A motor operation status monitoring method based on multi-sensor fusion

Abstract

The present invention discloses a motor operation status monitoring method based on multi-sensor fusion, and relates to the field of motor monitoring technology. The method comprises: optimizing the warning threshold of the servo motor according to the working operation data and the predetermined handling plan on the cloud analysis platform, and determining the dynamic warning threshold; at the edge judgment node, based on the dynamic warning threshold, performing the servo motor operation status monitoring in the first time window. The technical problem that it is difficult to dynamically adjust the indicator warning threshold of the servo motor according to the actual working scene of the handling robot in the prior art, resulting in a high false alarm or missed alarm rate of the servo motor abnormal state warning, is solved. By adaptively adjusting the indicator warning threshold of the servo motor according to the actual working scene of the handling robot, the accuracy and reliability of the warning threshold setting can be improved, thereby reducing the probability of false alarm or missed alarm of the motor abnormal warning, and achieving the technical effect of improving the accuracy of the motor abnormal warning.

Description

Technical Field

[0001] The present invention relates to the technical field of motor monitoring, and in particular to a motor operating status monitoring method based on multi-sensor fusion. Background Art

[0002] In the field of modern industrial automation, handling robots are indispensable key equipment, and their stable operation is crucial to ensure the continuity and efficiency of the production line. However, with the increasingly complex and changeable production environment, the operating conditions faced by the servo motors of handling robots when performing predetermined handling tasks are becoming more and more diverse, including different load conditions, working environment temperature fluctuations, vibration interference, and real-time changes in speed and position control requirements. Traditional motor operation status monitoring methods often rely on preset fixed warning thresholds, which are usually set based on theoretical calculations or empirical data and are difficult to accurately reflect the dynamic characteristics of the motor in actual work. Therefore, under complex and changeable operating conditions, these fixed thresholds may lead to false alarms (i.e., erroneous alarms when the motor is operating normally) or missed alarms (i.e., failure to detect in time when the motor fails or is abnormal), thereby affecting production efficiency and equipment safety. Summary of the invention

[0003] The present application provides a motor operation status monitoring method based on multi-sensor fusion, which solves the technical problem in the prior art that it is difficult to dynamically adjust the indicator warning threshold of the servo motor according to the actual working scenario of the handling robot, resulting in a high false alarm or missed alarm rate of the servo motor abnormal status warning.

[0004] In view of the above problems, the present application provides a motor operation status monitoring method based on multi-sensor fusion.

[0005] The present application provides a motor operation status monitoring method based on multi-sensor fusion, the method comprising:

[0006] Read the working operation data and the predetermined handling plan of the handling robot within the first time window; on the cloud analysis platform, optimize the warning threshold analysis of the servo motor of the handling robot according to the working operation data and the predetermined handling plan, and determine the dynamic warning threshold, wherein the dynamic warning threshold includes a first current threshold, a first torque threshold, a first temperature threshold set, a first vibration threshold set, a first speed threshold and a first position threshold; transmit the first current threshold, the first torque threshold, the first temperature threshold set, the first vibration threshold set, the first speed threshold and the first position threshold to the edge judgment node of the sensor monitoring array; at the edge judgment node, perform the servo motor operation status monitoring within the first time window based on the first current threshold, the first torque threshold, the first temperature threshold set, the first vibration threshold set, the first speed threshold and the first position threshold.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] First, read the working operation data and the predetermined handling plan of the handling robot within the first time window. Then, on the cloud analysis platform, optimize the warning threshold of the servo motor of the handling robot according to the working operation data and the predetermined handling plan, and determine the dynamic warning threshold, wherein the dynamic warning threshold includes a first current threshold, a first torque threshold, a first temperature threshold set, a first vibration threshold set, a first speed threshold and a first position threshold. Then, the first current threshold, the first torque threshold, the first temperature threshold set, the first vibration threshold set, the first speed threshold and the first position threshold are transmitted to the edge judgment node of the sensor monitoring array. Finally, at the edge judgment node, based on the first current threshold, the first torque threshold, the first temperature threshold set, the first vibration threshold set, the first speed threshold and the first position threshold, the servo motor operation status monitoring within the first time window is performed. The invention solves the technical problem in the prior art that it is difficult to dynamically adjust the indicator warning threshold of the servo motor according to the actual working scenario of the handling robot, resulting in a high false alarm or missed alarm rate of the servo motor abnormal state warning. By adaptively adjusting the indicator warning threshold of the servo motor according to the actual working scenario of the handling robot, the accuracy and reliability of the warning threshold setting can be improved, thereby reducing the probability of false alarm or missed alarm of the motor abnormality warning, and achieving the technical effect of improving the accuracy of the motor abnormality warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A schematic diagram of a flow chart of a motor operating status monitoring method based on multi-sensor fusion provided in an embodiment of the present application;

