Identification method for abnormal vibration of robot joint

Abstract

The invention discloses a robot joint vibration abnormity identification method comprising the following steps: S1, determining the motion range of a to-be-tested robot joint, and enabling the motion range of the to-be-tested robot joint to be as large as possible in a limited setting range; s2, configuration parameters of a motion module are determined, and the maximum speed, the maximum acceleration and the maximum jerk of the to-be-tested robot joint are made to be consistent with those of the to-be-tested robot joint in actual work; s3, data acquisition and verification: through internal communication connected with a driver, obtaining a time domain signal that the joint torque of the to-be-tested robot periodically fluctuates along with the change of an input angle position, and feeding back real-time information of the joint torque to an upper computer through a data transmission system to obtain the assembly condition of a speed reducer; and S4, analyzing the abnormal vibration data of the robot joint and outputting a calculation result. Any sensor does not need to be added, the structures of the speed reducer and the robot do not need to be modified, the method is suitable for any operation condition of the robot, the testing process is convenient, and the requirement for experiment testers is low.

Description

Technical Field

[0001] The invention relates to a method for identifying abnormal vibration of robot joints, and belongs to the technical field of robot joint vibration identification. Background Art

[0002] With the rapid development of robotics technology, industrial robots have been widely used in manufacturing, assembly, medical and other fields. In these applications, the joints of robots are the key part of their motion control, and their performance directly affects the overall efficiency and reliability of the robot. However, abnormal joint vibration often affects the stability and accuracy of robots when performing complex tasks, resulting in operational errors, equipment damage and reduced production efficiency.

[0003] Traditional robots rely mainly on manual inspection and experience to judge abnormal vibration. This method is not only inefficient, but also easily leads to delayed discovery of potential problems, increasing maintenance costs and safety risks. Therefore, it is particularly necessary to develop an efficient method for identifying joint vibration abnormalities.

[0004] An effective vibration identification method can timely identify abnormal conditions by real-time monitoring and analyzing the vibration characteristics of joints, thereby providing a reliable basis for robot maintenance. By reducing joint vibration, the accuracy and stability of the robot can be improved, its service life can be extended, and the failure rate can be reduced. In addition, with the advancement of intelligent manufacturing and Industry 4.0, autonomous monitoring and intelligent maintenance of robots will become an important trend in future development. Therefore, the proposed identification method not only meets the current industry needs, but also has broad application prospects.

[0005] The Chinese utility model patent with the authorization announcement number CN216207555U discloses a reducer assembly and quality inspection device based on vibration audio, which consists of a motion module, a detection module, a sound acquisition module, a data acquisition display and motion control module, a horizontal workbench, etc. The device drives the ball screw through a servo motor to drive the slide up and down to achieve the docking of the three-phase asynchronous motor and the reducer. The three-phase asynchronous motor drives the reducer to rotate through a flange coupling. The sound sensor collects the audio signal when the reducer is running. The data acquisition display and motion control module processes the audio signal to determine whether the reducer quality is qualified and display the detection results and sound waveform. The above patent has the following disadvantages:

[0006] 1. The device does not mention the vibration reduction measures for the three-phase asynchronous motor and reducer during operation, nor does it discuss the impact of external noise on the signal collected by the sound sensor. Insufficient vibration reduction and noise interference may cause the audio signal collected by the sound sensor to contain a large amount of irrelevant noise, affecting the accuracy of the detection results.

[0007] 2. The judgment is mainly made through the audio signal collected by the sound sensor. However, due to the different specific layout and angle of the sensor, the intensity of the collected audio signal is also different, which affects the stability of the detection result.

[0008] 3. The details of the data processing algorithm are not described in detail, and only the high-frequency signal is filtered out by a filter. The complexity and accuracy of the data processing algorithm directly affect the reliability of the test results. If the algorithm is too simple, it may not be able to effectively distinguish between normal vibration and abnormal vibration during the operation of the reducer. Summary of the invention

[0009] The purpose of the present invention is to provide a method for identifying abnormal vibration of robot joints, so as to solve the technical defect of poor accuracy of robot joint vibration detection in the prior art.

