pre-processing results of the leakage current signal, surface temperature change signal and reflected electromagnetic wave signal to obtain statistical features and spectral features; and data fusion is performed on the statistical features and spectral features to obtain a feature vector; The diagnosis module inputs the feature vector into a multi-branch deep neural network to process the elements in the feature vector, and obtains output values containing regression branches: performance degradation coefficients and defect type probabilities through an output layer; According to the performance degradation coefficients and the defect type probability output, a signal is determined.

Images