Motor driving system fault diagnosis method applying data analysis technology

Abstract

The invention discloses a motor driving system fault diagnosis method applying a data analysis technology, and relates to the technical field of motor driving system fault diagnosis, and the method comprises the steps: S10, obtaining a fault feature data set of a motor driving system; s20, selecting a fault diagnosis method of the motor driving system; s30, the fault reason of the motor driving system is determined; and S40, according to the determination condition of the fault reason, performing maintenance processing on the motor driving system. According to the method, early-stage potential faults can be rapidly and effectively monitored in combination with an early-stage precision prediction result, the fault diagnosis effect of the system is improved, whether points formed according to the adjusted feature data fall on a regression straight line or not is judged, some feature data capable of reflecting tiny faults are recognized, and the fault diagnosis accuracy is improved. The recognition process is visual, and the probability of feature data loss can be reduced.

Description

Technical Field

[0001] The invention relates to the technical field of motor drive system fault diagnosis, and in particular to a motor drive system fault diagnosis method using data analysis technology. Background Art

[0002] The motor drive system is mainly composed of motors, controllers, sensors and power supplies. The motor is the core component of the drive system, including DC motors, AC motors and other types; the controller is responsible for controlling the speed and direction of the motor to achieve precise control; the sensor is used to read the motor's speed, temperature, current and other parameters; the power supply provides energy to power the entire drive system.

[0003] The existing technology monitors motor drive system faults through the extreme value method, but it is difficult to effectively monitor some potential faults. Potential faults refer to faults that will occur if the running equipment continues to be used at a certain time without taking preventive maintenance and adjustment measures. At the same time, the existing technology cannot effectively monitor some minor faults, and it is difficult to place data showing nonlinear changes in a mathematical model, that is, it is impossible to effectively analyze data showing nonlinear changes, thereby reducing the diagnostic accuracy of faults.

[0004] Therefore, people are in urgent need of a motor drive system fault diagnosis method that applies data analysis technology to solve the above technical problems. Summary of the invention

[0005] The object of the present invention is to provide a motor drive system fault diagnosis method using data analysis technology to solve the problems raised in the prior art.

[0006] To achieve the above object, the present invention provides the following technical solution: a motor drive system fault diagnosis method using data analysis technology, the method comprising: S10: collecting the real-time working current value of each switch element in the motor control circuit through the current sensor installed in the motor drive system, collecting the real-time torque value of the drive motor through the controller installed in the motor drive system, and acquiring the fault feature data set of the motor drive system according to the correlation between the working current value and the torque value; S20: classifying the fault feature data set according to the rotation direction of the drive motor, determining whether the fault feature data in each type of fault feature data set presents a linear change, and selecting a fault diagnosis method for the motor drive system in combination with the smoothness corresponding to the connecting line segments between adjacent fault feature data; S30: based on the diagnosis results of each fault diagnosis method, predicting the accuracy of each fault cause analysis result, and determining the fault cause of the motor drive system based on the prediction result; S40: Perform maintenance on the motor drive system according to the determination of the cause of the fault.

[0007] Further, the S10 includes: S101: According to the acquisition time of the working current value and the torque value, the working current value and the torque value are matched, and the matching result is: (Ut1, Ut2, Ut3, Ut4, Ft), (Ut1, Ut2, Ut3, Ut4, Ft) represents the characteristic data of the motor drive system at time t, wherein Ut1, Ut2, Ut3, Ut4 represent the working current values ​​of the switch elements numbered 1, 2, 3, 4 in the drive motor at time t, respectively, and Ft represents the torque value of the drive motor at time t; S102: According to the rotation direction of the drive motor and the positive and negative conditions of the torque value of the drive motor, the fault characteristic data and the non-fault characteristic data are identified, which is conducive to identifying some fault data with obvious abnormal characteristics; S103: When the drive motor rotates forward, the non-fault characteristic data (Ut1, Ut2, Ut3, Ut4, Ft) is adjusted, and the adjustment result is: (Ut1, Ft); when the drive motor rotates reversely, the non-fault characteristic data is adjusted, and the adjustment result is: (Ut2, |Ft|); A rectangular coordinate system is constructed with the working current I of the driving motor as the horizontal coordinate and the torque T of the driving motor as the vertical coordinate. The regression line equation T=K*I is expressed in the rectangular coordinate system. The points (Ut1, Ft) and (Ut2, |Ft|) are expressed in the rectangular coordinate system. The points that do not fall on the regression line are obtained, and the original data corresponding to the obtained points are recorded as the fault feature data, where K represents the torque coefficient of the driving motor; Judging whether the points formed by the adjusted characteristic data fall on the regression line can identify some characteristic data that can reflect minor faults. Compared with the method of identifying fault characteristic data by extreme value method, the calculation amount is small and the display is more intuitive. S104: Using the set of fault characteristic data as a fault characteristic data set of the motor drive system.

