Motor fault diagnosis method and device, vehicle and storage medium

Abstract

The invention provides a motor fault diagnosis method and device, a vehicle and a storage medium. The method comprises the following steps: determining that a motor has a preliminary fault according to an inductance signal and / or an operation signal of the motor; and diagnosing the fault type of the motor according to the inductance signal and the operation signal. The method only depends on an existing sensor, does not need to use a vibration signal, firstly determines whether the motor has a preliminary fault according to the inductance signal and / or current and winding temperature and other operation signals of the motor obtained by the high-frequency current signal, and then determines whether the motor has the preliminary fault according to the preliminary fault diagnosis result as required or directly according to the inductance signal and the operation signal. And the joint hierarchical fusion diagnosis method is selected to further diagnose the fault type of the motor, so that the accuracy of the diagnosis process, the accuracy of the diagnosis result and the diagnosis efficiency are improved, and the diagnosis cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment detection, and in particular to a motor fault diagnosis method, a motor fault diagnosis device, a vehicle and a computer-readable storage medium. Background Art

[0002] In related technologies, motor fault diagnosis methods are usually implemented using a single vibration signal or current signal, or by collecting vibration signals, three-phase current signals, and temperature signals and performing multimodal data fusion on these three signals.

[0003] However, the above method has problems such as inaccurate diagnostic process, inaccurate diagnostic results, low diagnostic efficiency and high cost. For example, vibration signals are easily affected by external vibrations and noise, resulting in inaccurate diagnostic results, and it is impossible to diagnose multiple faults at the same time, resulting in low diagnostic efficiency. The need to add vibration sensors leads to high costs. Summary of the Invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art.

[0005] To this end, an object of the present invention is to provide a motor fault diagnosis method, which improves the precision of the diagnosis process, the accuracy of the diagnosis results and the diagnosis efficiency, and reduces the diagnosis cost.

[0006] To this end, a second object of the present invention is to provide a fault diagnosis device for a motor.

[0007] To this end, a third object of the present invention is to provide a vehicle.

[0008] To this end, a fourth object of the present invention is to provide a computer-readable storage medium.

[0009] In order to achieve the above-mentioned purpose, an embodiment of the first aspect of the present invention proposes a motor fault diagnosis method, which includes: determining that the motor has a preliminary fault based on the inductance signal and / or operating signal of the motor; and diagnosing the fault type of the motor based on the inductance signal and the operating signal.

[0010] The motor fault diagnosis method according to an embodiment of the present invention relies only on existing sensors and does not require the use of vibration signals. It first determines whether the motor has a preliminary fault based on the motor's inductance signal and / or operating signals such as current and winding temperature obtained from the high-frequency current signal. Then, based on the preliminary fault diagnosis result, as needed or directly based on the inductance signal and operating signal, a joint hierarchical fusion diagnosis method is selected to further diagnose the motor's fault type, thereby improving the accuracy of the diagnosis process, the accuracy of the diagnosis results, and the diagnosis efficiency, and reducing the diagnosis cost.

[0011] In some embodiments, diagnosing the fault type of the motor based on the inductance signal and the operating signal includes: determining a timing characteristic signal based on the operating signal and the inductance signal; determining a fusion characteristic signal based on the timing characteristic signal and the operating signal; and determining the fault type of the motor based on the fusion characteristic signal.

[0012] In some embodiments, the operating signal includes: a current signal, and determining the timing characteristic signal based on the operating signal and the inductance signal includes: determining a first characteristic signal based on the current signal; determining a second characteristic signal based on the inductance signal; and determining a timing characteristic signal based on the first characteristic signal and the second characteristic signal.

[0013] In some embodiments, the operating signal includes a winding temperature signal, and determining the fused characteristic signal based on the timing characteristic signal and the operating signal includes: determining a third characteristic signal based on the winding temperature signal; and determining the fused characteristic signal based on the timing characteristic signal and the third characteristic signal.

[0014] In some embodiments, determining the first characteristic signal based on the current signal includes: processing the current signal based on wavelet transform to determine wavelet coefficients at different frequencies; determining a one-dimensional time domain signal of the current signal based on a combination of coefficients of the wavelet coefficients; determining a two-dimensional time-frequency graph based on the one-dimensional time domain signal; and determining the first characteristic signal based on the two-dimensional time-frequency graph.

[0015] In some embodiments, determining that a preliminary fault has occurred in the motor based on the inductance signal and / or operating signal of the motor includes: determining a fundamental frequency amplitude based on the inductance signal; and determining that a preliminary fault has occurred in the motor when the fundamental frequency amplitude is greater than a preset frequency threshold and / or the operating signal is abnormal.

[0016] In some embodiments, before determining that the motor has a preliminary fault based on the inductance signal and / or operation signal of the motor, the method further includes: injecting a high-frequency rotational pulsating voltage signal into the d-axis and obtaining a current signal during the operation of the motor; determining a d-axis current signal in the dq coordinate system based on the current signal; and obtaining the inductance signal of the motor based on the d-axis current signal.

