Servo motor fault diagnosis method and system
Through data sensing and intelligent analysis of servo motors, the problems of traditional fault diagnosis need to be shut down and lack of intelligence are solved, real-time monitoring and accurate diagnosis of servo motor failures are realized, and production efficiency is improved.
Patent Information
- Application Number
- CN202410962399.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-18
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-07-18
AI Technical Summary
Fault diagnosis of traditional servo motors requires shutdown inspection, which affects production efficiency, and based on manual experience, lack of intelligence and automation, resulting in untimely and accurate fault detection and handling.
Through multiple sensing devices, data sensing of the servo motor is generated, operation sensing data sets are performed, feature extraction and cluster analysis are performed, fault diagnosis is performed in combination with data distribution maps, fault signals are generated and intelligent processing is performed, and servo motor failures are achieved. Accurate diagnosis and prediction.
Real-time monitoring and accurate diagnosis of servo motor faults is achieved, timely and accurate in fault detection is improved, the impact of shutdown inspection is reduced, and production efficiency is improved.
Smart Images

Figure CN118501692B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of motor analysis, and in particular to a servo motor fault diagnosis method and system. Background Art
[0002] With the development of industrial automation, the servo motor system transmission chain, as an important component, has been widely used in the fields of automobiles, wind power generation, industrial machine tools, etc. The fault diagnosis technology of the servo motor system transmission chain is crucial to avoid the economic losses and personnel safety threats caused by the gradual increase of mechanical failures. Although the servo motor fault diagnosis technology has been paid attention to, there are still some shortcomings. The fault diagnosis of traditional motors requires shutdown inspection, which affects production efficiency. In addition, it lacks intelligence and automation based on manual experience, resulting in technical problems such as untimely and inaccurate fault detection and processing. Summary of the invention
[0003] The present application provides a servo motor fault diagnosis method and system to solve the technical problems that the fault diagnosis of traditional motors requires shutdown inspection, which affects production efficiency, and lacks intelligence and automation based on manual experience, resulting in untimely and inaccurate fault detection and processing. The application achieves the technical effect of accurately diagnosing and predicting servo motor faults through real-time monitoring of motor operation data, feature extraction, intelligent learning, etc.
[0004] The present application provides a servo motor fault diagnosis method, which is applied to a servo motor fault diagnosis system, including: performing data sensing on a target servo motor through multiple sensing devices to generate an operation sensing data set; performing feature extraction on the operation sensing data set to generate multiple sensing data features, and determining multiple correlation coefficients of the operation sensing data set based on the multiple sensing data features; performing cluster analysis on the operation sensing data set according to the multiple correlation coefficients, and generating a data distribution map based on the cluster analysis results; performing fault diagnosis on the target servo motor in combination with the data distribution map to generate fault diagnosis data; generating a fault signal based on the fault diagnosis data, and performing intelligent processing on the target servo motor according to the fault signal.
[0005] In a possible implementation, feature extraction is performed on the operating sensor data set to generate multiple sensor data features, and the following processing is performed: feature analysis is performed on the operating sensor data set according to the sensing duration to generate sensor data time domain features; feature analysis is performed on the operating sensor data set by wavelet transform to generate sensor data frequency domain features; feature selection is performed on the sensor data time domain features and the sensor data frequency domain features based on information gain to generate the multiple sensor data features.
[0006] In a possible implementation, multiple correlation coefficients of the running sensor data set are determined according to the multiple sensor data features, and the following processing is performed: multiple feature sequences are determined based on the multiple sensor data features, and the multiple feature sequences include a reference sequence and a comparison sequence; each time node is calculated according to the reference sequence and the comparison sequence to generate multiple absolute differences, and each absolute difference in the multiple absolute differences corresponds to a time node; extreme value screening is performed based on the multiple absolute differences to generate extreme value differences; the sensor data resolution is set, and association calculation is performed based on the absolute difference, the extreme value difference, and the sensor data resolution to generate the multiple correlation coefficients.
[0007] In a possible implementation, cluster analysis is performed on the operation sensor data set according to the multiple correlation coefficients, and a data distribution map is generated according to the cluster analysis results, and the following processing is performed: a feature matrix is constructed based on the multiple correlation coefficients; clustering is performed on the operation sensor data set according to the feature matrix to obtain a cluster analysis result, wherein the cluster analysis result includes multiple data clusters, and each data cluster within the multiple data clusters has a cluster label; and according to the cluster labels, the data distribution map is constructed through the multiple data clusters according to the data distribution law.
