A new energy automobile part intelligent testing method

Abstract

The application relates to the technical field of motor testing, and specifically discloses a new energy automobile part intelligent testing method, which comprises the following steps: defining a plurality of fault types of a motor, and establishing fault combinations; collecting a first signal and extracting standard features; collecting a second signal and extracting judgment features; collecting a third signal and extracting real-time features; judging whether the motor is abnormal based on the real-time features and the standard features, and judging whether the motor is faulty based on the real-time features and the judgment features when the motor is abnormal; if the motor is faulty, the testing is ended; if the motor is not faulty, it is determined that the pretightening force is abnormal, and the bolt is retightened; after the bolt is retightened, it is judged whether there is a pretightening force dispersion risk. The application realizes intelligent identification of motor abnormalities and faults and bolt pretightening force state monitoring, and ensures that the testing process is stable and reliable.

Description

Technical Field

[0001] This invention relates to the field of motor testing technology, and more specifically to an intelligent testing method for new energy vehicle components. Background Technology

[0002] Testing of new energy vehicle components is a core component in ensuring the safety and performance of the entire vehicle and a key support for promoting high-quality development of the industry. It focuses on core components such as batteries, motors, and electronic controls, covering multi-dimensional testing including functionality, durability, and safety. It monitors key parameters such as voltage and temperature by simulating complex operating conditions. Relying on technologies such as sensors and big data, it achieves automated and intelligent testing, providing data support for optimized component design and strengthening the quality defense line of the industrial chain.

[0003] Vibration testing of motor drive systems needs to simulate the bumps and jolts of a vehicle in motion. In traditional testing, workers tighten the bolts securing the motor to the test bench to the standard torque, neglecting the natural decay of the preload during vibration. As the test progresses (e.g., after 100 hours of vibration), the bolts loosen, causing the motor mounting gap to increase. The actual vibration intensity it withstands will be lower than the set value, but the testing system still records data based on the initial vibration parameters, ultimately overestimating the motor's fatigue resistance.

[0004] To address the aforementioned issues, the bolts need to be retightened. However, even after retightening the same group of bolts, the actual preload may vary significantly (i.e., preload dispersion), preventing the vibration state from returning to the baseline level. Motor mounting typically uses 4-8 bolts. Even when using the same torque wrench during retightening, dispersion can occur due to factors such as differences in friction coefficients, varying degrees of oil or oxidation on bolt threads and contact surfaces, leading to deviations in the preload converted from the same torque. In this case, the system may trigger another warning, creating an ineffective cycle of "retightening-warning-retightening again." Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent testing method for new energy vehicle components, thereby solving the aforementioned technical problems.

[0006] The objective of this invention can be achieved through the following technical solutions: A method for intelligent testing of new energy vehicle components includes the following steps: Several fault types of motors are predefined. The fault types are combined to obtain several different fault combinations. At least one different fault type exists in each different fault combination. When there is no fault in the motor, the vibration signal during the operation of the motor is collected as the first signal, and the features of the first signal are extracted as the standard features of the motor. When the motor meets any fault combination, the vibration signal during the motor operation is collected and recorded as the second signal. The features of the second signal are extracted as the judgment features of the fault combination. During the vibration test of the motor, the vibration signal during the motor operation is collected and recorded as the third signal, and the features of the third signal are extracted as the real-time features of the motor. Based on real-time features and standard features, determine whether there is an abnormality in the motor, and when there is an abnormality in the motor, determine whether there is a fault in the motor based on real-time features and judgment features; If the motor is faulty, the test ends; If the motor is not faulty, then the preload is determined to be abnormal, and the bolts are retightened. After retightening the bolts, determine if there is a risk of preload dispersion; If there is a risk of preload dispersion, the bolts should be tightened a second time. If there is no risk of preload dispersion, continue with vibration testing of the motor.

[0007] As a further aspect of the present invention: determining whether the motor has any abnormalities based on real-time features and standard features, including: Obtain the standard feature A1 and the real-time feature A2 corresponding to the i-th feature, and obtain the absolute difference between the judgment feature A1 and the real-time feature A2 as the first deviation of the i-th feature; The first deviation of all types of features is obtained. If the first deviation is greater than the preset first deviation threshold, the motor is judged to be abnormal. The first deviation of different types of features corresponds to different first deviation thresholds.