[0017] In some embodiments, acquiring the inductance signal of the motor according to the d-axis current signal includes: acquiring a high-frequency current signal in the d-axis current signal based on a low-pass filter; and acquiring the inductance signal according to the high-frequency current signal.

[0018] In some embodiments, before obtaining the inductance signal of the motor, the method further includes: determining that a phase loss fault occurs in the motor when the current signal of the motor is unbalanced and the winding temperature signal of the motor exceeds a preset temperature range.

[0019] In order to achieve the above-mentioned purpose, an embodiment of the second aspect of the present invention proposes a motor fault diagnosis device, comprising: a controller, used to determine that the motor has a preliminary fault based on the inductance signal and / or operating signal of the motor; an edge module, connected to the controller, used to diagnose the fault type of the motor based on the inductance signal and the operating signal.

[0020] According to the fault diagnosis device of the motor according to the embodiment of the present invention, only existing sensors are relied upon, and vibration signals are not required. First, the inductance signal and / or operating signals such as current and winding temperature of the motor are obtained based on the high-frequency current signal to determine whether a preliminary fault has occurred in the motor. Then, according to the preliminary fault diagnosis result, a joint hierarchical fusion diagnosis method is selected to further diagnose the fault type of the motor as required or directly based on the inductance signal and operating signal, thereby improving the accuracy of the diagnosis process, the accuracy of the diagnosis results and the diagnosis efficiency, and reducing the diagnosis cost.

[0021] In order to achieve the above-mentioned object, an embodiment of a third aspect of the present invention provides a vehicle, comprising: a fault diagnosis device for a motor as described in the above-mentioned embodiment.

[0022] According to an embodiment of the present invention, a vehicle equipped with a motor fault diagnosis device of the above embodiment relies only on existing sensors and does not need to use vibration signals. It first determines whether the motor has a preliminary fault based on the motor's inductance signal and / or operating signals such as current and winding temperature obtained from the high-frequency current signal. Then, based on the preliminary fault diagnosis result, as needed or directly based on the inductance signal and operating signal, a joint hierarchical fusion diagnosis method is selected to further diagnose the motor fault type, thereby improving the accuracy of the diagnosis process, the accuracy of the diagnosis results, and the diagnosis efficiency, and reducing the diagnosis cost.

[0023] In order to achieve the above-mentioned purpose, an embodiment of the fourth aspect of the present invention proposes a computer-readable storage medium, on which a fault diagnosis program for a motor is stored. When the fault diagnosis program for the motor is executed by a processor, a device equipped with the fault diagnosis program for the motor implements the fault diagnosis method for the motor as described in the above-mentioned embodiment.

[0024] Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description which follows, or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments with reference to the following drawings, in which: Figure 1 is a flow chart of a motor fault diagnosis method according to one embodiment of the present invention; Figure 2 is a block diagram of a motor fault diagnosis device according to an embodiment of the present invention; Figure 3 is a schematic diagram of multi-level fusion diagnosis according to one embodiment of the present invention; Figure 4 is a flowchart of preliminary diagnosis of motor faults according to one embodiment of the present invention; Figure 5 is a block diagram of a pulse voltage injection method according to one embodiment of the present invention; Figure 6 is a flow chart of a motor fault diagnosis method according to another embodiment of the present invention; Figure 7 is a block diagram of a motor fault diagnosis device according to another embodiment of the present invention; Figure 8 is a block diagram of a vehicle according to one embodiment of the present invention.

[0026] Reference numerals: Motor state analysis 60; Motor signal acquisition 61; Main controller 62; Edge device 63; A fault diagnosis device 100 for a motor; Controller 98; Edge module 99; Vehicle 101. DETAILED DESCRIPTION

[0027] The embodiments described with reference to the drawings are exemplary, and embodiments of the present invention are described in detail below.

[0028] Motors are essential equipment in industrial production and daily life, and their operating status directly affects system stability and efficiency. During long-term operation, motors may experience various faults, such as bearing failure, winding short circuits, and rotor imbalance.

[0029] For example, motor fault diagnosis methods typically use a single vibration or current signal. However, these methods cannot diagnose multiple faults simultaneously and require the addition of vibration sensors, resulting in inaccurate diagnostics, high diagnostic costs, and low diagnostic efficiency.

[0030] For example, the Kalman filter algorithm is used to suppress and remove noise in vibration signals, three-phase current signals, and temperature signals; a sample feature set is established; SimCLR comparative learning is used to enhance the representation ability and quality of different modal features of each signal source in the sample feature set; combined with the multi-head attention mechanism network framework, the weighted features are linearly combined or spliced to generate the final fusion feature representation to complete motor fault diagnosis.

[0031] However, after collecting vibration, current, and temperature signals, the above method directly enters data fusion diagnosis. The data fusion diagnosis process consumes a large amount of CPU resources, resulting in high diagnostic costs. In addition, the vibration signals used are easily affected by external vibration and noise. Early faults may not be easy to detect and may take a long time to discover, resulting in inaccurate diagnostic results.

