New energy vehicle drive motor bearing fault diagnosis method based on data analysis
By monitoring and analyzing the wear index of the bearings of the new energy vehicle drive motor at different speeds, determining the lossless cycle, and collecting and comparing the sound signals during this cycle, the problem of the fault signal being covered by noise during high-speed operation is solved, and the motor is realized Accurate diagnosis of bearing failure.
Patent Information
- Application Number
- CN202411434168.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-15
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-10-15
AI Technical Summary
In the prior art, it is difficult to distinguish fault signals from background noise when the bearings of new energy vehicle drive motors are faulty when running at high speed, resulting in difficulty in diagnosis.
By monitoring the wear index of motor bearings at different speeds, determining the lossless cycle, collecting and sorting sound signals during this cycle, establishing a lossless signal set, comparing the current sound signals in real time to judge the fault, and troubleshooting the normal signal components to identify the fault signal.
Improves the accuracy and efficiency of fault detection, and can accurately identify potential faults of motor bearings in noisy environments.
Smart Images
Figure CN118936884B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of fault diagnosis, and in particular to a new energy vehicle drive motor bearing fault diagnosis method based on data analysis. Background Art
[0002] New energy vehicle drive motor bearings are key components in new energy vehicle drive motors. Their main function is to support the motor rotor and bear the load to ensure the normal operation and efficient performance of the motor. Bearings play a vital role in the operation of new energy vehicles, not only affecting the performance of the motor, but also directly related to the driving experience and safety of the vehicle.
[0003] The fault diagnosis of motor bearings aims to detect and identify abnormal conditions of motor bearings during operation to ensure the normal operation and safety of the equipment; fault diagnosis methods include abnormal rotation sound analysis, current spectrum analysis, intelligent learning method and adversarial network method.
[0004] In the prior art, abnormal rotation sound analysis is one of the most commonly used bearing fault diagnosis methods. When the inner or outer ring of the bearing is damaged, it will generate vibrations at high speeds, thereby producing a sound of a specific frequency. This sound can generally be used as a fault signal to analyze whether the bearing is faulty. However, motors often generate a lot of noise when running at high speeds, and the fault signal will be submerged in the strong background noise, making fault diagnosis more difficult. Summary of the invention
[0005] The purpose of the present invention is to provide a new energy vehicle drive motor bearing fault diagnosis method based on data analysis to solve the above technical problems.
[0006] The purpose of the present invention can be achieved through the following technical solutions:
[0007] The new energy vehicle drive motor bearing fault diagnosis method based on data analysis includes the following steps:
[0008] Step S1: Select several vehicle samples and set several different speed values to obtain the wear index of the bearing of the motor of the vehicle sample after the motor runs for T time at different speed values, where T is the preset monitoring time; obtain the lossless period of the bearing according to the wear index corresponding to each speed value;
[0009] Step S2: When the motor is in a lossless cycle, the speed and sound signal of the motor are obtained in real time. Each acquisition duration of Tmin is recorded as an acquisition cycle, where Tmin is a preset minimum duration. The timestamp corresponding to each acquisition in the acquisition cycle is recorded as a sampling point, and each sampling point is numbered. The sound signal is a time domain signal composed of the amplitude at each sampling point.