[0011] Figure 2 A schematic diagram of a flow chart for determining a dynamic warning threshold in a motor operation status monitoring method based on multi-sensor fusion provided in an embodiment of the present application. DETAILED DESCRIPTION

[0012] The present application provides a motor operation status monitoring method based on multi-sensor fusion, which solves the technical problem in the prior art that it is difficult to dynamically adjust the indicator warning threshold of the servo motor according to the actual working scenario of the handling robot, resulting in a high false alarm or missed alarm rate of the servo motor abnormal status warning.

[0013] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0014] It should be noted that the terms "including" and "having" 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 explicitly listed, but may include other steps or modules that are not explicitly listed or are inherent to these processes, methods, products or devices.

[0015] Examples, such as Figure 1 As shown, the embodiment of the present application provides a motor operation status monitoring method based on multi-sensor fusion, wherein the method includes:

[0016] Read the working operation data and the predetermined handling plan of the handling robot within the first time window.

[0017] Furthermore, the work operation data at least includes real-time load, continuous working time and operating environment factors, and the predetermined transportation plan at least includes predetermined cargo characteristics, predetermined transportation path and predetermined task requirements.

[0018] In the first time window, the working operation data and the scheduled handling plan of the handling robot are read. Among them, the working operation data includes the real-time load (the weight of the goods carried by the handling robot at the current moment), the continuous working time (the continuous working time of the handling robot) and the operating environment factors (ambient temperature, humidity, air pressure, etc.); the scheduled handling plan refers to the handling task that the handling robot needs to complete, and the scheduled handling plan includes the scheduled cargo characteristics (the physical properties of the goods to be handled, such as weight, volume, shape, material, and whether it is fragile, perishable, and other special requirements), the scheduled handling path (the specific path of the handling robot from the starting point to the end point) and the scheduled task requirements. The scheduled task requirements include the execution time window and the continuous task frequency. The execution time window refers to the specified time for the handling robot to perform the scheduled task, and the continuous task frequency refers to the frequency at which the handling robot continuously performs the handling task.

[0019] On the cloud analysis platform, the warning threshold optimization analysis of the servo motor of the handling robot is performed according to the working operation data and the predetermined handling plan, and the dynamic warning threshold is determined, wherein the dynamic warning threshold includes a first current threshold, a first torque threshold, a first temperature threshold set, a first vibration threshold set, a first speed threshold and a first position threshold.

[0020] The collected working operation data and the scheduled handling plan are transmitted to the cloud analysis platform. The cloud analysis platform will dynamically optimize and analyze the warning threshold of the servo motor according to various factors such as the current workload, ambient temperature, vibration interference, etc., to ensure that the warning threshold can adapt to the changes in the motor's operating status in real time, thereby improving the accuracy and timeliness of the warning. After analysis, the cloud analysis platform will determine a set of dynamic warning thresholds, including the first current threshold, the first torque threshold, the first temperature threshold set (which may include thresholds of multiple temperature sensors), the first vibration threshold set (which may also include thresholds of multiple vibration sensors), the first speed threshold (the speed of the servo motor) and the first position threshold (the position of the servo motor). These thresholds will be dynamically adjusted according to the actual operation of the motor and the requirements of the scheduled handling plan to ensure the accuracy and timeliness of the warning.