[0008] In a possible implementation, the target servo motor is diagnosed for faults in combination with the data distribution graph, and the following processing is performed: the data distribution graph is traversed for trend analysis to generate a data change trend; abnormal data points are identified based on the data change trend, and analysis is performed according to the correlation coefficient based on the abnormal data points to generate a plurality of associated abnormal data; fault analysis is performed based on the plurality of associated abnormal data to generate fault diagnosis type information; anomaly calculation is performed based on the plurality of associated abnormal data to obtain an abnormal indicator data set, the abnormal indicator data set includes the amount of abnormal data, the area of the abnormal region, and the degree of abnormal deviation; according to the amount of abnormal data and the area of the abnormal region, an abnormality level is divided according to the degree of abnormal deviation to determine the fault diagnosis level information; according to the plurality of associated abnormal data, the fault tracing is performed in combination with the data distribution graph to determine the fault diagnosis source information; the fault diagnosis type information, the fault diagnosis level information, and the fault diagnosis source information are added to the fault diagnosis data.
[0009] In a possible implementation, a fault signal is generated based on the fault diagnosis data, and the following processing is performed: an abnormal trend of real-time operating data is generated based on the fault diagnosis type information, the fault diagnosis level information, the fault diagnosis source information and the data change trend; when the abnormal trend is less than a preset change threshold, based on the data change trend, the fault diagnosis source is extracted and added to the fault signal, and the fault signal is updated to a warning signal; when the abnormal trend is greater than or equal to the preset change threshold, the fault occurrence time is recorded, the fault diagnosis type information, the fault diagnosis level information, and the fault diagnosis source information are extracted and added to the fault signal, and the fault signal is updated to an alarm signal.
[0010] In a possible implementation, the target servo motor is intelligently processed according to the fault signal, and the following processing is performed: if the fault signal is a warning signal, preventive measures are formulated, and the target servo motor is troubleshooted according to the preventive measures; if the fault signal is an alarm signal, a maintenance strategy is formulated, the maintenance strategy is executed, and the target servo motor is repaired according to the fault diagnosis data.
[0011] The present application also provides a servo motor fault diagnosis system, including: a data sensing module, the data sensing module is used to perform data sensing on a target servo motor through multiple sensing devices to generate an operation sensing data set; a feature extraction module, the feature extraction module is used to perform feature extraction on the operation sensing data set to generate multiple sensing data features, and determine multiple correlation coefficients of the operation sensing data set based on the multiple sensing data features; a cluster analysis module, the cluster analysis module is used to perform cluster analysis on the operation sensing data set according to the multiple correlation coefficients, and generate a data distribution map according to the cluster analysis results; a fault diagnosis module, the fault diagnosis module is used to perform fault diagnosis on the target servo motor in combination with the data distribution map to generate fault diagnosis data; an intelligent processing module, the intelligent processing module is used to generate a fault signal based on the fault diagnosis data, and perform intelligent processing on the target servo motor according to the fault signal.
[0012] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0013] The servo motor fault diagnosis method and system provided in the present application relate to the field of motor analysis technology, and solve the technical problems that the fault diagnosis of traditional motors requires shutdown inspection, which affects production efficiency, and lacks intelligence and automation based on manual experience, resulting in untimely and inaccurate fault detection and processing. The method and system achieve the technical effect of accurately diagnosing and predicting servo motor faults through real-time monitoring of motor operation data, feature extraction, intelligent learning, etc. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.
[0015] Figure 1 A schematic diagram of a servo motor fault diagnosis method according to an embodiment of the present invention;
[0016] Figure 2 A schematic diagram of the structure of a servo motor fault diagnosis system provided in an embodiment of the present application. DETAILED DESCRIPTION
[0017] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.
[0018] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.
[0019] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.
[0020] The present application provides a servo motor fault diagnosis method, which is applied to a servo motor fault diagnosis system, such as Figure 1 As shown, the method includes:
[0021] Step A100, performing data sensing on the target servo motor through multiple sensing devices to generate an operation sensing data set;
[0022] According to the operating characteristics of the servo motor and the parameters that need to be monitored, multiple sensors with high precision, stability and reliability can be selected based on measuring the motor's speed, position, temperature, current, voltage, etc., and the sensors can be installed at appropriate locations of the servo motor to ensure that the sensors can accurately measure the required parameters. At the same time, the sampling frequency, data format, storage method, etc. of multiple sensors are verified. Furthermore, the servo motor is started to operate according to the predetermined working conditions, and the target servo motor is sensed in real time through multiple sensors. The speed sensing data, position sensing data, temperature sensing data, current sensing data, and voltage sensing data of the target servo motor can be obtained, and the operation sensing data set of the target servo motor is generated based on the speed sensing data, position sensing data, temperature sensing data, current sensing data, and voltage sensing data.