[0010] The acquired sound signals are sorted and numbered according to the corresponding rotation speeds to obtain a lossless signal set; and a rotation speed set K is obtained from the lossless signal sequence, denoted by K={Sp 1 , Sp 2 , ..., Sp n}, Sp 1 is the first speed, n is the total number of speeds;
[0011] Step S3: when the motor is not in a lossless cycle, the speed and sound signal of the motor are obtained in real time during operation, and recorded as the current speed Sp´ and the current sound signal Vio´;
[0012] If Sp´∉K, then select from K a tree that satisfies the first constraint
[0013] An element of, where i∈[1, n] and i is a positive integer, denoted as Sp max ; and select the second constraint Min(J)∩Sp i <The element of Sp´, denoted as Sp min ; Obtain Sp respectively in the lossless signal set max and Sp min The corresponding sound signal is denoted as Vio max and Vio min ;
[0014] According to Vio max and Vio min Get analog sound signal Vio 0 The acquisition process is: obtain the amplitude y[m]=(x max [m]+x min [m]) / 2, where x max [m] for Vio max The amplitude at the mth sampling point, x min [m] for Vio min The amplitude at the mth sampling point on the previous page; According to the amplitude at each sampling point, the analog sound signal Vio is obtained 0 ;
[0015] Step S4: Compare the current sound signal Vio ' and the simulated sound signal Vio 0 , according to the comparison result, it is determined whether the bearing has a fault; if there is a fault, the simulated sound signal Vio is excluded from the current sound signal Vio ' 0 , the time domain signal after elimination is recorded as the fault signal.
[0016] As a further solution of the present invention: the process of setting the plurality of different speed values includes:
[0017] Set the standard speed value Sp 0 =10000rpm, and set a speed interval threshold S, the setting range of the speed interval threshold is S∈[50rpm, 100rpm]; take the standard speed value as the initial value, decrease downward and increase upward the initial value according to the speed interval threshold, and obtain several speed values.
[0018] As a further solution of the present invention: the process of obtaining the wear index of the bearing of the vehicle sample includes:
[0019] Before driving, image data of a bearing of a vehicle sample is obtained, wherein the image data includes photographed images of an inner ring and an outer ring; the image data is rasterized to obtain a plurality of pixel points, and each pixel point is numbered; a pixel value of each pixel point is obtained, and associated with the number of the corresponding pixel point to obtain an initial pixel value sequence;
[0020] Acquire new image data of the bearing of the vehicle sample after the driving, perform rasterization processing on the new image data to obtain a number of new pixel points, and number each new pixel point; obtain a pixel value of each new pixel point, and associate it with the number of the corresponding new pixel point to obtain a new pixel value sequence;
[0021] In the initial pixel value sequence and the new pixel value sequence, select |R v -R´ v |≥Rt, the number of elements num, where R v is the pixel value of the vth pixel in the initial pixel value sequence, R´ v is the pixel value of the vth new pixel in the new pixel value sequence, Rt is the preset difference threshold; then the wear index is obtained , where N is the total number of vehicle samples, NUM is the total number of pixels, is the num / NUM of the e-th vehicle sample.
[0022] As a further solution of the present invention: the process of obtaining the lossless period of the bearing includes:
[0023] A regression model is established according to the wear index corresponding to each speed value, and the maximum speed value MaxSp corresponding to the wear index of 0 is obtained in the regression model; since the first operation of the vehicle's motor, the vehicle's speed is monitored in real time, and the average speed of the vehicle's motor is obtained in real time, and the timestamp corresponding to the average speed equal to MaxSp is recorded as the fault time node, and the time period between the timestamp corresponding to the first operation of the motor and the fault time node is recorded as the damage-free period of the bearing.
[0024] As a further solution of the present invention: the process of obtaining the rotation speed and sound signal of the motor when it is running includes:
[0025] When there are two rotational speeds that are the same but the sound signals corresponding to the two rotational speeds are different, the acquisition times of the two rotational speeds are compared, and the sound signal corresponding to the rotational speed with the earlier acquisition time is selected as the sound signal at that rotational speed.
[0026] As a further solution of the present invention: if Sp´∈K, the sound signal when the rotation speed is Sp´ is directly obtained from the lossless signal set.