[0021] Furthermore, if Figure 2 As shown, the warning threshold optimization analysis of the servo motor of the handling robot is performed according to the working operation data and the predetermined handling plan to determine the dynamic warning threshold, including:

[0022] A multidimensional optimization operator is configured, wherein the multidimensional optimization operator includes at least a support vector machine, a BP neural network and a random decision forest; a first optimization analysis unit, a second optimization analysis unit and a third optimization analysis unit are constructed based on the support vector machine, the BP neural network and the random decision forest respectively; according to a predetermined fusion strategy, the first optimization analysis unit, the second optimization analysis unit and the third optimization analysis unit are integrated to construct a threshold optimization analyzer; real-time load, continuous working time, operating environment factors, predetermined cargo characteristics, predetermined transportation routes and predetermined task requirements are input into the threshold optimization analyzer, and the dynamic warning threshold is output.

[0023] Specifically, different machine learning algorithms are used to configure multidimensional optimization operators, which include support vector machines, BP neural networks, and random decision forests. Based on support vector machines, BP neural networks, and random decision forests, a first optimization analysis unit, a second optimization analysis unit, and a third optimization analysis unit are constructed respectively. The first optimization analysis unit is used to analyze the relationship between key features in the work operation data (such as real-time load, continuous working time, and operating environment factors) and the performance of the servo motor, and preliminarily determine the range of the warning threshold. The second optimization analysis unit can learn complex nonlinear relationships and consider the impact of more factors (such as the characteristics of the scheduled goods, the scheduled handling path) on the warning threshold, and further refine the determination of the warning threshold. The third optimization analysis unit is used to verify and optimize the outputs of the first two units. As a result, the third optimization analysis unit improves the stability and accuracy of the analysis by integrating the prediction results of multiple decision trees, and is able to process more input variables (including scheduled task requirements, etc.); according to the predetermined fusion strategy, the first optimization analysis unit, the second optimization analysis unit and the third optimization analysis unit are integrated to construct a threshold optimization analyzer; input variables such as real-time load, continuous working time, operating environment factors, scheduled cargo characteristics, scheduled handling path and scheduled task requirements are input into the threshold optimization analyzer, and the threshold optimization analyzer will use the integration advantages of multi-dimensional optimization operators to conduct a comprehensive analysis of these input data and output dynamic warning thresholds, including a first current threshold, a first torque threshold, a first temperature threshold set, a first vibration threshold set, a first speed threshold and a first position threshold.

[0024] Furthermore, based on the support vector machine, the BP neural network and the random decision forest, a first optimization analysis unit, a second optimization analysis unit and a third optimization analysis unit are respectively constructed, including:

[0025] Configure multi-dimensional monitoring indicators, wherein the multi-dimensional monitoring indicators include working current, output torque, motor temperature, component vibration, motor speed and motor position, the motor temperature includes at least coil temperature, bearing temperature and housing temperature, and the component vibration includes at least bearing vibration, motor shaft vibration and housing vibration; use the equipment attribute information of the handling robot and the motor attribute information of the servo motor as retrieval constraints, use motor state monitoring and the multi-dimensional monitoring indicators as retrieval guides, perform related information retrieval based on industrial big data, and obtain a sample data set; divide the sample data set into nine equal parts, and select nine times with replacement to construct a first training set, and select three times in sequence to obtain three training sets; supervise the training of the first optimization analysis unit, the second optimization analysis unit and the third optimization analysis unit through the three training sets, and output the first optimization analysis unit, the second optimization analysis unit and the third optimization analysis unit that meet the convergence conditions, wherein the first optimization analysis unit, the second optimization analysis unit and the third optimization analysis unit are respectively marked with a first convergence degree, a second convergence degree and a third convergence degree.