[0010] To solve the above problems, the technical solution adopted by the present invention is: a method for identifying abnormal vibration of robot joints, comprising the following steps:

[0011] S1. Determine the range of motion of the robot joint to be tested, so that its range of motion is as large as possible within the range of the limit setting, and remove other joints or keep other joints in a static state;

[0012] S2. Determine the motion module parameter configuration so that the maximum speed, maximum acceleration and maximum jerk of the robot joint to be tested are consistent with those during actual operation;

[0013] S3, data acquisition and verification, through the internal communication connected to the driver, obtain the time domain signal of the periodic fluctuation of the joint torque of the robot to be tested with the change of the input angular position, and feed back the real-time information of the joint torque to the upper computer through the data transmission system to obtain the assembly status of the reducer;

[0014] S4. Analyze the abnormal vibration data of the robot joints and output the calculation results.

[0015] As a further improvement of the present invention, S1 adopts the method of disassembling other joints or keeping other joints in a static state so as to ensure that only the robot joint to be tested performs a single joint movement during the test.

[0016] As a further improvement of the present invention, S1 removes other non-test robot joints to perform abnormal vibration test on the robot joints.

[0017] When testing the robot J1, only the base component, the swivel component, the J1 reducer and the J1 motor are retained, and J1 is guaranteed to operate normally, with the base fixed;

[0018] When testing the robot J2, keep the swivel assembly, arm assembly, J2 reducer and J2 motor, and ensure that J2 can operate normally. The contact surface between the swivel assembly and the J1 reducer is fixed.

[0019] When testing the robot J3, keep the big arm component, motor base component, J3 reducer and J3 motor, and ensure that J3 can operate normally, and the contact surface between the big arm component and the J2 reducer is fixed;

[0020] When testing robot J4, keep the motor base, forearm, J4 reducer and J4 motor, and ensure that J4 can operate normally. The contact surface between the motor base and J3 reducer is fixed.

[0021] When testing robot J5, keep the forearm component, wrist body component, J5 reducer and J5 motor, and ensure that J5 can operate normally, and the contact surface between the forearm component and the J4 reducer is fixed;

[0022] When testing the robot J6, retain the wrist body, J6 reducer and J6 motor, ensure that J6 can operate normally, and fix the contact surface between the wrist body and the J5 reducer.

[0023] As a further improvement of the present invention, S1 is to keep other non-tested robot joints in a stationary state to perform the robot joint vibration abnormality test.

[0024] When testing robot J1, the J2-J6 axis joints need to be adjusted to the specified position and remain stationary;

[0025] When testing robot J2, the J1 axis joint and J3-J6 axis joints need to be adjusted to the specified position and remain stationary;

[0026] When testing robot J3, the J1-J2 axis joints and J4-J6 axis joints need to be adjusted to the specified position and remain stationary;

[0027] When testing robot J4, the J1-J3 axis joints and J5-J6 axis joints need to be adjusted to the specified position and remain stationary;

[0028] When testing robot J5, the J1-J4 axis joints and J6 axis joints need to be adjusted to the specified position and remain stationary;

[0029] When testing robot J6, the J1-J5 axis joints need to be adjusted to the specified position and remain stationary.

[0030] As a further improvement of the present invention, in S3, a time domain signal of the joint torque of the robot to be tested fluctuating in 20 cycles with the change of the input angular position is obtained for analysis.

[0031] As a further improvement of the present invention, S4 includes,

[0032] S4.1, performing wavelet transform noise reduction processing on the real-time torque signal of the joint obtained in step S3;

[0033] S4.2. Analyze the data to see if they have ensured the stability of the data collection cycle. Some data may have deviations in the data cycle results due to objective factors. Remove the singular data and check if the test data conforms to the normal distribution.