[0008] Furthermore, the specific method for S102 to identify fault characteristic data and non-fault characteristic data is: when Ut1=Ut3≠0 and Ut2=Ut4=0, it indicates that the drive motor rotates forward. If Ft≤0 at this time, it means that (Ut1, Ut2, Ut3, Ut4, Ft) is the fault characteristic data. If Ft>0 at this time, it means that (Ut1, Ut2, Ut3, Ut4, Ft) is non-fault characteristic data. When Ut1=Ut3=0 and Ut2=Ut4≠0, it indicates that the drive motor rotates reversely. If Ft≥0 at this time, it means that (Ut1, Ut2, Ut3, Ut4, Ft) is the fault characteristic data. If Ft<0 at this time, it means that (Ut1, Ut2, Ut3, Ut4, Ft) is non-fault characteristic data.

[0009] Further, the S20 includes: S201: Acquire the acquisition time of the fault characteristic data. If the drive motor is in a forward rotation state at the acquired acquisition time point, put the fault characteristic data into a set M; if the drive motor is in a reverse rotation state at the acquired acquisition time point, put the fault characteristic data into a set N; S202: In a rectangular coordinate system, mark the adjustment results corresponding to the fault feature data stored in the set M, and connect the marked points using line segments. If the smoothness of the connecting line segment = 1, and the smoothness = exp(a), then the proportional analysis method is used as the fault diagnosis method of the motor drive system. If the smoothness of the connecting line segment is less than 1, then the similar analysis method is used as the fault diagnosis method of the motor drive system, where a represents the angle between the straight line where the connecting line segment is located and the regression line, and exp(a) represents e to the power of a. By determining the smoothness of each connecting line segment, the fault diagnosis method of the motor drive system in different time periods is determined. Considering that the data showing nonlinear changes and the data showing linear changes may be caused by different fault causes, the above method can accurately analyze the fault cause of the motor drive system. S203: Based on the method in S202 and according to the fault feature data stored in the set N, a fault diagnosis method for the motor drive system is selected.

[0010] Furthermore, the specific analysis process of the ratio analysis method is as follows: A point is randomly selected on the connecting line segment A, and the selected point is recorded as (Ut11, Ft1). On the regression line, the matching point of the selected point is determined, and the matching point is (Ut11, K*Ut11). Ft1 / (K*Ut11) is used as the proportional analysis coefficient between the connecting line segment A and the regression line. The rated load R of the drive motor is obtained, and the load E of the drive motor is obtained at time t1. If E / R=Ft1 / (K*Ut11), the connection state of the load connected to the drive motor is adjusted by the controller. After the adjustment, the load connected to the drive motor is ≤R. At this time, if the re-collected feature data is identified as non-fault feature data, it indicates that there is no fault in the motor drive system, and the abnormal cause of the motor drive system is overload of the drive motor. If E / R≠Ft1 / (K*Ut11) or E / R=Ft1 / (K*Ut11) and E<R, it indicates that there is a fault in the motor drive system, and the cause of the fault of the motor drive system is a fault in the drive motor transmission device or a fault in the internal parts of the drive motor.