[0026] The multi-dimensional monitoring indicators comprehensively cover the key parameters of servo motor operation, including working current, output torque, motor temperature (divided into coil temperature, bearing temperature and housing temperature), component vibration (bearing vibration, motor shaft vibration and housing vibration), motor speed and motor position. These indicators can reflect the motor's operating status, health and potential faults. The equipment attribute information of the handling robot and the motor attribute information of the servo motor are used as retrieval constraints, combined with the motor status monitoring requirements and multi-dimensional monitoring indicators as retrieval guides, to retrieve related information from the industrial big data platform; the raw data obtained through data retrieval is cleaned, preprocessed and other steps to form a sample data set that can be used for model training; the sample data set is divided into nine equal parts, and nine parts are randomly selected with replacement, each time one part is selected as the test set (or verification set), and the remaining eight parts are used as training sets; in order to train the three optimization analysis units, three parts (i.e., one-third of the data) are randomly selected from the above nine parts of data each time as the training set, and a total of three selections are performed to obtain three different training sets; the above three training sets are used for classification The support vector machine, BP neural network and random decision forest are supervised and trained separately; during the training process, by adjusting the parameters of the model (such as the kernel function of SVM, the number of layers and nodes of the BP neural network, the number and depth of trees of the random decision forest, etc.), the model can accurately identify and predict the state of the servo motor; the training process will continue until the preset convergence conditions are met, including the training error is less than a certain threshold, the performance on the verification set is no longer significantly improved, or the preset training rounds are reached; after the convergence conditions are met, the trained first optimization analysis unit, the second optimization analysis unit and the third optimization analysis unit are output, and their convergence degrees (first convergence degree, second convergence degree, third convergence degree) are respectively marked to reflect the training effect of the model.

[0027] Further, obtain a sample data set, including:

[0028] A retrieval data set is obtained, and first retrieval data is randomly selected, wherein the first retrieval data includes first working operation data, a first handling plan and a first warning threshold; taking the device attribute information and the motor attribute information as device constraints, and taking the first working operation data, the first handling plan and the first warning threshold as similarity comparison constraints, historical motor monitoring event retrieval is performed to obtain a first monitoring event set; according to the first monitoring event set, a warning misjudgment and warning missed judgment analysis is performed to determine a first misjudgment rate and a first missed judgment rate; if the first misjudgment rate is less than a predetermined misjudgment threshold and the first missed judgment rate is less than a predetermined missed judgment threshold, the first retrieval data is added to the sample data set.

[0029] A collection containing multiple types of data is retrieved from the industrial big data platform, which is the retrieval data set; a piece of data is randomly selected from the retrieval data set as the first retrieval data, and the first retrieval data includes the first working operation data (such as real-time load, continuous working time, operating environment factors, etc.), the first handling plan (such as predetermined cargo characteristics, predetermined handling path, predetermined task requirements, etc.) and the first warning threshold (such as current threshold, torque threshold, temperature threshold, etc.); the equipment attribute information of the handling robot and the motor attribute information of the servo motor are used as equipment constraints, and the first working operation data, the first handling plan and the first warning threshold are used as similarity comparison constraints. Search is performed in the historical motor monitoring event database to find historical data that are similar or related to the current first retrieval data in multiple dimensions. Monitoring events, these historical monitoring events constitute the first monitoring event set; each monitoring event in the first monitoring event set is analyzed for warning false positives and warning missed positives, a warning false positive refers to an erroneous triggering of a warning (i.e., there is actually no fault but the system issues a warning), and a warning missed positive refers to a fault that fails to issue a timely warning (i.e., there is actually a fault but the system does not issue a warning), by analyzing these events, the first false positive rate and the first missed positive rate can be calculated; the calculated first false positive rate and the first missed positive rate are compared with the predetermined false positive threshold and the missed positive threshold, if the first false positive rate is less than the predetermined false positive threshold and the first missed positive rate is less than the predetermined missed positive threshold, it means that the scene represented by the first retrieval data has high data quality and warning accuracy, so this data can be added to the sample data set.

[0030] Furthermore, setting a predetermined fusion strategy includes:

[0031] The first credibility, the second credibility and the third credibility are configured according to the first convergence degree, the second convergence degree and the third convergence degree; and the predetermined fusion strategy is set based on the first credibility, the second credibility and the third credibility.

[0032] Specifically, the first credibility, the second credibility and the third credibility are configured according to the first convergence, the second convergence and the third convergence of the first optimization analysis unit, the second optimization analysis unit and the third optimization analysis unit respectively; according to the credibility weight of each unit, the output results of each unit are weighted averaged to obtain a predetermined fusion strategy. A predetermined fusion strategy based on the convergence credibility is set to effectively integrate the output results of different optimization analysis units and improve the accuracy and reliability of servo motor state monitoring and prediction.