[0023] Executing step A200, performing feature extraction on the operation sensor data set to generate a plurality of sensor data features, and determining a plurality of correlation coefficients of the operation sensor data set according to the plurality of sensor data features;
[0024] In one possible implementation, step A200 further includes step A210, performing feature analysis on the running sensor data set according to the sensing duration to generate time domain features of the sensor data; executing step A220, performing feature analysis on the running sensor data set through wavelet transform to generate frequency domain features of the sensor data; executing step A230, performing feature selection on the time domain features of the sensor data and the frequency domain features of the sensor data based on information gain to generate the multiple sensor data features.
[0025] According to the sensing duration (such as seconds, minutes, hours, etc.), the continuous operation sensor data set is divided into multiple time periods (such as data per minute as a time period), and the data in each time period is statistically analyzed to extract the time domain features of the sensor data related to the operating state of the servo motor. For example, the mean, variance, standard deviation, peak value and other statistics of the data can be calculated to reflect the distribution and fluctuation of the data, and finally the features extracted in different time periods are combined to form a time domain feature set of sensor data containing multiple time periods.
[0026] Furthermore, the operating sensor data set is subjected to wavelet transform to convert the time domain signal into the frequency domain. The wavelet transform can provide localized information of time and frequency at the same time. By analyzing the energy distribution of different frequency components, the frequency domain features of the sensor data related to the operating state of the servo motor are extracted. For example, the energy or power, bandwidth, frequency center of gravity, and spectrum entropy in a specific frequency band can be calculated, and information gain is used as an indicator to evaluate the importance of features. Information gain measures the degree of change in the uncertainty (or purity) of the data set before and after the feature is used. The greater the change, the greater the contribution of the feature to classification or prediction. The time domain features and frequency domain features of the sensor data are sorted according to the size of the information gain, and the features with larger information gain are selected as multiple sensor data features for output, providing strong support for subsequent fault diagnosis and performance evaluation.
[0027] In a possible implementation, step A200 further includes step A240, determining multiple feature sequences based on the multiple sensor data features, the multiple feature sequences include a reference sequence and a comparison sequence; executing step A250, calculating each time node according to the reference sequence and the comparison sequence to generate multiple absolute differences, each absolute difference in the multiple absolute differences corresponds to a time node; executing step A260, performing extreme value screening based on the multiple absolute differences to generate extreme value differences; executing step A270, setting the sensor data resolution, performing correlation calculation based on the absolute difference, the extreme value difference, and the sensor data resolution, and generating the multiple correlation coefficients.
[0028] Based on multiple sensor data features, multiple feature sequences are selected. The multiple feature sequences may include a reference sequence (for example, data under a standard operating state) and multiple comparison sequences (for example, data at different times or under different conditions). Based on each feature, the values of the reference sequence and the comparison sequence at each time node are compared, and the absolute difference between these values is calculated to obtain a series of absolute differences corresponding to different time nodes. The maximum value, i.e., the maximum difference, and the minimum value, i.e., the minimum difference, are found among all the calculated absolute differences. At the same time, the sensor data resolution is set according to the actual application requirements or the specific circumstances of the data. The sensor data resolution can be regarded as the threshold of the error or variation range considered when calculating the correlation coefficient. The grey correlation analysis formula is as follows:
[0029] ;
[0030] in, is the correlation coefficient of the kth time node, k is a positive integer, is the value of the reference sequence at the kth time node, is the value of the ith comparison sequence at the kth time node, is the sensor data resolution, which can be a small positive number, i.e. 0.5. The sensor data resolution is used to control the influence of extreme values on the correlation coefficient. is the minimum of all absolute differences, is the maximum of all absolute differences.
[0031] Furthermore, the correlation coefficient can be calculated for each time node according to the above grey correlation analysis formula, so as to obtain a set of correlation coefficients corresponding to different time nodes, and the multiple correlation coefficients are used to evaluate the correlation between the reference sequence and the comparison sequence.
[0032] Executing step A300, performing cluster analysis on the operation sensor data set according to the multiple correlation coefficients, and generating a data distribution graph according to the cluster analysis result;
[0033] In one possible implementation, step A300 further includes step A310, constructing a feature matrix based on the multiple correlation coefficients; executing step A320, performing clustering on the running sensor data set according to the feature matrix to obtain a clustering analysis result, wherein the clustering analysis result includes multiple data clusters, and each data cluster within the multiple data clusters has a cluster label; executing step A330, constructing the data distribution graph according to the data distribution law through the multiple data clusters according to the cluster labels.