[0011] Furthermore, the specific analysis process of the same type of analysis method is as follows: Suppose the starting point of the connecting line segment B is (Ut21, Ft2) and the end point is (Utˆ1, Ftˆ). Draw a ray from the point (Ut21, Ft2) as the starting point. The ray is parallel to the regression line. The ray model is T´=K*I+b (I≥Ut21), where b represents the bias of the ray model and T´ represents the dependent variable of the ray model. G=|Ftˆ-(K*Utˆ1+b)| / b is used as the homogeneous analysis index between the connecting line segment B and the regression line. According to the relationship between E / R and Ft2 / (K*Ut21), a preliminary analysis is performed on the fault cause of the motor drive system. If the preliminary analysis result is a motor drive system fault and the analyzed fault cause occurs for the first time, the associated fault causes of the analyzed fault cause are obtained, and the correlation coefficient between the analyzed fault cause and each associated fault cause is obtained. If the obtained correlation coefficient = G, the associated fault cause corresponding to the corresponding correlation coefficient is also the fault cause of the motor drive system; If the preliminary analysis result shows that there is no fault in the motor drive system, the torque values ​​collected by the torque sensor in the motor drive system at times t2 and tˆ are obtained, and the absolute value Y of the difference between the obtained torque values ​​is calculated. If |Ftˆ-Ft2|=Y, it means that the torque sensor in the motor drive system is faulty. If |Ftˆ-Ft2|≠Y, it means that the current sensor in the motor drive system is faulty or the power supply voltage is unstable.

[0012] Further, the S30 includes: S301: Analyze the fault causes of the motor drive system in the time period corresponding to each connecting line segment by using a proportional analysis method and a fault analysis method; S302: predicting the accuracy of the analysis results of each fault cause according to the occurrence time period of each fault cause; S303: According to the prediction result in S302, the fault cause of the motor drive system is determined.

[0013] Furthermore, the specific method of predicting the accuracy of each fault cause analysis result in S302 is: The initial occurrence time fi and elimination time gi of the fault cause numbered i are obtained, and the duration ci of the fault cause is calculated within the time period [fi,gi]; Ⅰ. When the fault cause of the motor drive system is unstable sensor or unstable power supply, the accuracy of the fault cause analysis result numbered i is predicted according to V1i=log2{[(gi-fi-ci) / hi]+1}, where hi represents the duration of the sensor or power supply being in an unstable state within the time period [fi,gi]; Ⅱ. When the fault cause of the motor drive system is not the fault cause in Ⅰ, the accuracy of the fault cause analysis result numbered i is predicted according to V2i=ci / (gi-fi), where i=1,2,…,m, representing the number corresponding to each fault cause, and m represents the total number of fault causes existing in the motor drive system.

[0014] The accuracy of the analysis results of each fault cause is predicted based on the duration of each fault cause within the corresponding fault cause occurrence time period. This process takes into account the impact of the continuous instability of the sensor or power supply on the accuracy of the fault cause analysis results, thereby improving the accuracy of determining the fault cause of the motor drive system.

[0015] When the accuracy of the fault cause analysis result increases with time, it can be determined that the fault cause is a potential fault of the motor drive system, thereby improving the system's efficiency in identifying potential faults. Furthermore, the specific method of determining the fault cause of the motor drive system in S303 is: If 0.9≤the accuracy of the analysis result of the fault cause numbered i≤1, the fault cause numbered i is the fault cause of the motor drive system; If 0≤the accuracy of the analysis result of the fault cause numbered i<0.9, the fault cause numbered i is not the fault cause of the motor drive system.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention determines whether the points formed by the adjusted feature data fall on the regression line, and can identify some feature data that can reflect minor faults. Compared with the extreme value method, the present invention has higher recognition efficiency and the recognition process is visualized, which is beneficial to reduce the probability of feature data loss.

[0017] 2. In the process of determining the cause of the fault of the motor drive system, the present invention predicts the accuracy of the analysis results of the cause of each fault according to the duration of each fault in the corresponding fault time period. During the prediction process, it can combine the early accuracy prediction results to quickly and effectively monitor the potential faults that exist in the early stage, thereby improving the fault diagnosis effect of the system.

