Industrial robot joint motor torque fluctuation coefficient online detection method and system

Abstract

The present application relates to the field of motor torque fluctuation coefficient online monitoring, in particular to an industrial robot joint motor torque fluctuation coefficient online detection method and system, comprising the following steps: modifying a phase coil of a permanent magnet motor, a data acquisition module acquiring three-phase stator currents of the permanent magnet motor; a data processing module performing FFT spectrum analysis on the collected three-phase stator currents to obtain fundamental currents and harmonic currents; a modeling module constructing an electric signal and torque fluctuation coefficient correlation model according to the obtained fundamental currents and harmonic currents; predicting the torque fluctuation coefficient trend based on machine learning according to the electric signal and torque fluctuation coefficient correlation model to obtain a prediction result of the torque fluctuation coefficient; and a motor evaluation module acquiring a failure threshold of the torque fluctuation coefficient and comparing the prediction result of the torque fluctuation coefficient with the failure threshold to perform real-time health evaluation and early warning on the torque fluctuation coefficient of the motor.

Description

Technical Field

[0001] This invention relates to the field of online monitoring of motor torque fluctuation coefficient, specifically to a method and system for online detection of torque fluctuation coefficient of joint motors in industrial robots. Background Technology

[0002] Industrial robots play a crucial role in intelligent manufacturing, and their manufacturing precision directly determines the quality of the final product. The focus has shifted to long-term performance monitoring and prediction of overall machine precision, as this is essential for ensuring product consistency and quality. In industrial robots, servo motors are considered one of the core components because they control the robot's critical movements. The performance indicators of servo motors, including torque ripple coefficient and speed ripple coefficient, have a direct and significant impact on the overall performance of the robot. The speed ripple coefficient can typically be obtained by online monitoring of the encoder output inside the motor, thus allowing for real-time understanding of the robot's speed control performance.

[0003] However, measuring the torque ripple coefficient of a servo motor is relatively complex. The traditional method uses a torque sensor to measure the maximum / minimum values ​​of the motor torque to obtain the torque ripple coefficient. Due to the large size of the sensor, this method typically requires offline measurement with the robot shut down. Another method is to indirectly calculate the torque based on the input current signal of the servo motor during industrial robot operation. Considering that the current signal can be directly read from the robot controller, this method offers a potential solution for online measurement of the torque ripple coefficient.

[0004] The core of the traditional torque ripple coefficient method based on electrical signals is to calculate the flux linkage based on the back electromotive force. However, the back electromotive force is usually calculated based on the reverse drag method, which also requires the machine to be stopped during operation and still cannot achieve online measurement. Summary of the Invention

[0005] The purpose of this invention is to provide an online monitoring method and system for the torque fluctuation coefficient of an industrial robot joint motor. This invention fully considers that the torque fluctuation coefficient of the motor is an important parameter affecting the accuracy of the joint robot, and also considers that the method of measuring torque may have large errors and affect the operation of the joint robot. Therefore, it proposes to calculate the torque fluctuation coefficient by means of mechanism derivation.

[0006] The technical solution adopted by this invention to achieve the above objectives is: an online monitoring method for the torque fluctuation coefficient of an industrial robot joint motor, comprising the following steps:

[0007] 1) Permanent magnet servo motor current acquisition:

[0008] A phase coil of a permanent magnet motor was modified to simulate an inter-turn short-circuit fault, forming a fault injection verification experiment; the data acquisition module obtained the three-phase stator current i of the permanent magnet motor through a current transformer device. u i v i w And send it to the data processing module;

[0009] 2) Signal Analysis and Processing: The data processing module processes the three-phase stator current i collected in step 1). u i v i w Perform FFT spectrum analysis to obtain the fundamental current and harmonic current;

[0010] 3) The modeling module constructs a correlation model between the electrical signal and the torque fluctuation coefficient based on the acquired fundamental current and harmonic current;

[0011] 4) Based on the correlation model between electrical signal and torque fluctuation coefficient, the trend of torque fluctuation coefficient is predicted by machine learning, and the prediction result of torque fluctuation coefficient is obtained.