[0027] As a further solution of the present invention: the process of determining whether the bearing is faulty includes:
[0028] The current sound signal Vio ' and the simulated sound signal Vio 0 are all recorded as initial signals; the initial signals are converted from the time domain to the frequency domain by fast Fourier transform to obtain frequency domain signals; the frequency domain signals are processed by a bandpass filter bank, and the processed frequency domain signals are subjected to discrete cosine transform to extract signal feature vectors;
[0029] Get the current sound signal Vio 'and the analog sound signal Vio 0 signal feature vector, and obtain the cosine similarity of the two signal feature vectors; if the cosine similarity is less than a preset similarity threshold, the bearing is not faulty; otherwise, the bearing is faulty.
[0030] As a further solution of the present invention: if there is no fault in the bearing, the current speed and the current sound signal of the motor during operation in the next acquisition cycle are continuously monitored.
[0031] Beneficial effects of the present invention:
[0032] In the prior art, abnormal rotation sound analysis is one of the most commonly used bearing fault diagnosis methods. When the inner or outer ring of the bearing is damaged, it will generate vibrations at high speeds, thereby producing a sound of a specific frequency. This sound can generally be used as a fault signal to analyze whether the bearing is faulty. However, motors often generate a lot of noise when running at high speeds, and the fault signal will be submerged in the strong background noise, making fault diagnosis more difficult.
[0033] Compared with the prior art, the present invention determines the lossless period of the bearing by monitoring the wear index of the motor bearing at different speeds; through this step, the time interval in which the bearing does not fail, that is, the lossless period, is determined, and the sound signal generated by the motor during this time interval is a signal mixed with a fault-free signal; within the known lossless period, the sound signal of the motor during operation is collected and associated with the corresponding speed to establish a lossless signal set; this step provides a comparison benchmark for subsequent fault judgment, and the sound signal generated by the motor during the lossless period can be used as a standard for subsequent judgment; when the motor is not in the lossless period, the current speed and sound signal are collected in real time; at this time, the collected sound signal is likely to be a sound signal mixed with a fault signal; the current sound signal obtained in real time is compared with the signal in the lossless signal set, and it is judged whether the bearing has a fault based on the difference; if a fault exists, the normal signal component in the current sound signal is excluded, so as to accurately identify the fault signal; through precise signal acquisition, analysis and comparison, the potential fault of the new energy vehicle drive motor bearing can be effectively diagnosed, and the accuracy and efficiency of fault detection can be improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] The present invention will be further described below in conjunction with the accompanying drawings.
[0035] Figure 1 This is a flow chart of a method for diagnosing a bearing fault of a new energy vehicle drive motor based on data analysis of the present invention. DETAILED DESCRIPTION
[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0037] See also Figure 1 As shown, the present invention is a new energy vehicle drive motor bearing fault diagnosis method based on data analysis, comprising the following steps:
[0038] Step S1: Select several vehicle samples and set several different speed values to obtain the wear index of the bearing of the motor of the vehicle sample after the motor runs for T time at different speed values, where T is the preset monitoring time; obtain the lossless period of the bearing according to the wear index corresponding to each speed value;
[0039] It should be noted that the specific value of the monitoring time depends on many factors, including the purpose of the experiment, the type of vehicle, the specifications of the motor, and the expected conditions of use. Generally, T should be set long enough to accurately measure the surface wear of the bearing at different speeds, but the efficiency of the experiment and the availability of resources should also be taken into account.