[0018] 3. The present invention identifies characteristic data showing nonlinear changes through the smoothness of each connecting line segment, divides the characteristic data showing nonlinear changes into two parts, and places the characteristic data showing nonlinear changes in a mathematical model, which is conducive to accurately analyzing the cause of the failure of the motor drive system. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 The present invention is a schematic diagram of the working process of a motor drive system fault diagnosis method using data analysis technology. DETAILED DESCRIPTION

[0020] Next, the technical solution of the present invention will be clearly and completely explained through specific case analysis to ensure that those skilled in the art can implement the present invention according to this embodiment.

[0021] like Figure 1 As shown, the present invention provides the following technical solution, a motor drive system fault diagnosis method using data analysis technology, the method comprising: S10: collecting the real-time working current value of each switch element in the motor control circuit through the current sensor installed in the motor drive system, collecting the real-time torque value of the drive motor through the controller installed in the motor drive system, and acquiring the fault feature data set of the motor drive system according to the correlation between the working current value and the torque value; S10 includes: S101: According to the acquisition time of the working current value and the torque value, the working current value and the torque value are matched, and the matching result is: (Ut1, Ut2, Ut3, Ut4, Ft), (Ut1, Ut2, Ut3, Ut4, Ft) represents the characteristic data of the motor drive system at time t, wherein Ut1, Ut2, Ut3, Ut4 represent the working current values ​​of the switch elements numbered 1, 2, 3, 4 in the drive motor at time t, respectively, and Ft represents the torque value of the drive motor at time t; S102: According to the rotation direction of the drive motor and the positive and negative conditions of the torque value of the drive motor, the fault characteristic data and the non-fault characteristic data are identified. The specific identification method is: when Ut1=Ut3≠0 and Ut2=Ut4=0, it means that the drive motor rotates in the forward direction. If Ft≤0 at this time, it means that (Ut1, Ut2, Ut3, Ut4, Ft) is the fault characteristic data. If Ft>0 at this time, it means that (Ut1, Ut2, Ut3, Ut4, Ft) is the non-fault characteristic data. When Ut1=Ut3=0 and Ut2=Ut4≠0, it means that the drive motor rotates in the reverse direction. If Ft≥0 at this time, it means that (Ut1, Ut2, Ut3, Ut4, Ft) is the fault characteristic data. If Ft<0 at this time, it means that (Ut1, Ut2, Ut3, Ut4, Ft) is the non-fault characteristic data. S103: When the drive motor rotates forward, the non-fault characteristic data (Ut1, Ut2, Ut3, Ut4, Ft) is adjusted, and the adjustment result is: (Ut1, Ft); when the drive motor rotates reversely, the non-fault characteristic data is adjusted, and the adjustment result is: (Ut2, |Ft|); A rectangular coordinate system is constructed with the working current I of the driving motor as the horizontal coordinate and the torque T of the driving motor as the vertical coordinate. The regression line equation T=K*I is expressed in the rectangular coordinate system. The points (Ut1, Ft) and (Ut2, |Ft|) are expressed in the rectangular coordinate system. The points that do not fall on the regression line are acquired, and the original data corresponding to the acquired points are recorded as the fault characteristic data, wherein K represents the torque coefficient of the driving motor, and the original data of the acquired point refers to the characteristic data corresponding to the acquired point before adjustment. S104: using the set of fault characteristic data as a fault characteristic data set of the motor drive system; S20: classifying the fault feature data set according to the rotation direction of the drive motor, determining whether the fault feature data in each type of fault feature data set presents a linear change, and selecting a fault diagnosis method for the motor drive system in combination with the smoothness corresponding to the connecting line segments between adjacent fault feature data; The S20 includes: S201: Acquire the acquisition time of the fault characteristic data. If the drive motor is in a forward rotation state at the acquired acquisition time point, put the fault characteristic data into a set M; if the drive motor is in a reverse rotation state at the acquired acquisition time point, put the fault characteristic data into a set N; S202: In a rectangular coordinate system, mark the adjustment results corresponding to the fault feature data stored in the set M, and connect the marked points using line segments. If the smoothness of the connecting line segment = 1, and the smoothness = exp(a), then the proportional analysis method is used as the fault diagnosis method of the motor drive system. If the smoothness of the connecting line segment is less than 1, then the same type analysis method is used as the fault diagnosis method of the motor drive system, wherein a represents the angle between the straight line where the connecting line segment is located and the regression line, exp(a) represents e to the power of a, and the regression line refers to the straight line represented by the regression line equation; The specific analysis process of the ratio analysis method is as follows: A point is randomly selected on the connecting line segment A, the smoothness of the connecting line segment A = 1, and the selected point is recorded as (Ut11, Ft1). On the regression line, the matching point of the selected point is determined, and the matching point is (Ut11, K*Ut11). Ft1 / (K*Ut11) is used as the proportional analysis coefficient between the connecting line segment A and the