[0012] 5) The motor evaluation module obtains the failure threshold of the torque fluctuation coefficient, and compares the predicted result of the torque fluctuation coefficient with the failure threshold to perform real-time health assessment and early warning of the motor's torque fluctuation coefficient.

[0013] The modification of a certain phase coil of the permanent magnet motor specifically involves:

[0014] Remove a phase coil from the motor and connect wires at turns ratios of 0%, 2%, 5%, 10%, and 15%. Put the coil back into the motor and extend the wires to the outside of the motor.

[0015] During the experiment, the wires extending from 0% outside the motor were connected to other wires to simulate inter-turn short circuit faults.

[0016] The three-phase stator current i u i v i w Perform FFT spectrum analysis to obtain the fundamental current and harmonic current, i.e.:

[0017]

[0018] Where t is time, i1, w1 and These represent the amplitude, frequency, and phase of the fundamental current, i. k w k and These represent the amplitude, frequency, and phase of the higher-order harmonic current of phase u; i n w n and These represent the amplitude, frequency, and phase of the higher-order harmonic current of phase v; i m w m and These represent the amplitude, frequency, and phase of the higher-order harmonic current of phase w, respectively.

[0019] Step 3) specifically includes:

[0020] 1-1) The modeling module obtains the q-axis current i of the permanent magnet motor based on the fundamental and harmonic currents. q ;

[0021] 1-2) Current i through the q-axis of the permanent magnet motor q Obtain the torque ripple coefficient K tb ;

[0022] 1-3) Establish the three-phase stator current i through the mechanism model u i v i w and torque ripple coefficient K tb The correlation model between them is used, and the model is corrected by the actual measured torque ripple coefficient, thereby improving the model accuracy.

[0023] Step 1-1) specifically includes:

[0024] The three-phase stator stationary coordinate system is transformed into a two-phase rotor rotating coordinate system, i.e., the dp coordinate system, in order to decouple the influence caused by the rotor position.

[0025] By combining the formulas for the fundamental and harmonic currents of the three-phase stator current with the transformation matrix formula, the q-axis current parameter i under the two-phase rotor system is obtained. q ,Right now:

[0026]

[0027] in, θ r It is the angle between the d-axis in the two-phase rotor coordinate system and the a-axis in the three-phase stator stationary coordinate system.

[0028] Steps 1-2) are specifically as follows:

[0029] In the dq coordinate system, the electromagnetic torque equation is:

[0030]

[0031] For a convex permanent magnet synchronous motor, the magnetic circuit structure has L d =L q The characteristics of this, the simplified electromagnetic torque equation, are as follows:

[0032]

[0033] The permanent magnet motor uses i d The control method of 0 ensures that the direct-axis component of the three-phase current in the dq coordinate system is 0, i.e., the excitation current is 0; the permanent magnet is unaffected by the excitation current and the magnetic flux remains unchanged. If the current is constant, then the q-axis current i q and electromagnetic torque T em The relationship is linear; therefore, the torque ripple coefficient is obtained, i.e.:

[0034]

[0035] Among them, K tb T is the torque ripple coefficient. max electromagnetic torque T em The maximum value, T min electromagnetic torque T em The minimum value;

[0036] q-axis current i q and electromagnetic torque T em The relationship is linear, and the torque ripple coefficient is:

[0037]

[0038] Among them, K tb Let i be the torque ripple coefficient. max For the q-axis current i q The maximum value of i min For the q-axis current i q The minimum value.

[0039] Step 4) specifically includes:

[0040] Based on the correlation model between electrical signal and torque fluctuation coefficient, the torque fluctuation coefficient is obtained. Through accelerated life test of motor, a time series data of the torque fluctuation coefficient of motor throughout its entire life is obtained.