[0032] Therefore, the motor fault diagnosis method of the embodiment of the present invention only relies on existing sensors and does not need to use vibration signals. It first determines whether the motor has a preliminary fault based on the inductance signal and / or current and winding temperature and other operating signals of the motor obtained according to the high-frequency current signal. Then, according to the preliminary fault diagnosis result, as needed or directly based on the inductance signal and operating signal, a joint hierarchical fusion diagnosis method is selected to further diagnose the fault type of the motor, thereby improving the accuracy of the diagnosis process, the accuracy of the diagnosis results and the diagnosis efficiency, and reducing the diagnosis cost.

[0033] The following combination Figures 1-6 A fault diagnosis method for a motor according to an embodiment of the present invention is described.

[0034] like Figure 1 FIG. 1 is a flow chart of a motor fault diagnosis method according to an embodiment of the present invention. The motor fault diagnosis method according to the embodiment of the present invention comprises at least step S1 and step S2.

[0035] Step S1: determining whether a preliminary fault occurs in the motor based on the inductance signal and / or operation signal of the motor.

[0036] In an embodiment, Figure 2 Figure 2 shows a block diagram of a motor fault diagnosis device according to an embodiment of the present invention. Operating signals include three-phase current and winding temperature. The motor inductance signal is extracted by processing the collected motor current signal and filtering it with a filter.

[0037] A preliminary fault of the motor is determined based on the inductance signal and / or operation signal of the motor. For example, in the motor signal acquisition 61, the three-phase current and winding temperature are collected by sensors. In the motor state analysis 60, it is determined whether the current motor has a phase loss fault in the preliminary fault by observing whether the three-phase current is balanced, whether the temperature of the motor winding is within the normal range, and whether there is abnormal temperature rise. Specifically, if the winding temperature signal of the motor exceeds the preset temperature range, it is considered that the motor winding temperature is abnormal and there is abnormal temperature rise. When the current signal of the motor is unbalanced and the winding temperature signal of the motor exceeds the preset temperature range, it can be determined that the motor has a phase loss fault.

[0038] The extracted inductance signal is processed in the edge device 63. Using Fourier transform, the continuous inductance signal Convert to the frequency domain and extract the amplitude of the fundamental frequency ,When the fundamental frequency amplitude is greater than the maximum value of the normal fundamental frequency range, it is determined that the motor has a preliminary fault.

[0039] The preliminary fault of the motor is determined based on the inductance signal and / or operation signal of the motor. If the preliminary judgment is abnormal, the fusion diagnosis can be entered according to the demand or directly selected. Otherwise, the fusion diagnosis will not be entered. Since the fusion diagnosis process consumes a lot of CPU resources, the preliminary judgment process can save CPU resources to a certain extent. In addition, when the vehicle is driving, the environment in which the motor is located is relatively complex. The extracted features are easily affected by noise, which makes it easy to misjudge and difficult to accurately judge the motor fault. The threshold is used to determine whether the motor is faulty. The multi-level fusion method (fusion diagnosis) is combined, that is, traditional motor diagnosis is combined with deep learning diagnosis, to enhance the accuracy of motor diagnosis in complex environments, and timely remind users to repair and maintain the motor to avoid accidents.

[0040] Step S2: diagnose the fault type of the motor according to the inductance signal and the operation signal.

[0041] In the embodiment, it is understandable that after the motor has a preliminary fault, it is still not clear whether the motor has fault types such as eccentricity and / or demagnetization faults. Therefore, a multi-level fusion diagnosis method can be used for further diagnosis as needed or directly based on the inductance signal and the operating signal, so as to achieve a joint judgment based on the preliminary fault diagnosis results of the motor and the multi-level fusion diagnosis results, thereby improving the accuracy of fault diagnosis, and achieving simultaneous judgment of whether the motor has multiple faults, thereby improving the efficiency of fault diagnosis. No additional sensors are added, and only existing sensors are relied upon to realize the diagnosis of motor rotor eccentricity, demagnetization, phase loss and other faults. No vibration sensor is relied upon, which saves costs and avoids the vibration signal being easily affected by external vibration and noise, and the problem that early faults may not be easy to detect and may take a long time to discover. The inductance signal is obtained using a high-frequency current signal, which can detect smaller eccentricities and is sensitive to early faults. The degree and position of the rotor eccentricity can be more accurately judged through changes in the inductance signal.

[0042] Among them, the multi-level fusion diagnosis method can collect data on the motor in normal state and under faults such as eccentricity and demagnetization by collecting current and temperature signals at different speeds, torques, and different ambient temperatures, and produce data for training. At the same time, it can collect data on vehicle faults during use, continuously update the data set of the multi-level fusion diagnosis method, and further optimize the multi-level fusion diagnosis method.

[0043] The motor fault diagnosis method according to an embodiment of the present invention relies only on existing sensors and does not require the use of vibration signals. It first determines whether the motor has a preliminary fault based on the motor's inductance signal and / or operating signals such as current and winding temperature obtained from the high-frequency current signal. Then, based on the preliminary fault diagnosis result, as needed or directly based on the inductance signal and operating signal, a joint hierarchical fusion diagnosis method is selected to further diagnose the motor's fault type, thereby improving the accuracy of the diagnosis process, the accuracy of the diagnosis results, and the diagnosis efficiency, and reducing the diagnosis cost.