regression line. Ut11 represents the working current value of the switch element numbered 1 in the drive motor at time t1, and Ft1 represents the torque value of the drive motor at time t1. The rated load R of the drive motor is obtained, and the load E of the drive motor is obtained at time t1. If E / R=Ft1 / (K*Ut11), the connection state of the load connected to the drive motor is adjusted by the controller. After the adjustment, the load connected to the drive motor is ≤R. At this time, if the re-collected feature data is identified as non-fault feature data, it means that there is no fault in the motor drive system, and the abnormal cause of the motor drive system is overload of the drive motor. If E / R≠Ft1 / (K*Ut11) or E / R=Ft1 / (K*Ut11) and E<R, it means that there is a fault in the motor drive system, and the fault cause of the motor drive system is a fault in the drive motor transmission device or a fault in the internal parts of the drive motor; The specific analysis process of the same type analysis method is as follows: Assume that the starting point of the connecting line segment B is (Ut21, Ft2), the end point is (Utˆ1, Ftˆ), the smoothness of the connecting line segment B is less than 1, and a ray is drawn from the point (Ut21, Ft2) as the starting point. The ray is parallel to the regression line. The ray model is T´=K*I+b (I≥Ut21), where b represents the bias of the ray model, T´ represents the dependent variable of the ray model, and G=|Ftˆ-(K*Utˆ1+b)| / b is used as the homogeneous analysis index of the connecting line segment B and the regression line. Ut21 and Utˆ1 represent the working current values ​​of the switch element numbered 1 in the drive motor at t2 and tˆ respectively, and Ft2 and Ftˆ represent the torque values ​​of the drive motor at t2 and tˆ respectively; According to the relationship between E / R and Ft2 / (K*Ut21), a preliminary analysis is performed on the fault cause of the motor drive system. If the preliminary analysis result is a motor drive system fault and the analyzed fault cause appears for the first time, the associated fault causes of the analyzed fault cause are obtained. The associated fault causes refer to other fault causes induced by the analyzed fault cause. The correlation coefficient between the analyzed fault cause and each associated fault cause is obtained. The method for obtaining the correlation coefficient is: randomly select a fault cause from the associated fault causes. When the fault cause of the motor drive system is the analyzed fault cause and the selected fault cause, the real-time torque value P of the motor drive system is obtained. When the fault cause of the motor drive system is only the analyzed fault cause, the real-time torque value Q of the motor drive system is obtained. The correlation coefficient = (PQ) / (QW), wherein W represents the real-time torque value obtained when the motor drive system does not fail. If the obtained correlation coefficient = G, the associated fault cause corresponding to the corresponding correlation coefficient is also the fault cause of the motor drive system. If the preliminary analysis result shows that there is no fault in the motor drive system, the torque values ​​collected by the torque sensor in the motor drive system at time t2 and tˆ are obtained, and the absolute value Y of the difference between the obtained torque values ​​is calculated. If |Ftˆ-Ft2|=Y, it means that the torque sensor in the motor drive system is faulty. If |Ftˆ-Ft2|≠Y, it means that the current sensor in the motor drive system is faulty or the power supply voltage is unstable. S203: In a rectangular coordinate system, mark the adjustment results corresponding to the fault feature data stored in the set N, and connect the marked points using line segments. If the smoothness of the connecting line segment = 1, and the smoothness = exp(a), then the proportional analysis method is used as the fault diagnosis method of the motor drive system. If the smoothness of the connecting line segment is less than 1, then the same type analysis method is used as the fault diagnosis method of the motor drive system, wherein a represents the angle between the straight line where the connecting line segment is located and the regression line, exp(a) represents e to the power of a, and the regression line refers to the straight line represented by the regression line equation; S30: based on the diagnosis results of each fault diagnosis method, predicting the accuracy of each fault cause analysis result, and determining the fault cause of the motor drive system based on the prediction result; S30 includes: S301: Analyze the fault causes of the motor drive system in the time period corresponding to each connecting line segment by using a proportional analysis method and a fault analysis method; S302: Predicting the accuracy of the analysis results of each fault cause according to the time period of occurrence of each fault cause, the specific method is: The initial occurrence time fi and elimination time gi of the fault cause numbered i are obtained, and the duration ci of the fault cause is calculated within the time period [fi,gi]; Ⅰ. When the fault cause of the motor drive system is unstable sensor or unstable power supply, the accuracy of the fault cause analysis result numbered i is predicted according to V1i=log2{[(gi-fi-ci) / hi]+1}, where hi represents the duration of the sensor or power supply being in an unstable state within the time period [fi,gi], and log2{[(gi-fi-ci) / hi]+1} represents the logarithm with base 2; Ⅱ. When the fault cause of the motor drive system is not one of the fault causes in Ⅰ, the accuracy of the fault cause analysis result numbered i is predicted according to V2i=ci / (gi-fi), where i=1,2,…,m, represents the number corresponding to each fault cause, m represents the total number of fault causes existing in the motor drive system, and V2i represents the accuracy of the fault cause analysis result numbered i.