[0034] S4.3 Perform frequency domain analysis on the normally distributed data in S4.2;

[0035] S4.4. Extract specific frequency components, kurtosis, skewness, kurtosis, main frequency, main frequency amplitude, secondary frequency and secondary frequency amplitude from the frequency spectrum, calculate the degree of deviation from the relevant data of the robot joint without vibration abnormality, and determine whether the vibration of the tested robot joint is abnormal;

[0036] S4.5. Output the test results of the joint vibration of the tested robot.

[0037] As a further improvement of the present invention, S4.4 includes,

[0038] S4.4.1. Perform local maximum detection on the signal, traverse the frequency domain signal data points, compare the value of each point with its adjacent points, determine whether the current point is a local maximum, determine the local maximum value as the peak value, and obtain the amplitude corresponding to the main frequency and the secondary frequency according to the peak value;

[0039] S4.4.2. Perform kurtosis and skewness analysis to determine whether the data has a longer right tail, a longer left tail, or a symmetrical distribution;

[0040] S4.4.3. Perform kurtosis analysis to determine the distribution characteristics of the primary and secondary frequencies and detect outliers. Calculate the kurtosis of the data based on the amplitude sequence to measure the thickness of the tail of the distribution and the sharpness of the peak.

[0041] S4.4.4. Calculate the degree of deviation between the extracted kurtosis, main frequency, main frequency amplitude, secondary frequency, secondary frequency amplitude and other data and the characteristic component data of the robot joint without vibration abnormality, and determine whether the vibration of the tested robot joint is abnormal.

[0042] As a further improvement of the present invention, in S4.4.4,

[0043] First determine whether there is a multiple relationship between the main frequency and the secondary frequency. If there is no multiple relationship, it is directly determined that the vibration of the joint of the tested robot is abnormal. If there is a multiple relationship, then determine whether the kurtosis and skewness of the secondary frequency are abnormal. If there is no abnormality, it is determined that the vibration of the joint of the tested robot is abnormal. If there is no abnormality, it is determined that the vibration of the joint of the tested robot is normal.

[0044] In summary, the beneficial effects of the present invention are as follows: the present invention does not require any additional sensors, does not require any modifications to the reducer and the structure of the robot, is applicable to any operating conditions of the robot, has a convenient testing process, and has low requirements for experimental testers.

[0045] The present invention avoids the use of additional sound sensors through internal communication connected to the driver to obtain the electrical signal of the periodic fluctuation of the joint torque with the change of the input angular position, effectively avoiding the influence of external noise on the signal collected by the external sensor, and there is no need to consider the specific layout and angle of the sensor, thereby ensuring the stability of the detection result.

[0046] The present invention provides a quantitative index of joint vibration by intercepting and calculating data, effectively distinguishing normal vibration and abnormal vibration during the operation of the reducer, and avoiding misjudgment that is easily caused by human ear judgment.

[0047] The present invention does not require the design of additional testing devices, and only requires ensuring that the joints can move normally and that the configuration parameters are consistent with the actual working configuration parameters. The testing process is simple, time-saving and low-cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flow chart of the present invention.

[0049] Figure 2 It is a schematic diagram of abnormal torque signal of joint vibration.

[0050] Figure 3 This is the spectrum characteristic diagram of joint vibration without abnormality.

[0051] Figure 4 It is the abnormal spectrum characteristic of joint vibration Figure 1 .

[0052] Figure 5 It is the abnormal spectrum characteristic of joint vibration Figure 2 .

[0053] Figure 6 It is a schematic flow chart of the method for identifying abnormal joint vibration of the present invention. DETAILED DESCRIPTION

[0054] The specific implementation of the present invention is further described below in conjunction with the accompanying drawings.

[0055] The method for identifying abnormal vibration of robot joints of the present invention is as follows: Figure 1 As shown, the following steps are included:

[0056] S1. Determine the range of motion of the robot joint to be tested, so that its range of motion is as large as possible within the range of the limit setting. In order to ensure that the test data is reasonable and not interfered with, an abnormal joint vibration identification test is performed on a single joint. Specifically, the test can be performed by disassembling other joints that are not tested or by keeping other joints that are not tested in a static state to ensure that only the robot joint to be tested makes a single joint movement during the test.