[0044] In some embodiments, diagnosing the fault type of the motor based on the inductance signal and the operation signal includes: determining a timing characteristic signal based on the operation signal and the inductance signal; determining a fusion characteristic signal based on the timing characteristic signal and the operation signal; and determining the fault type of the motor based on the fusion characteristic signal.

[0045] In an embodiment, Figure 3The figure shows a schematic diagram of multi-level fusion diagnosis of an embodiment of the present invention. The operating signals include three-phase current and winding temperature; the time series feature signal is determined based on the three-phase current signal and inductance signal in the operating signal, combined with the pre-trained residual network (Resnet convolutional neural network), self-attention mechanism, residual connection and layer normalization and feedforward neural network; the time series feature signal and the winding temperature in the operating signal are fused in combination with the cross-attention mechanism to obtain a fused feature signal; the fused feature signal is passed through a fully connected layer (linear) and a probabilistic normalization function (softmax) to output a multi-level fusion diagnosis result, the preliminary diagnosis result of the motor is jointly judged with the multi-level fusion diagnosis result, and the specific type of motor fault is finally determined, such as eccentricity fault and demagnetization fault, so as to achieve accurate judgment of the motor fault type. Compared with the method of using a single signal for fault diagnosis, the embodiment of the present invention fuses current, inductance signal and temperature signal, adopts a multi-modal fusion strategy for data fusion joint diagnosis, and uses the cross-self-attention mechanism to fuse current, Inductance signals and temperature characteristics are enhanced to enhance the extracted features and the accuracy of the fault diagnosis system. It is also possible to diagnose multiple faults at the same time. At the same time, the data set is continuously updated using the fault data collected in real time to further optimize the diagnosis system.

[0046] In some embodiments, the operating signal includes: a current signal, and determining a timing characteristic signal based on the operating signal and the inductance signal includes: determining a first characteristic signal based on the current signal; determining a second characteristic signal based on the inductance signal; and determining a timing characteristic signal based on the first characteristic signal and the second characteristic signal.

[0047] In an embodiment, Figure 3 As shown, the first characteristic signal is determined according to the current signal, for example, the time-frequency characteristics of the current signal are extracted, and the first characteristic signal is obtained by combining the pre-trained residual network (Resnet convolutional neural network) and the like; the pre-trained residual network (Resnet convolutional neural network) is used to extract The features of the spectrum graph are used to obtain the second feature signal, which is used to prepare for feature fusion and determine the type of motor fault. First, the self-attention mechanism is used to fuse the first feature signal and the second feature signal, and then the residual connection and normalization layer (add&norm) and feedforward neural network (feed forward) module are used to further extract the time series features to obtain the time series feature signal, which is used to prepare for determining whether it is a feature of the fault type.

[0048] In some embodiments, the operating signal includes a winding temperature signal, and determining a fused characteristic signal based on the timing characteristic signal and the operating signal includes: determining a third characteristic signal based on the winding temperature signal; and determining a fused characteristic signal based on the timing characteristic signal and the third characteristic signal.

[0049] In an embodiment, Figure 3 As shown, the operating signal includes a winding temperature signal, and the third characteristic signal is determined based on the winding temperature signal. For example, the winding temperature is encoded and processed, and an MLP (Multilayer Perceptron) such as a feedforward artificial neural network is used for feature extraction to obtain the third characteristic signal. A cross-attention mechanism (multi-head attention) is used to fuse the time series feature signal and the third feature signal to obtain a fused feature signal, so as to adjust the model to pay more attention to important features by calculating the attention weight.

[0050] In some embodiments, determining the first characteristic signal based on the current signal includes: determining the wavelet coefficients at different frequencies by processing the current signal based on the wavelet transform; determining the one-dimensional time domain signal of the current signal based on the coefficient combination of the wavelet coefficients; determining the two-dimensional time-frequency graph based on the one-dimensional time domain signal; and determining the first characteristic signal based on the two-dimensional time-frequency graph.

[0051] In an embodiment, Figure 2 As shown, the collected current signal (for example, three-phase current) and winding temperature are transmitted to the edge device 63. In the edge device 63, the collected current signal is first processed by continuous wavelet transform. In the time domain, the wavelet is moved in time and the window signals at different positions are compared one by one to obtain the wavelet coefficients. In the frequency domain, the length of the wavelet is stretched or compressed to change the length and frequency of the wavelet to achieve the wavelet coefficients at different frequencies. The count of the wavelet coefficients is , in, is the wavelet coefficient, is the original input signal, a is the scale parameter, b is the time parameter, Odd functions of wavelet transform; The wavelet coefficients at different frequencies are determined by the current signal to prepare for determining the time domain signal; the wavelet coefficients at different frequencies are combined to extract the time-frequency characteristics of the current signal and obtain the one-dimensional time domain signal of the current signal; the one-dimensional time domain signal characteristics are converted into a two-dimensional time-frequency graph; the first characteristic signal is determined based on the two-dimensional time-frequency graph, for example, Figure 3 As shown in the figure, a pre-trained residual network (Resnet convolutional neural network) is used to extract the features of the converted two-dimensional time-frequency graph to obtain the first feature signal, in preparation for feature fusion and judgment of the motor fault type.