[0022] S303: According to the prediction result in S302, the fault cause of the motor drive system is determined, and the specific method is as follows: If 0.9≤the accuracy of the analysis result of the fault cause numbered i≤1, the fault cause numbered i is the fault cause of the motor drive system; If 0≤the accuracy of the fault cause analysis result numbered i<0.9, the fault cause numbered i is not the fault cause of the motor drive system; S40: Perform maintenance on the motor drive system according to the determination of the cause of the fault.

[0023] Example 1: Assume that the fault cause of the motor drive system is sensor fault, and the specific fault cause is sensor instability. The number corresponding to the sensor fault is 2. The initial occurrence time of the fault cause numbered 2 is f2=12:00, and the elimination time is g2=12:20. In the time period [12:00, 12:20], the duration of the sensor in the unstable state is h2=10 minutes, and the duration of the fault cause numbered 2 is c2=15 minutes. Then, the accuracy of the fault cause analysis result numbered 2 is: V12=log2{[(g2-f2-c2) / h2]+1}=log2{[5 / 10]+1}=0.59; Since V12=0.59<0.9, the fault cause numbered 2 is not a fault cause of the motor drive system.

Claims

1. A motor drive system fault diagnosis method using data analysis technology, characterized in that: The method comprises: S10: collecting the real-time working current value of each switch element in the motor control circuit through the current sensor installed in the motor drive system, collecting the real-time torque value of the drive motor through the controller installed in the motor drive system, and acquiring the fault feature data set of the motor drive system according to the correlation between the working current value and the torque value; S20: classifying the fault feature data set according to the rotation direction of the drive motor, determining whether the fault feature data in each type of fault feature data set presents a linear change, and selecting a fault diagnosis method for the motor drive system in combination with the smoothness corresponding to the connecting line segments between adjacent fault feature data; S30: based on the diagnosis results of each fault diagnosis method, predicting the accuracy of each fault cause analysis result, and determining the fault cause of the motor drive system based on the prediction result; S40: Perform maintenance on the motor drive system according to the determination of the cause of the fault.