[0057] In this step, if you disassemble other robot joints not under test to test the abnormal vibration of the robot joints,

[0058] When performing an abnormal vibration test on the robot's J1-axis joint, only the base component, swivel component, J1 reducer and J1 motor are retained, and the J1-axis joint is guaranteed to operate normally, wherein the base component is fixed, the swivel component is rotatably arranged on the base component, the swivel component is driven by the J1 reducer and the J1 motor to rotate relative to the base component, and the rotation centerline of the swivel component is the J1 axis.

[0059] When performing an abnormal vibration test on the robot's J2-axis joint, retain the swivel assembly, arm assembly, J2 reducer and J2 motor, and ensure that the J2-axis joint can operate normally. The contact surface between the swivel assembly and the J1 reducer is fixed to ensure that the swivel assembly is fixed. The J2 reducer and J2 motor drive the arm assembly to rotate, and the rotation centerline of the arm assembly relative to the swivel assembly is the J2 axis.

[0060] When performing an abnormal vibration test on the robot's J3-axis joint, retain the boom component, motor base component, J3 reducer and J3 motor, and ensure that the J3-axis joint can operate normally. The contact surface between the boom component and the J2 reducer is fixed to fix the boom component. The motor base component is driven by the J3 reducer and the J3 motor to rotate relative to the boom component, wherein the rotation centerline of the motor base component relative to the boom component is the J3 axis.

[0061] When performing an abnormal vibration test on the robot's J4-axis joint, retain the motor base assembly, forearm assembly, J4 reducer and J4 motor, and ensure that the J4-axis joint can operate normally. The contact surface between the motor base assembly and the J3 reducer is fixed to fix the motor base assembly. The forearm assembly is driven by the J4 reducer and the J4 motor to rotate relative to the motor base assembly. The rotation centerline of the forearm assembly relative to the motor base assembly is the J4 axis.

[0062] When performing an abnormal vibration test on the robot's J5-axis joint, retain the forearm component, wrist body component, J5 reducer and J5 motor, and ensure that the J5-axis joint can operate normally. The contact surface between the forearm component and the J4 reducer is fixed to fix the forearm component. The wrist body component is driven by the J5 reducer and the J5 motor to rotate relative to the forearm component. The rotation centerline of the wrist body component relative to the forearm component is the J5 axis.

[0063] When performing an abnormal vibration test on the robot's J6-axis joint, retain the wrist body, J6 reducer and J6 motor, and ensure that the J6-axis joint can operate normally. The contact surface between the wrist body and the J5 reducer is fixed, the wrist body is fixed, and the J6 reducer and J6 motor drive the wrist body to rotate itself. The rotation center of the wrist body itself is the J6 axis, which is perpendicular to the J5 axis.

[0064] In this step, if the robot joint vibration abnormality test is performed while keeping other non-tested robot joints stationary,

[0065] When the robot's J1 axis joint is tested for abnormal vibration, the J2-J6 axis joints need to be adjusted to the specified position and remain stationary.

[0066] When performing an abnormal vibration test on the robot's J2 axis joint, the J1 axis joint and the J3-J6 axis joints need to be adjusted to the specified position and remain stationary.

[0067] When performing an abnormal vibration test on the robot's J3 axis joint, the J1-J2 axis joints and the J4-J6 axis joints need to be adjusted to the specified position and remain stationary.

[0068] When performing an abnormal vibration test on the robot's J4 axis joint, the J1-J3 axis joints and the J5-J6 axis joints need to be adjusted to the specified position and remain stationary.

[0069] When performing an abnormal vibration test on the robot's J5-axis joint, the J1-J4-axis joints and the J6-axis joints need to be adjusted to the specified position and remain stationary.