[0052] In some embodiments, determining whether a preliminary fault has occurred in the motor is based on the inductance signal and / or operating signal of the motor, including: determining the fundamental frequency amplitude based on the inductance signal; determining that a preliminary fault has occurred in the motor when the fundamental frequency amplitude is greater than a preset frequency threshold and / or the operating signal is abnormal.

[0053] In an embodiment, the fundamental frequency amplitude is determined based on the inductance signal, for example, Figure 2 As shown, the extracted inductance signal is processed in the edge device 63. Using Fourier transform, the continuous inductance signal Convert to the frequency domain and extract the amplitude of the fundamental frequency , in preparation for determining whether the motor has a preliminary fault; when the fundamental frequency amplitude is greater than the preset frequency threshold, and / or abnormal operating signals such as three-phase current imbalance and winding temperature exceeding the normal temperature range appear, it is determined that the motor has a preliminary fault, providing a basis for judging whether to perform fusion diagnosis. For example, Figure 4 FIG. 1 is a flowchart of a preliminary diagnosis of a motor fault according to an embodiment of the present invention. The preliminary diagnosis of a motor fault includes at least steps S10 - S14 .

[0054] Step S10: Acquire an inductance signal.

[0055] Step S11, extracting the fundamental frequency amplitude.

[0056] Step S12: Determine whether the fundamental frequency amplitude is greater than a preset frequency threshold. If so, proceed to step S13; otherwise, proceed to step S14.

[0057] Step S13: The motor is faulty.

[0058] Step S14: The motor is normal.

[0059] In some embodiments, before determining that a preliminary fault has occurred in the motor based on the inductance signal and / or operating signal of the motor, the method further includes: injecting a high-frequency rotating pulse voltage signal into the d-axis and obtaining a current signal during the operation of the motor; determining a d-axis current signal in the dq coordinate system based on the current signal; and obtaining the inductance signal of the motor based on the d-axis current signal.

[0060] In an embodiment, Figure 2 As shown, taking the permanent magnet synchronous motor as an example, the current signal, such as the three-phase current signal, is set to 、 and After the motor state analysis 60, it is necessary to further diagnose the fault type of the motor, and it is necessary to use the high-frequency pulse voltage injection method. Specifically, under the condition of ignoring the motor eddy current, hysteresis loss and magnetic steel saturation, a coordinate system is established on the permanent magnet synchronous motor rotor, and the direct axis is set as the d axis and the quadrature axis is set as the q axis. The mathematical model of the permanent magnet synchronous motor is established as follows: , in, 、 、 、 、 、 are the d and q axis voltage, current and flux respectively, is the winding resistance, is the electrical angular frequency; The magnetic flux of d and q axes is , in, 、 is the d and q axis inductance, is the magnetic flux of the permanent magnet; The high-frequency pulse voltage injection method is used in the main controller 62 to inject a high-frequency rotating pulse voltage signal into the d-axis. Since this signal is much higher than the fundamental frequency of the motor operation, the influence of the motor winding and back electromotive force can be ignored. Simplified to , in, is the d-axis inductance, is the d-axis current; Determining whether a motor fault occurs based on the motor's mathematical model can improve the accuracy of fault diagnosis. Because the motor is in a complex environment during vehicle driving, the extracted features are easily affected by noise, leading to misjudgment and making it difficult to accurately determine motor faults.

[0061] The injected voltage high-frequency rotation pulse voltage signal is , in, 、 are the amplitude and frequency of the injected signal, respectively.

[0062] By injecting a high-frequency rotating pulse voltage signal into the d-axis, preparation is made for obtaining the inductance signal; obtaining the current signal during the operation of the motor, for example, using a current sensor to collect three-phase current in the motor signal acquisition 61, obtaining the current signal during the operation of the motor, and preparing data for fault judgment; combining Figure 5 FIG. 1 is a block diagram of a pulse voltage injection method according to an embodiment of the present invention, wherein: is the d-axis current signal, The q-axis current signal, PI, and proportional-integral controller are used to dynamically adjust the error signal after high-frequency signal injection to optimize the control system's response speed and steady-state accuracy. HPF is a filter, and PMSM is a permanent magnet synchronous motor. SVPWM generates a PWM signal based on the voltage command output by the control algorithm, driving the inverter to output the target voltage vector and regulating the PMSM's torque and speed. Based on the d-axis current signal, the motor's inductance signal is obtained using filters and Clark transforms, paving the way for accurate motor fault diagnosis based on the inductance signal.

[0063] In some embodiments, obtaining the inductance signal of the motor according to the d-axis current signal includes: obtaining a high-frequency current signal in the d-axis current signal based on a low-pass filter; and obtaining the inductance signal according to the high-frequency current signal.