[0034] Each sensor data feature sequence is regarded as a row of the feature matrix, and the time node or feature point is regarded as a column. The corresponding positions of the feature matrix are filled with the multiple correlation coefficients obtained by the above calculation, and the correlation coefficient of each time node will become the value of the corresponding column in the feature matrix. Furthermore, based on clustering algorithms, such as K-means, hierarchical clustering, DBSCAN, etc., they are applied to the feature matrix to cluster the operating sensor data set. The operating sensor data set is automatically divided into multiple data clusters according to the clustering algorithm, and a unique cluster label is assigned to each data cluster.
[0035] Furthermore, based on each cluster, the corresponding sensor data points are extracted, and the data distribution of each cluster is plotted using visualization methods such as scatter plots or heat maps. This means that a data distribution graph is plotted for each data cluster based on the clustering results and cluster labels. The data distribution graph can be represented by a two-dimensional or higher-dimensional visualization, and in the data distribution graph, data points of different clusters can be distinguished by using different colors or marks, so as to quickly obtain the data distribution pattern of each cluster.
[0036] Execute step A400, perform fault diagnosis on the target servo motor in combination with the data distribution diagram, and generate fault diagnosis data;
[0037] In a possible implementation, step A400 further includes step A410, traversing the data distribution graph to perform trend analysis and generate a data change trend; executing step A420, identifying abnormal data points based on the data change trend, and analyzing the abnormal data points according to the correlation coefficient to generate multiple associated abnormal data;
[0038] By combining the above-mentioned visualized data distribution graph, statistical methods such as moving average, exponential smoothing, etc. or machine learning models are used to fit and predict the trend of data series, so as to obtain the data change trend. Further, according to the data distribution graph and the data change trend, one or more abnormal thresholds can be set based on statistical methods (such as Z-score, IQR rule, etc.) or based on business experience and expert judgment, and the processed data is traversed to compare each data point with the abnormal threshold. When there is any data point that exceeds or falls below the set abnormal threshold, it is marked as an abnormal data point. According to the abnormal data point, the correlation coefficient is analyzed, which means calculating the correlation coefficient between the abnormal data point and other data points. The correlation coefficient represents the degree of correlation between the data points. Based on the data characteristics, the correlation threshold is set. The correlation threshold is used to determine the data point that has sufficient correlation with the abnormal data point, so as to regard it as the associated abnormal data. All data points are traversed to find the data points whose correlation coefficient with the abnormal data point exceeds the correlation threshold. These data points are marked as associated abnormal data, so as to obtain multiple associated abnormal data.
[0039] Execute step A430, perform fault analysis according to the multiple associated abnormal data, and generate fault diagnosis type information; execute step A440, perform abnormality calculation based on the multiple associated abnormal data, and obtain an abnormal indicator data set, wherein the abnormal indicator data set includes the amount of abnormal data, the area of the abnormal region, and the abnormal deviation;
[0040] First, key features are extracted from the associated abnormal data. The extracted key features are directly related to the fault type, such as timestamps, sensor readings, location information, etc. A classifier is trained according to the key features using a machine learning or deep learning algorithm. Multiple associated abnormal data are used as inputs, and a fault classification model is applied for classification. Each associated abnormal data is assigned a fault diagnosis type information, and the classification result is verified. The verification process can verify the accuracy of the result through manual inspection or comparison with other known fault cases.
[0041] Furthermore, anomaly calculation based on the multiple associated abnormal data refers to counting the number of data points in the associated abnormal data set, which is used to characterize the frequency of abnormal events occurring in a specific time period or a specific area, and then calculating the area ratio covered by the multiple abnormal data points in the data distribution diagram, and obtaining the area of the abnormal area, which is used to obtain the distribution and diffusion range of the abnormal event. Furthermore, the degree of deviation of the associated abnormal data point relative to the normal data or trend line is calculated by calculating the difference between the abnormal data point and the mean, median or trend line of the normal data. The abnormal deviation degree is used to characterize the severity of the abnormal event and the potential impact on the system performance. Finally, the calculated abnormal data volume, abnormal area and abnormal deviation degree are integrated into a data set to form an abnormal indicator data set, and the fault diagnosis degree is divided into different levels according to the value of the quantitative indicator, such as minor, moderate and severe.