[0041] Based on machine learning or deep learning models, by using historical time-series data of torque fluctuation coefficients and known time-series data of torque fluctuation coefficients of currently running motors, the future time-series data of torque fluctuation coefficients of the motor are predicted, thus completing the future time-series prediction of torque fluctuation coefficients and obtaining the prediction results of torque fluctuation coefficients.

[0042] The failure threshold of the torque fluctuation coefficient includes: alarm threshold and shutdown threshold.

[0043] Step 5) specifically includes:

[0044] 2-1) The motor evaluation module, based on national and industry standards, obtains a repeatability accuracy alarm threshold within ±0.02mm and a shutdown threshold within ±0.05mm.

[0045] 2-2) After obtaining the alarm threshold and shutdown threshold for repeatability accuracy, an acceleration experiment is conducted on the industrial robot through kinematic and mechanistic analysis.

[0046] 2-3) In the experiment, the torque fluctuation coefficient and repeatability were measured simultaneously to obtain the mapping relationship between the torque fluctuation coefficient and repeatability.

[0047] 2-4) By mapping the torque fluctuation coefficient and repeatability accuracy, the alarm threshold and shutdown threshold of the torque fluctuation coefficient are obtained.

[0048] An online detection system for torque fluctuation coefficient of joint motor of industrial robot includes: a data acquisition module, a data processing module, a modeling module and a motor evaluation module;

[0049] The data acquisition module is used to acquire the three-phase stator current of the motor under test through the current transformer device and send it to the data processing module.

[0050] The data processing module is used to perform FFT spectrum analysis on the collected three-phase stator currents to obtain the fundamental current and harmonic current, and send them to the modeling module.

[0051] The modeling module is used to obtain the current i of the q-axis of the permanent magnet motor based on the fundamental current and harmonic current. q And thus obtain the torque ripple coefficient K. tb The three-phase stator current i is established through a mechanism model. u i v i w and torque ripple coefficient K tb The correlation model between them is used, and the model is corrected by the actual measured torque ripple coefficient;

[0052] The motor evaluation module is used to obtain the failure threshold of the torque fluctuation coefficient, and compare the predicted result of the torque fluctuation coefficient with the failure threshold to perform real-time health evaluation and early warning of the motor's torque fluctuation coefficient.

[0053] The present invention has the following beneficial effects and advantages:

[0054] 1. This method fully considers the problem that the torque fluctuation coefficient is difficult to measure directly. By analyzing the mechanism, the three-phase stator current is collected to calculate the torque fluctuation coefficient, which is of great significance for avoiding unexpected downtime caused by motor failure and ultimately achieving the maintenance of industrial equipment.

[0055] 2. This invention integrates mechanism and data-driven methods, combining the mechanism model of servo motor performance with the analysis of actual data, thereby improving the accuracy of monitoring.

[0056] 3. This invention addresses the needs of industrial robots for overall precision degradation and health management by introducing joint motor health assessment and early warning. Through real-time analysis of electrical signals, it monitors the torque fluctuation coefficient of the servo motor and predicts its trend. This helps to identify potential problems in advance and achieve prediction and health management of overall robot failures. Attached Figure Description

[0057] Figure 1 This is a flowchart of the non-invasive online monitoring method for the performance of robot joint servo motors according to the present invention;

[0058] Figure 2 This is a schematic diagram of fault injection in the motor stator coil of the present invention;

[0059] Figure 3 This is a schematic diagram of the three-phase stator current measurement scheme for the motor of the present invention;

[0060] Figure 4 This is a graph showing the degradation curve of the torque ripple coefficient of the present invention.

[0061] Figure 5 This is a graph showing the relationship between the torque fluctuation coefficient and the repeatability accuracy of the present invention.

[0062] Figure 6 This is a flowchart illustrating the method for obtaining the alarm threshold and shutdown threshold of the torque fluctuation coefficient according to the present invention. Detailed Implementation

[0063] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings and embodiments. 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 protection scope of the present invention.