[0064] In an embodiment, a low-pass filter such as a Butterworth filter is used to process the d-axis current signal. The amplitude-frequency characteristic of the filter is: , The transfer function is calculated based on the cutoff frequency, stopband edge frequency, passband edge frequency, stopband minimum attenuation value, and passband maximum oscillation value of the designed filter. The minimum duration of signal injection is calculated based on the transfer function. The minimum injection duration is set as c. According to the calculated minimum duration of signal injection, the interval signal injection method is used. According to the injected voltage high-frequency rotating pulse voltage signal , then there exists , The collected three-phase current is converted to the dq coordinate system through Clark and Park transformation. The d-axis signal is first filtered using the designed Butterworth bandpass filter, and then the high-frequency current signal in the d-axis current signal is screened out using a low-pass filter. At the same time, the Butterworth high-pass filter filters the interference signal and restores the real control signal to avoid affecting the motor control, so as to improve the accuracy of high-frequency current signal acquisition and prepare for obtaining the inductance signal. By increasing the injection signal duration and interval strategy, the impact of signal injection on motor control is reduced; the high-frequency current signal is used as the inductance signal and set as , and at the same time obtain The data is transmitted to the edge device 63 in the vehicle for further processing to achieve accurate fault diagnosis through the acquired inductance signal. It does not rely on vibration sensors, saves costs, and avoids the vibration signal being easily affected by external vibration and noise. Early faults may be difficult to detect and may take a long time to discover. In addition, the inductance signal is obtained using a high-frequency current signal, which can detect smaller eccentricities and is sensitive to early faults. The degree and position of the rotor eccentricity can be judged more accurately through changes in the inductance signal.

[0065] In some embodiments, before obtaining the inductance signal of the motor, the method further includes: when the current signal of the motor is unbalanced and the winding temperature signal of the motor exceeds a preset temperature range, determining that a phase loss fault occurs in the motor.

[0066] In an embodiment, Figure 2 As shown, before obtaining the inductance signal of the motor, it is necessary to perform a motor state analysis 60 based on the collected three-phase current and winding temperature, that is, to perform a preliminary diagnosis of the motor state. When the vehicle is driven normally, the three-phase current should be equal. When a phase failure occurs in the motor, the three-phase current will be unbalanced or too large, the winding temperature will rise rapidly, and the motor speed will drop rapidly. Therefore, by observing whether the three-phase current is balanced, whether the temperature of the motor winding is within the normal range, and whether there is an abnormal temperature rise phenomenon, it is determined whether the current motor has a phase failure. For example, if the motor winding temperature signal exceeds the preset temperature range, it is considered that the motor winding temperature is abnormal and there is an abnormal temperature rise phenomenon. When the motor current signal is unbalanced and the motor winding temperature signal is If the signal exceeds the preset temperature range, it can be determined that the motor has a phase loss fault. It is understandable that after the motor has a phase loss fault, it cannot be ruled out that the motor does not have eccentricity and or demagnetization faults, etc., so it can be further diagnosed as needed or directly continue to use the multi-level fusion diagnosis method to achieve a joint judgment based on the preliminary diagnosis results of the motor and the multi-level fusion diagnosis results, so as to achieve simultaneous judgment whether the motor has multiple faults, thereby improving the fault diagnosis efficiency, and first make a preliminary judgment. If the preliminary judgment is abnormal, it can be selected as needed or directly, then enter the fusion diagnosis, otherwise, do not enter the fusion diagnosis. Since the fusion diagnosis process consumes a lot of CPU resources, the preliminary judgment process can save CPU resources to a certain extent.

[0067] Reference below Figure 6 The motor fault diagnosis method according to the embodiment of the present invention is described in detail.

[0068] like Figure 6 FIG. 1 is a flow chart of a motor fault diagnosis method according to another embodiment of the present invention. The method according to the embodiment of the present invention at least includes steps S30 to S45.

[0069] Step S30: injecting a high-frequency rotation pulse voltage signal into the d-axis and acquiring a current signal during the operation of the motor.

[0070] In step S31 , the current signal of the motor is unbalanced and the winding temperature signal of the motor exceeds a preset temperature range, and it is determined that a phase loss fault occurs in the motor.

[0071] Step S32: determining a d-axis current signal in a dq coordinate system according to the current signal.

[0072] Step S33 : obtaining a high-frequency current signal in the d-axis current signal based on a low-pass filter.

[0073] Step S34: obtaining an inductance signal according to the high-frequency current signal.

[0074] Step S35: determining the fundamental frequency amplitude according to the inductance signal.

[0075] Step S36: When the fundamental frequency amplitude is greater than a preset frequency threshold and / or the operating signal is abnormal, it is determined that a preliminary fault occurs in the motor.

[0076] Step S37 : Processing the current signal based on wavelet transform to determine wavelet coefficients at different frequencies.

[0077] Step S38: determining a one-dimensional time domain signal of the current signal according to the coefficient combination of the wavelet coefficients.

[0078] Step S39: determining a two-dimensional time-frequency diagram according to the one-dimensional time-domain signal.

[0079] Step S40: determining a first characteristic signal according to the two-dimensional time-frequency diagram.

[0080] Step S41: determining a second characteristic signal according to the inductance signal.

[0081] Step S42: determining a timing characteristic signal according to the first characteristic signal and the second characteristic signal.

[0082] Step S44: determining a third characteristic signal according to the winding temperature signal.

[0083] Step S44: determining a fused characteristic signal according to the time series characteristic signal and the third characteristic signal.

