A method for evaluating the health of new energy vehicle motors based on big data

Abstract

The present invention provides a method for evaluating the health of new energy vehicle motors based on big data. Starting from the efficiency of the motor and effectively utilizing the big data of the actual vehicle, the instantaneous efficiency of the motor under different working conditions is analyzed. The characteristic threshold data for distinguishing the working conditions and the threshold for efficiency screening are obtained through the historical data of the electric vehicle. The efficiency of the sliding interval is then screened and counted using the threshold. On this basis, the arithmetic mean filtering method is used to weaken the influence of the noise data. Finally, a weighted average method is used to comprehensively give an evaluation result of the health of the motor under various working conditions. This method overcomes the shortcomings of the low efficiency and long time consumption of the existing motor health diagnosis, provides a solution for the rapid online safety inspection of new energy vehicles, and also provides an effective reference for the offline safety inspection of new energy vehicles.

Description

Technical Field

[0001] The present invention belongs to the technical field of new energy vehicle motor performance detection, and specifically relates to a new energy vehicle motor health evaluation method based on big data. Background Art

[0002] With the promotion and popularization of new energy vehicles, in order to adapt to the various complex working conditions faced by vehicles, it is necessary to conduct timely testing of the performance of on-board motors to reflect the current health status of the motors, so as to detect potential faults as early as possible and reduce the probability of accidents. However, in the prior art, for the actual performance testing of motors used in new energy vehicles, a series of testing processes and performance experiments need to be carried out on each vehicle separately, which is very inefficient and cannot complete efficient and rapid testing of massive on-board motors, thus limiting the timeliness and effectiveness of obtaining the health status of the motors. Summary of the invention

[0003] In view of the technical problems existing in the above-mentioned field, the present invention provides a method for evaluating the health of a new energy vehicle motor based on big data, which specifically includes the following steps:

[0004] Step 1: The vehicle terminal extracts the actual vehicle operation data under different working conditions and uploads it to the platform;

[0005] Step 2: The platform extracts various characteristic data that affect the health status of the motor from the received real vehicle operation data, including: motor current, voltage, vehicle speed, motor speed, torque, etc.;

[0006] Step 3, dividing the driving conditions corresponding to the vehicle according to the vehicle speed, calculating the instantaneous efficiency of the motor corresponding to each frame under different driving conditions, and obtaining a box plot of the instantaneous efficiency of the motor under different conditions;

[0007] Step 4: Using the minimum value in the motor instantaneous efficiency box diagram calculated for the same vehicle model, determine the motor efficiency threshold of the vehicle model under the corresponding working condition and establish a threshold library;

[0008] Step 5: For the target vehicle to be evaluated, within a specific frame time window under each working condition, calculate and count the number of healthy frames in which the instantaneous efficiency of the motor is not less than the corresponding threshold in the threshold library;

[0009] Step 6: Perform a motor health score based on the number of healthy frames of the target vehicle under different working conditions, and determine the actual current health status of the vehicle motor based on the score.

[0010] Furthermore, in step 2, after the platform receives the actual vehicle operation data, it first performs data cleaning, data integration, data transformation, and data reduction in sequence before extracting feature data, removes outliers, splits data, and reorganizes data, and processes the original data into a format that can be directly used and processed later.

[0011] Furthermore, in step 3, the driving conditions corresponding to the vehicle are divided based on the vehicle speed, specifically including:

[0012] For multiple vehicles of the same model in actual operation, the speed data in their historical driving segments are monitored offline and a speed box plot is drawn. The upper and lower quartiles in the box plot and the speed difference corresponding to each segment are extracted, and the vehicle operating conditions are divided into multiple different types based on the speed and speed difference.

[0013] The instantaneous efficiency η of the motor corresponding to each frame in step 3 can be calculated based on the following formula:

[0014]

[0015] Where v (r / min) is the motor speed, T (N·m) is the motor torque, U (V) is the motor controller input voltage, and I (A) is the motor controller input current.

[0016] Furthermore, step 4 determines the motor efficiency threshold, specifically including:

[0017] For each of n vehicles of the same model, the minimum instantaneous efficiency η of the motor under the same working condition is i (i=1, 2, ..., n) and take the average value to obtain the motor efficiency threshold:

[0018]

[0019] Furthermore, before step 6, step 5 is continuously repeated multiple times for different working conditions of the target vehicle to be evaluated, and the healthy frame numbers counted each time under the same working condition are subjected to arithmetic average filtering:

[0020]

[0021] In the formula, J i is the number of healthy frames with normal motor efficiency obtained by the ith statistics under a certain working condition, N is the number of time windows for calculation and statistics, and Y j is the result obtained by arithmetic mean filtering;

[0022] The filtered result is used for the health score in the subsequent step 6.