[0070] When performing an abnormal vibration test on the robot's J6 axis joint, the J1-J5 axis joints need to be adjusted to the specified position and remain stationary.

[0071] S2. Determine the motion module parameter configuration. To ensure that the vibration anomaly during the normal operation of the robot joint is fully simulated, the joint configuration parameters during the test must be consistent with the actual working configuration parameters, that is, the maximum speed, maximum acceleration and maximum jerk of the robot joint to be tested are consistent with those during actual operation.

[0072] S3, data acquisition and verification, through the internal communication connected to the driver, the sensor inside the joint obtains the electrical signal, i.e., the real-domain signal, of the joint torque of the robot to be tested that fluctuates periodically with the change of the input angular position, and feeds back the real-time information of the joint torque to the upper computer through data transmission, so that the upper computer can know the assembly status of the reducer. Taking into account the instability of the control system itself, and the fact that a too short acquisition cycle may cause signal distortion, and a too long acquisition cycle will affect the background calculation time, in order to completely restore the original signal without reducing the test efficiency, the present invention obtains the time domain signal of the joint torque of the robot to be tested that fluctuates with the change of the input angular position in 20 cycles for analysis and use, and intercepts the real-time torque signal of the abnormal joint vibration working section according to the vibration working section during the robot movement, such as Figure 2 shown.

[0073] S4. Analyze the abnormal vibration data of the robot joints and output the calculation results, including

[0074] S4.1. Perform wavelet transform denoising processing on the real-time torque signal of the joint obtained in step S3. The wavelet transform denoising processing is a prior art and will not be described in detail in the present invention.

[0075] S4.2. Analyze the data to see if the stability of the data acquisition cycle is guaranteed. Some data may be deviated from the data cycle results because the tested robot joints have not yet stabilized when they just started to move. Eliminate the singular data that occasionally appear on the image and deviate greatly from the normal trajectory. Check whether the test data conforms to the normal distribution. If the singular data is cleared, the test data will be normally distributed. You can also make the robot joints move for a while and then test them after the movement is stable to avoid generating singular data.

[0076] S4.3, use fast Fourier transform to perform frequency domain analysis on the normally distributed data in S4.2;

[0077]

[0078] Where M(k) is the frequency domain signal, M(n) is the time domain signal, and N is the signal length.

[0079] S4.4. Extract the frequency components, kurtosis, skewness, kurtosis, main frequency, main frequency amplitude, secondary frequency and secondary frequency amplitude from the spectrum, calculate the degree of deviation from the relevant data of the robot joints without vibration abnormalities, and determine whether the vibration of the tested robot joints is abnormal, including:

[0080] S4.4.1. Perform local maximum detection on the signal, traverse the frequency domain signal data points, compare the value of each point with its neighboring points, and determine whether the current point is a local maximum. Specifically, for a data point X(n), if the following conditions are met:

[0081] X(n)>X(n-1) and X(n)>X(n+1)

[0082] Then X(n) is regarded as the peak value, the peak value and its horizontal coordinate value are found, the peak position vector is indexed and conditionally screened, and the amplitude corresponding to the main frequency and the secondary frequency is obtained.

[0083] S4.4.2. Perform skewness analysis to quantify shape characteristics. Skewness is used to determine the degree of deviation of the data from the mean. A skewness greater than 0 is positive skewness, which means that the right tail of the data is longer and the distribution is tilted to the right. A skewness less than 0 is negative skewness, which means that the left tail of the data is longer and the distribution is tilted to the left. A skewness of 0 indicates a symmetrical distribution.

[0084]

[0085] Where S is the skewness, n is the sample size, x i is the data point, is the sample mean, and s is the sample standard deviation.

[0086] S4.4.3. Perform kurtosis analysis to determine the distribution characteristics of the primary and secondary frequencies and detect outliers. Calculate the kurtosis of the data based on the amplitude, measure the thickness of the tail of the distribution and the sharpness of the peak, and calculate the ratio of the fourth-order central moment of the data sequence to the square of the variance.