[0008] As a further aspect of the present invention: determining whether a motor has a fault based on real-time features and judgment features, including: The judgment features of the Xth fault combination are used as the target features; Obtain the target feature B1 and the real-time feature B2 corresponding to the j-th feature, and obtain the absolute difference between the target feature B1 and the real-time feature B2 as the second deviation of the j-th feature; The second deviation of all types of features is weighted and summed to obtain the judgment value of the Xth fault combination; Obtain the judgment value corresponding to all fault combinations. If there is a judgment value less than the preset judgment value threshold, then the motor is judged to be faulty.

[0009] As a further aspect of the present invention: determining whether there is a risk of preload dispersion includes: The vibration signals collected by the vibration sensors at the preset positions are acquired. Different vibration sensors correspond to different bolts. The vibration signals collected by a single vibration sensor are converted into vibration values ​​and the curve of vibration value change over time is plotted. Obtain the peaks on the curve and record the timestamps corresponding to the peaks. Sort the peaks according to the timeline to obtain the peak sorting. The timestamp corresponding to the k-th peak in the peak sorting is recorded as the target stamp. The median of the target stamps is taken as the standard stamp. If the difference between the target stamp and the standard stamp is greater than the preset timestamp deviation, the number of the first abnormal bits is increased by 1. The number of the first abnormal bits is initially 0, and k is a positive integer. The total number N of the first abnormal position is counted. If the total number N is greater than the preset first total number threshold, it is determined that there is a risk of preload dispersion.

[0010] As a further aspect of the present invention: the secondary tightening of the bolts includes: If the difference between the target stamp and the standard stamp of peak sorting i is greater than the preset timestamp deviation, then the kth bit in peak sorting i will be regarded as the second abnormal bit. The total number M of the second abnormal position of peak sorting i is counted. If the total number M is greater than the preset second total number threshold, the bolt corresponding to peak sorting i is determined to be an abnormal bolt, and the abnormal bolt is tightened again.

[0011] As a further aspect of the present invention: in the process of obtaining the peak points on the curve: Divide the curve into several sub-curves with a preset time span, obtain all the peak points on a single sub-curve, and record them as candidate peak points; The candidate peak with the largest corresponding vibration value is selected as the target peak. Candidate peaks of the sub-curve other than the target peak are no longer selected as peaks. In the peak sorting, only the target peaks of different sub-curves exist.

[0012] As a further aspect of the present invention, it also includes: If the motor has a fault, the total duration of the vibration test is recorded. If the total duration is less than the preset total duration threshold, the motor is judged to have poor performance, and the corresponding fault combination is reported.

[0013] As a further aspect of the present invention: the total number of times a single bolt is retightened is recorded; if the total number exceeds a preset threshold, an error message is sent to a preset administrator.

[0014] The beneficial effects of this invention compared to the prior art are as follows: (1) By establishing a multi-layer feature comparison mechanism of standard features, judgment features and real-time features of motor, this invention can identify abnormal operation and fault status of motor in a timely manner during the test, realize dynamic monitoring of motor performance changes, avoid test result distortion caused by vibration intensity deviation, and thus ensure the reliability and consistency of motor fatigue performance test.

[0015] (2) This invention performs time series analysis on the vibration signal peaks of bolts to determine the risk of preload dispersion and performs secondary re-tightening on abnormal bolts, thereby achieving self-correction and dispersion control of re-tightening quality, avoiding ineffective re-tightening operations, significantly improving the consistency and reliability of bolt connections, and thus ensuring the accuracy and safety of the entire vibration test process. Attached Figure Description

[0016] The invention will now be further described with reference to the accompanying drawings.

[0017] Figure 1 This is a flowchart illustrating an intelligent testing method for new energy vehicle components according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 As shown, this invention is an intelligent testing method for new energy vehicle components, comprising the following steps: Several fault types of motors are predefined. The fault types are combined to obtain several different fault combinations. At least one different fault type exists in each different fault combination. It should be noted that several possible fault types are pre-defined based on common failure modes of motors during operation, such as rotor imbalance, bearing wear, stator eccentricity, loose shaft, or winding short circuit. Each fault type corresponds to different vibration characteristics or signal changes. By superimposing these individual fault types in different combinations, various fault combinations can be obtained. For example, combining rotor imbalance with bearing wear forms a composite fault, and combining stator eccentricity with loose shaft forms another composite fault. Different fault combinations contain at least one different fault type, thus covering a variety of potential abnormal situations of the motor under complex operating conditions.