[0084] Step S45: determining the fault type of the motor according to the fused characteristic signal.

[0085] According to the method of the embodiment of the present invention, only existing sensors are relied upon, without the need to use vibration signals. First, the inductance signal and / or operating signals such as current and winding temperature of the motor are obtained based on the high-frequency current signal to determine whether a preliminary fault has occurred in the motor. Then, based on the preliminary fault diagnosis result, as required or directly based on the inductance signal and operating signal, a joint hierarchical fusion diagnosis method is selected to further diagnose the fault type of the motor, thereby improving the accuracy of the diagnosis process, the accuracy of the diagnosis results and the diagnosis efficiency, and reducing the diagnosis cost.

[0086] Reference below Figure 7 A fault diagnosis device for a motor according to an embodiment of the present invention will be described.

[0087] like Figure 7 FIG2 is a block diagram of a motor fault diagnosis device according to another embodiment of the present invention. The motor fault diagnosis device 100 according to this embodiment of the present invention includes a controller 98 for determining a preliminary motor fault based on the motor's inductance signal and / or operating signal; and an edge module 99 connected to the controller 98 for diagnosing the motor fault type based on the inductance signal and operating signal.

[0088] According to the motor fault diagnosis device 100 of the embodiment of the present invention, only existing sensors are relied on, and vibration signals are not required. First, the inductance signal and / or operating signals such as current and winding temperature of the motor are obtained based on the high-frequency current signal to determine whether a preliminary fault has occurred in the motor. Then, according to the preliminary fault diagnosis result, a joint hierarchical fusion diagnosis method is selected to further diagnose the fault type of the motor as needed or directly based on the inductance signal and the operating signal, thereby improving the accuracy of the diagnosis process, the accuracy of the diagnosis results and the diagnosis efficiency, and reducing the diagnosis cost.

[0089] In some embodiments, the edge module 99 is used to: diagnose the fault type of the motor based on the inductance signal and the operation signal, including: determining the timing characteristic signal based on the operation signal and the inductance signal; determining the fusion characteristic signal based on the timing characteristic signal and the operation signal; and determining the fault type of the motor based on the fusion characteristic signal.

[0090] In some embodiments, the edge module 99 is used for: the operating signal includes: a current signal, and determining a timing characteristic signal based on the operating signal and the inductance signal, including: determining a first characteristic signal based on the current signal; determining a second characteristic signal based on the inductance signal; determining a timing characteristic signal based on the first characteristic signal and the second characteristic signal.

[0091] In some embodiments, the edge module 99 is used for: the operating signal includes a winding temperature signal, and determining a fused characteristic signal based on the timing characteristic signal and the operating signal, including: determining a third characteristic signal based on the winding temperature signal; determining a fused characteristic signal based on the timing characteristic signal and the third characteristic signal.

[0092] In some embodiments, the edge module 99 is used to: determine the first characteristic signal based on the current signal, including: determining the wavelet coefficients at different frequencies by processing the current signal based on the wavelet transform; determining the one-dimensional time domain signal of the current signal based on the coefficient combination of the wavelet coefficients; determining the two-dimensional time-frequency graph based on the one-dimensional time domain signal; and determining the first characteristic signal based on the two-dimensional time-frequency graph.

[0093] In some embodiments, the controller 98 is used to: determine whether the motor has a preliminary fault based on the inductance signal and / or operating signal of the motor, including: determining the fundamental frequency amplitude based on the inductance signal; when the fundamental frequency amplitude is greater than a preset frequency threshold and / or the operating signal is abnormal, determining that the motor has a preliminary fault.

[0094] In some embodiments, the controller 98 is used to: before determining that a preliminary fault has occurred in the motor based on the inductance signal and / or operating signal of the motor, it also includes: injecting a high-frequency rotating pulse voltage signal into the d-axis and obtaining a current signal during the operation of the motor; determining the d-axis current signal in the dq coordinate system based on the current signal; and obtaining the inductance signal of the motor based on the d-axis current signal.

[0095] In some embodiments, the controller 98 is configured to obtain an inductance signal of the motor based on the d-axis current signal, including obtaining a high-frequency current signal in the d-axis current signal based on a low-pass filter; and obtaining an inductance signal based on the high-frequency current signal.

[0096] In some embodiments, the controller 98 is configured to: before obtaining the inductance signal of the motor, further include: determining that a phase loss fault occurs in the motor when the current signal of the motor is unbalanced and the winding temperature signal of the motor exceeds a preset temperature range.

[0097] According to the motor fault diagnosis device 100 of the embodiment of the present invention, only existing sensors are relied on, and vibration signals are not required. First, the inductance signal and / or operating signals such as current and winding temperature of the motor are obtained based on the high-frequency current signal to determine whether a preliminary fault has occurred in the motor. Then, according to the preliminary fault diagnosis result, a joint hierarchical fusion diagnosis method is selected to further diagnose the fault type of the motor as needed or directly based on the inductance signal and the operating signal, thereby improving the accuracy of the diagnosis process, the accuracy of the diagnosis results and the diagnosis efficiency, and reducing the diagnosis cost.