[0040] For example, if the purpose of the experiment is to evaluate the performance of the bearing under extreme conditions, T may need to be set to a larger value to ensure sufficient stress and wear, then T = k*t 0 , where t 0 is a basic time unit (such as hours, days, months, etc.), and k is a coefficient adjusted according to experimental needs;
[0041] As a preferred embodiment of the present invention, the process of setting the plurality of different speed values includes:
[0042] Set the standard speed value Sp 0 =10000rpm, and set a speed interval threshold S, the setting range of the speed interval threshold is S∈[50rpm, 100rpm]; take the standard speed value as the initial value, decrease downward and increase upward the initial value according to the speed interval threshold, and obtain a plurality of speed values;
[0043] It can be understood that the standard speed value Sp 0 As a base point, the initial value is adjusted up and down according to the speed interval threshold S; for example, if S=60rpm is set, several new speed values can be generated, which will start from 10000rpm, decrease downward (for example: 9940rpm, 9880rpm, 9820rpm...) and increase upward (for example: 10060rpm, 10120rpm, 10180rpm...);
[0044] As a preferred embodiment of the present invention, the process of obtaining the wear index of the bearing of the vehicle sample includes:
[0045] Before driving, image data of a bearing of a vehicle sample is obtained, wherein the image data includes photographed images of an inner ring and an outer ring; the image data is rasterized to obtain a plurality of pixel points, and each pixel point is numbered; a pixel value of each pixel point is obtained, and associated with the number of the corresponding pixel point to obtain an initial pixel value sequence;
[0046] Acquire new image data of the bearing of the vehicle sample after the driving, perform rasterization processing on the new image data to obtain a number of new pixel points, and number each new pixel point; obtain a pixel value of each new pixel point, and associate it with the number of the corresponding new pixel point to obtain a new pixel value sequence;
[0047] In the initial pixel value sequence and the new pixel value sequence, select |R v -R´ v |≥Rt, the number of elements num, where R v is the pixel value of the vth pixel in the initial pixel value sequence, R´ v is the pixel value of the vth new pixel in the new pixel value sequence, Rt is the preset difference threshold; then the wear index is obtained , where N is the total number of vehicle samples, NUM is the total number of pixels, is the num / NUM of the e-th vehicle sample;
[0048] It can be understood that before and after the vehicle is driven, image data of the inner and outer rings of the bearing are obtained respectively, and these images are used to record the state of the bearing before and after use; the image data is converted into a pixel array through rasterization processing; each pixel point is numbered and its pixel value is recorded to form an initial pixel value sequence; after the vehicle is driven, the above image acquisition and processing steps are repeated to obtain a new pixel value sequence; by comparing the pixel value sequences before and after driving, points with large pixel value changes can be identified; a difference threshold Rt is set to measure the significance of the pixel value change; the wear index is calculated, and the wear index reflects the average wear degree of the bearing during driving;
[0049] It is worth noting that the acquisition of image data requires that the environmental factors such as light acquired each time be kept consistent as much as possible to reduce the judgment of the degree of bearing wear by external factors;
[0050] As a preferred embodiment of the present invention, the process of obtaining the lossless period of the bearing includes:
[0051] A regression model is established according to the wear index corresponding to each speed value, and the maximum speed value MaxSp corresponding to the wear index being 0 is obtained in the regression model; the speed of the vehicle is monitored in real time from the first operation of the motor of the vehicle, and the average speed of the motor of the vehicle is obtained in real time, and the timestamp corresponding to the time when the average speed is equal to MaxSp is recorded as the fault time node, and the time period between the timestamp corresponding to the first operation of the motor and the fault time node is recorded as the damage-free period of the bearing;
[0052] It is understandable that there is a close relationship between the wear index of the motor and its speed. By establishing a regression model, the wear condition of the motor at different speeds can be predicted. First, it is necessary to collect the wear index data of the motor at different speed values. Using the collected data, a regression model between the wear index and the speed is established. The regression model is generally nonlinear, and the specific form depends on the characteristics of the data. Through the established regression model, the speed value corresponding to the wear index of 0 can be found, that is, the maximum speed value MaxSp. This value indicates that the motor will not theoretically wear out if the speed does not exceed this value. Starting from the first operation of the motor, its speed is monitored in real time and the average speed is calculated. The purpose of monitoring is to understand the speed of the motor in actual operation so as to compare it with the model prediction. When the monitored average speed reaches MaxSp, this time point is recorded as the fault time node. The fault time node indicates that the motor is about to enter a wear state. The time period between the time when the motor first runs and the fault time node is defined as the lossless period of the bearing. During this period, there is no fault signal in the sound signal generated by the motor.
