Vehicle, fault detection method and detection device for driving motor of vehicle, and storage medium

Abstract

The invention discloses a vehicle, a fault detection method and device for a driving motor of the vehicle and a storage medium, and the method comprises the steps: obtaining a time domain current signal of the driving motor of the vehicle during operation, and recognizing the current operation state of the driving motor; based on the operation state and the time-domain current signal, determining a target conversion mode for converting into a frequency-domain current signal; and determining the frequency spectrum characteristic of the frequency domain current signal by using the target transformation mode, detecting whether the driving motor has a fault based on the frequency spectrum characteristic, and identifying the current fault type of the driving motor when detecting that the driving motor has the fault. According to the detection method, the defects that an existing driving motor rotor eccentricity detection technology is large in difficulty, low in accuracy and the like can be overcome, and whether the driving motor breaks down or not and the fault type when the driving motor breaks down can be accurately detected.

Description

Technical Field

[0001] The present application relates to the technical field of vehicles, and in particular, to a method for detecting a fault of a driving motor of a vehicle, a computer-readable storage medium, a vehicle, and a device for detecting a fault of a driving motor of a vehicle. Background Art

[0002] Based on the requirements of environmental protection and energy conservation, and the continuous breakthroughs in key technologies such as vehicle batteries and driving motors, the proportion of new energy vehicles in the automobile stock market is increasing. Ideally, the magnetic flux density in the air gap of the motor is evenly distributed and the magnetic field intensity remains constant. However, in reality, the misalignment of the stator and rotor axes is caused by various factors such as errors in the manufacturing and assembly process of the motor, external force impacts on rough roads, and the imbalance of the stator and rotor masses, resulting in eccentricity. The eccentricity of the motor rotor not only causes distortion of the air gap magnetic field and deteriorates the motor performance, but when the eccentricity is too large, it may even cause a series of transmission fault problems such as abnormal noise, motor failure, and bearing fracture due to the mutual friction between the stator and rotor.

[0003] In the current related technologies, a first sound signal recorded in real time during the operation of the device is obtained, and the first sound signal is preprocessed to obtain a first signal-to-noise ratio and first voiceprint data corresponding to the first sound signal; the first signal-to-noise ratio and first voiceprint data are input into a pre-trained target deep neural network model to obtain a device status label corresponding to the first sound signal, and the device status of the rotating device is determined based on the device status label, so as to timely identify the device fault of the rotating device. However, this method of identifying faults by sound signals has high detection technical difficulty and low accuracy. Summary of the Invention

[0004] The present application aims to at least solve one of the technical problems in the related technologies to some extent. To this end, the first object of the present application is to propose a method for detecting a fault of a driving motor of a vehicle, which obtains a time-domain current signal of the driving motor of the vehicle during operation, identifies the current operating state of the driving motor, determines a target transformation method for converting the time-domain current signal into a frequency-domain current signal based on the operating state and the time-domain current signal, uses the target transformation method to determine the spectral characteristics of the frequency-domain current signal, and detects whether the driving motor is faulty based on the spectral characteristics, and when it is detected that the driving motor has a fault, identifies the current fault type of the driving motor, so as to solve the problems of high detection technical difficulty and low accuracy of the existing driving motor rotor eccentricity detection technology, and can accurately detect whether the driving motor has a fault and the fault type when a fault occurs.

[0005] The second object of the present application is to propose a computer-readable storage medium.

[0006] The third object of the present application is to propose a vehicle.

[0007] The fourth object of the present application is to propose a device for detecting faults in a driving motor of a vehicle.

[0008] To achieve the above object, an embodiment of the first aspect of the present application proposes a method for detecting faults in a driving motor of a vehicle, the method comprising: acquiring a time-domain current signal of the driving motor of the vehicle during operation, and identifying a current operating state of the driving motor; determining a target transformation method for transforming into a frequency-domain current signal based on the operating state and the time-domain current signal; using the target transformation method to determine spectral characteristics of the frequency-domain current signal, and detecting whether the driving motor is faulty based on the spectral characteristics, and identifying a current fault type of the driving motor when it is detected that the driving motor has a fault.

[0009] According to the method for detecting faults in a driving motor of a vehicle according to an embodiment of the present application, a time-domain current signal of the driving motor of the vehicle during operation is acquired, and a current operating state of the driving motor is identified. A target transformation method for transforming into a frequency-domain current signal is determined based on the operating state and the time-domain current signal. The spectral characteristics of the frequency-domain current signal are determined using the target transformation method, and it is detected whether the driving motor is faulty based on the spectral characteristics. When it is detected that the driving motor has a fault, a current fault type of the driving motor is identified. Thus, the method can solve the disadvantages of the existing driving motor rotor eccentricity detection technology, such as high detection difficulty and low accuracy, and can accurately detect whether the driving motor has a fault and the fault type when a fault occurs.

[0010] In addition, the method for detecting faults in a driving motor of a vehicle according to the above embodiment of the present application may further have the following additional technical features:

[0011] According to an embodiment of the present application, the operating state includes a first operating state and a second operating state. Identifying the current operating state of the driving motor includes: determining an autocorrelation value of the autocorrelation function at a preset lag amount based on the current signal and the autocorrelation function; determining that the operating state is the first operating state when the autocorrelation value is greater than a preset threshold; and determining that the operating state is the second operating state when the autocorrelation value is less than the preset threshold.

[0012] According to an embodiment of the present application, the target transformation method includes a fast Fourier transform method and a discrete wavelet transform method. The method for determining the target transformation method for converting the time-domain current signal into a frequency-domain current signal based on the operating state and the time-domain current signal includes: when the operating state is the first operating state, determining that the target transformation method for converting the time-domain current signal into the frequency-domain current signal is the fast Fourier transform method; when the operating state is the second operating state, determining that the target transformation method for converting the time-domain current signal into the frequency-domain current signal is the discrete wavelet transform method.

[0013] According to an embodiment of the present application, the spectral characteristics include the frequency and amplitude of the harmonics in the frequency-domain current signal, and / or the energy distribution of different frequency bands in the frequency-domain current signal. The method for determining the spectral characteristics of the frequency-domain current signal by using the target transformation method includes: determining the frequency and the amplitude by the fast Fourier transform method; and / or determining the energy distribution by the discrete wavelet transform method.

[0014] According to an embodiment of the present application, the method for detecting whether the drive motor is faulty based on the spectral characteristics and identifying the current fault type of the drive motor when it is detected that the drive motor is faulty includes: inputting the spectral characteristics into a preset motor fault model, where the motor fault model includes a first fault model and a second fault model, and the first fault model and the second fault model are used to identify whether the drive motor is faulty and the fault type based on the spectral characteristics; determining the confidence level of the frequency-domain current signal belonging to each fault type based on the first fault model; when the confidence level is greater than or equal to a preset confidence threshold, determining whether the drive motor is faulty based on the spectral characteristics and the first fault model, and identifying the current fault type of the drive motor when it is detected that the drive motor is faulty; when the confidence level is less than the preset confidence threshold, determining whether the drive motor is faulty based on the spectral characteristics and the second fault model, and identifying the current fault type of the drive motor when it is detected that the drive motor is faulty.