[0064] This invention relates to an online monitoring system for the torque fluctuation coefficient of joint motors in industrial robots based on electrical signals, such as... Figure 1 The diagram shown is a flowchart of the method of the present invention. To verify the effectiveness of the proposed method, the present invention provides an online monitoring method for the torque fluctuation coefficient of an industrial robot joint motor, comprising the following steps:

[0065] 1) Permanent magnet servo motor current acquisition:

[0066] A phase coil of a permanent magnet motor was modified to simulate an inter-turn short-circuit fault, forming a fault injection verification experiment; the data acquisition module obtained the three-phase stator current i of the permanent magnet motor through a current transformer device. u i v i w And send it to the data processing module;

[0067] Modify the coil of a certain phase of the motor.

[0068] like Figure 2 As shown, a phase coil of the motor is removed, and wires are connected at turns ratios of 0%, 2%, 5%, 10%, and 15%. The coil is then put back into the motor, and the wires are extended to the outside of the motor. During the experiment, the wires extended at the 0% position are connected to other wires outside the motor to simulate an inter-turn short circuit fault.

[0069] like Figure 3 As shown, the motor under test was debugged on the experimental platform, and a current transformer was connected in series between the three-phase power supply and the motor. The three-phase stator current i of the motor under test with different inter-turn short-circuit fault levels was obtained through the current transformer. u i v i w .

[0070] 2) Signal Analysis and Processing (FFT):

[0071] The data processing module processes the three-phase stator current i collected in step 1). u i v i w Perform FFT spectrum analysis to obtain the fundamental current and harmonic current;

[0072] Three-phase stator current i u i v i w The fundamental current and harmonic currents are equal to the vector sum of the fundamental current and all harmonic currents. By performing an FFT on the three-phase current parameters, the fundamental and harmonic currents are obtained, yielding the following formula:

[0073]

[0074] Where t is time, i1, w1 and These represent the amplitude, frequency, and phase of the fundamental current, i. k w k and These represent the amplitude, frequency, and phase of the higher-order harmonic current of phase u. n w n and These represent the amplitude, frequency, and phase of the higher-order harmonic current of phase v, respectively. m w m and These represent the amplitude, frequency, and phase of the higher-order harmonic current of phase w, respectively.

[0075] 3) The modeling module constructs a correlation model between the electrical signal and the torque ripple coefficient based on the acquired fundamental and harmonic currents, including the following steps:

[0076] 1-1) The modeling module obtains the q-axis current i of the permanent magnet motor based on the fundamental and harmonic currents. q ;

[0077] Transforming the ABC coordinate system into the dq coordinate system decouples the influence of rotor position. By combining the fundamental and harmonic current formulas of the three-phase stator current with the transformation matrix formula, the q-axis current parameter i in the dq coordinate system can be obtained. q As shown in the formula:

[0078]

[0079] Since constant power conversion is used, then, θ r Let be the angle between the d-axis in the dq coordinate system and the a-axis in the ABC coordinate system.

[0080] 1-2) Current i through the q-axis of the permanent magnet motor q Obtain the torque ripple coefficient K tb ;

[0081] In the dq coordinate system, the electromagnetic torque equation is as follows:

[0082]

[0083] For a convex permanent magnet synchronous motor, the magnetic circuit structure has L d =L q The simplified electromagnetic torque equation, based on the characteristics of [the electromagnetic torque equation], is as follows:

[0084]

[0085] The permanent magnet synchronous motor used in this invention is i d The control method of 0 ensures that the direct-axis component of the three-phase current in the dq coordinate system is zero, i.e., the excitation current is zero. The permanent magnet is unaffected by the excitation current, and the magnetic flux remains unchanged. It is a constant. Therefore, the q-axis current i q and electromagnetic torque T em The relationship is linear.

[0086] The formula for calculating the torque ripple coefficient is as follows:

[0087]

[0088] In the formula K tb T is the torque ripple coefficient. max electromagnetic torque T em The maximum value, T min electromagnetic torque T em The minimum value.

