Multi-mode switching monitoring system based on level analysis

Abstract

The invention relates to a multi-mode switching monitoring system based on level analysis, and the system comprises a damper which is disposed at the output end of a motor, is located between the motor and a load, and outputs damping to change the load; the data monitoring module is used for acquiring parameters during operation of the motor according to a passive monitoring mode or an active monitoring mode; the analysis module is used for carrying out feature extraction on the motor operation parameters, carrying out anomaly comparison, matching and executing corresponding analysis schemes in a chain analysis strategy according to a preset selection strategy, and matching a next analysis scheme based on a comparison result of a previous analysis scheme; the simulation module is used for reproducing and obtaining the reference parameters in the analysis scheme, the obtained reference parameter waveform serves as the comparison basis of abnormality judgment, and the abnormality type of the motor is accurately recognized through cooperation of damper adjustment, multi-mode monitoring and a progressive analysis scheme.

Description

Technical Field

[0001] This invention relates to the field of industrial motor condition monitoring technology, specifically a multi-mode switching monitoring system based on level position analysis. Background Technology

[0002] In key industrial fields such as intelligent manufacturing, rail transit, and new energy equipment, the motor system, as the core power output unit, directly determines the production efficiency, operational safety, and maintenance costs of the entire equipment through its operational stability and reliability. Therefore, motor condition monitoring technology has become a core research direction for industrial equipment health management (PHM) systems.

[0003] With the rapid iteration of microelectromechanical systems (MEMS) sensing technology and edge computing technology, motor monitoring systems have evolved from early offline fault diagnosis relying on manual methods, such as periodic shutdowns for disassembly and inspection, and lubricating oil spectral analysis, to online real-time monitoring systems. The monitoring dimensions have also expanded from single electrical parameters, such as stator three-phase current and bus voltage, to the collaborative sensing of multiple physical quantities, covering motor vibration acceleration, stator winding temperature, output torque fluctuations, load speed, and transmission link noise. Meanwhile, to achieve precise multi-level control of power transmission between the motor and load, dampers, due to their advantages such as strong anti-electromagnetic interference capability, stable transmission efficiency, good adaptability to low-temperature environments, and low maintenance costs, are widely integrated into motor transmission chains requiring gear switching, becoming a key transmission node connecting the motor output shaft and the load actuator. Therefore, by detecting the speed and torque of the damper, the speed and torque of the motor's output shaft can be obtained.

[0004] Existing multi-sensor monitoring solutions generally adopt independent parameter analysis or parallel processing modes without logical connections, resulting in a large amount of redundant data processing and invalid calculations during the analysis process. This not only prolongs the time required for anomaly location but also easily leads to an increased fault misjudgment rate due to interference from irrelevant parameters. Furthermore, the analysis process takes a long time and cannot meet the high reliability and high operating requirements of industrial motors. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the present invention aims to accurately identify single, compound, and novel anomalies in motors through the synergy of damper adjustment, multi-mode monitoring, and progressive analysis schemes, thereby achieving intelligent and efficient monitoring and anomaly determination of motor operating status.

[0006] This invention provides a multi-mode handover monitoring system based on level bit resolution, comprising:

[0007] A damper, mounted at the output of the motor and located between the motor and the load, provides output damping to change the load. Dampers provide an intervention-based adjustment means for anomaly verification, enabling the adjustment of motor output torque without interrupting production. Through damping adjustment, the force state of the motor can be precisely controlled, providing an operable physical intervention basis for multi-mode monitoring and anomaly analysis. The data monitoring module acquires motor operating parameters according to either passive or active monitoring modes. This data monitoring module acquires motor operating parameters through two monitoring modes, and performs comprehensive and differentiated data collection on the motor status. The analysis module extracts features from the motor operating parameters and performs anomaly comparisons. Based on a preset selected strategy, it matches and executes the corresponding analysis scheme within the chain analysis strategy. The analysis module matches the next analysis scheme based on the comparison results of the previous analysis scheme. By using feature extraction, anomaly comparison, and chain analysis strategy matching, a closed-loop analysis system from parameter acquisition to anomaly type determination is achieved. The system dynamically selects the first, second, or third analysis scheme based on parameter characteristics, and links subsequent schemes based on previous results to form a progressive analysis logic, avoiding the limitations of a single analysis mode. Furthermore, the chain strategy reduces redundant analysis steps. The simulation module is used to analyze the reproduction of the scheme to obtain benchmark parameters, and the obtained benchmark parameter waveforms are used as the comparison basis for anomaly judgment. By reproducing the waveform of the benchmark parameters as a comparison basis, cross-scenario verification of theoretical data and actual data was achieved.