[0015] According to an embodiment of the present application, the first fault model is a K-nearest neighbor classification algorithm model, which is trained based on the spectral characteristics of the frequency-domain current signal when the drive motor is operating normally and the spectral characteristics of the frequency-domain current signal with fault type labels when the drive motor is operating with faults. The second fault model is a long short-term memory neural network algorithm model, which is trained based on the spectral characteristics of the frequency-domain current signal screened by the K-nearest neighbor classification algorithm model and the spectral characteristics of the frequency-domain current signal when the drive motor is actually operating.

[0016] According to an embodiment of the present application, after identifying the current fault type of the drive motor, the method further includes: when the current fault type is a bearing damage fault, issuing a fault warning of bearing damage and determining a maintenance measure of stopping for unpacking inspection; when the current fault type is a stator winding fault, issuing a fault warning of stator winding fault and determining a maintenance measure of adjusting the vehicle driving mode; when the current fault type is a rotor eccentricity fault, issuing a fault warning of rotor eccentricity fault and determining a maintenance measure of correcting the dynamic balance of the rotor.

[0017] To achieve the above object, an embodiment of the second aspect of the present application provides a computer-readable storage medium, on which a program is stored, and when the program is executed by a processor, the above-mentioned drive motor fault detection method for a vehicle is implemented.

[0018] According to the computer-readable storage medium of the embodiment of the present application, by implementing the above-mentioned drive motor fault detection method for a vehicle when executed, it can solve the problems of high difficulty and low accuracy in the existing drive motor rotor eccentricity detection technology, and can accurately detect whether the drive motor fails and the fault type when a failure occurs.

[0019] To achieve the above object, a vehicle provided by an embodiment of the third aspect of the present application includes a memory, a processor, and a program stored on the memory and executable on the processor. When the processor executes the program, the above-mentioned drive motor fault detection method for a vehicle is implemented.

[0020] According to the vehicle of the embodiment of the present application, by implementing the above-mentioned drive motor fault detection method for a vehicle, it can solve the problems of high difficulty and low accuracy in the existing drive motor rotor eccentricity detection technology, and can accurately detect whether the drive motor fails and the fault type when a failure occurs.

[0021] To achieve the above object, an embodiment of the fourth aspect of the present application provides a drive motor fault detection device for a vehicle, the device includes: an acquisition module, configured to acquire a time-domain current signal of the drive motor of the vehicle during operation and identify the current operating state of the drive motor; a determination module, configured to determine a target transformation method for transforming into a frequency-domain current signal based on the operating state and the time-domain current signal; a detection module, configured to use the target transformation method to determine the spectral characteristics of the frequency-domain current signal, and detect whether the drive motor fails based on the spectral characteristics, and when it is detected that the drive motor fails, identify the current fault type of the drive motor.

[0022] A drive motor fault detection device for a vehicle according to an embodiment of the present application, an acquisition module is configured to acquire a time-domain current signal of the drive motor of the vehicle during operation, and identify the current operating state of the drive motor, a determination module is configured to determine a target transformation method for transforming into a frequency-domain current signal based on the operating state and the time-domain current signal, a detection module is configured to determine the spectral characteristics of the frequency-domain current signal by using the target transformation method, and detect whether the drive motor is faulty based on the spectral characteristics, and identify the current fault type of the drive motor when it is detected that the drive motor has a fault. Thus, the device can solve the disadvantages of the existing drive motor rotor eccentricity detection technology, such as high detection difficulty and low accuracy, and can accurately detect whether the drive motor has a fault and the fault type when a fault occurs.

[0023] Additional aspects and advantages of the present application will be given in part in the following description, become apparent in part from the following description, or be learned through the practice of the present application. Brief Description of the Drawings

[0024] Figure 1 is a flowchart of a drive motor fault detection method for a vehicle according to an embodiment of the present application;

[0025] Figure 2 is a flowchart of a drive motor fault detection method for a vehicle according to a specific example of the present application;

[0026] Figure 3 is a block diagram of a vehicle according to an embodiment of the present application;

[0027] Figure 4 is a block diagram of a drive motor fault detection device for a vehicle according to an embodiment of the present application. Detailed Description of the Embodiments

[0028] Embodiments of the present application will be described in detail below. Examples of the embodiments are shown in the drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary and are intended to explain the present application, and should not be construed as limiting the present application.

[0029] The drive motor fault detection method, computer-readable storage medium, vehicle, and drive motor fault detection device for a vehicle proposed by embodiments of the present application will be described below with reference to the drawings.

[0030] Figure 1 is a flowchart of a drive motor fault detection method for a vehicle according to an embodiment of the present application.

[0031] As Figure 1 shown, the drive motor fault detection method for a vehicle according to an embodiment of the present application may include the following steps:

[0032] S1. Obtain the time-domain current signal of the vehicle's drive motor during operation, and identify the current operating state of the drive motor;

[0033] S2. Determine the target transformation method for converting to the frequency-domain current signal based on the operating state and the time-domain current signal;

[0034] S3. Use the target transformation method to determine the spectral characteristics of the frequency-domain current signal, and detect whether the drive motor is faulty based on the spectral characteristics. When a fault of the drive motor is detected, identify the current fault type of the drive motor.

[0035] Specifically, first obtain the time-domain current signal of the vehicle's drive motor during operation. For example, the time-domain current signal of the drive motor during operation can be collected in real time through current sensors installed on the vehicle. These sensors can record the change of current over time and generate a time series data. In addition, the collected current signal may contain noise and interference, and preprocessing such as filtering (removing high-frequency noise) and denoising (smoothing the signal) is also required to obtain the real-time current signal that can be used for data analysis.

[0036] And identify the current operating state of the drive motor. For example, the road condition ahead can be identified through an in-vehicle camera. When the vehicle is driving on a smooth road, it can be determined that the motor is in a stable state. When the vehicle is driving on a bumpy road, the motor is subjected to external impacts and the current signal shows transient changes, so it can be determined that the motor is in an unstable state. Another example is to calculate the autocorrelation function of the preprocessed current signal. The autocorrelation function can reflect the periodicity and stability of the signal, and the current operating state of the drive motor can be determined based on the magnitude relationship between the value of the autocorrelation function at a preset lag and the preset threshold.

[0037] After determining the time-domain current signal of the drive motor during operation and identifying the current operating state of the drive motor, the target transformation method for converting it into a frequency-domain current signal can be determined based on the operating state and the time-domain current signal. That is, the frequency-domain current signal is the signal obtained by converting the time-domain current signal through a mathematical transformation, which reflects the amplitude and energy distribution of different frequency components in the signal. The target transformation method refers to the specific method of converting the time-domain current signal into a frequency-domain current signal. Different transformation methods are applicable to different signal characteristics and operating states. For example, the target transformation method can be the fast Fourier transform, which is suitable for analyzing stable and periodically changing signals, or the discrete wavelet transform, which is suitable for analyzing unstable, transiently changing or non-periodic signals, or the short-time Fourier transform, which is suitable for analyzing non-stationary signals and provides joint time and frequency analysis, or the Hilbert-Huang transform, which is suitable for analyzing complex non-stationary signals. Thus, by analyzing the characteristics of the time-domain current signal (such as the autocorrelation function, the stationarity of the signal, etc.), it is judged whether the current operating state of the motor is stable or unstable, and then the most suitable target transformation method is selected according to the operating state and the characteristics of the time-domain current signal. For example, when the operating state is a stable state, selecting the fast Fourier transform as the target transformation method can effectively extract the fundamental frequency and the frequencies and amplitudes of each harmonic in the signal. When the operating state is an unstable state, selecting the discrete wavelet transform or the Hilbert-Huang transform can capture the transient changes and local characteristics of the signal for analyzing non-stationary signals.