2. The motor drive system fault diagnosis method using data analysis technology according to claim 1 is characterized in that: The S10 includes: S101: According to the acquisition time of the working current value and the torque value, the working current value and the torque value are matched, and the matching result is: (Ut1, Ut2, Ut3, Ut4, Ft), (Ut1, Ut2, Ut3, Ut4, Ft) represents the characteristic data of the motor drive system at time t, wherein Ut1, Ut2, Ut3, Ut4 represent the working current values ​​of the switch elements numbered 1, 2, 3, 4 in the drive motor at time t, respectively, and Ft represents the torque value of the drive motor at time t; S102: Identifying fault characteristic data and non-fault characteristic data according to the rotation direction of the drive motor and the positive and negative conditions of the torque value of the drive motor; S103: When the drive motor rotates forward, the non-fault characteristic data (Ut1, Ut2, Ut3, Ut4, Ft) is adjusted, and the adjustment result is: (Ut1, Ft); when the drive motor rotates reversely, the non-fault characteristic data is adjusted, and the adjustment result is: (Ut2, |Ft|); A rectangular coordinate system is constructed with the working current I of the driving motor as the horizontal coordinate and the torque T of the driving motor as the vertical coordinate. The regression line equation T=K*I is expressed in the rectangular coordinate system. The points (Ut1, Ft) and (Ut2, |Ft|) are expressed in the rectangular coordinate system. The points that do not fall on the regression line are obtained, and the original data corresponding to the obtained points are recorded as the fault feature data, where K represents the torque coefficient of the driving motor; S104: Using the set of fault characteristic data as a fault characteristic data set of the motor drive system.

3. The motor drive system fault diagnosis method using data analysis technology according to claim 2 is characterized in that: The specific method for S102 to identify fault characteristic data and non-fault characteristic data is: when Ut1=Ut3≠0 and Ut2=Ut4=0, it means that the drive motor rotates forward. If Ft≤0 at this time, it means that (Ut1, Ut2, Ut3, Ut4, Ft) is the fault characteristic data. If Ft>0 at this time, it means that (Ut1, Ut2, Ut3, Ut4, Ft) is non-fault characteristic data. When Ut1=Ut3=0 and Ut2=Ut4≠0, it means that the drive motor rotates reversely. If Ft≥0 at this time, it means that (Ut1, Ut2, Ut3, Ut4, Ft) is the fault characteristic data. If Ft<0 at this time, it means that (Ut1, Ut2, Ut3, Ut4, Ft) is non-fault characteristic data.

4. The motor drive system fault diagnosis method using data analysis technology according to claim 3 is characterized by: The S20 includes: S201: Acquire the acquisition time of the fault characteristic data. If the drive motor is in a forward rotation state at the acquired acquisition time point, put the fault characteristic data into a set M; if the drive motor is in a reverse rotation state at the acquired acquisition time point, put the fault characteristic data into a set N; S202: In a rectangular coordinate system, mark the adjustment results corresponding to the fault feature data stored in the set M, and connect the marked points using line segments. If the smoothness of the connecting line segment = 1, and the smoothness = exp(a), then the proportional analysis method is used as the fault diagnosis method of the motor drive system. If the smoothness of the connecting line segment is less than 1, then the similar analysis method is used as the fault diagnosis method of the motor drive system, where a represents the angle between the straight line where the connecting line segment is located and the regression line, and exp(a) represents e to the power of a. S203: Based on the method in S202 and according to the fault feature data stored in the set N, a fault diagnosis method for the motor drive system is selected.