[0020] When there is no fault in the motor, the vibration signal during the operation of the motor is collected as the first signal, and the features of the first signal are extracted as the standard features of the motor. When the motor meets any fault combination, the vibration signal during the motor operation is collected and recorded as the second signal. The features of the second signal are extracted as the judgment features of the fault combination. During the vibration test of the motor, the vibration signal during the motor operation is collected and recorded as the third signal, and the features of the third signal are extracted as the real-time features of the motor. Understandably, features refer to key information parameters extracted from motor vibration signals that characterize the motor's operating state. These parameters reflect the signal's variation patterns in the time and frequency domains, including but not limited to the vibration signal's amplitude, root mean square value, kurtosis, skewness, frequency distribution, dominant frequency component, spectral energy ratio, and energy characteristics of the envelope signal. These features reflect the differences in the motor's mechanical response under different operating conditions. For example, during normal motor operation, the frequency distribution of the vibration signal is relatively stable, while when bearing wear or rotor imbalance occurs, the feature parameters will show significant fluctuations or shifts. Therefore, feature extraction can achieve an accurate description of vibration modes under different conditions, providing a quantitative basis for anomaly identification and fault diagnosis.

[0021] Based on real-time features and standard features, determine whether there is an abnormality in the motor, and when there is an abnormality in the motor, determine whether there is a fault in the motor based on real-time features and judgment features; A preferred embodiment of the present invention determines whether a motor has an abnormality based on real-time features and standard features, including: Obtain the standard feature A1 and the real-time feature A2 corresponding to the i-th feature, and obtain the absolute difference between the judgment feature A1 and the real-time feature A2 as the first deviation of the i-th feature; The first deviation of all types of features is obtained. If the first deviation is greater than the preset first deviation threshold, the motor is judged to be abnormal. The first deviation of different types of features corresponds to different first deviation thresholds.

[0022] It is important to note that when diagnosing motor anomalies, real-time features are compared one-to-one with standard features. Here, 'i' represents any feature, such as amplitude. A1 and A2 represent the amplitude of the standard feature and the real-time feature, respectively. The difference between A1 and A2 is calculated to obtain the first deviation of that amplitude feature, reflecting the gap between the current state and the normal state. After obtaining the first deviations of different features, these deviations are compared with their respective set thresholds. If the deviation of a feature exceeds the corresponding threshold, it indicates that the operating state corresponding to that feature has deviated from the standard state, which may represent a change in the motor structure or connection status.

[0023] Another preferred embodiment of the present invention determines whether a motor has a fault based on real-time features and judgment features, including: The judgment features of the Xth fault combination are used as the target features; Obtain the target feature B1 and the real-time feature B2 corresponding to the j-th feature, and obtain the absolute difference between the target feature B1 and the real-time feature B2 as the second deviation of the j-th feature; The second deviation of all types of features is weighted and summed to obtain the judgment value of the Xth fault combination; Obtain the judgment value corresponding to all fault combinations. If there is a judgment value less than the preset judgment value threshold, then the motor is judged to be faulty.

[0024] It is worth noting that when performing fault diagnosis, the judgment feature of the Xth fault combination is used as the target feature. Taking the jth feature as an example (j represents any one of the features), when the jth feature is a frequency distribution, B1 and B2 are the frequency distributions of the target feature and the real-time feature, respectively. By comparing the difference between B1 and B2, the second deviation of this feature is obtained, which is used to characterize the degree of similarity between the current state and the fault combination on the jth feature. After weighted summation of the second deviations obtained from all features, the judgment value of the Xth fault combination is formed. This judgment value represents the overall similarity between the real-time feature and the target feature. After obtaining the judgment values ​​corresponding to all fault combinations, if there is a judgment value less than the preset judgment value threshold, it means that the real-time feature is relatively close to the feature of the fault combination, that is, the current motor operating state is consistent with the mode of the fault combination, thus it can be determined that the motor has a fault. At the same time, there may be multiple judgment values ​​less than the preset judgment value threshold. The fault combination corresponding to the smallest judgment value is taken as the fault combination corresponding to the motor.

[0025] It should be noted that if the motor has a fault, the test will end and the total duration of the vibration test will be recorded. If the total duration is less than the preset total duration threshold, the motor is judged to have poor performance and the corresponding fault combination will be reported.

[0026] If the motor is not faulty, then the preload is determined to be abnormal, and the bolts are retightened. After retightening the bolts, determine if there is a risk of preload dispersion; If there is a risk of preload dispersion, the bolts should be tightened a second time. If there is no risk of preload dispersion, continue with vibration testing of the motor.