[0038] After determining the target transformation method, the spectral characteristics of the frequency-domain current signal can be determined using the target transformation method, so as to detect whether the drive motor is faulty based on the spectral characteristics, and identify the current fault type of the drive motor in the case of detecting that the drive motor has a fault. Among them, the spectral characteristics refer to the characteristics in the frequency-domain current signal. For example, it can include the frequencies and amplitudes of harmonics, as well as the energy distribution of different frequency bands. These characteristics can reflect the operating state and potential faults of the motor. For example, a comprehensive feature vector can be constructed by combining multiple characteristics such as frequency, amplitude, and energy distribution for fault detection and classification, and compared with the combined feature vector constructed during the normal operation of the drive motor to determine whether a fault has occurred. The fusion of multiple features can improve the accuracy and robustness of fault detection. Another example is to input the extracted spectral characteristics into a preset fault model. The fault model is a model established by learning the relationship between the input data and the fault type labels and can identify whether the drive motor is faulty. And in the case of determining that the drive motor has a fault, the type of the fault that has occurred is further determined in order to take corresponding maintenance measures.

[0039] Thus, it can effectively improve the accuracy and reliability of fault diagnosis and ensure the stable operation of the motor system.

[0040] According to an embodiment of the present application, the operating state includes a first operating state and a second operating state. Identifying the current operating state of the drive motor includes: determining the autocorrelation value of the autocorrelation function at a preset lag amount based on the current signal and the autocorrelation function; determining that the operating state is the first operating state when the autocorrelation value is greater than a preset threshold; and determining that the operating state is the second operating state when the autocorrelation value is less than the preset threshold. Wherein, the preset lag amount and the preset threshold can be determined according to the actual situation.

[0041] Specifically, during the operation of the drive motor, accurately identifying its operating state is crucial for fault diagnosis, performance optimization, and the stability and reliability of the system. The current operating state of the drive motor can include a first operating state (such as a stable state) and a second operating state (an unstable state). The stable state means that the drive motor operates under normal and ideal working conditions, and its various parameters are relatively stable and predictable; while the unstable state may imply potential fault hazards, external interference, or a special operating mode of the drive motor, which requires timely attention and handling.

[0042] When identifying the current operating state of the drive motor, the autocorrelation value of the autocorrelation function at a preset lag amount can be determined based on the current signal and the autocorrelation function, and the operating state can be determined according to the magnitude relationship between the autocorrelation value and the preset threshold. For example, the existing motor current sensor of the vehicle can be used to collect the stator current signal of the motor in real time. The current signal is a direct reflection of the operating state of the motor and contains rich information, such as the load condition, operating frequency, fault characteristics, etc. of the motor. The collected current signal is usually a time series data, which reflects the current change of the motor at different time points. In addition, the collected current signal needs to be preprocessed, such as filtering and denoising, to remove possible interference signals and noise to ensure the accuracy of subsequent analysis. For example, a low-pass filter can be used to remove high-frequency noise, or a wavelet denoising method can be used to process the signal to obtain a clearer and usable current signal.

[0043] Calculate the autocorrelation function of the preprocessed current signal. The autocorrelation function is a statistic that measures the similarity of a signal with itself at different time lags and can effectively reflect the characteristics of the signal, such as periodicity and stationarity. The definition of the autocorrelation function is: Wherein, R xx [m] represents the autocorrelation value at a lag amount of m, E[x] represents the expected value, m is the lag amount (the distance between two time points), N is the length of the time series of the current signal, and x is the current value. Thus, a preset lag amount m can be selected to calculate the autocorrelation value R xx[m], the preset hysteresis m can be determined according to factors such as the operating characteristics of the motor, the sampling frequency of the signal, and the actual application requirements. For example, for a current signal with a relatively high sampling frequency, a relatively small hysteresis can be selected to capture the short-term correlation of the signal; while for a signal with a relatively low sampling frequency, a larger hysteresis can be selected to analyze the long-term correlation of the signal.

[0044] Thus, the autocorrelation value R xx [m] can be compared with the preset threshold to determine the current operating state of the motor. When the autocorrelation value is greater than the preset threshold, the operating state can be determined as the first operating state, which indicates that at the preset hysteresis m, the current signal has a high similarity with itself, the signal is relatively stable, and the motor is in a stable operating state. In this state, the parameters of the driving motor are relatively stable, the operation is relatively normal, and there are generally no obvious faults or abnormal conditions. When the autocorrelation value is less than the preset threshold, the operating state can be determined as the second operating state. This shows that the similarity of the current signal at the preset hysteresis m is low, the signal has large fluctuations or changes, and the driving motor is in an unstable operating state. The unstable state may be caused by various factors, such as sudden changes in the load of the motor, rotor eccentricity, bearing damage, external interference, etc., and further analysis and diagnosis are required to determine the specific cause of the fault.

[0045] Therefore, by identifying the current operating state of the driving motor through the method based on the current signal and the autocorrelation function, it can effectively determine whether the motor is in a stable state or an unstable state. This method is simple, efficient, easy to implement, and has strong adaptability and practicality. Accurate identification of the operating state is of great significance for aspects such as motor fault diagnosis, system control and optimization, and equipment maintenance and management, which helps to improve the overall performance and reliability of the motor system and ensure the normal operation of the equipment.

[0046] According to an embodiment of the present application, the target transformation method includes the fast Fourier transform method and the discrete wavelet transform method. Determining the target transformation method for converting the time-domain current signal into the frequency-domain current signal based on the operating state and the time-domain current signal includes: when the operating state is the first operating state, determining the target transformation method for converting the time-domain current signal into the frequency-domain current signal as the fast Fourier transform method; when the operating state is the second operating state, determining the target transformation method for converting the time-domain current signal into the frequency-domain current signal as the discrete wavelet transform method.

[0047] Specifically, in the fault diagnosis and operation status monitoring of the drive motor, converting the time-domain current signal into a frequency-domain signal is an important analysis method. Frequency-domain analysis can more clearly reveal the frequency components and characteristics in the current signal, which helps to identify the fault types and operation status of the motor. That is, the target transformation methods may include the fast Fourier transform method and the discrete wavelet transform method. The fast Fourier transform is suitable for analyzing stable and periodically varying signals and can decompose the signal into sine wave components of different frequencies. It is applicable to the analysis of the current signal of the motor in a stable operation state (the first operation state). In the stable state, for example, in the motor operation state with periodic changes, the current signal of the motor usually has relatively stable periodicity and frequency characteristics. The fast Fourier transform can effectively extract these characteristics to help identify the normal operation state of the motor or some common fault modes, such as rotor eccentricity, stator winding faults, etc. The discrete wavelet transform is suitable for analyzing unstable, transiently changing or non-periodic signals and can provide detailed information about the signal at different time scales. It is applicable to the analysis of the current signal of the motor in an unstable operation state (the second operation state). In the unstable state, for example, at the moment when the drive motor starts or during non-periodic mutations, the current signal of the motor may be affected by factors such as external interference, load mutation or internal faults, showing transient changes or non-periodic characteristics. The discrete wavelet transform can capture these transient changes and local characteristics, which helps to identify sudden faults or complex fault modes of the motor.