[0042] Execute step A450, perform abnormal level classification according to the abnormal deviation according to the abnormal data volume and the abnormal area, and determine fault diagnosis level information; execute step A460, perform fault tracing according to the multiple associated abnormal data and the data distribution map, and determine fault diagnosis source information;
[0043] Execute step A470 to add the fault diagnosis type information, the fault diagnosis level information, and the fault diagnosis source information to the fault diagnosis data.
[0044] The standard of abnormal level can be set based on the thresholds of abnormal data volume, abnormal region area and abnormal deviation, and then a comprehensive abnormal index is calculated by weighted sum of abnormal data volume, abnormal region area and abnormal deviation, and the comprehensive abnormal index is compared with the set abnormal level standard to determine the abnormal level of each associated abnormal data, and determine the fault diagnosis level information according to the abnormal level of the associated abnormal data. The fault diagnosis level information is a classification label used to characterize the severity of the fault, such as minor, moderate, and severe.
[0045] Furthermore, multiple associated abnormal data are analyzed through a data distribution map to obtain the correlation relationship between the multiple associated abnormal data, and then the data distribution map is used to assist in fault tracing, which means finding possible fault sources or problem areas by observing the position and distribution of abnormal data points in the data distribution map, and determining the source information of the fault diagnosis that is directly related to the occurrence of the fault based on the analysis of the associated abnormal data and the observation of the data distribution map, and finally integrating the fault diagnosis type information, fault diagnosis level information and fault diagnosis source information to form complete fault diagnosis data.
[0046] Next, step A500 is executed to generate a fault signal based on the fault diagnosis data, and intelligently process the target servo motor according to the fault signal.
[0047] In one possible implementation, step A500 further includes step A510, generating an abnormal trend of real-time operating data based on the fault diagnosis type information, the fault diagnosis level information, the fault diagnosis source information and the data change trend; executing step A520, when the abnormal trend is less than a preset change threshold, based on the data change trend, extracting the fault diagnosis source and adding it to the fault signal, and updating the fault signal to a warning signal; executing step A530, when the abnormal trend is greater than or equal to the preset change threshold, recording the time of occurrence of the fault, extracting the fault diagnosis type information, the fault diagnosis level information, and the fault diagnosis source information and adding them to the fault signal, and updating the fault signal to an alarm signal.
[0048] An abnormal trend model is constructed by combining fault diagnosis type information, fault diagnosis level information, fault diagnosis source information and data change trend. The abnormal trend model can be a dynamic scoring system. The abnormal trend model includes the weights of the above-mentioned fault diagnosis type information, fault diagnosis level information, fault diagnosis source information, and their influence on the abnormal trend of the data. The operating data is collected in real time, and its deviation from the normal state is calculated. The historical data and current diagnostic information are combined to generate a real-time abnormal trend curve.
[0049] To judge abnormal changes based on the abnormal trend of real-time operation data, means to first set a preset change threshold based on historical abnormal change data. The preset change threshold is used to distinguish abnormal situations that require early warning and alarm. When the abnormal trend score calculated in real time or the slope of the abnormal trend curve is less than the preset change threshold, it is considered that the abnormality in the real-time data is gradually appearing, and the trend analysis shows that there is a possibility of further deterioration, that is, the abnormal situation at this time is slowly developing or stabilizing. Then, based on the data change trend, the fault diagnosis source is extracted and added to an early warning signal. The early warning signal can be used to notify relevant operators to pay attention to potential problems and take preventive measures.
[0050] When the real-time calculated abnormal trend score or the slope of the abnormal trend curve reaches or exceeds the preset change threshold, it is considered that there are significant outliers in the real-time data, and multiple monitoring indicators are abnormal at the same time, indicating that the abnormal situation is rapidly deteriorating or a serious fault has occurred. At this time, the system needs to record the time of the fault occurrence, and extract the fault diagnosis type information, fault diagnosis level information, and fault diagnosis source information, and add them to an alarm signal. The alarm signal should contain sufficient information so that the operator can quickly understand the fault situation and take emergency measures.
[0051] In a possible implementation, step A500 further includes step A540, if the fault signal is a warning signal, formulating preventive measures, and performing fault troubleshooting on the target servo motor according to the preventive measures;
[0052] When the fault signal is determined to be a warning signal, the fault diagnosis source information in the warning signal is analyzed to determine the possible causes of the fault, and based on the analysis results, targeted preventive measures are formulated. The preventive measures may include adjusting operating parameters, optimizing operating procedures, strengthening equipment maintenance, etc., and the preventive measures are used to reduce the risk of faults and ensure the normal operation of the target servo motor. Further, based on the preventive measures, the target servo motor is troubleshooted, which means checking the operating status, temperature, vibration and other key indicators of the servo motor to confirm whether there is any abnormality. For example, the servo motor can be disassembled for internal inspection to find potential fault points. Based on the troubleshooting results, corresponding measures are taken to eliminate potential faults or reduce the risk of failure.