[0027] In another preferred embodiment of the present invention, determining whether there is a risk of preload dispersion includes: The vibration signals collected by the vibration sensors at the preset positions are acquired. Different vibration sensors correspond to different bolts. The vibration signals collected by a single vibration sensor are converted into vibration values ​​and the curve of vibration value change over time is plotted. Obtain the peaks on the curve and record the timestamps corresponding to the peaks. Sort the peaks according to the timeline to obtain the peak sorting. The timestamp corresponding to the k-th peak in the peak sorting is recorded as the target stamp. The median of the target stamps is taken as the standard stamp. If the difference between the target stamp and the standard stamp is greater than the preset timestamp deviation, the number of the first abnormal bits is increased by 1. The number of the first abnormal bits is initially 0, and k is a positive integer. The total number N of the first abnormal position is counted. If the total number N is greater than the preset first total number threshold, it is determined that there is a risk of preload dispersion.

[0028] Understandably, time-series analysis of vibration signals from different bolt locations can reflect the synchronicity of force on each bolt under vibration loading. When the preload of the bolts is uniformly distributed, the vibration peaks collected by each sensor appear relatively concentrated in time, with small differences in the timestamps corresponding to the peaks. However, when the preload of some bolts is low or decays rapidly, their corresponding vibration responses will be earlier or later, leading to an increase in the dispersion of the peak time distribution. By comparing the deviations of each peak timestamp from the corresponding standard timestamp and counting the number of outliers, the consistency of preload among bolts can be quantified. This method uses the time characteristics of vibration response as an indirect measurement means, allowing for the determination of force dispersion without directly detecting the preload, thus verifying the uniformity of the preload after re-tightening. This avoids connection loosening and test errors caused by uneven preload, ensuring the stability and data accuracy of the motor during vibration testing. Different vibration sensors correspond to different bolts, indicating that an independent vibration sensor is installed near each critical bolt at the connection between the motor and the test bench to collect local vibration signals at that location.

[0029] In a preferred embodiment, the secondary tightening of the bolts includes: If the difference between the target stamp and the standard stamp of peak sorting i is greater than the preset timestamp deviation, then the kth bit in peak sorting i will be regarded as the second abnormal bit. The total number M of the second abnormal position of peak sorting i is counted. If the total number M is greater than the preset second total number threshold, the bolt corresponding to peak sorting i is determined to be an abnormal bolt, and the abnormal bolt is tightened again.

[0030] It is worth noting that by re-analyzing the deviations between each timestamp in the peak point sorting and the standard timestamp, bolts that still exhibit uneven stress or abnormal response after retightening can be identified. When the vibration peak point time distribution of a bolt continuously deviates from the overall synchronization range, it indicates that the preload of the bolt has not recovered to a stable level, possibly due to residual differences caused by factors such as frictional resistance, contact surface deformation, or thread slippage. Marking these time points with large deviations as the second anomaly, and judging the bolt condition by statistically analyzing the number of anomalies, ensures that a secondary retightening operation is only performed when there is indeed a significant deviation. This method utilizes the temporal characteristics of vibration response to re-verify the retightening effect, making the stress on each bolt more balanced, avoiding vibration inconsistencies caused by some connection points remaining loose, thereby ensuring the structural stability and data reliability of the motor during the testing process.

[0031] It should be noted that during the process of obtaining the peak points on the curve: Divide the curve into several sub-curves with a preset time span, obtain all the peak points on a single sub-curve, and record them as candidate peak points; The candidate peak with the largest corresponding vibration value is selected as the target peak. Candidate peaks of the sub-curve other than the target peak are no longer selected as peaks. In the peak sorting, only the target peaks of different sub-curves exist.

[0032] By dividing the vibration signal curve into several sub-curves according to a preset time span, and selecting only the peak with the largest vibration value in each sub-curve as the target peak, the stability and representativeness of peak extraction can be ensured. Vibration signals often exhibit noise interference or localized minor fluctuations under complex operating conditions. Directly identifying all peaks in the entire signal segment can easily lead to an excessive number of peaks or uneven distribution, resulting in time series misalignment. Segmenting by time span controls the number of peaks within each time interval, while retaining only the point with the largest vibration intensity represents the main response characteristics of that interval, avoiding sorting confusion caused by missed or excessive detections.

[0033] Record the total number of times a single bolt is tightened twice. If the total number exceeds the preset threshold, send an error message to the preset administrator.