[0048] Therefore, when determining the target transformation method for converting the time-domain current signal into a frequency-domain current signal according to the operation state and the time-domain current signal, judge the current operation state. When the operation state is the first operation state, it can be determined that the target transformation method for converting the time-domain current signal into a frequency-domain current signal is the fast Fourier transform method. When the operation state is the second operation state, it can be determined that the target transformation method for converting the time-domain current signal into a frequency-domain current signal is the discrete wavelet transform method.

[0049] That is to say, in the first operating state, the current signal of the drive motor usually has stable periodicity and frequency characteristics. For example, when the drive motor is operating normally, its current signal may be a stable sine wave or a periodic signal close to a sine wave, with a relatively fixed frequency and small amplitude variation. In this case, the fast Fourier transform is selected as the target transform method. The fast Fourier transform can effectively decompose the time-domain current signal into sine wave components of different frequencies and extract the frequency spectrum of the signal. Through the analysis of the fast Fourier transform, the fundamental frequency, harmonic frequencies, and their amplitudes in the current signal can be clearly seen, so as to judge whether the operating state of the motor is normal and whether there are common faults such as rotor eccentricity and stator winding faults. In the second operating state, the current signal of the drive motor usually exhibits unstable, transient changes, or non-periodic characteristics. For example, when the drive motor starts suddenly, the load changes suddenly, or a sudden fault occurs, the current signal may show transient peaks, frequency mutations, or non-periodic fluctuations. In this case, the discrete wavelet transform is selected as the target transform method. The discrete wavelet transform can provide detailed information about the signal at different time scales and capture the transient changes and local characteristics of the signal.

[0050] Therefore, this method of selecting the target transform method according to the operating state can improve the accuracy and reliability of the drive motor fault diagnosis and provide strong support for the stable operation and maintenance of the motor system.

[0051] According to an embodiment of the present application, the spectral characteristics include the frequencies and amplitudes of the harmonics in the frequency-domain current signal, and / or the energy distribution of different frequency bands in the frequency-domain current signal. Determining the spectral characteristics of the frequency-domain current signal by using the target transform method includes: determining the frequencies and amplitudes by the fast Fourier transform method; and / or determining the energy distribution by the discrete wavelet transform method.

[0052] Specifically, the spectral characteristics may include the frequencies and amplitudes of the harmonics in the frequency-domain current signal, or may include the energy distribution of different frequency bands in the frequency-domain current signal, or may include both the frequencies and amplitudes of the harmonics in the current signal and the energy distribution of different frequency bands in the frequency-domain current signal. That is, the spectral characteristics refer to the manifestation form of the signal in the frequency domain, which describes the amplitude (amplitude) and energy distribution of different frequency components in the signal. The frequencies and amplitudes of the harmonics represent the intensity (amplitude) and specific frequency values of different frequency components in the signal, and the energy distribution of different frequency bands represents the degree of energy concentration of the signal in different frequency bands.

[0053] When determining the spectral characteristics of the frequency-domain current signal by using the target transform method, the frequencies and amplitudes can be determined by the fast Fourier transform method. For example, extract the motor current signal x[n] with a time series length of N (N is a power of 2), where n = 0, 1,..., N - 1, and its discrete Fourier transform is defined as: Among them, X[k] is the frequency-domain feature, is the rotation factor, and j is the imaginary unit. The fast Fourier transform divides the current signal with a time series length of N into two sub-discrete Fourier transforms according to odd and even numbers, each with a size of N / 2. Calculate their X[k] respectively and finally synthesize them. The frequency-domain feature can be obtained as follows: Among them, Its frequency-domain feature includes the frequencies and amplitudes of each harmonic component.

[0054] Or, when determining the spectral characteristics of the frequency-domain current signal using the target transformation method, the energy distribution can also be determined through the discrete wavelet transform method. That is, the discrete wavelet transform is a multi-resolution decomposition that decomposes the input signal into two frequency bands: low frequency and high frequency. Finally, wavelet coefficients in different frequency bands can be obtained. The discrete wavelet transform can be defined as: c j.k = ∑ n x[n] * ψ j.k [n], where Among them, among them, c j.k is the wavelet coefficient, which is the wavelet transform value of the signal at scale j and position k. x[n] is the discrete signal in the time-domain signal, and ψ j.k [n] is the discrete wavelet basis function, j is the scale parameter that controls the stretching of the wavelet function, k is the position parameter that controls the translation of the wavelet function, and ψ(t) is the mother wavelet function. The mother wavelet function ψ(t) adopts the Daubechies4 (db4) form, and its low-pass (scale) and high-pass (wavelet) filter coefficients are: low-pass filter coefficient h(n): High-pass filter coefficient g(n): g(n) = (-1) n h[1 - n]. Assume a discrete time signal x[n] with a length of N. For the j-th layer of decomposition, the approximation coefficient (low frequency) a j [k] is expressed as: a j [k] = ∑ n x[n]h[2k - n], and the detail coefficient (high frequency) d j [k] is expressed as: d j [k] = ∑ n x[n]g[2k - n]. Thus, by obtaining the approximation coefficient (low frequency) and the detail coefficient (high frequency) in each decomposition process, both of which are projections of the original signal on a specific frequency band, the sum of their squares (energy) can represent the energy distribution of the signal on the corresponding frequency band.

[0055] In addition, when the operating state of the motor over a period of time includes both a stable state (the first operating state) and an unstable state (the second operating state), the frequency and amplitude can be determined by means of fast Fourier transform, and the energy distribution can be determined by means of discrete wavelet transform, so as to obtain the spectral characteristics (frequency, amplitude, and energy distribution in different frequency bands) of the current frequency-domain signal.

[0056] Thus, by using the target transform method to determine the spectral characteristics of the frequency-domain current signal, the operating state and potential fault characteristics of the motor can be analyzed comprehensively and accurately. The fast Fourier transform can effectively extract the harmonic frequencies and amplitudes in the signal and is suitable for analyzing the current signal of the motor in the stable operating state; the discrete wavelet transform can capture the energy distribution in different frequency bands of the signal and is suitable for analyzing the current signal of the motor in the unstable operating state. Combining the spectral characteristic analysis of the fast Fourier transform and the discrete wavelet transform can improve the accuracy and reliability of motor fault diagnosis and provide strong support for the stable operation and maintenance of the motor system.

[0057] According to an embodiment of the present application, based on the spectral characteristics, it is detected whether the drive motor is faulty, and when it is detected that the drive motor has a fault, the current fault type of the drive motor is identified, including: inputting the spectral characteristics into a preset motor fault model, where the motor fault model includes a first fault model and a second fault model, and the first fault model and the second fault model are used to identify whether the drive motor is faulty and the fault type based on the spectral characteristics; determining the confidence level of the frequency-domain current signal belonging to each fault type based on the first fault model; when the confidence level is greater than or equal to a preset confidence threshold, determining whether the drive motor is faulty based on the spectral characteristics and the first fault model, and when it is detected that the drive motor has a fault, identifying the current fault type of the drive motor; when the confidence level is less than the preset confidence threshold, determining whether the drive motor is faulty based on the spectral characteristics and the second fault model, and when it is detected that the drive motor has a fault, identifying the current fault type of the drive motor. Among them, the preset confidence threshold can be determined according to the actual situation.