[0089] q-axis current i q and electromagnetic torque T em Since the relationship is linear, the formula for calculating the torque ripple coefficient can also be written as follows:

[0090]

[0091] Among them, K tb Let i be the torque ripple coefficient. max For the q-axis current i q The maximum value of i min For the q-axis current i q The minimum value.

[0092] 1-3) Establish the three-phase stator current i using the mechanism model u i v i w and torque ripple coefficient K tb The correlation model between them was established, and the model accuracy was improved by correcting the model with the actual measured torque fluctuation coefficient.

[0093] 4) Based on the correlation model between electrical signals and torque fluctuation coefficients, and using historical time-series data of torque fluctuation coefficients and known time-series data of the torque fluctuation coefficients of currently operating motors, the trend of torque fluctuation coefficients is predicted using a machine learning / deep learning model, thus obtaining the prediction result of the torque fluctuation coefficient. The method for establishing the above machine learning / deep learning model is as follows:

[0094] Accelerated life testing of motors yields time-series data on the torque fluctuation coefficient over the entire motor's lifespan, which is then used to train machine learning / deep learning models. For example, when employing a multi-scale local-global feature learning network (MLN) model, based on the time-series data of the torque fluctuation coefficient over the entire motor's lifespan, the model can uncover the temporal patterns inherent in the historical data of the input torque fluctuation coefficient. To fully utilize the fundamental information in the time series, the model employs a multi-scale branching structure to model different latent patterns separately. Each pattern is extracted using a combination of interactive learning convolution and causal frequency augmentation to capture local features and global correlations, thereby enabling future time-series prediction of the torque fluctuation coefficient.

[0095] 5) The motor evaluation module obtains the failure threshold of the torque fluctuation coefficient, and compares the predicted result of the torque fluctuation coefficient with the failure threshold to perform real-time health assessment and early warning of the motor's torque fluctuation coefficient.

[0096] like Figure 4As shown, the torque fluctuation coefficient increases with the increase of motor usage time. The motor operates in the healthy zone, alarm zone and shutdown zone at different times. The trend prediction results of the aforementioned torque fluctuation coefficient are compared with the corresponding failure threshold to complete the joint motor health assessment and early warning.

[0097] Repeatability accuracy refers to the consistency of the actual pose after responding to the same command pose n times from the same direction. The threshold for torque ripple coefficient depends on the threshold for repeatability accuracy, such as... Figure 5 As shown, the failure threshold (alarm threshold, shutdown threshold) of the torque fluctuation coefficient can be determined by studying the correlation between the torque fluctuation coefficient of the industrial robot's joint motor and the repeatability of the industrial robot. The failure threshold of repeatability can be referred to the definition in "GB / T 12642-2013 Industrial Robot Performance Specification and Test Methods". There are two methods to obtain the alarm threshold and shutdown threshold of repeatability. The first method is based on national and industry standards, requiring the repeatability alarm threshold to be within ±0.02mm and the shutdown threshold to be within ±0.05mm. The other method is to obtain it according to the actual needs of the enterprise. After obtaining the alarm threshold and shutdown threshold of repeatability, as... Figure 6 As shown, accelerated experiments were conducted on industrial robots using kinematic and mechanistic analysis methods. During the experiments, the torque fluctuation coefficient and repeatability were measured simultaneously to obtain the mapping relationship between the torque fluctuation coefficient and repeatability. Finally, the alarm threshold and shutdown threshold of the torque fluctuation coefficient were obtained through the mapping.

[0098] This invention takes into account that industrial robot joint motors typically employ convex permanent magnet synchronous motors, whose magnetic circuit structure has L... d =L q The characteristics, and usually adopt i d The control method of 0 ensures that the direct-axis component of the three-phase current in the dq coordinate system is 0, i.e., the excitation current is 0. Under this condition, the permanent magnet is unaffected by the excitation current and the magnetic flux remains unchanged. It is a constant. Therefore, the q-axis current i q and electromagnetic torque T em For a linear relationship, the torque ripple coefficient can be calculated using the q-axis current i. q This allows for online monitoring of the torque ripple coefficient of the servo motor without requiring the acquisition of back electromotive force, thus enabling the robot to operate without interrupting its operation.