[0008] Furthermore, the data monitoring module also includes a torque-speed sensor, a piezoelectric acceleration sensor, a current sensor, a voltage sensor, and a temperature sensor, specifically: The input terminal of the torque and speed sensor is connected to the output terminal of the damper, and the output terminal is connected to the load to obtain the torque and speed parameters of the motor operation; the piezoelectric acceleration sensor is installed on the motor to obtain vibration parameters; the current sensor and the voltage sensor are installed in the power supply circuit of the motor to obtain voltage and current parameters; the temperature sensor is built into the motor to obtain the internal temperature parameters of the motor; and the power parameters of the motor are calculated using the voltage and current parameters of the motor.

[0009] It enables the acquisition of electrical, mechanical, and temperature parameters of motors from all dimensions, providing multi-physical data support for the analysis module and avoiding the one-sidedness of a single parameter.

[0010] Furthermore, the analysis module is configured with a selected strategy, specifically: The pre-built database stores multiple anomaly types, and each anomaly type is pre-matched with corresponding abnormal waveform features and associated influence parameters. The abnormal waveform features reflect the parameter waveforms of the motor when the corresponding anomaly type is running, and the associated influence parameters reflect the parameters affected by the anomaly type. The analysis module obtains the motor operating parameters as real-time waveforms of regular parameters, extracts the features of the real-time waveforms of regular parameters and compares them with the abnormal waveform features of the corresponding parameters in the database to obtain the similarity. The abnormal type with the highest similarity is taken as the dominant abnormality, and the highest similarity is judged. The module also switches from passive monitoring mode to active monitoring mode. If the similarity is higher than the first high threshold, then execute the first analysis scheme; If the similarity is lower than the first low threshold, then the second analysis scheme is executed; If the similarity is between the first high threshold and the first low threshold, then the third analysis scheme is executed.

[0011] By comparing real-time waveform features with a database to obtain similarity scores, and matching analysis schemes according to similarity intervals, dynamic selection of analysis paths is achieved. This provides the system with decision-making logic for hierarchical processing based on anomaly characteristics, avoiding blindly executing complex analyses. High-similarity anomalies directly enter the first analysis scheme, low-similarity anomalies enter the second analysis scheme, and medium-similarity anomalies enter the third analysis verification, reducing invalid operations and improving overall analysis efficiency.

[0012] Furthermore, the first analysis scheme specifically includes: The abnormal waveforms are filtered by matching the characteristics of real-time waveforms with the characteristics of abnormal waveforms in the database based on the similarity. The abnormal waveforms with the highest similarity are used to preliminarily determine the corresponding abnormal types. Based on the determined anomaly type, the associated impact parameters are retrieved from the pre-built database, and in the active monitoring mode, the associated impact parameters are gradually adjusted to obtain enhanced anomaly waveform features. Specifically, the enhanced anomaly waveform features are the actual waveforms after amplification of the real-time waveforms of regular parameters that are highly similar to the anomaly waveforms in the database. The simulation module is used to reproduce the simulation baseline waveform after adjusting the associated influence parameters, and the similarity is calculated by comparing it with the enhanced abnormal waveform features. If the similarity is higher than the second high threshold when comparing similarity in the selected strategy, the preliminary anomaly type is accurate; if the similarity is lower than the second low threshold, the second analysis scheme is executed; if the similarity is between the second high threshold and the second low threshold, the third analysis scheme is executed.

[0013] By initially matching anomaly types, adjusting correlation parameters to amplify features, and then comparing with simulation benchmarks, we can accurately verify highly similar anomalies, quickly identify specific anomaly types, and avoid over-analysis.

[0014] Furthermore, the second analysis scheme specifically includes: Establish a baseline pattern for parameter fluctuations in a pre-built database. Specifically, the baseline pattern for parameter fluctuations is the synergistic pattern between parameter fluctuations and related influencing parameters. The fluctuation pattern of the waveform of the parameter to be analyzed is compared with the parameter fluctuation benchmark pattern to determine whether it conforms to the parameter fluctuation benchmark pattern. If it conforms, it is determined to be a normal fluctuation; if it does not conform, it is determined to be an anomaly, and the anomaly confirmation sub-strategy is executed. The simulation module is used to reproduce the waveform of the simulated parameter of the parameter to be analyzed, which is determined to be fluctuating normally. The deviation is compared with the waveform of the parameter to be analyzed. If the waveform deviation is determined to be lower than the deviation threshold, the passive monitoring mode is switched; otherwise, the third analysis scheme is executed.