[0058] Specifically, when detecting whether the drive motor is faulty according to the spectral characteristics and identifying the current fault type of the drive motor in the case of detecting a fault in the drive motor, first input the spectral characteristics into a preset motor fault model. The motor fault model includes a first fault model and a second fault model, and the first fault model and the second fault model are used to identify whether the drive motor is faulty and the fault type based on the spectral characteristics. That is to say, the role of the first fault model and the second fault model is to identify whether the drive motor is faulty and the fault type based on the spectral characteristics. The first fault model is used to initially judge the confidence level of the time-domain current signal belonging to each fault type, and the second fault model is used to further analyze the spectral characteristics to determine whether the drive motor is faulty and identify the current fault type when the confidence level of the first fault model is relatively low.

[0059] The confidence level of the frequency-domain current signal belonging to each fault type can be determined according to the first fault model. The confidence level is a probability value, indicating the confidence level of the model in the current signal belonging to a certain fault type. Compare the confidence level output by the first fault model with a preset confidence threshold. The preset confidence threshold is a preset value used to judge whether the confidence level of the model is high enough, so as to decide whether to directly use the result of the first fault model. If the confidence level is greater than or equal to the preset confidence threshold, it means that the judgment of the fault type by the first fault model is relatively reliable, and directly use the result of the first fault model to determine whether the drive motor is faulty and identify the current fault type. If the confidence level is less than the preset confidence threshold, it means that the judgment of the fault type by the first fault model is not reliable enough, and it is necessary to further use the second fault model for more accurate analysis.

[0060] For example, the preset confidence threshold is 100%. When the confidence level of the frequency-domain current signal belonging to a certain fault type determined by the first fault model is 100%, there is no need to perform further fault analysis through the second fault model. When the confidence level of the frequency-domain current signal belonging to a certain fault type determined by the first fault model is less than 100%, for example, for sample A, 40% belongs to fault type 1; 30% belongs to fault type 2; 30% belongs to fault type 3. The first fault model will determine the fault type of output sample A as 1, and its confidence level is 40%, and the fault judgment accuracy rate is relatively low. At this time, it is necessary to input the data of sample A into the second fault model to output a motor fault and confirm that the fault type is one of 1, 2, or 3. This integrated mechanism can improve efficiency and increase the accuracy of motor fault determination at the same time.

[0061] Thus, by combining the first fault model and the second fault model, efficient detection and accurate identification of drive motor faults can be achieved. The first fault model is used to preliminarily judge the fault type and its confidence level. When the confidence level is high enough, the result of the first fault model is directly used; when the confidence level is low, the second fault model is further used for more accurate analysis. This integrated method fully utilizes the advantages of the two models, improves the accuracy and reliability of fault diagnosis, and provides strong support for the stable operation and maintenance of the motor system.

[0062] According to an embodiment of the present application, the first fault model is a K-nearest neighbor classification algorithm model, which is trained based on the spectral characteristics of the frequency-domain current signal when the drive motor is operating normally and the spectral characteristics of the frequency-domain current signal with fault type labels when the drive motor is operating with faults. The second fault model is a long short-term memory neural network algorithm model, which is trained based on the spectral characteristics of the frequency-domain current signal screened by the K-nearest neighbor classification algorithm model and the spectral characteristics of the frequency-domain current signal when the drive motor is actually operating.

[0063] Specifically, the first fault model is a K-nearest neighbor classification algorithm model, which is trained based on the spectral characteristics of the frequency-domain current signal when the drive motor is operating normally and the spectral characteristics of the frequency-domain current signal with fault type labels when the drive motor is operating with faults. That is to say, the spectral characteristics of the frequency-domain current signal during normal operation are extracted from the current signals collected when the drive motor is in a normal operating state. Through fast Fourier transform and discrete wavelet transform, the time-domain current signal is converted into a frequency-domain signal to obtain spectral characteristics, including the frequency and amplitude of harmonics, and the energy distribution in different frequency bands. The spectral characteristics of the frequency-domain current signal during fault operation are extracted from the current signals collected when the drive motor is in different fault states. Similarly, through fast Fourier transform and discrete wavelet transform processing, spectral characteristics are obtained, and these data are labeled with clear fault types for the model to learn the signal characteristics in different fault states. For example, the labels can be "rotor eccentricity", "stator winding fault", "bearing damage", etc.

[0064] The K-nearest neighbor classification algorithm model calculates the distance (such as Euclidean distance) between the input spectral characteristics and the samples in the training dataset, finds the K nearest neighbor samples, and votes according to the fault type labels of these K neighbor samples to obtain the confidence levels of the current input signal belonging to various fault types. If the confidence level is higher than the preset threshold, the K-nearest neighbor classification algorithm model can directly determine whether the motor is faulty and identify the fault type. When establishing the K-nearest neighbor classification algorithm model, the frequency-domain signal features can be divided into a 70% training set and a 30% validation set. The Euclidean distance is used to evaluate the classification of fault data, and the optimal K value of the model is gradually confirmed based on the cross-validation method. When the fault classification accuracy is greater than 80%, it is considered that the K-nearest neighbor classification algorithm model is established.

[0065] The specific steps may include: 1. Initializing the K-nearest neighbor classification algorithm model: Selecting a suitable implementation method of the K-nearest neighbor classification algorithm and determining the initial value range of K. For example, K can be tried from 1 to 20. 2. Distance calculation: For each sample in the training set, calculate its Euclidean distance (or other common distance calculations) from the sample to be classified. 3. Finding the nearest neighbors: According to the calculated distances, find the K samples that are closest to the sample to be classified. 4. Class voting: Count the frequencies of various labels appearing in these K nearest neighbor samples, and use the label with the highest frequency as the predicted label of the sample to be classified. 5. Cross-validation and K value optimization, Cross-validation analysis: Further divide the training set into several subsets (such as 5-fold or 10-fold). Each time, use 4 / 5 or 9 / 10 of them as the training subset, and the remaining 1 / 5 or 1 / 10 as the validation subset. K value traversal: Within the preset K value range, successively take different K values and perform cross-validation for each K value. Model evaluation: In each cross-validation, use the K-nearest neighbor classification algorithm model with the current K value to predict the validation subset, and calculate evaluation metrics such as accuracy, recall rate, and F1 score to comprehensively evaluate the model performance. Determining the optimal K value: Compare the evaluation metrics corresponding to different K values, and select the K value that makes the model performance optimal (such as the highest accuracy, the largest F1 score, etc.) as the optimal K value. 6. Use the determined optimal K value to re-input the entire dataset into the K-nearest neighbor classification algorithm model to obtain the preliminarily screened fault data.