[0023] Furthermore, the motor health score in step 6 specifically includes:

[0024] For a certain working condition, the corresponding health score H is calculated based on the following formula:

[0025]

[0026] Where p is the frame value when the number of healthy frames is less than 5, that is, Y j (<5); M is the number of frames executed and N is the total number of frames in each time window in the calculation and statistics;

[0027] Based on the motor efficiency map of the target vehicle to be evaluated, combined with the motor states corresponding to different working conditions, including low speed and low torque, low speed and high torque, low torque and high speed, etc., corresponding weights are assigned to different working conditions; the weighted sum of the health scores of all working conditions is calculated to obtain the total motor health score:

[0028] Score=α1H1+α2H2+…α m H m +…+α n H n

[0029] In the formula, α m is the weight of the mth working condition, Hm is the health score of the mth (m=1, 2, ..., n) working condition, and n is the total number of divided working conditions.

[0030] The above-mentioned method for evaluating the health of new energy vehicle motors based on big data provided by the present invention starts with the efficiency of the motor and effectively utilizes the big data of the actual vehicle to analyze the instantaneous efficiency of the motor under different working conditions. Through the historical data of the electric vehicle, the characteristic threshold data for distinguishing the working conditions and the threshold for efficiency screening are obtained, and then the efficiency of the sliding interval is screened and counted using the threshold. On this basis, the arithmetic mean filtering method is used to weaken the influence of the noise data. Finally, a weighted average method is used to comprehensively give an evaluation result of the health of the motor under various working conditions. This method overcomes the shortcomings of the existing motor health diagnosis, which is inefficient and time-consuming, and provides a solution for the rapid online safety inspection of new energy vehicles. It also provides an effective reference for the offline safety inspection of new energy vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 The overall flow chart of the method provided by the present invention is as follows. DETAILED DESCRIPTION

[0032] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. 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.

[0033] The present invention provides a method for evaluating the health of new energy vehicle motors based on big data, such as Figure 1 As shown, the specific steps include:

[0034] Step 1: The vehicle terminal extracts the actual vehicle operation data under different working conditions and uploads it to the platform;

[0035] Step 2: The platform extracts various characteristic data that affect the health status of the motor from the received real vehicle operation data, including: motor current, voltage, vehicle speed, motor speed, torque, etc.;

[0036] Step 3, dividing the driving conditions corresponding to the vehicle according to the vehicle speed, calculating the instantaneous efficiency of the motor corresponding to each frame under different driving conditions, and obtaining a box plot of the instantaneous efficiency of the motor under different conditions;

[0037] Step 4: Using the minimum value in the motor instantaneous efficiency box diagram calculated for the same vehicle model, determine the motor efficiency threshold of the vehicle model under the corresponding working condition and establish a threshold library;

[0038] Step 5: For the target vehicle to be evaluated, within a specific frame time window under each working condition, calculate and count the number of healthy frames in which the instantaneous efficiency of the motor is not less than the corresponding threshold in the threshold library;

[0039] Step 6: Perform a motor health score based on the number of healthy frames of the target vehicle under different working conditions, and determine the actual current health status of the vehicle motor based on the score.

[0040] In a preferred embodiment of the present invention, after the platform receives the actual vehicle operation data in step 2, it first performs data preprocessing methods such as data cleaning, data integration, data transformation, and data reduction in sequence before extracting feature data, eliminates outliers, splits data, and reorganizes data, and processes the original data into a format that can be directly used and processed later.

[0041] In a preferred embodiment of the present invention, with 47 Yutong buses as samples in step 3, the speed data in the historical driving segments are monitored offline and a speed box plot is drawn. The upper and lower quartiles in the box plot and the speed difference corresponding to each segment are extracted. The vehicle operating conditions are divided into 4 different types based on the speed and the speed difference, and a corresponding operating condition-speed threshold library is formed, as shown in Table 1:

[0042] Table 1 Classification of driving conditions of new energy vehicles

[0043]

[0044] The above-mentioned working condition classification method is particularly suitable for some commercial operation routes with specific driving routes and a large number of samples, which can make the working condition classification more accurate and targeted.