[0015] By comparing the waveform under analysis with the baseline pattern of parameter fluctuations and verifying it through simulation reproduction, the distinction between false anomalies and true anomalies is realized, thereby filtering out external interference, such as suspected anomalies caused by normal load fluctuations, and reducing the number of cases that are mistakenly judged as anomalies.

[0016] Furthermore, the specific details of the anomaly confirmation sub-strategy are as follows: Feature extraction is performed on the waveforms of the parameters to be analyzed that do not conform to the fluctuation benchmark law to generate anomaly feature vectors; The associated influencing parameters are adjusted based on the active monitoring mode to obtain the waveform difference before and after adjustment, and the abnormal response intensity is calculated. If the abnormal response intensity is higher than the response threshold, the third analysis scheme will be executed based on the abnormal response intensity.

[0017] The anomaly confirmation sub-strategy verifies genuine anomalies, clarifies the linkage conditions between the second and third analysis schemes, avoids all parameters that do not conform to the baseline rules from entering the third analysis scheme, and improves the operating efficiency of the monitoring system.

[0018] Furthermore, the third analysis scheme is specifically as follows: The pre-built database pre-stores combinations of anomaly types and corresponding combined collaborative waveform features, as well as combined association influence parameters associated with each group of anomaly types. The characteristics of the waveform to be analyzed are compared with the characteristics of the combined and coordinated waveforms, and the combination of anomaly types corresponding to the maximum similarity is used as the criterion for judgment. If the similarity is greater than the third highest threshold, then the corresponding combination of anomaly types is determined; If the similarity is below the third lowest threshold, it is treated as a new anomaly type and stored. If the similarity is between the third high threshold and the third low threshold, the corresponding combined correlation influence parameter is adjusted through active monitoring mode to obtain the characteristics of the amplified waveform to be analyzed. The characteristics are then compared again with the combined collaborative waveform features corresponding to the abnormal type combinations in the database. If the similarity of all abnormal type combinations is lower than the third low threshold, it is determined to be a new abnormal type and written into the database.

[0019] The third analysis scheme mainly achieves the identification of abnormal type combinations and new abnormalities by matching abnormal type combination features, adjusting combination correlation parameters to amplify features, and then determining known combinations or new types, thus breaking through the analysis limitations of single abnormalities.

[0020] Furthermore, the passive monitoring mode specifically includes: The data monitoring module acquires actual parameters according to a preset benchmark sampling frequency, summarizes and records the actual parameters at intervals, collects real-time waveforms of regular parameters, and generates log reports periodically based on historical records.

[0021] The passive monitoring mode is used for low-intensity monitoring during normal operation and when there is no abnormal risk to the motor. It reduces resource consumption when there is no abnormal risk to the motor, while retaining basic status data.

[0022] Furthermore, the active monitoring mode specifically includes: The active monitoring mode specifically includes scenarios that only collect data without intervention and scenarios that collect data and intervene. The scenario of collecting data without intervention specifically refers to the following: the data monitoring module collects parameters at a sampling frequency higher than the benchmark, records the parameters point by point, and records the real-time waveform of the parameters; The specific acquisition and intervention scenario is as follows: when the triggering condition is met, the intervention operation is initiated. The intervention operation is carried out by adjusting the associated influencing parameters and simultaneously acquiring parameters at a sampling frequency higher than the benchmark, recording the parameters point by point, and recording the real-time waveform of the parameters. The triggering conditions include the preliminary determination of the anomaly type in the first analysis scheme, the execution of the anomaly confirmation sub-strategy, and the analysis parameters whose similarity is between the third high threshold and the third low threshold in the third analysis scheme.

[0023] The proactive monitoring mode enables differentiated data collection and intervention under abnormal conditions. The monitoring strategy is dynamically adjusted according to the abnormal situation. The data collection and intervention scenario is for scenarios after the abnormality is identified. The abnormal characteristics are enhanced by adjusting the correlation parameters to analyze the abnormality.