[0033] The first current threshold, the first torque threshold, the first temperature threshold set, the first vibration threshold set, the first rotation speed threshold and the first position threshold are transmitted to an edge judgment node of a sensor monitoring array.

[0034] The first current threshold, the first torque threshold, the first temperature threshold set, the first vibration threshold set, the first speed threshold and the first position threshold are transmitted from the cloud analysis platform to the edge judgment node of the sensor monitoring array. After receiving the threshold parameters, the edge judgment node will use the pre-set threshold parameters to compare and judge these data.

[0035] Furthermore, the servo motor operating status monitoring within the first time window is performed, and before that, the method further includes:

[0036] A sensor monitoring array is built based on the multi-dimensional monitoring indicators; the servo motor is monitored at a fixed point using the sensor monitoring array, and the monitoring data set is transmitted to the edge judgment node, wherein the monitoring data set includes real-time working current, real-time output torque, real-time coil temperature, real-time bearing temperature, real-time casing temperature, real-time bearing vibration data, real-time motor shaft vibration data, real-time casing vibration data, real-time motor speed and real-time motor position.

[0037] Before performing the running status monitoring of the servo motor within the first time window, it is necessary to build a complete sensor monitoring array and use this array to perform fixed-point monitoring of the servo motor to collect comprehensive real-time data. Specifically, according to the multi-dimensional monitoring indicators, select appropriate sensors, for example, use current sensors to monitor real-time working current, torque sensors to monitor real-time output torque, temperature sensors to monitor the temperature of different parts, vibration sensors to monitor the vibration of different positions, speed sensors and position sensors to monitor the speed and position of the motor respectively, and build the selected sensors into a sensor monitoring array according to the predetermined layout and connection method to ensure full coverage of the servo motor, and then collect sufficient monitoring data; deploy each sensor in the sensor monitoring array to the corresponding position of the servo motor, these positions should be the key points that can accurately reflect the motor status, such as motor shaft, bearing, coil, etc.; start the sensor and start collecting real-time data of the servo motor, including real-time working current, real-time output torque, real-time coil temperature, real-time bearing temperature, real-time shell temperature, real-time bearing vibration data, real-time motor shaft vibration data, real-time shell vibration data, real-time motor speed and real-time motor position, etc.; the edge judgment node receives the monitoring data set from the sensor monitoring array.

[0038] At the edge judgment node, the servo motor operating status monitoring within the first time window is performed based on the first current threshold, the first torque threshold, the first temperature threshold set, the first vibration threshold set, the first speed threshold and the first position threshold.

[0039] The edge judgment node matches and compares the received real-time monitoring data with the preset thresholds (the first current threshold, the first torque threshold, the sub-thresholds in the first temperature threshold set, the sub-thresholds in the first vibration threshold set, the first speed threshold and the first position threshold); based on the comparison results, the operating status of the servo motor in the current time window is judged. If all monitoring indicators remain within their respective threshold ranges, the motor is considered to be in normal operating condition; if any one or more indicators exceed their corresponding thresholds, it may indicate that the motor is abnormal or faulty.

[0040] Further, performing the servo motor operating status monitoring within the first time window includes:

[0041] Based on the first current threshold, the first torque threshold, the first temperature threshold set, the first vibration threshold set, the first speed threshold and the first position threshold, early warning judgments are respectively performed on the real-time working current, the real-time output torque, the real-time coil temperature, the real-time bearing temperature, the real-time housing temperature, the real-time bearing vibration data, the real-time motor shaft vibration data, the real-time housing vibration data, the real-time motor speed and the real-time motor position, and indicators that do not meet the threshold are set as early warning indicators;

[0042] If the number of the warning indicators is not 0, a motor warning signal is generated based on the warning indicators and a warning is issued.