[0098] Reference below Figure 8 A vehicle according to an embodiment of the present invention will be described.

[0099] like Figure 8 FIG. 1 is a block diagram of a vehicle according to an embodiment of the present invention. A vehicle 101 includes: a motor fault diagnosis device 100 according to the above embodiment.

[0100] According to the vehicle 100 of the embodiment of the present invention, the vehicle is equipped with a motor fault diagnosis device 100 as in the above embodiment, which only relies on existing sensors and does not need to use vibration signals. It first determines whether the motor has a preliminary fault based on the inductance signal and / or current and winding temperature and other operating signals of the motor obtained according to the high-frequency current signal. Then, based on the preliminary fault diagnosis result, according to demand or directly based on the inductance signal and operating signal, a joint hierarchical fusion diagnosis method is selected to further diagnose the fault type of the motor, thereby improving the accuracy of the diagnosis process, the accuracy of the diagnosis results and the diagnosis efficiency, and reducing the diagnosis cost.

[0101] The following describes a computer-readable storage medium according to an embodiment of the present invention.

[0102] The computer-readable storage medium of an embodiment of the present invention stores a motor fault diagnosis program on the computer-readable storage medium. When the motor fault diagnosis program is executed by a processor, the device installed with the motor fault diagnosis program implements the motor fault diagnosis method as described in the above embodiment.

[0103] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "illustrative embodiments," "example," "specific example," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with the embodiment or example is included in at least one embodiment or example of the present invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example.

[0104] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the claims and their equivalents.

Claims

1. A method for diagnosing a motor fault, characterized in that: include: determining, based on an inductance signal and / or an operation signal of the motor, that a preliminary fault occurs in the motor; The fault type of the motor is diagnosed according to the inductance signal and the operation signal.

2. The motor fault diagnosis method according to claim 1, characterized in that: The diagnosing the fault type of the motor according to the inductance signal and the operation signal includes: determining a timing characteristic signal according to the operation signal and the inductance signal; Determine a fusion feature signal according to the timing feature signal and the operation signal; The fault type of the motor is determined according to the fused characteristic signal.

3. The motor fault diagnosis method according to claim 2, characterized in that: The operation signal includes a current signal, and determining a timing characteristic signal according to the operation signal and the inductance signal includes: determining a first characteristic signal according to the current signal; determining a second characteristic signal according to the inductance signal; A timing characteristic signal is determined according to the first characteristic signal and the second characteristic signal.

4. The motor fault diagnosis method according to claim 2 or 3, characterized in that: The operation signal includes a winding temperature signal, and determining a fusion characteristic signal according to the timing characteristic signal and the operation signal includes: determining a third characteristic signal according to the winding temperature signal; The fused characteristic signal is determined according to the timing characteristic signal and the third characteristic signal.

5. The motor fault diagnosis method according to claim 3, characterized in that: The determining of the first characteristic signal according to the current signal includes: Processing the current signal based on wavelet transform to determine wavelet coefficients at different frequencies; Determine a one-dimensional time domain signal of the current signal according to a coefficient combination of the wavelet coefficients; Determine a two-dimensional time-frequency diagram according to the one-dimensional time-domain signal; The first characteristic signal is determined according to the two-dimensional time-frequency diagram.

6. The motor fault diagnosis method according to claim 1, characterized in that: Determining that a preliminary fault occurs in the motor according to the inductance signal and / or the operation signal of the motor includes: determining a fundamental frequency amplitude according to the inductance signal; When the fundamental frequency amplitude is greater than a preset frequency threshold and / or the operating signal is abnormal, it is determined that a preliminary fault occurs in the motor.

7. The motor fault diagnosis method according to claim 1, characterized in that: Before determining that a preliminary fault occurs in the motor according to the inductance signal and / or the operation signal of the motor, the method further includes: Injecting a high-frequency rotational pulse voltage signal into the d-axis and acquiring a current signal during operation of the motor; Determine a d-axis current signal in a dq coordinate system according to the current signal; An inductance signal of the motor is obtained according to the d-axis current signal.

8. The motor fault diagnosis method according to claim 7, characterized in that: The acquiring the inductance signal of the motor according to the d-axis current signal includes: acquiring a high-frequency current signal from the d-axis current signal based on a low-pass filter; The inductance signal is obtained according to the high-frequency current signal.

9. The motor fault diagnosis method according to claim 1 or 7, characterized in that: Before obtaining the inductance signal of the motor, the method further includes: The current signal of the motor is unbalanced and the winding temperature signal of the motor exceeds a preset temperature range, and it is determined that a phase loss fault occurs in the motor.

10. A motor fault diagnosis device, characterized in that: include: a controller, configured to determine that a preliminary fault occurs in the motor based on an inductance signal and / or an operation signal of the motor; The edge module is connected to the controller and is used to diagnose the fault type of the motor according to the inductance signal and the operation signal.

11. A vehicle, characterized in that: include: The motor fault diagnosis device as claimed in claim 10.

12. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a motor fault diagnosis program, and when the motor fault diagnosis program is executed by the processor, the device installed with the motor fault diagnosis program implements the motor fault diagnosis method according to any one of claims 1 to 9.

Images