[0024] The beneficial effects of this invention are: This invention solves the problem of the inability to adjust the actual load in industrial scenarios due to production constraints by setting a damper between the motor and the load and using its output damping to equivalently simulate load changes. This allows the system to change the motor output torque by adjusting the damping without interrupting production, thereby realizing functions such as amplifying abnormal features and verifying the root cause. This not only ensures production continuity but also provides an operable physical intervention basis for multi-mode monitoring and progressive analysis, filling the technical gap in anomaly verification under fixed load scenarios.

[0025] This invention uses a passive monitoring mode to adapt to normal conditions to reduce resource consumption, and an active monitoring mode to focus on abnormal conditions to capture details. Furthermore, the analysis module uses similarity judgment to hierarchically link the first, second, and third analysis schemes, forming a closed-loop logic from hierarchical processing to chain verification. This avoids the limitations of a single mode and ensures that resources are prioritized for high-value analysis, significantly improving the accuracy of anomaly identification and overall operational efficiency.

[0026] The present invention incorporates an analysis of an anomaly type combination feature library in the third analysis scheme, which solves the problem that single fault analysis cannot cover compound anomalies. Furthermore, the automatic storage mechanism for new anomaly types endows the system with self-learning capabilities, allowing the recognition range to be expanded as the system accumulates data during operation. Attached Figure Description

[0027] Figure 1 This is a logic diagram of the chain analysis scheme under the multi-mode switching monitoring system based on level position analysis of the present invention.

[0028] Figure 2 This is a logic diagram of the monitoring mode switching and anomaly analysis interaction in the multi-mode switching monitoring system based on level position analysis of the present invention. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] It should be noted that when a component is described as "fixed to" another component, it can be directly on the other component or may have a component in between. When a component is considered "connected to" another component, it can be directly connected to the other component or may have a component in between. When a component is considered "set on" another component, it can be directly set on the other component or may have a component in between. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0031] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0032] like Figure 1 and Figure 2 As shown, this invention provides a multi-mode switching monitoring system based on level bit analysis, comprising: A damper is installed at the output of the motor and located between the motor and the load. The damper outputs damping to change the load. Dampers provide an intervention-based adjustment means for anomaly verification, enabling the adjustment of motor output torque without interrupting production. Through damping adjustment, the force state of the motor can be precisely controlled, providing an operable physical intervention basis for multi-mode monitoring and anomaly analysis. The data monitoring module acquires motor operating parameters based on either passive or active monitoring modes. This data monitoring module acquires motor operating parameters through two monitoring modes, and performs comprehensive and differentiated data collection on the motor status. The analysis module extracts features from motor operating parameters and performs anomaly comparisons. Based on a preset selected strategy, it matches and executes the corresponding analysis plan within the cascading analysis strategy. The analysis module matches the next analysis plan based on the comparison results of the previous analysis plan. By using feature extraction, anomaly comparison, and chain analysis strategy matching, a closed-loop analysis system from parameter acquisition to anomaly type determination is achieved. The system dynamically selects the first, second, or third analysis scheme based on parameter characteristics, and links subsequent schemes based on previous results to form a progressive analysis logic, avoiding the limitations of a single analysis mode. Furthermore, the chain strategy reduces redundant analysis steps. The simulation module is used to analyze and reproduce baseline parameters in the scheme, and obtain the waveforms of the baseline parameters as a comparison basis for anomaly judgment. By reproducing the waveform of the benchmark parameters as a comparison basis, cross-scenario verification of theoretical data and actual data was achieved.

[0033] It should be noted that the damper used in this solution is a conventional device already existing in the field of motor drives. Its inherent properties allow for a stable output of damping, a characteristic already well-established in motor load connection and transmission adjustment scenarios, and the relevant technology is common knowledge. Furthermore, this solution does not make any improvements or innovations to the damper's structure; its core function is solely as a mounting platform, accommodating torque sensors, photoelectric encoders, and other detection components installed between the motor and the load to achieve close-range, accurate acquisition of key parameters during motor operation under load. Given that its structure is a conventional design, and the focus of this invention is on the system architecture, analysis strategy, and monitoring mode linkage logic for multi-mode motor switching monitoring, rather than the damper's own construction, there is no need to elaborate on its specific structure.