[0053] Execute step A550. If the fault signal is an alarm signal, formulate a maintenance strategy, execute the maintenance strategy, and perform fault maintenance processing on the target servo motor according to the fault diagnosis data.
[0054] When the fault signal is determined to be an alarm signal, the fault diagnosis type information, fault diagnosis level information and fault diagnosis source information in the alarm signal are analyzed, and a maintenance strategy is formulated based on the analysis results. The maintenance strategy may include determining the maintenance method, replacing parts, adjusting parameters, etc. The maintenance strategy is used to quickly restore the normal operation of the servo motor and reduce the impact of the fault on the system. Furthermore, according to the maintenance strategy, the target servo motor is repaired. Exemplarily, the servo motor can be disassembled for internal repair or replacement of faulty parts. During the maintenance process, the maintenance steps and the information of the replaced parts can also be recorded for subsequent analysis and reference, that is, the time, type, level, source and other information of the fault, as well as the maintenance strategy and processing process are recorded, and the fault record is saved in the database for subsequent query and analysis. Finally, according to the results and feedback of the fault handling, the fault diagnosis and maintenance process is continuously optimized to improve the stability and reliability of the system.
[0055] The embodiments of the present application solve the technical problems that the fault diagnosis of traditional motors requires shutdown inspection, which affects production efficiency, and lacks intelligence and automation based on manual experience, resulting in untimely and inaccurate fault detection and processing. The technical effect of accurately diagnosing and predicting servo motor faults is achieved through real-time monitoring of motor operation data, feature extraction, intelligent learning, etc.
[0056] In the above, refer to Figure 1The servo motor fault diagnosis method according to the embodiment of the present invention is described in detail. Figure 2 A servo motor fault diagnosis system according to an embodiment of the present invention is described.
[0057] The servo motor fault diagnosis system according to the embodiment of the present invention is used to solve the technical problems that the fault diagnosis of traditional motors requires shutdown inspection, which affects production efficiency, and lacks intelligence and automation based on manual experience, resulting in untimely and accurate fault detection and processing, and achieves the technical effect of accurately diagnosing and predicting servo motor faults through real-time monitoring of motor operation data, feature extraction, intelligent learning, etc. The servo motor fault diagnosis system includes: a data sensing module 10, a feature extraction module 20, a cluster analysis module 30, a fault diagnosis module 40, and an intelligent processing module 50.
[0058] A data sensing module 10, wherein the data sensing module 10 is used to perform data sensing on a target servo motor through a plurality of sensing devices to generate an operation sensing data set;
[0059] A feature extraction module 20, the feature extraction module 20 is used to extract features from the operation sensor data set, generate a plurality of sensor data features, and determine a plurality of correlation coefficients of the operation sensor data set according to the plurality of sensor data features;
[0060] A cluster analysis module 30, the cluster analysis module 30 is used to perform cluster analysis on the operation sensor data set according to the multiple correlation coefficients, and generate a data distribution map according to the cluster analysis results;
[0061] A fault diagnosis module 40, wherein the fault diagnosis module 40 is used to perform fault diagnosis on the target servo motor in combination with the data distribution diagram to generate fault diagnosis data;
[0062] The intelligent processing module 50 is used to generate a fault signal based on the fault diagnosis data, and perform intelligent processing on the target servo motor according to the fault signal.
[0063] The specific configuration of the feature extraction module 20 will be described in detail below. As described above, the feature extraction of the operation sensor data set is performed to generate multiple sensor data features. The feature extraction module 20 may further include: a first feature analysis unit for performing feature analysis on the operation sensor data set according to the sensing duration to generate sensor data time domain features; a second feature analysis unit for performing feature analysis on the operation sensor data set through wavelet transform to generate sensor data frequency domain features; a feature selection unit for performing feature selection on the sensor data time domain features and the sensor data frequency domain features based on information gain to generate the multiple sensor data features.