[0043] When performing operating status monitoring of the servo motor within the first time window, the edge judgment node will perform detailed early warning judgments on various real-time monitoring data based on preset thresholds (first current threshold, first torque threshold, first temperature threshold set, first vibration threshold set, first speed threshold and first position threshold). Specifically, the edge judgment node will compare each monitoring data with the corresponding threshold; if the real-time working current exceeds the first current threshold, the indicator will be set as a warning indicator; if the real-time output torque exceeds the first torque threshold, it will also be set as a warning indicator; for temperature data, the edge judgment node will check whether the real-time coil temperature, real-time bearing temperature and real-time shell temperature exceed the corresponding sub-threshold in the first temperature threshold set, and if so, they will be set as warning indicators; for vibration data, the edge judgment node will check whether the real-time bearing vibration data, real-time motor shaft vibration data and real-time shell vibration data exceed the corresponding sub-threshold in the first vibration threshold set, and set warning indicators accordingly; if the real-time motor speed exceeds the first speed threshold, or the real-time motor position does not meet the requirements of the first position threshold, they will also be set as warning indicators respectively; after completing the warning judgment of all monitoring data, the edge judgment node will count the number of warning indicators. If the number of warning indicators is not 0, that is, there is at least one monitoring data that exceeds its corresponding threshold, then the edge judgment node will generate a motor warning signal based on these warning indicators; once the motor warning signal is generated, the warning operation will be immediately executed, such as sending a warning message.

[0044] In summary, the embodiments of the present application have at least the following technical effects:

[0045] First, read the working operation data and the predetermined handling plan of the handling robot within the first time window. Then, on the cloud analysis platform, optimize the warning threshold of the servo motor of the handling robot according to the working operation data and the predetermined handling plan, and determine the dynamic warning threshold, wherein the dynamic warning threshold includes a first current threshold, a first torque threshold, a first temperature threshold set, a first vibration threshold set, a first speed threshold and a first position threshold. Then, the first current threshold, the first torque threshold, the first temperature threshold set, the first vibration threshold set, the first speed threshold and the first position threshold are transmitted to the edge judgment node of the sensor monitoring array. Finally, at the edge judgment node, based on the first current threshold, the first torque threshold, the first temperature threshold set, the first vibration threshold set, the first speed threshold and the first position threshold, the servo motor operation status monitoring within the first time window is performed. The invention solves the technical problem in the prior art that it is difficult to dynamically adjust the indicator warning threshold of the servo motor according to the actual working scenario of the handling robot, resulting in a high false alarm or missed alarm rate of the servo motor abnormal state warning. By adaptively adjusting the indicator warning threshold of the servo motor according to the actual working scenario of the handling robot, the accuracy and reliability of the warning threshold setting can be improved, thereby reducing the probability of false alarm or missed alarm of the motor abnormality warning, and achieving the technical effect of improving the accuracy of the motor abnormality warning.

[0046] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0047] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0048] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A motor operation status monitoring method based on multi-sensor fusion, characterized in that: The method comprises: Read the working operation data and the predetermined handling plan of the handling robot within the first time window; On the cloud analysis platform, an early warning threshold optimization analysis of the servo motor of the handling robot is performed according to the work operation data and the predetermined handling plan to determine a dynamic early warning threshold, wherein the dynamic early warning threshold includes a first current threshold, a first torque threshold, a first temperature threshold set, a first vibration threshold set, a first speed threshold and a first position threshold; Transmitting the first current threshold, the first torque threshold, the first temperature threshold set, the first vibration threshold set, the first speed threshold, and the first position threshold to an edge judgment node of a sensor monitoring array; At the edge judgment node, based on the first current threshold, the first torque threshold, the first temperature threshold set, the first vibration threshold set, the first speed threshold and the first position threshold, the servo motor operating state monitoring within the first time window is performed; Among them, performing an optimization analysis of the warning threshold of the servo motor of the handling robot according to the working operation data and the predetermined handling plan to determine the dynamic warning threshold includes: Configure a multi-dimensional optimization operator, wherein the multi-dimensional optimization operator includes at least a support vector machine, a BP neural network, and a random decision forest; Based on the support vector machine, BP neural network and random decision forest, a first optimization analysis unit, a second optimization analysis unit and a third optimization analysis unit are respectively constructed; According to a predetermined fusion strategy, the first optimization analysis unit, the second optimization analysis unit and the third optimization analysis unit are integrated to construct a threshold optimization analyzer; The real-time load, continuous working time, operating environment factors, planned cargo characteristics, planned transportation path and planned task requirements are input into the threshold optimization analyzer, and the dynamic warning threshold is output.