[0053] Step S2: When the motor is in a lossless cycle, the speed and sound signal of the motor are obtained in real time. Each acquisition duration of Tmin is recorded as an acquisition cycle, where Tmin is a preset minimum duration. The timestamp corresponding to each acquisition in the acquisition cycle is recorded as a sampling point, and each sampling point is numbered. The sound signal is a time domain signal composed of the amplitude at each sampling point.
[0054] The acquired sound signals are sorted and numbered according to the corresponding rotation speeds to obtain a lossless signal set; and a rotation speed set K is obtained from the lossless signal sequence, denoted by K={Sp 1 , Sp 2 , ..., Sp n}, Sp 1 is the first speed, n is the total number of speeds;
[0055] It can be understood that, first of all, it is necessary to obtain the speed and sound signal of the motor in real time, which are key indicators for evaluating the motor status; a minimum time length Tmin is set, and each time this time length passes, it is recorded as an acquisition cycle, in order to ensure the consistency and comparability of the data; the lossless signal set includes the speed recorded in all lossless cycles and their corresponding sound signals;
[0056] As a preferred embodiment of the present invention, the process of obtaining the rotation speed and sound signal of the motor when it is running includes:
[0057] When there are two rotation speeds that are the same but the sound signals corresponding to the two rotation speeds are different, the acquisition times of the two rotation speeds are compared, and the sound signal corresponding to the rotation speed with the earlier acquisition time is selected as the sound signal at the rotation speed;
[0058] It should be noted that when two speeds are the same but the corresponding sound signals are different, it is necessary to first compare the acquisition time of the two speeds. Among the two different sound signals, the sound signal corresponding to the speed with the earlier acquisition time is selected as the representative sound signal at the speed; because within the lossless period, the earlier the sound signal collected at the sampling point, the closer it is to the situation where the bearing is not worn;
[0059] Step S3: when the motor is not in a lossless cycle, the speed and sound signal of the motor are obtained in real time during operation, and recorded as the current speed Sp´ and the current sound signal Vio´;
[0060] If Sp´∉K, then select from K a tree that satisfies the first constraint
[0061] An element of, where i∈[1, n] and i is a positive integer, denoted as Sp max ; and select the second constraint Min(J)∩Sp i <The element of Sp´, denoted as Sp min ; Obtain Sp respectively in the lossless signal set max and Sp min The corresponding sound signal is denoted as Vio max and Vio min ;
[0062] According to Vio max and Vio min Get analog sound signal Vio 0 The acquisition process is: obtain the amplitude y[m]=(x max [m]+x min [m]) / 2, where x max [m] for Vio max The amplitude at the mth sampling point, x min [m] for Vio min The amplitude at the mth sampling point on the previous page; According to the amplitude at each sampling point, the analog sound signal Vio is obtained 0 ;
[0063] It can be understood that if the obtained speed Sp 'is not in the preset speed set K, find the speed in the speed set K that is closest to the current speed Sp 'but greater than Sp ', and the speed that is closest to but less than Sp '; and obtain Sp in the lossless signal set max and Sp minCorresponding sound signal Vio max and Vio min ; Then at each sampling point, take Vio max and Vio min The average value of the amplitude at the sampling point is taken as the amplitude of the analog sound signal at the sampling point;
[0064] As a preferred embodiment of the present invention, if Sp´∈K, then the sound signal when the rotation speed is Sp´ is directly obtained from the lossless signal set;
[0065] Step S4: Compare the current sound signal Vio ' and the simulated sound signal Vio 0 , according to the comparison result, it is determined whether the bearing has a fault; if there is a fault, the simulated sound signal Vio is excluded from the current sound signal Vio ' 0 , the time domain signal after elimination is recorded as the fault signal.