[0066] The second fault model is a long short-term memory neural network algorithm model, which is trained based on the spectral characteristics of the frequency-domain current signal screened by the K-nearest neighbor classification algorithm model and the spectral characteristics of the frequency-domain current signal during the actual operation of the drive motor. That is to say, for the spectral characteristics of the frequency-domain current signal after the K-nearest neighbor classification algorithm model, this part of the data is preliminarily screened by the K-nearest neighbor classification algorithm model. The K-nearest neighbor classification algorithm model identifies samples that may have faults from a large number of frequency-domain current signals and marks their fault types. These screened and marked data are used as part of the training set of the long short-term memory neural network algorithm model to help the long short-term memory neural network algorithm model learn more complex fault feature patterns. For the spectral characteristics of the frequency-domain current signal during actual operation, this part of the data is extracted from the current signals collected during the actual operation of the drive motor. Through fast Fourier transform and discrete wavelet transform processing, the spectral characteristics are obtained. These data reflect the real signal characteristics of the motor during actual operation and are used to train the long short-term memory neural network algorithm model to enable it to adapt to complex situations in the actual operation environment.

[0067] When establishing the long short-term memory neural network algorithm model, the obtained current signals of the K-nearest neighbor classification algorithm model can be randomly divided into three parts: 70% of the current signals as the training set, 15% as the validation set, and 15% as the test set. By adjusting the model structure parameters (number of layers, number of units) and training parameters (learning rate, batch size, number of iterations, etc.) of the long short-term neural network model, the model error is reduced. Finally, if the fault judgment accuracy rate of the actual motor fault type is greater than 90%, it is considered that the motor fault analysis model based on artificial intelligence is established.

[0068] It should be noted that in another embodiment of the present application, the first fault model can also be trained based on other supervised classification algorithms. For example, if the data volume is small and the features are relatively linear, the model can be trained based on logistic regression or naive Bayes algorithms. In addition to being trained based on the long short-term memory neural network algorithm, the second fault model can also be trained based on feedforward neural networks, recurrent neural networks, or convolutional neural networks. And in the case where the input is a time series (such as a frequency-domain sequence at multiple time points (the frequency-domain sequence of the motor current signal)), the long short-term memory network algorithm has better effects than other types of neural networks (feedforward neural networks, recurrent neural networks, convolutional neural networks) in capturing the dynamic temporal relationships in the time series, that is, the prediction results are also more accurate.

[0069] According to an embodiment of the present application, after identifying the current fault type of the drive motor, the drive motor fault detection method of the vehicle further includes: in the case where the current fault type is a bearing damage fault, issuing a fault warning of bearing damage and determining a maintenance measure of parking and unpacking for inspection; in the case where the current fault type is a stator winding fault, issuing a fault warning of stator winding fault and determining a maintenance measure of adjusting the vehicle driving mode; in the case where the current fault type is a rotor eccentricity fault, issuing a fault warning of rotor eccentricity fault and determining a maintenance measure of correcting the dynamic balance of the rotor.

[0070] Specifically, after identifying the current fault type of the drive motor, corresponding fault warnings can also be issued according to different fault types, and targeted maintenance measures can be taken. The fault types may include but are not limited to bearing damage, stator winding fault, rotor eccentricity, etc. According to the identified fault type, a corresponding fault warning can be issued. The purpose of the fault warning is to notify relevant personnel or systems in a timely manner that there may be a fault in the drive motor and further measures need to be taken, and corresponding maintenance measures are determined according to the fault type. Different fault types may require different handling methods to ensure the safe operation of the motor and the stability of the system.

[0071] For example, in the case where the current fault type is a bearing damage fault, a fault warning of bearing damage can be issued. The warning can be carried out through system prompts, alarm sounds, indicator lights, etc. to notify the driver or maintenance personnel that there may be a problem with the motor bearing. The warning information should clearly indicate that the fault type is "bearing damage" and prompt that immediate measures need to be taken. The maintenance measures may include parking and unpacking for inspection, that is, it is recommended to stop the vehicle immediately and conduct unpacking inspection on the motor. Bearing damage may cause abnormal operation of the motor and even lead to more serious faults, so it is necessary to check and replace the damaged bearing in time. It also includes emergency stop. If obvious abnormal vibration or noise occurs during the operation of the motor, emergency stop measures should be taken immediately to avoid further damage to the motor or other components.

[0072] In the case where the current fault type is a stator winding fault, a fault warning of stator winding fault can be issued. The warning information should clearly indicate that the fault type is "stator winding fault" and prompt that measures need to be taken. The warning can be carried out through system prompts, alarm sounds, indicator lights, etc. to notify the driver or maintenance personnel that there may be a problem with the motor stator winding. And the vehicle driving mode can be adjusted. It is recommended to adjust the vehicle driving mode to reduce the load on the motor. For example, switch to a low-power mode or limit the vehicle speed to reduce the load on the motor and avoid overheating of the stator winding or insulation aging due to long-term high-load operation. And regular inspection and maintenance are recommended. It is recommended to regularly inspect and maintain the motor, especially to detect the insulation performance of the stator winding. If insulation aging or damage is found, the winding should be repaired or replaced in time.

[0073] When the current fault type is rotor eccentricity fault, a fault warning for rotor eccentricity fault is issued. The warning message should clearly indicate that the fault type is "rotor eccentricity" and prompt that measures need to be taken. The warning can be given through system prompts, alarm sounds, indicator lights, etc., to notify the driver or maintenance personnel that the motor rotor may have an eccentricity problem. And correct the dynamic balance of the rotor. It is recommended to perform dynamic balance correction on the rotor. Rotor eccentricity may cause increased motor vibration and noise, affecting the operating efficiency and lifespan of the motor. By correcting the dynamic balance of the rotor, vibration and noise can be reduced, the operating stability of the motor can be improved, and it is recommended to regularly monitor the operating status of the motor, especially the vibration and noise levels. If it is found that the rotor eccentricity problem persists or worsens, further inspection and repair should be carried out in a timely manner.

[0074] In addition, a detailed maintenance record can be established, recording the time of each fault warning, the fault type, the maintenance measures taken, and the maintenance results, and regularly tracking the operating status of the motor to evaluate the effectiveness of the maintenance measures to ensure the safe and stable operation of the motor.

[0075] Thus, after identifying the current fault type of the drive motor, corresponding fault warnings are issued according to different fault types, and targeted maintenance measures are taken. The combination of such fault warnings and maintenance measures can effectively reduce the impact of motor faults on vehicle operation, improve the reliability and safety of the system, and at the same time extend the service life of the motor.

[0076] The following combines Figure 2 to describe the detection method of this application.

[0077] As a specific example, the drive motor fault detection method for the vehicle of this application may include the following steps:

[0078] S101, obtain the time-domain current signal of the drive motor of the vehicle during operation.

[0079] S102, determine the autocorrelation value of the autocorrelation function at a preset lag amount based on the current signal and the autocorrelation function.

[0080] S103, determine whether the autocorrelation value is greater than a preset threshold. If so, execute step S104; if not, execute step S110.

[0081] S104, determine the operating state as the first operating state, and determine that the target transformation method for converting the time-domain current signal into a frequency-domain current signal is the fast Fourier transform method.

[0082] S105, determine the frequencies and amplitudes of the harmonics in the frequency-domain current signal through the fast Fourier transform method.