2. The motor operation status monitoring method based on multi-sensor fusion according to claim 1 is characterized in that: The work operation data at least includes real-time load, continuous working time and operating environment factors, and the predetermined transportation plan at least includes predetermined cargo characteristics, predetermined transportation path and predetermined task requirements.

3. The motor operation status monitoring method based on multi-sensor fusion according to claim 1 is characterized in that: Based on the support vector machine, the BP neural network and the random decision forest, a first optimization analysis unit, a second optimization analysis unit and a third optimization analysis unit are respectively constructed, including: Configure multi-dimensional monitoring indicators, wherein the multi-dimensional monitoring indicators include working current, output torque, motor temperature, component vibration, motor speed and motor position, the motor temperature includes at least coil temperature, bearing temperature and housing temperature, and the component vibration includes at least bearing vibration, motor shaft vibration and housing vibration; Taking the equipment attribute information of the handling robot and the motor attribute information of the servo motor as search constraints, taking the motor state monitoring and the multi-dimensional monitoring indicators as search guides, performing related information search based on industrial big data, and obtaining a sample data set; The sample data set is equally divided into nine parts, and the parts are selected nine times with replacement to construct a first training set, and the parts are selected three times in sequence to obtain three training sets; The first optimization analysis unit, the second optimization analysis unit and the third optimization analysis unit are supervisedly trained respectively through the three training sets, and the first optimization analysis unit, the second optimization analysis unit and the third optimization analysis unit that meet the convergence conditions are output, wherein the first optimization analysis unit, the second optimization analysis unit and the third optimization analysis unit are respectively marked with a first convergence degree, a second convergence degree and a third convergence degree.

4. The motor operation status monitoring method based on multi-sensor fusion according to claim 3 is characterized in that: Get a sample dataset, including: Acquire a search data set, and randomly select first search data, wherein the first search data includes first work operation data, a first handling plan, and a first warning threshold; Taking the device attribute information and the motor attribute information as device constraints, taking the first working operation data, the first handling plan and the first warning threshold as similarity comparison constraints, performing historical motor monitoring event retrieval to obtain a first monitoring event set; Performing early warning misjudgment and early warning missed judgment analysis according to the first monitoring event set to determine a first misjudgment rate and a first missed judgment rate; If the first false positive rate is less than a predetermined false positive threshold and the first missed positive rate is less than a predetermined missed positive threshold, the first search data is added to the sample data set.

5. The motor operation status monitoring method based on multi-sensor fusion according to claim 3 is characterized in that: Set the predefined convergence strategy, including: configuring a first credibility, a second credibility, and a third credibility according to the first convergence degree, the second convergence degree, and the third convergence degree; The predetermined fusion strategy is set based on the first credibility, the second credibility and the third credibility.

6. The motor operation status monitoring method based on multi-sensor fusion according to claim 3 is characterized in that: Performing the servo motor operating status monitoring within the first time window, before that also includes: Building a sensor monitoring array based on the multi-dimensional monitoring indicators; The sensor monitoring array is used to perform fixed-point monitoring on the servo motor, and the monitoring data set is transmitted to the edge judgment node, wherein the monitoring data set includes real-time working current, real-time output torque, real-time coil temperature, real-time bearing temperature, real-time casing temperature, real-time bearing vibration data, real-time motor shaft vibration data, real-time casing vibration data, real-time motor speed and real-time motor position.

7. The motor operation status monitoring method based on multi-sensor fusion according to claim 6 is characterized in that: Executing the servo motor operating status monitoring within the first time window includes: Based on the first current threshold, the first torque threshold, the first temperature threshold set, the first vibration threshold set, the first speed threshold and the first position threshold, early warning judgments are respectively performed on the real-time working current, the real-time output torque, the real-time coil temperature, the real-time bearing temperature, the real-time housing temperature, the real-time bearing vibration data, the real-time motor shaft vibration data, the real-time housing vibration data, the real-time motor speed and the real-time motor position, and indicators that do not meet the threshold are set as early warning indicators; If the number of the warning indicators is not 0, a motor warning signal is generated based on the warning indicators and a warning is issued.

Images