[0066] As a preferred embodiment of the present invention, the process of determining whether the bearing has a fault includes:
[0067] The current sound signal Vio ' and the simulated sound signal Vio 0 are all recorded as initial signals; the initial signals are converted from the time domain to the frequency domain by fast Fourier transform to obtain frequency domain signals; the frequency domain signals are processed by a bandpass filter bank, and the processed frequency domain signals are subjected to discrete cosine transform to extract signal feature vectors;
[0068] Get the current sound signal Vio 'and the analog sound signal Vio 0 signal feature vectors, and obtain the cosine similarity of the two signal feature vectors; if the cosine similarity is less than a preset similarity threshold, the bearing is not faulty; otherwise, the bearing is faulty;
[0069] It can be understood that the Fast Fourier Transform is an efficient algorithm for converting a signal from its original time domain form to a frequency domain representation; in the frequency domain, the individual frequency components of the signal and their corresponding amplitudes become clearly visible for further analysis; the frequency domain signal is processed through a set of bandpass filters that allow signals in a specific frequency range to pass through while attenuating signals of other frequencies; this step helps remove noise and emphasizes specific frequency components associated with bearing faults; the processed frequency domain signal is then subjected to a discrete cosine transform (DCT); DCT is a transform commonly used in signal and image processing that helps extract the eigenvectors of a signal that characterize the main characteristics of the signal; the cosine similarity between the signal eigenvectors of the current sound signal and the simulated sound signal is calculated; the cosine similarity is an indicator that measures the degree of proximity of the directions of two vectors, and the closer its value is to 1, the more similar the two signal eigenvectors are, that is, the more similar the current sound signal and the simulated sound signal are;
[0070] As a preferred embodiment of the present invention, if there is no bearing fault, the current rotation speed and the current sound signal of the motor during operation in the next acquisition cycle continue to be monitored.
[0071] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.
Claims
1. New energy vehicle drive motor bearing fault diagnosis method based on data analysis, It is characterized in that The following steps are involved: Step S1: Select several vehicle samples and set several different speed values to obtain the wear index of the bearing of the motor of the vehicle sample after the motor runs for T time at different speed values, where T is the preset monitoring time; According to the wear index corresponding to each speed value, the loss-free cycle of the bearing is obtained; Step S2: When the motor is in a lossless cycle, the speed and sound signal of the motor are obtained in real time. Each acquisition duration of Tmin is recorded as an acquisition cycle, where Tmin is a preset minimum duration. The timestamp corresponding to each acquisition in the acquisition cycle is recorded as a sampling point, and each sampling point is numbered. The sound signal is a time domain signal composed of the amplitude at each sampling point. The acquired sound signals are sorted and numbered according to the corresponding rotation speeds to obtain a lossless signal set; and a rotation speed set K is obtained from the lossless signal sequence, denoted by K={Sp 1 , Sp 2 , ..., Sp n }, Sp 1 is the first speed, n is the total number of speeds; Step S3: when the motor is not in a lossless cycle, the speed and sound signal of the motor are obtained in real time during operation, and recorded as the current speed Sp´ and the current sound signal Vio´; If Sp´∉K, then select from K a tree that satisfies the first constraint An element of, where i∈[1, n] and i is a positive integer, denoted as Sp max ; and select the second constraint Min(J)∩Sp i <The element of Sp´ is denoted as Sp min ; Obtain Sp respectively in the lossless signal set max and Sp min The corresponding sound signal is denoted as Vio max and Vio min ; According to Vio max and Vio min Get analog sound signal Vio 0 The acquisition process is: obtain the amplitude y[m]=(x max [m]+x min [m]) / 2, where x max [m] for Vio max The amplitude at the mth sampling point, x min [m] for Vio min The amplitude at the mth sampling point on the previous page; According to the amplitude at each sampling point, the analog sound signal Vio is obtained 0 ; Step S4: Compare the current sound signal Vio ' and the simulated sound signal Vio 0 , determine whether the bearing is faulty based on the comparison results; If there is a fault, the analog sound signal Vio is excluded from the current sound signal Vio'. 