[0083] S106. Determine the confidence levels of the frequency-domain current signal belonging to each fault type based on the first fault model.

[0084] S107. Determine whether the confidence level is greater than or equal to a preset confidence threshold. If yes, execute step S108; if no, execute step S112.

[0085] S108. Determine whether the drive motor is faulty based on frequency, amplitude, energy distribution, and the K-nearest neighbor classification algorithm model. When a fault of the drive motor is detected, identify the current fault type of the drive motor.

[0086] S109. When the current fault type is a bearing damage fault, issue a fault warning for bearing damage and determine the maintenance measure of stopping for disassembly and inspection; when the current fault type is a stator winding fault, issue a fault warning for stator winding fault and determine the maintenance measure of adjusting the vehicle driving mode; when the current fault type is a rotor eccentricity fault, issue a fault warning for rotor eccentricity fault and determine the maintenance measure of correcting the rotor dynamic balance.

[0087] S110. Determine that the operating state is the second operating state and determine that the target transformation method for converting the time-domain current signal into the frequency-domain current signal is the discrete wavelet transform method.

[0088] S111. Determine the energy distribution of different frequency bands in the frequency-domain current signal through the discrete wavelet transform method and enter step S106.

[0089] S112. Determine whether the drive motor is faulty based on frequency, amplitude, energy distribution, and the long short-term memory neural network algorithm model. When a fault of the drive motor is detected, identify the current fault type of the drive motor.

[0090] In summary, according to the drive motor fault detection method of the vehicle in the embodiments of the present application, obtain the time-domain current signal of the drive motor of the vehicle during operation, identify the current operating state of the drive motor, determine the target transformation method for converting into the frequency-domain current signal based on the operating state and the time-domain current signal, use the target transformation method to determine the spectral characteristics of the frequency-domain current signal, and detect whether the drive motor is faulty based on the spectral characteristics. When a fault of the drive motor is detected, identify the current fault type of the drive motor. Thus, this method can solve the problems of high difficulty and low accuracy in the existing drive motor rotor eccentricity detection technology, and can accurately detect whether the drive motor fails and the fault type when a fault occurs.

[0091] Corresponding to the above embodiments, the present application also proposes a computer-readable storage medium.

[0092] The computer-readable storage medium of the embodiment of the present application stores a program thereon, and when the program is executed by a processor, the above-mentioned driving motor fault detection method of the vehicle is implemented.

[0093] According to the computer-readable storage medium of the embodiment of the present application, by implementing the above-mentioned driving motor fault detection method of the vehicle, it is possible to solve the problems of high difficulty and low accuracy in the existing driving motor rotor eccentricity detection technology, and accurately detect whether the driving motor has a fault and the type of fault when a fault occurs.

[0094] Corresponding to the above embodiment, the present application also proposes a vehicle.

[0095] As Figure 3 shown, the vehicle 200 of the embodiment of the present application may include: a memory 210, a processor 220, and a program stored on the memory 210 and executable on the processor 220. When the processor 220 executes the program, the above-mentioned driving motor fault detection method of the vehicle is implemented.

[0096] According to the vehicle of the embodiment of the present application, by implementing the above-mentioned driving motor fault detection method of the vehicle, it is possible to solve the problems of high difficulty and low accuracy in the existing driving motor rotor eccentricity detection technology, and accurately detect whether the driving motor has a fault and the type of fault when a fault occurs.

[0097] Corresponding to the above embodiment, the present application also proposes a driving motor fault detection device for a vehicle.

[0098] As Figure 4 shown, the driving motor fault detection device 100 of the embodiment of the present application for a vehicle includes: an acquisition module 110, a determination module 120, and a detection module 130.

[0099] Among them, the acquisition module 110 is used to acquire the time-domain current signal of the driving motor of the vehicle during operation and identify the current operating state of the driving motor. The determination module 120 is used to determine the target transformation method for converting the time-domain current signal into a frequency-domain current signal based on the operating state and the time-domain current signal. The detection module 130 is used to determine the spectral characteristics of the frequency-domain current signal by using the target transformation method, and detect whether the driving motor has a fault based on the spectral characteristics, and identify the current fault type of the driving motor when it is detected that the driving motor has a fault.

[0100] According to an embodiment of the present application, the operating state includes a first operating state and a second operating state. The acquisition module 110 identifies the current operating state of the driving motor, and specifically is used to: determine the autocorrelation value of the autocorrelation function at a preset lag amount based on the current signal and the autocorrelation function; determine that the operating state is the first operating state when the autocorrelation value is greater than the preset threshold; and determine that the operating state is the second operating state when the autocorrelation value is less than the preset threshold.

[0101] According to an embodiment of the present application, the target transformation method includes a fast Fourier transform method and a discrete wavelet transform method. The determination module 120 determines the target transformation method for converting the time-domain current signal into a frequency-domain current signal based on the operating state and the time-domain current signal, specifically: when the operating state is the first operating state, determining that the target transformation method for converting the time-domain current signal into a frequency-domain current signal is the fast Fourier transform method; when the operating state is the second operating state, determining that the target transformation method for converting the time-domain current signal into a frequency-domain current signal is the discrete wavelet transform method.

[0102] According to an embodiment of the present application, the spectral characteristics include the frequencies and amplitudes of the harmonics in the frequency-domain current signal, and / or the energy distribution of different frequency bands in the frequency-domain current signal. The detection module 130 determines the spectral characteristics of the frequency-domain current signal by using the target transformation method, specifically: determining the frequencies and the amplitudes by using the fast Fourier transform method; and / or determining the energy distribution by using the discrete wavelet transform method.

[0103] According to an embodiment of the present application, the detection module 130 detects whether the drive motor is faulty based on the spectral characteristics, and when it is detected that the drive motor has a fault, identifies the current fault type of the drive motor, specifically: inputting the spectral characteristics into a preset motor fault model, where the motor fault model includes a first fault model and a second fault model, and the first fault model and the second fault model are used to identify whether the drive motor is faulty and the fault type based on the spectral characteristics; determining the confidence levels of the frequency-domain current signal belonging to each fault type based on the first fault model; when the confidence level is greater than or equal to a preset confidence threshold, determining whether the drive motor is faulty based on the spectral characteristics and the first fault model, and when it is detected that the drive motor has a fault, identifying the current fault type of the drive motor; when the confidence level is less than the preset confidence threshold, determining whether the drive motor is faulty based on the spectral characteristics and the second fault model, and when it is detected that the drive motor has a fault, identifying the current fault type of the drive motor.

[0104] According to an embodiment of the present application, the first fault model is a K-nearest neighbor classification algorithm model, which is trained based on the spectral characteristics of the frequency-domain current signal when the drive motor is operating normally and the spectral characteristics of the frequency-domain current signal with fault type labels when the drive motor is operating with a fault. The second fault model is a long short-term memory neural network algorithm model, which is trained based on the spectral characteristics of the frequency-domain current signal screened by the K-nearest neighbor classification algorithm model and the spectral characteristics of the frequency-domain current signal when the drive motor is actually operating.