5. The motor drive system fault diagnosis method using data analysis technology according to claim 4 is characterized in that: The specific analysis process of the ratio analysis method is as follows: A point is randomly selected on the connecting line segment A, and the selected point is recorded as (Ut11, Ft1). On the regression line, the matching point of the selected point is determined, and the matching point is (Ut11, K*Ut11). Ft1 / (K*Ut11) is used as the proportional analysis coefficient between the connecting line segment A and the regression line. The rated load R of the drive motor is obtained, and the load E of the drive motor is obtained at time t1. If E / R=Ft1 / (K*Ut11), the connection state of the load connected to the drive motor is adjusted by the controller. After the adjustment, the load connected to the drive motor is ≤R. At this time, if the re-collected feature data is identified as non-fault feature data, it indicates that there is no fault in the motor drive system, and the abnormal cause of the motor drive system is overload of the drive motor. If E / R≠Ft1 / (K*Ut11) or E / R=Ft1 / (K*Ut11) and E<R, it indicates that there is a fault in the motor drive system, and the cause of the fault of the motor drive system is a fault in the drive motor transmission device or a fault in the internal parts of the drive motor.

6. The motor drive system fault diagnosis method using data analysis technology according to claim 5, characterized in that: The specific analysis process of the same type of analysis method is as follows: Suppose the starting point of the connecting line segment B is (Ut21, Ft2) and the end point is (Utˆ1, Ftˆ). Draw a ray from the point (Ut21, Ft2) as the starting point. The ray is parallel to the regression line. The ray model is T´=K*I+b (I≥Ut21), where b represents the bias of the ray model and T´ represents the dependent variable of the ray model. G=|Ftˆ-(K*Utˆ1+b)| / b is used as the homogeneous analysis index between the connecting line segment B and the regression line. According to the relationship between E / R and Ft2 / (K*Ut21), a preliminary analysis is performed on the fault cause of the motor drive system. If the preliminary analysis result is a motor drive system fault and the analyzed fault cause occurs for the first time, the associated fault causes of the analyzed fault cause are obtained, and the correlation coefficient between the analyzed fault cause and each associated fault cause is obtained. If the obtained correlation coefficient = G, the associated fault cause corresponding to the corresponding correlation coefficient is also the fault cause of the motor drive system; If the preliminary analysis result shows that there is no fault in the motor drive system, the torque values ​​collected by the torque sensor in the motor drive system at times t2 and tˆ are obtained, and the absolute value Y of the difference between the obtained torque values ​​is calculated. If |Ftˆ-Ft2|=Y, it means that the torque sensor in the motor drive system is faulty. If |Ftˆ-Ft2|≠Y, it means that the current sensor in the motor drive system is faulty or the power supply voltage is unstable.

7. The motor drive system fault diagnosis method using data analysis technology according to claim 6 is characterized by: The S30 includes: S301: Analyze the fault causes of the motor drive system in the time period corresponding to each connecting line segment by using a proportional analysis method and a fault analysis method; S302: predicting the accuracy of the analysis results of each fault cause according to the occurrence time period of each fault cause; S303: According to the prediction result in S302, the fault cause of the motor drive system is determined.

8. The motor drive system fault diagnosis method using data analysis technology according to claim 7, characterized in that: The specific method of predicting the accuracy of each fault cause analysis result in S302 is: The initial occurrence time fi and elimination time gi of the fault cause numbered i are obtained, and the duration ci of the fault cause is calculated within the time period [fi,gi]; Ⅰ. When the fault cause of the motor drive system is unstable sensor or unstable power supply, the accuracy of the fault cause analysis result numbered i is predicted according to V1i=log2{[(gi-fi-ci) / hi]+1}, where hi represents the duration of the sensor or power supply being in an unstable state within the time period [fi,gi]; Ⅱ. When the fault cause of the motor drive system is not the fault cause in Ⅰ, the accuracy of the fault cause analysis result numbered i is predicted according to V2i=ci / (gi-fi), where i=1,2,…,m, representing the number corresponding to each fault cause, and m represents the total number of fault causes existing in the motor drive system.

9. The motor drive system fault diagnosis method using data analysis technology according to claim 8, characterized in that: The specific method of determining the fault cause of the motor drive system in S303 is: If 0.9≤the accuracy of the analysis result of the fault cause numbered i≤1, the fault cause numbered i is the fault cause of the motor drive system; If 0≤the accuracy of the analysis result of the fault cause numbered i<0.9, the fault cause numbered i is not the fault cause of the motor drive system.

Images