[0034] Furthermore, the data monitoring module includes a torque and speed sensor, a piezoelectric acceleration sensor, a current sensor, a voltage sensor, and a temperature sensor, specifically: The output end of the motor is connected to the input end of the damper. The output end of the damper is connected to one end of the torque and speed sensor through a coupling. The other end of the torque and speed sensor is also connected to the load through a coupling. The dynamic torque and speed parameters of the motor under load are collected in real time. During installation, it is necessary to ensure that the motor, damper, torque and speed sensor and load are coaxially connected. The setting of the damper avoids the sensor being directly mounted on the motor shaft end, which is susceptible to the acquisition deviation caused by the vibration of the motor body. The piezoelectric accelerometer is installed at two radial positions, horizontal and vertical, on the motor base. It can be installed using a magnetic base to acquire vibration parameters generated by gear meshing and bearing rotation. A current sensor can be installed at the contactor output terminal of the motor power supply circuit to collect the current parameters during motor operation, and a voltage sensor can be connected in parallel at the fuse end of the power supply circuit to obtain the voltage parameters during motor operation. Temperature sensors are built into the motor to obtain internal temperature parameters. For example, they are built into the tooth gap of the motor stator core to collect the motor winding temperature. Furthermore, using the voltage and current parameters of the motor obtained above, the power parameters of the motor are calculated using the power calculation formula.

[0035] It enables the acquisition of electrical, mechanical, and temperature parameters of motors from all dimensions, providing multi-physical data support for the analysis module and avoiding the one-sidedness of a single parameter.

[0036] This invention includes a passive monitoring mode and an active monitoring mode, the details of which are as follows: The passive monitoring mode is as follows: the data monitoring module acquires actual parameters according to the preset benchmark sampling frequency, summarizes and records the actual parameter intervals and the real-time waveforms of the acquired parameters, and generates log reports periodically based on historical records.

[0037] The passive monitoring mode is used for low-intensity monitoring during normal operation and when there is no abnormal risk to the motor. It reduces resource consumption when there is no abnormal risk to the motor, while retaining basic status data.

[0038] The active monitoring mode specifically includes a data collection without intervention scenario and a data collection and intervention scenario. The data collection without intervention scenario specifically refers to the data monitoring module collecting parameters at a sampling frequency higher than the baseline, recording the parameters point by point, and recording the real-time waveform of the parameters. The specific process of collecting and intervening in the scenario is as follows: when the triggering conditions are met, the intervention operation is initiated. The intervention operation is carried out by adjusting the related influencing parameters and simultaneously collecting parameters at a sampling frequency higher than the baseline, recording the parameters point by point, and recording the real-time waveform of the parameters. Switch to active monitoring mode to collect data without intervention. When the triggering conditions are met, the intervention operation will be initiated. The triggering conditions include the preliminary determination of the anomaly type in the first analysis plan, the execution of the anomaly confirmation sub-strategy, and the analysis parameters whose similarity is between the third high threshold and the third low threshold in the third analysis plan.

[0039] The proactive monitoring mode enables differentiated data collection and intervention under abnormal conditions. The monitoring strategy is dynamically adjusted according to the abnormal situation. The data collection and intervention scenario is for scenarios after the abnormality is identified. The abnormal characteristics are enhanced by adjusting the correlation parameters to analyze the abnormality.

[0040] Furthermore, the analysis module is configured with selected strategies, specifically: The pre-built database stores various anomaly types, and for each anomaly type, it pre-matches corresponding anomaly waveform features and associated impact parameters. The anomaly waveform features reflect the motor's parameter waveforms during operation under the corresponding anomaly type, and the associated impact parameters reflect the parameters affected by the anomaly type. The system includes a mapping between anomaly types and anomaly waveform features, storing unique parameter waveform features for each known anomaly. These features are extracted from numerous historical fault cases and directly reflect the anomaly's parameter behavior during runtime. Another feature is the mapping between anomaly types and associated impact parameters, recording which associated parameters significantly affect each anomaly, providing clear targets for strengthening anomaly analysis in subsequent plans. The analysis module obtains the motor operating parameters as real-time waveforms of regular parameters, extracts the features of the real-time waveforms of regular parameters and compares them with the abnormal waveform features of the corresponding parameters in the database to obtain the similarity. The abnormal type with the highest similarity is taken as the dominant abnormality, and the highest similarity is judged. The system then switches from passive monitoring mode to active monitoring mode. If the similarity is higher than the first high threshold, then execute the first analysis scheme; If the similarity is lower than the first low threshold, then the second analysis scheme is executed; If the similarity is between the first high threshold and the first low threshold, then the third analysis scheme is executed.