[0099] This invention integrates mechanism-based and data-driven methods, combining theoretical understanding of servo motor performance with analysis of actual data to improve monitoring accuracy. Furthermore, the system introduces joint motor health assessment and early warning for industrial robot machining precision. Through real-time analysis of electrical signals, it monitors the torque fluctuation coefficient of the servo motor and predicts its trend. This helps identify potential problems early, enabling prediction and health management of overall robot failures. Finally, this technology is not only applicable to the health management of the entire robot but also holds promise for playing a crucial role in intelligent operation and maintenance and intelligent manufacturing. Through real-time monitoring and prediction, manufacturing enterprises can improve robot performance and availability, reduce maintenance costs, and achieve more efficient production. This invention represents an innovation in the field of industrial robot performance monitoring and maintenance, and has potentially significant implications for intelligent manufacturing and industrial automation.

[0100] The above description is merely an embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, extensions, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A method for online detection of torque fluctuation coefficient of joint motor in industrial robots, characterized in that, Includes the following steps: 1) Permanent magnet servo motor current acquisition: A phase coil of the permanent magnet motor was modified to simulate an inter-turn short-circuit fault, forming a fault injection verification experiment; the data acquisition module obtained the three-phase stator current of the permanent magnet motor through a current transformer device. , , And send it to the data processing module; The modification of a certain phase coil of the permanent magnet motor specifically involves: Remove a phase coil from the motor and connect wires at turns ratios of 0%, 2%, 5%, 10%, and 15% respectively. Put the coil back into the motor and extend the wires to the outside of the motor. During the experiment, the wires extending from 0% outside the motor were connected to other wires to simulate inter-turn short circuit faults. 2) Signal Analysis and Processing: The data processing module processes the three-phase stator currents acquired in step 1). , , Perform FFT spectrum analysis to obtain the fundamental current and harmonic current; 3) The modeling module constructs a correlation model between the electrical signal and the torque fluctuation coefficient based on the acquired fundamental current and harmonic current; Step 3) specifically refers to: 1-1) The modeling module obtains the q-axis current of the permanent magnet motor based on the fundamental and harmonic currents. ; 1-2) Current through the q-axis of the permanent magnet motor Obtain the torque ripple coefficient ; 1-3) Establishing three-phase stator current through mechanistic model , , and torque ripple coefficient The correlation model between them is used, and the model is corrected by the actual measured torque ripple coefficient, thereby improving the model accuracy; 4) Based on the correlation model between electrical signal and torque fluctuation coefficient, the trend of torque fluctuation coefficient is predicted by machine learning, and the prediction result of torque fluctuation coefficient is obtained. Step 4) specifically involves: Based on the correlation model between electrical signal and torque fluctuation coefficient, the torque fluctuation coefficient is obtained. Through accelerated life test of motor, a time series data of the torque fluctuation coefficient of motor throughout its entire life is obtained. Based on machine learning or deep learning models, by using historical time-series data of torque fluctuation coefficient and known time-series data of torque fluctuation coefficient of currently running motor, the future time-series data of torque fluctuation coefficient of the motor is predicted, thus completing the future time-series prediction of torque fluctuation coefficient and obtaining the prediction result of torque fluctuation coefficient. 5) The motor evaluation module obtains the failure threshold of the torque fluctuation coefficient, and compares the predicted results of the torque fluctuation coefficient with the failure threshold to perform real-time health assessment and early warning of the motor's torque fluctuation coefficient. Step 5) specifically involves: 2-1) The motor evaluation module, based on national and industry standards, obtains a repeatability alarm threshold within ±0.02mm and a shutdown threshold within ±0.05mm. 2-2) After obtaining the alarm threshold and shutdown threshold for repeatability accuracy, an acceleration experiment is conducted on the industrial robot through kinematic and mechanistic analysis. 2-3) In the experiment, the torque fluctuation coefficient and repeatability were measured simultaneously to obtain the mapping relationship between the torque fluctuation coefficient and repeatability. 2-4) By mapping the torque fluctuation coefficient and repeatability accuracy, the alarm threshold and shutdown threshold of the torque fluctuation coefficient are obtained.