[0087]

[0088] Where K represents the kurtosis, n is the sample size, and x i is the data point, is the sample mean, and s is the sample standard deviation.

[0089] S4.4.4. Compare the extracted kurtosis, main frequency, main frequency amplitude, secondary frequency, secondary frequency amplitude and other data with the characteristic component data of the robot joints without vibration abnormalities, and determine whether the vibration of the robot joints under test is abnormal. The specific determination method is to first determine whether there is a multiple relationship between the main frequency and the secondary frequency. If there is no multiple relationship, directly determine that the vibration of the robot joints under test is abnormal. If there is a multiple relationship, then determine whether the kurtosis and skewness of the secondary frequency are abnormal. If there is an abnormality, determine that the vibration of the robot joints under test is abnormal. If there is no abnormality, determine that the vibration of the robot joints under test is normal.

[0090] S4.5. Output the test results of the vibration of the robot joints being tested, and feed back the calculated results to the host computer. If there is no abnormal vibration in the robot joints, "pass" will be displayed; if there is abnormal vibration in the robot joints, "fail" will be displayed.

[0091] The present invention automatically extracts real-time torque signals through internal communication connected to the driver, performs frequency domain analysis on the intercepted sampled data using fast Fourier transform, extracts specific frequency components in the spectrum, quantifies the shape characteristics of each frequency in the spectrum diagram, determines the distribution characteristics of the main frequency and the secondary frequency, identifies the frequency domain components of the fault characteristics, and thus performs abnormal joint vibration characteristic detection. The present invention can identify vibration characteristics in a timely manner without the aid of external sensors, intercept potential faults caused by assembly abnormalities during the installation and adjustment stage, and improve the consistency and stability of joint assembly.

[0092] The parts not specifically described in the above description are all prior art, or can be implemented by prior art. Moreover, the specific implementation cases described in the present invention are only preferred implementation cases of the present invention, and are not used to limit the scope of implementation of the present invention. That is, all equivalent changes and modifications made according to the content of the patent scope of the present invention should be regarded as the technical scope of the present invention.

Claims

1. A method for identifying abnormal vibration of a robot joint, characterized in that: The steps include: S1. Determine the motion range of the robot joint to be tested, so that the motion range is as large as possible within the range of the limit setting, and keep only the robot joint to be tested performing a single joint motion during the test; S2. Determine the motion module configuration parameters so that the maximum speed, maximum acceleration and maximum jerk of the robot joint to be tested are consistent with those during actual operation; S3, data acquisition and verification, through the internal communication connected to the driver, obtain the time domain signal of the periodic fluctuation of the joint torque of the robot to be tested with the change of the input angular position, and feed back the real-time information of the joint torque to the upper computer through the data transmission system to obtain the assembly status of the reducer; S4. Analyze the abnormal vibration data of the robot joints and output the calculation results.

2. The method for identifying abnormal joint vibration of a robot according to claim 1, characterized in that: S1 adopts the method of disassembling other joints or keeping other joints in a static state to ensure that only the robot joint to be tested performs a single joint movement during the test.

3. The method for identifying abnormal joint vibration of a robot according to claim 2, characterized in that: S1 takes the approach of disassembling other robot joints not under test to conduct abnormal vibration test on the robot joints. When testing the robot J1, only the base component, the swivel component, the J1 reducer and the J1 motor are retained, and J1 is guaranteed to operate normally, with the base fixed; When testing the robot J2, keep the swivel assembly, arm assembly, J2 reducer and J2 motor, and ensure that J2 can operate normally. The contact surface between the swivel assembly and the J1 reducer is fixed. When testing the robot J3, keep the big arm component, motor base component, J3 reducer and J3 motor, and ensure that J3 can operate normally, and the contact surface between the big arm component and the J2 reducer is fixed; When testing robot J4, keep the motor base, forearm, J4 reducer and J4 motor, and ensure that J4 can operate normally. The contact surface between the motor base and J3 reducer is fixed. When testing robot J5, keep the forearm component, wrist body component, J5 reducer and J5 motor, and ensure that J5 can operate normally, and the contact surface between the forearm component and the J4 reducer is fixed; When testing the robot J6, retain the wrist body, J6 reducer and J6 motor, ensure that J6 can operate normally, and fix the contact surface between the wrist body and the J5 reducer.