[0064] The specific configuration of the feature extraction module 20 will be described in detail below. As described above, the multiple correlation coefficients of the running sensor data set are determined according to the multiple sensor data features, and the feature extraction module 20 may further include: a sequence determination unit is used to determine multiple feature sequences based on the multiple sensor data features, and the multiple feature sequences include reference sequences and comparison sequences; a first calculation unit is used to calculate each time node according to the reference sequence and the comparison sequence to generate multiple absolute differences, and each absolute difference in the multiple absolute differences corresponds to a time node; a screening unit is used to perform extreme value screening based on the multiple absolute differences to generate extreme value differences; a second calculation unit is used to set the sensor data resolution, perform association calculation based on the absolute difference, the extreme value difference, and the sensor data resolution, and generate the multiple correlation coefficients.
[0065] The specific configuration of the cluster analysis module 30 will be described in detail below. As described above, the operation sensor data set is clustered according to the multiple correlation coefficients, and a data distribution diagram is generated according to the cluster analysis results. The cluster analysis module 30 may further include: a matrix construction unit for constructing a feature matrix based on the multiple correlation coefficients; a clustering unit for clustering the operation sensor data set according to the feature matrix to obtain a cluster analysis result, wherein the cluster analysis result includes multiple data clusters, and each data cluster in the multiple data clusters has a cluster label; and a first construction unit for constructing the data distribution diagram according to the data distribution law through the multiple data clusters according to the cluster labels.
[0066] The specific configuration of the fault diagnosis module 40 will be described in detail below. As described above, the target servo motor is fault diagnosed in combination with the data distribution map to generate fault diagnosis data. The fault diagnosis module 40 may further include: a trend analysis unit for traversing the data distribution map to perform trend analysis and generate data change trends; an abnormal identification unit for identifying abnormal data points based on the data change trend, and analyzing the abnormal data points according to the correlation coefficient to generate multiple associated abnormal data; a fault analysis unit for performing fault analysis based on the multiple associated abnormal data to generate fault diagnosis type information; a third calculation unit for performing abnormal calculation based on the multiple associated abnormal data to obtain an abnormal index data set, the abnormal index data set including the amount of abnormal data, the area of the abnormal region, and the degree of abnormal deviation; a level division unit for performing abnormal level division according to the abnormal deviation according to the amount of abnormal data, the area of the abnormal region, and the degree of abnormal deviation to determine the fault diagnosis level information; a fault tracing unit for performing fault tracing according to the multiple associated abnormal data in combination with the data distribution map to determine the fault diagnosis source information; and a first adding unit for adding the fault diagnosis type information, the fault diagnosis level information, and the fault diagnosis source information to the fault diagnosis data.
[0067] The specific configuration of the intelligent processing module 50 will be described in detail below. As described above, based on the fault diagnosis data, a fault signal is generated, and the intelligent processing module 50 may further include: an abnormal trend unit for generating an abnormal trend of real-time operation data based on the fault diagnosis type information, the fault diagnosis level information, the fault diagnosis source information and the data change trend; a first judgment unit for extracting the fault diagnosis source and adding it to the fault signal based on the data change trend when the abnormal trend is less than a preset change threshold, and updating the fault signal to a warning signal; a second judgment unit for recording the fault occurrence time, extracting the fault diagnosis type information, the fault diagnosis level information, and the fault diagnosis source information and adding them to the fault signal when the abnormal trend is greater than or equal to the preset change threshold, and updating the fault signal to an alarm signal.
[0068] The specific configuration of the intelligent processing module 50 will be described in detail below. As described above, the target servo motor is intelligently processed according to the fault signal, and the intelligent processing module 50 may further include: a third judgment unit for formulating preventive measures if the fault signal is a warning signal, and performing fault troubleshooting on the target servo motor according to the preventive measures; a fourth judgment unit for formulating a maintenance strategy if the fault signal is an alarm signal, executing the maintenance strategy, and performing fault maintenance on the target servo motor according to the fault diagnosis data.
[0069] The servo motor fault diagnosis system provided in the embodiment of the present invention can execute the servo motor fault diagnosis method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.
[0070] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.
[0071] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.