2. The method for online detection of torque fluctuation coefficient of industrial robot joint motor according to claim 1, characterized in that, The three-phase stator current , , Perform FFT spectrum analysis to obtain the fundamental current and harmonic current, i.e.: ； in, For time, and These represent the amplitude, frequency, and phase of the fundamental current, respectively. and These represent the amplitude, frequency, and phase of the higher-order harmonic current of phase u, respectively. and These represent the amplitude, frequency, and phase of the higher-order harmonic current of phase v, respectively. and These represent the amplitude, frequency, and phase of the higher-order harmonic current of phase w, respectively.

3. The method for online detection of torque fluctuation coefficient of industrial robot joint motor according to claim 1, characterized in that, Step 1-1) specifically involves: The three-phase stator stationary coordinate system is transformed into a two-phase rotor rotating coordinate system, i.e., the dp coordinate system, in order to decouple the influence caused by the rotor position. By combining the formulas for the fundamental and harmonic currents of the three-phase stator current with the transformation matrix formula, the q-axis current parameters under the two-phase rotor system can be obtained. ,Right now: ； in, , It is the angle between the d-axis in the two-phase rotor coordinate system and the a-axis in the three-phase stator stationary coordinate system.

4. The method for online detection of torque fluctuation coefficient of industrial robot joint motor according to claim 1, characterized in that, Steps 1-2) are specifically as follows: In the dq coordinate system, the electromagnetic torque equation is: ； For a convex permanent magnet synchronous motor, the magnetic circuit structure has The characteristics of this, the simplified electromagnetic torque equation, are: ； Permanent magnet motors use The control method ensures that the direct-axis component of the three-phase current in the dq coordinate system is 0, i.e., the excitation current is 0; the permanent magnet is unaffected by the excitation current and the magnetic flux remains unchanged, i.e. If the current is constant, then the q-axis current... and electromagnetic torque The relationship is linear; therefore, the torque ripple coefficient is obtained, i.e.: ； in, For torque ripple coefficient, electromagnetic torque The maximum value, electromagnetic torque The minimum value; q-axis current and electromagnetic torque The relationship is linear, and the torque ripple coefficient is: ； in, For torque ripple coefficient, q-axis current The maximum value, q-axis current The minimum value.

5. The method for online detection of torque fluctuation coefficient of industrial robot joint motor according to claim 1, characterized in that, The failure threshold of the torque fluctuation coefficient includes: alarm threshold and shutdown threshold.

6. The detection system for an online detection method of torque fluctuation coefficient of an industrial robot joint motor according to claim 1, characterized in that, The testing system includes: a data acquisition module, a data processing module, a modeling module, and a motor evaluation module; The data acquisition module is used to acquire the three-phase stator current of the motor under test through the current transformer device and send it to the data processing module. The data processing module is used to perform FFT spectrum analysis on the collected three-phase stator currents to obtain the fundamental current and harmonic current, and send them to the modeling module. The modeling module is used to obtain the q-axis current of the permanent magnet motor based on the fundamental current and harmonic current. And thus obtain the torque ripple coefficient. Three-phase stator current was established through a mechanistic model. , , and torque ripple coefficient The correlation model between them is used, and the model is corrected by the actual measured torque ripple coefficient; The motor evaluation module is used to obtain the failure threshold of the torque fluctuation coefficient, and compare the predicted result of the torque fluctuation coefficient with the failure threshold to perform real-time health assessment and early warning of the motor's torque fluctuation coefficient.

Images