[0034] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A new energy vehicle part intelligent testing method, characterized in that, Includes the following steps: Several fault types of motors are predefined. The fault types are combined to obtain several different fault combinations. At least one different fault type exists in each different fault combination. When there is no fault in the motor, the vibration signal during the operation of the motor is collected as the first signal, and the features of the first signal are extracted as the standard features of the motor. When the motor meets any fault combination, the vibration signal during the motor operation is collected and recorded as the second signal. The features of the second signal are extracted as the judgment features of the fault combination. During the vibration test of the motor, the vibration signal during the motor operation is collected and recorded as the third signal, and the features of the third signal are extracted as the real-time features of the motor. Based on real-time features and standard features, determine whether there is an abnormality in the motor, and when there is an abnormality in the motor, determine whether there is a fault in the motor based on real-time features and judgment features; If the motor is faulty, the test ends; If the motor is not faulty, then the preload is determined to be abnormal, and the bolts are retightened. After retightening the bolts, determine if there is a risk of preload dispersion; If there is a risk of preload dispersion, the bolts should be tightened a second time. If there is no risk of preload dispersion, continue with vibration testing of the motor.

2. The intelligent testing method for new energy vehicle parts according to claim 1, characterized in that, Based on real-time features and standard features, determine whether the motor has any abnormalities, including: Obtain the standard feature A1 and the real-time feature A2 corresponding to the i-th feature, and obtain the absolute difference between the judgment feature A1 and the real-time feature A2 as the first deviation of the i-th feature; The first deviation of all types of features is obtained. If the first deviation is greater than the preset first deviation threshold, the motor is judged to be abnormal. The first deviation of different types of features corresponds to different first deviation thresholds.

3. The intelligent testing method for new energy vehicle parts according to claim 1, characterized in that, Determining whether a motor has a fault based on real-time features and judgment features includes: The judgment features of the Xth fault combination are used as the target features; Obtain the target feature B1 and the real-time feature B2 corresponding to the j-th feature, and obtain the absolute difference between the target feature B1 and the real-time feature B2 as the second deviation of the j-th feature; The second deviation of all types of features is weighted and summed to obtain the judgment value of the Xth fault combination; Obtain the judgment value corresponding to all fault combinations. If there is a judgment value less than the preset judgment value threshold, then the motor is judged to be faulty.

4. The intelligent testing method for new energy vehicle parts according to claim 1, characterized in that, Determining whether there is a risk of preload dispersion includes: The vibration signals collected by the vibration sensors at the preset positions are acquired. Different vibration sensors correspond to different bolts. The vibration signals collected by a single vibration sensor are converted into vibration values ​​and the curve of vibration value change over time is plotted. Obtain the peaks on the curve and record the timestamps corresponding to the peaks. Sort the peaks according to the timeline to obtain the peak sorting. The timestamp corresponding to the k-th peak in the peak sorting is recorded as the target stamp. The median of the target stamps is taken as the standard stamp. If the difference between the target stamp and the standard stamp is greater than the preset timestamp deviation, the number of the first abnormal bits is increased by 1. The number of the first abnormal bits is initially 0, and k is a positive integer. The total number N of the first abnormal position is counted. If the total number N is greater than the preset first total number threshold, it is determined that there is a risk of preload dispersion.

5. The intelligent testing method for new energy vehicle parts according to claim 4, characterized in that, Secondary tightening of bolts includes: If the difference between the target stamp and the standard stamp of peak sorting i is greater than the preset timestamp deviation, then the kth bit in peak sorting i will be regarded as the second abnormal bit. The total number M of the second abnormal position of peak sorting i is counted. If the total number M is greater than the preset second total number threshold, the bolt corresponding to peak sorting i is determined to be an abnormal bolt, and the abnormal bolt is tightened again.

6. The intelligent testing method for new energy vehicle parts according to claim 4, characterized in that, In the process of obtaining the peak points on the curve: Divide the curve into several sub-curves with a preset time span, obtain all the peak points on a single sub-curve, and record them as candidate peak points; The candidate peak with the largest corresponding vibration value is selected as the target peak. Candidate peaks of the sub-curve other than the target peak are no longer selected as peaks. In the peak sorting, only the target peaks of different sub-curves exist.

7. The intelligent testing method for new energy vehicle parts according to claim 3, characterized in that, Also includes: If the motor has a fault, the total duration of the vibration test is recorded. If the total duration is less than the preset total duration threshold, the motor is judged to have poor performance, and the corresponding fault combination is reported. 8.The intelligent testing method for new energy automobile parts according to claim 1, characterized in that, Record the total number of times a single bolt is tightened twice. If the total number exceeds the preset threshold, send an error message to the preset administrator.

Images