Claims
1. A servo motor fault diagnosis method, characterized in that: The method comprises: Performing data sensing on the target servo motor through multiple sensing devices to generate an operation sensing data set; Performing feature extraction on the operation sensor data set to generate a plurality of sensor data features, and determining a plurality of correlation coefficients of the operation sensor data set according to the plurality of sensor data features; Performing cluster analysis on the operation sensor data set according to the multiple correlation coefficients, and generating a data distribution graph according to the cluster analysis result; Performing fault diagnosis on the target servo motor in combination with the data distribution diagram to generate fault diagnosis data; generating a fault signal based on the fault diagnosis data, and performing intelligent processing on the target servo motor according to the fault signal; Performing feature analysis on the running sensor data set according to the sensing duration to generate a time domain feature of the sensor data; Performing feature analysis on the operating sensor data set by wavelet transform to generate frequency domain features of the sensor data; Performing feature selection on the time domain features of the sensor data and the frequency domain features of the sensor data based on information gain to generate the plurality of sensor data features; Determine a plurality of feature sequences based on the plurality of sensor data features, wherein the plurality of feature sequences include a reference sequence and a comparison sequence; Calculating each time node according to the reference sequence and the comparison sequence to generate multiple absolute differences, each absolute difference in the multiple absolute differences corresponds to a time node; Perform extreme value screening based on the multiple absolute differences to generate extreme value differences; Setting the sensor data resolution, performing correlation calculation based on the absolute difference, the extreme value difference, and the sensor data resolution, and generating the multiple correlation coefficients; constructing a feature matrix based on the plurality of correlation coefficients; Performing clustering on the running sensor data set according to the feature matrix to obtain a cluster analysis result, wherein the cluster analysis result includes a plurality of data clusters, and each data cluster in the plurality of data clusters has a cluster label; According to the cluster labels, the data distribution graph is constructed using the multiple data clusters according to data distribution rules.
2. The servo motor fault diagnosis method according to claim 1, characterized in that: Performing fault diagnosis on the target servo motor in combination with the data distribution diagram to generate fault diagnosis data, the method comprising: Traversing the data distribution graph to perform trend analysis and generate data change trends; Identify abnormal data points based on the data change trend, analyze the abnormal data points according to the correlation coefficient, and generate multiple associated abnormal data; Performing fault analysis based on the plurality of associated abnormal data to generate fault diagnosis type information; Perform anomaly calculation based on the multiple associated abnormal data to obtain an abnormal index data set, wherein the abnormal index data set includes an abnormal data volume, an abnormal region area, and an abnormal deviation degree; According to the amount of abnormal data and the area of the abnormal region, the abnormal level is divided according to the abnormal deviation, and the fault diagnosis level information is determined; According to the plurality of associated abnormal data, combined with the data distribution diagram, fault tracing is performed to determine fault diagnosis source information; The fault diagnosis type information, the fault diagnosis level information, and the fault diagnosis source information are added to the fault diagnosis data.
3. The servo motor fault diagnosis method according to claim 2, characterized in that: Generating a fault signal based on the fault diagnosis data, the method includes: Generate an abnormal trend of real-time operation data based on the fault diagnosis type information, the fault diagnosis level information, the fault diagnosis source information and the data change trend; When the abnormal trend is less than a preset change threshold, based on the data change trend, the fault diagnosis source is extracted and added to the fault signal, and the fault signal is updated to a warning signal; When the abnormal trend is greater than or equal to the preset change threshold, the fault occurrence time is recorded, the fault diagnosis type information, the fault diagnosis level information, and the fault diagnosis source information are extracted and added to the fault signal, and the fault signal is updated to an alarm signal.
4. The servo motor fault diagnosis method according to claim 3, characterized in that: Intelligently processing the target servo motor according to the fault signal, the method comprising: If the fault signal is a warning signal, then formulate preventive measures, and perform fault troubleshooting on the target servo motor according to the preventive measures; If the fault signal is an alarm signal, a maintenance strategy is formulated, the maintenance strategy is executed, and the target servo motor is subjected to fault maintenance processing according to the fault diagnosis data.
5. Servo motor fault diagnosis system, characterized in that: The system is used to implement the servo motor fault diagnosis method according to any one of claims 1 to 4, and the system comprises: A data sensing module, wherein the data sensing module is used to perform data sensing on a target servo motor through a plurality of sensing devices to generate an operation sensing data set; A feature extraction module, the feature extraction module is used to extract features from the operation sensor data set, generate a plurality of sensor data features, and determine a plurality of correlation coefficients of the operation sensor data set according to the plurality of sensor data features; A cluster analysis module, the cluster analysis module is used to perform cluster analysis on the operation sensor data set according to the multiple correlation coefficients, and generate a data distribution map according to the cluster analysis results; A fault diagnosis module, the fault diagnosis module is used to perform fault diagnosis on the target servo motor in combination with the data distribution diagram to generate fault diagnosis data; An intelligent processing module is used to generate a fault signal based on the fault diagnosis data and perform intelligent processing on the target servo motor according to the fault signal.
Citation Information
Cited By
Servo motor fault diagnosis method and system based on data model analysis
CN119147963B
Rapid door protection braking device and method
CN121066480A