4. The method for identifying abnormal joint vibration of a robot according to claim 2, characterized in that: S1 takes the approach of keeping other non-test robot joints in a stationary state to perform abnormal vibration test on the robot joints. When testing robot J1, the J2-J6 axis joints need to be adjusted to the specified position and remain stationary; When testing robot J2, the J1 axis joint and J3-J6 axis joints need to be adjusted to the specified position and remain stationary; When testing robot J3, the J1-J2 axis joints and J4-J6 axis joints need to be adjusted to the specified position and remain stationary; When testing robot J4, the J1-J3 axis joints and J5-J6 axis joints need to be adjusted to the specified position and remain stationary; When testing robot J5, the J1-J4 axis joints and J6 axis joints need to be adjusted to the specified position and remain stationary; When testing robot J6, the J1-J5 axis joints need to be adjusted to the specified position and remain stationary.

5. The method for identifying abnormal joint vibration of a robot according to claim 1, characterized in that: In S3, the time domain signal of the joint torque of the robot to be tested fluctuating in 20 cycles with the change of the input angular position is obtained for analysis.

6. The method for identifying abnormal joint vibration of a robot according to claim 1, characterized in that: S4 includes, S4.1, performing wavelet transform noise reduction processing on the real-time torque signal of the joint obtained in step S3; S4.

2. Analyze the data to see if the stability of the data collection cycle is guaranteed, remove the singular data, and check if the test data conforms to the normal distribution; S4.3 Perform frequency domain analysis on the normally distributed data in S4.2; S4.

4. Extract specific frequency components, kurtosis, skewness, kurtosis, main frequency, main frequency amplitude, secondary frequency and secondary frequency amplitude from the frequency spectrum, calculate the degree of deviation from the relevant data of the robot joint without vibration abnormality, and determine whether the vibration of the tested robot joint is abnormal; S4.

5. Output the test results of the joint vibration of the tested robot.

7. The method for identifying abnormal joint vibration of a robot according to claim 6, characterized in that: S4.4 includes, S4.4.

1. Perform local maximum detection on the signal, traverse the frequency domain signal data points, compare the value of each point with its adjacent points, determine whether the current point is a local maximum, determine the local maximum value as the peak value, and obtain the amplitude corresponding to the main frequency and the secondary frequency according to the peak value; S4.4.

2. Perform skewness analysis to determine whether the data have a longer right tail, a longer left tail, or a symmetrical distribution; S4.4.

3. Perform kurtosis analysis to determine the distribution characteristics of the primary and secondary frequencies and detect outliers. Calculate the kurtosis of the data based on the amplitude to measure the thickness of the tail of the distribution and the sharpness of the peak. S4.4.

4. Calculate the degree of deviation between the extracted kurtosis, main frequency, main frequency amplitude, secondary frequency, secondary frequency amplitude and other data and the characteristic component data of the robot joint without vibration abnormality, and determine whether the vibration of the tested robot joint is abnormal.

8. The method for identifying abnormal joint vibration of a robot according to claim 6, characterized in that: In S4.4.4, First determine whether there is a multiple relationship between the main frequency and the secondary frequency. If there is no multiple relationship, it is directly determined that the vibration of the joint of the tested robot is abnormal. If there is a multiple relationship, then determine whether the kurtosis and skewness of the secondary frequency are abnormal. If there is an abnormality, it is determined that the vibration of the joint of the tested robot is abnormal. If there is no abnormality, it is determined that the vibration of the joint of the tested robot is normal.

Images