[0045] The instantaneous efficiency η of the motor corresponding to each frame in step 3 can be calculated based on the following formula:

[0046]

[0047] Where v (r / min) is the motor speed, T (N·m) is the motor torque, U (V) is the motor controller input voltage, and I (A) is the motor controller input current.

[0048] It can be seen from the instantaneous efficiency of the motor in each frame that the efficiency of the new energy vehicle motor is between 80% and 100% at most times. Due to the large amount of data and frequent changes in efficiency, it is not possible to directly see an obvious trend. Therefore, it is necessary to process the data through a corresponding algorithm to obtain a more intuitive motor efficiency situation. In a preferred embodiment of the present invention, step 4 determines the motor efficiency threshold, specifically including:

[0049] For each of n vehicles of the same model, the minimum instantaneous efficiency η of the motor under the same working condition is i (i=1, 2, ..., n) and take the average value to obtain the motor efficiency threshold:

[0050]

[0051] At this point, a threshold library for distinguishing abnormal efficiency can be established through training data (a total of 47 vehicles). When analyzing new test vehicles on this basis, the test vehicles can be added to the statistical cycle to update the threshold library in real time.

[0052] After the above calculation, the threshold value for distinguishing abnormal efficiency under training samples is On this basis, the efficiency in the window L (L = 100) is higher than the efficiency threshold That is, the number of normal efficiency, denoted by J i , i is the number of intervals. By calculating the number of normal efficiencies in each interval, it can be seen that the number of motors with efficiencies greater than 0.85 with an interval of 100 decreases with time, and presents an overall reduction trend. The longer the motor runs, the greater the loss of the motor, and the efficiency of the corresponding motor decreases, which is in line with expectations. However, there is still a lot of noise in the curve, and it is difficult to obtain an accurate result by directly counting the data. Therefore, in a preferred embodiment of the present invention, step 5 is continuously repeated multiple times for different working conditions of the target vehicle to be evaluated before step 6, and the number of healthy frames counted each time under the same working condition is subjected to arithmetic average filtering:

[0053]

[0054] In the formula, J i is the number of healthy frames with normal motor efficiency obtained by the ith statistics under a certain working condition, N is the number of time windows for calculation and statistics, and Y j is the result obtained by arithmetic mean filtering. Due to the large fluctuation of efficiency between each frame in actual statistics, there is random interference with approximate positive and negative deviations, which can be effectively filtered out by using the arithmetic mean filtering method. This filtering method can adjust the length N of the calculation interval according to the density of the data. When the N value is large, the smoothness of the signal is high, but the sensitivity of the model is low. When the N value is small, the smoothness of the signal is low, but the sensitivity of the model is high. Through actual vehicle verification and data analysis, taking N = 50 can ensure the smoothness of the data, and also maintain a certain sensitivity to extreme values, improving the overall stability and accuracy of the model.

[0055] The curve obtained after arithmetic mean filtering is relatively smooth, and the overall result is the same as expected, showing a trend of decreasing efficiency. Therefore, the accuracy of the result can be further improved by performing efficiency calculation on this basis.

[0056] The filtered result is used for the health score in the subsequent step 6.

[0057] After obtaining the number of healthy frames with normal motor efficiency under each working condition, the number of healthy frames less than 5 can be excluded from the screening range based on the 3σ principle. Here, the situation where the number of normal efficiency is less than 5 is determined as a reference indicator of the motor health status. Therefore, in a preferred embodiment of the present invention, the motor health score in step 6 specifically includes:

[0058] For a certain working condition, the corresponding health score H is calculated based on the following formula:

[0059]

[0060] Where p is the frame value when the number of healthy frames is less than 5, that is, Y j (<5); M is the number of frames executed and N is the total number of frames in each time window in the calculation and statistics;

[0061] The motor health calculated based on the four working conditions is weighted and statistically analyzed. The weight distribution is done through the motor efficiency map. The motor efficiency map shows that the efficiency is low at low speed and low torque, low speed and high torque, and low torque and high speed, while the efficiency is high at high speed and high torque. Therefore, the working condition of high speed and high torque can better reflect the attenuation of efficiency.