0 , the time domain signal after elimination is recorded as the fault signal; In step S1, the process of obtaining the wear index of the bearing of the vehicle sample includes: Before driving, image data of a bearing of a vehicle sample is obtained, wherein the image data includes photographed images of an inner ring and an outer ring; the image data is rasterized to obtain a plurality of pixel points, and each pixel point is numbered; a pixel value of each pixel point is obtained, and associated with the number of the corresponding pixel point to obtain an initial pixel value sequence; Acquire new image data of the bearing of the vehicle sample after the driving, perform rasterization processing on the new image data to obtain a number of new pixel points, and number each new pixel point; obtain a pixel value of each new pixel point, and associate it with the number of the corresponding new pixel point to obtain a new pixel value sequence; In the initial pixel value sequence and the new pixel value sequence, select |R v -R´ v |≥Rt, the number of elements num, where R v is the pixel value of the vth pixel in the initial pixel value sequence, R´ v is the pixel value of the vth new pixel in the new pixel value sequence, Rt is the preset difference threshold; then the wear index is obtained , where N is the total number of vehicle samples, NUM is the total number of pixels, is the num / NUM of the e-th vehicle sample; In step S1, the process of obtaining the lossless period of the bearing includes: A regression model is established according to the wear index corresponding to each speed value, and the maximum speed value MaxSp corresponding to the wear index of 0 is obtained in the regression model; since the first operation of the vehicle's motor, the vehicle's speed is monitored in real time, and the average speed of the vehicle's motor is obtained in real time, and the timestamp corresponding to the average speed equal to MaxSp is recorded as the fault time node, and the time period between the timestamp corresponding to the first operation of the motor and the fault time node is recorded as the damage-free period of the bearing.
2. The new energy vehicle drive motor bearing fault diagnosis method based on data analysis according to claim 1, It is characterized in that In step S1, the process of setting the plurality of different speed values includes: Set the standard speed value Sp 0 =10000rpm, and set a speed interval threshold S, the setting range of the speed interval threshold is S∈[50rpm, 100rpm]; take the standard speed value as the initial value, decrease downward and increase upward the initial value according to the speed interval threshold, and obtain several speed values.
3. The new energy vehicle drive motor bearing fault diagnosis method based on data analysis according to claim 1, It is characterized in that In step S2, the process of obtaining the speed and sound signal of the motor during operation includes: When there are two rotational speeds that are the same but the sound signals corresponding to the two rotational speeds are different, the acquisition times of the two rotational speeds are compared, and the sound signal corresponding to the rotational speed with the earlier acquisition time is selected as the sound signal at that rotational speed.
4. The new energy vehicle drive motor bearing fault diagnosis method based on data analysis according to claim 1, It is characterized in that In step S3, if Sp´∈K, the sound signal when the rotation speed is Sp´ is directly obtained from the lossless signal set.
5. The new energy vehicle drive motor bearing fault diagnosis method based on data analysis according to claim 1, It is characterized in that In step S4, the process of determining whether the bearing is faulty includes: The current sound signal Vio ' and the simulated sound signal Vio 0 are all recorded as initial signals; the initial signals are converted from the time domain to the frequency domain by fast Fourier transform to obtain frequency domain signals; the frequency domain signals are processed by a bandpass filter bank, and the processed frequency domain signals are subjected to discrete cosine transform to extract signal feature vectors; Get the current sound signal Vio 'and the analog sound signal Vio 0 signal feature vector, and obtain the cosine similarity of the two signal feature vectors; if the cosine similarity is less than a preset similarity threshold, the bearing is not faulty; otherwise, the bearing is faulty.
6. The new energy vehicle drive motor bearing fault diagnosis method based on data analysis according to claim 1, It is characterized in that In step S4, if there is no bearing fault, the current speed and the current sound signal of the motor during operation in the next acquisition cycle are continuously monitored.