[0105] According to an embodiment of the present application, the detection module 130 is further configured to: after identifying the current fault type of the drive motor, in the case where the current fault type is a bearing damage fault, issue a fault warning of bearing damage and determine a maintenance measure of stopping for unpacking inspection; in the case where the current fault type is a stator winding fault, issue a fault warning of stator winding fault and determine a maintenance measure of adjusting the vehicle driving mode; in the case where the current fault type is a rotor eccentricity fault, issue a fault warning of rotor eccentricity fault and determine a maintenance measure of correcting the rotor dynamic balance.

[0106] It should be noted that for the details not disclosed in the drive motor fault detection device of the vehicle in the embodiment of the present application, please refer to the details disclosed in the drive motor fault detection method of the vehicle in the embodiment of the present application, and will not be elaborated here specifically.

[0107] For the drive motor fault detection device of the vehicle according to the embodiment of the present application, the acquisition module is configured to acquire the time-domain current signal of the drive motor of the vehicle during operation and identify the current operating state of the drive motor, the determination module is configured to determine the target transformation method for transforming into the frequency-domain current signal based on the operating state and the time-domain current signal, and the detection module is configured to use the target transformation method to determine the spectral characteristics of the frequency-domain current signal and detect whether the drive motor is faulty based on the spectral characteristics, and in the case where it is detected that the drive motor has a fault, identify the current fault type of the drive motor. Thus, the device can solve the problems of high detection difficulty and low accuracy in the existing rotor eccentricity detection technology of the drive motor, and can accurately detect whether the drive motor has a fault and the fault type when a fault occurs.

[0108] It should be noted that the logic and / or steps represented in the flowchart or described otherwise herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in combination with these instruction execution systems, apparatuses, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in combination with an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection part with one or more wirings (electronic device), a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or otherwise processing as appropriate, and then stored in a computer memory.

[0109] It should be understood that each part of the present application can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0110] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0111] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In the description of this application, "a plurality of" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0112] In this application, unless otherwise clearly defined and limited, terms such as "installed", "connected", "coupled", "fixed", etc. shall be construed in a broad sense. For example, it may be a fixed connection, a detachable connection, or integrated; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the internal communication of two components or the interaction relationship between two components, unless otherwise clearly limited. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0113] Although the embodiments of this application have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting this application. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A method for detecting a fault in a vehicle driving motor, characterized in that: The method comprises: Acquiring a time domain current signal of a driving motor of the vehicle when it is running, and identifying a current running state of the driving motor; Determine a target conversion method for converting into a frequency domain current signal based on the operating state and the time domain current signal; The target transformation method is used to determine the frequency spectrum characteristics of the frequency domain current signal, and based on the frequency spectrum characteristics, it is detected whether the drive motor is faulty, and when it is detected that the drive motor is faulty, the current fault type of the drive motor is identified.

2. The method for detecting a fault in a driving motor of a vehicle according to claim 1, characterized in that: The operating state includes a first operating state and a second operating state, and the identifying the current operating state of the drive motor includes: Determining an autocorrelation value of the autocorrelation function at a preset hysteresis based on the current signal and the autocorrelation function; When the autocorrelation value is greater than a preset threshold, determining that the operating state is the first operating state; When the autocorrelation value is less than the preset threshold, the operating state is determined to be the second operating state.

3. The method for detecting a fault in a driving motor of a vehicle according to claim 2, characterized in that: The target transformation method includes a fast Fourier transform method and a discrete wavelet transform method. The target transformation method determined based on the operating state and the time-domain current signal to be converted into a frequency-domain current signal includes: When the operating state is the first operating state, determining that a target transformation method for converting the time-domain current signal into the frequency-domain current signal is the fast Fourier transform method; When the operating state is the second operating state, a target transformation method for converting the time-domain current signal into the frequency-domain current signal is determined to be the discrete wavelet transformation method.

4. The method for detecting a fault in a driving motor of a vehicle according to claim 3, characterized in that: The spectral characteristics include the frequency and amplitude of harmonics in the frequency domain current signal, and / or the energy distribution of different frequency bands in the frequency domain current signal. The method of determining the spectral characteristics of the frequency domain current signal by using the target transformation method includes: Determine the frequency and the amplitude by means of the fast Fourier transform; and / or, The energy distribution is determined by discrete wavelet transform.

5. The method for detecting a fault in a driving motor of a vehicle according to claim 1, characterized in that: The detecting whether the drive motor is faulty based on the frequency spectrum characteristics, and identifying the current fault type of the drive motor when the drive motor is detected to be faulty, comprises: Inputting the frequency spectrum characteristics into a preset motor fault model, wherein the motor fault model includes a first fault model and a second fault model, and the first fault model and the second fault model are used to identify whether the drive motor is faulty and the fault type based on the frequency spectrum characteristics; Determining the confidence level that the frequency-domain current signal belongs to each fault type based on the first fault model; In the case where the confidence is greater than or equal to a preset confidence threshold, determining whether the drive motor is faulty based on the frequency spectrum characteristics and the first fault model, and identifying a current fault type of the drive motor when a fault is detected in the drive motor; When the confidence is less than the preset confidence threshold, whether the drive motor is faulty is determined based on the frequency spectrum characteristics and the second fault model, and when a fault is detected in the drive motor, a current fault type of the drive motor is identified.

6. The method for detecting a fault in a driving motor of a vehicle according to claim 5, characterized in that: The first fault model is a K nearest neighbor classification algorithm model, which is trained based on the spectral characteristics of the frequency domain current signal when the drive motor is operating normally and the spectral characteristics of the frequency domain current signal with a fault type label when the drive motor is operating faultily. The second fault model is a long short-term memory neural network algorithm model, which is trained based on the spectral characteristics of the frequency domain current signal after being screened by the K nearest neighbor classification algorithm model and the spectral characteristics of the frequency domain current signal when the drive motor is actually operating.

7. The method for detecting a fault in a driving motor of a vehicle according to claim 1, characterized in that: After identifying the current fault type of the drive motor, the method further includes: In the case where the current fault type is a bearing damage fault, a bearing damage fault warning is issued, and maintenance measures such as parking and unpacking for inspection are determined; In the case where the current fault type is a stator winding fault, issuing a fault warning of the stator winding fault, and determining a maintenance measure for adjusting a vehicle driving mode; In the case where the current fault type is a rotor eccentricity fault, a fault warning of the rotor eccentricity fault is issued, and maintenance measures for correcting the dynamic balance of the rotor are determined.

8. A computer-readable storage medium, characterized in that: A program is stored thereon, and when the program is executed by a processor, a method for detecting a fault in a drive motor of a vehicle according to any one of claims 1 to 7 is implemented.

9. A vehicle, characterized in that: include: A memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, a method for detecting a fault in a drive motor of a vehicle according to any one of claims 1 to 7 is implemented.

10. A vehicle drive motor fault detection device, characterized in that: The device comprises: An acquisition module, used to acquire a time domain current signal of a driving motor of the vehicle when the driving motor is running, and to identify a current running state of the driving motor; A determination module, configured to determine a target conversion method for converting the current signal into a frequency domain current signal based on the operating state and the time domain current signal; A detection module is used to determine the frequency spectrum characteristics of the frequency domain current signal by using the target transformation method, and detect whether the drive motor is faulty based on the frequency spectrum characteristics, and identify the current fault type of the drive motor when a fault is detected in the drive motor.

Images