[0062] Therefore, for example, for the fourth working condition (vehicle speed is between 20 and 60 km / h, and the speed difference is more than 5 km / h), its weight should be correspondingly greater, and the weights of the remaining working conditions are correspondingly less. In this model, the weight of working condition 1 (vehicle speed is between 0 and 20 km / h, and the speed difference is within 5 km / h) is initially selected as 0.2, the weight of working condition 2 (vehicle speed is between 0 and 20 km / h, and the speed difference is more than 5 km / h) is 0.2, the weight of working condition 3 (vehicle speed is between 20 and 60 km / h, and the speed difference is less than 5 km / h) is 0.2, and the weight of working condition 4 (vehicle speed is between 20 and 60 km / h, and the speed difference is more than 5 km / h) is correspondingly larger, which is 0.4. The final total health score is:

[0063] Score=0.2H1+0.2H2+0.2H3+0.4H4

[0064] Among them, H j (j=1, 2, 3, 4) are the health scores of the four working conditions obtained after filtering. So far, the health results of the motor can be obtained by weighting the health of each working condition.

[0065] It should be understood that the size of the serial numbers of the steps in the embodiment of the present invention does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0066] Although 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 present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for evaluating the health of new energy vehicle motors based on big data, characterized in that: The specific steps include: Step 1: The vehicle terminal extracts the actual vehicle operation data under different working conditions and uploads it to the platform; Step 2: The platform extracts various characteristic data that affect the health status of the motor from the received real vehicle operation data, including: motor current, voltage, vehicle speed, motor speed and torque; Step 3, dividing the driving conditions corresponding to the vehicle according to the vehicle speed, calculating the instantaneous efficiency of the motor corresponding to each frame under different driving conditions, and obtaining a box plot of the instantaneous efficiency of the motor under different conditions; Step 4: Using the minimum value in the motor instantaneous efficiency box diagram calculated for the same vehicle model, determine the motor efficiency threshold of the vehicle model under the corresponding working condition and establish a threshold library; Step 5: For the target vehicle to be evaluated, within a specific frame time window under each working condition, calculate and count the number of healthy frames in which the instantaneous efficiency of the motor is not less than the corresponding threshold in the threshold library; Step 6: Perform a motor health score based on the number of healthy frames of the target vehicle under different working conditions, and determine the actual current health status of the vehicle motor based on the score.

2. The method according to claim 1, characterized in that: In step 2, after the platform receives the actual vehicle operation data, it first performs data cleaning, data integration, data transformation and data reduction in sequence before extracting feature data, removes outliers, splits data and reorganizes data, and processes the original data into a format that can be directly used and processed later.

3. The method according to claim 1, characterized in that: In step 3, the driving conditions corresponding to the vehicle are divided according to the vehicle speed, specifically including: For multiple vehicles of the same model in actual operation, the speed data in their historical driving segments are monitored offline and a speed box plot is drawn. The upper and lower quartiles in the box plot and the speed difference corresponding to each segment are extracted, and the vehicle operating conditions are divided into multiple different types based on the speed and speed difference. The instantaneous efficiency η of the motor corresponding to each frame is calculated based on the following formula: Where v is the motor speed, T is the motor torque, U is the motor controller input voltage, and I is the motor controller input current.

4. The method according to claim 1, characterized in that: Step 4 determines the motor efficiency threshold, specifically including: For each of n vehicles of the same model, the minimum instantaneous efficiency η of the motor under the same working condition is i , i = 1, 2, ..., n, and the motor efficiency threshold is obtained by taking the average value:

5. The method according to claim 1, characterized in that: Before step 6, step 5 is continuously repeated for multiple times for different working conditions of the target vehicle to be evaluated, and the healthy frame numbers counted each time under the same working condition are subjected to arithmetic average filtering: In the formula, J i is the number of healthy frames with normal motor efficiency obtained by the ith statistics under a certain working condition, N is the number of time windows for calculation and statistics, and Y j is the result obtained by arithmetic mean filtering; The filtered result is used for the health score in the subsequent step 6.

6. The method according to claim 5, characterized in that: The motor health score in step 6 specifically includes: For a certain working condition, the corresponding health score H is calculated based on the following formula: Where p is the number of healthy frames less than 5 after arithmetic mean filtering; M is the total number of frames in N time windows; Based on the motor efficiency map of the target vehicle to be evaluated, and in combination with the motor states corresponding to different working conditions, including low speed and low torque, low speed and high torque, and low torque and high speed, corresponding weights are assigned to different working conditions; the weighted sum of the health scores of all working conditions is calculated to obtain the total motor health score: Score=α1H1+α2H2+…α m H m +…+a q H q In the formula, α m is the weight of the mth operating condition, H m is the health score of the mth working condition, m = 1, 2, ..., q, q is the total number of divided working conditions.

Images