[0041] For example: The database stores the following information: Anomaly type A is bearing outer ring wear. The abnormal waveform characteristics of anomaly type A are a significant peak in the 23Hz frequency band, and the significant peak accounts for more than 30% of the total energy. An impact pulse occurs every 10ms. Its relevant influencing parameters are load rate and damper damping. Anomaly type B is rotor imbalance. The abnormal waveform characteristics of anomaly type B are a significant peak in the 24Hz frequency band, with the significant peak accounting for more than 50% of the total energy, and no impact pulse. Its relevant influencing parameter is the rotational speed. The preset first high threshold for similarity judgment is 85%, and the first low threshold is 60%. Vibration parameters obtained through the data monitoring module were used to extract real-time vibration waveform features, showing a peak frequency of 32% in the 23.2Hz band with an impact pulse occurring every 9.8ms, and a peak frequency of 18% in the 24Hz band. After similarity calculation, the similarity with the features of anomaly type A was 82%, and the similarity with the features of anomaly type B was 45%. If the highest similarity score is used as the dominant anomaly for analysis, then the dominant anomaly is anomaly type A with a similarity score of 82%, which is between the first high threshold of 85% and the first low threshold of 60%. This accurately guides the analysis to the third analysis scheme to continue analyzing the anomaly, realizing dynamic rotation of the analysis path, avoiding blindly executing complex analyses, and reducing redundant analyses.

[0042] If the vibration parameters collected by the data monitoring module show that the extracted real-time vibration waveform features a peak value of 35% in the 23.1Hz frequency band and an impact pulse every 10ms, which highly matches the abnormal characteristics of anomaly type A, and the similarity is 91% based on similarity calculation, then anomaly type A will be analyzed as the dominant anomaly. Since the similarity of 91% is greater than the first high threshold of 85%, it meets the triggering conditions of the first analysis scheme, and the first analysis scheme will be executed.

[0043] The first analysis plan, in detail, is as follows: The abnormal waveforms are filtered by matching the characteristics of real-time waveforms with the characteristics of abnormal waveforms in the database based on the similarity. The abnormal waveforms with the highest similarity are used to preliminarily determine the corresponding abnormal types. Using the highest similarity as a preliminary criterion ensures that the most likely option is selected from the known anomaly types, providing a clear direction for subsequent verification and avoiding aimless generalization analysis. Based on the determined anomaly type, the associated impact parameters are retrieved from the pre-built database. In the active monitoring mode, the associated impact parameters are gradually adjusted to obtain enhanced anomaly waveform features. Specifically, the enhanced anomaly waveform features are the actual waveforms of real-time waveforms of regular parameters that are highly similar to the anomaly waveforms in the database after being amplified. By actively adjusting the relevant correlation influence parameters, the characteristics of anomalies can be amplified, which solves the problem that the characteristics of early faults or minor anomalies are not obvious and improves the accuracy of subsequent verification. At the same time, the correlation influence parameters should be adjusted gradually, and a fixed value should be set for the magnitude of each adjustment to avoid the impact on the motor caused by sudden parameter changes and ensure the safe operation of the motor. The simulation module is used to reproduce the simulation baseline waveform after adjusting the associated influence parameters, and the similarity is calculated by comparing it with the enhanced abnormal waveform features. By comparing similarity in the selected strategy, if the similarity is higher than the second high threshold, the preliminarily determined anomaly type is accurate; if the similarity is lower than the second low threshold, the second analysis scheme is executed; if the similarity is between the second high threshold and the second low threshold, the third analysis scheme is executed.

[0044] By using secondary similarity assessment to form a closed loop from initial assessment to enhanced verification and then to final decision, the accuracy of anomaly type assessment is ensured, and possible misjudgments are corrected through scheme linkage.

[0045] For example, if the anomaly type is initially identified as A in the selected strategy, and the similarity of 91% is higher than the first high threshold of 85%, the first analysis plan is triggered. Firstly, the data monitoring module switches from passive monitoring mode to active monitoring mode, and retrieves the correlation parameters corresponding to the wear of the bearing outer ring from the database. By gradually adjusting the output damping of the damper, the initial damping is 0.03 kN. At m, the vibration peak value was 0.09 mm, and the 23.1 Hz peak value accounted for 35%. The first adjustment increased the damping to 0.04 kN. After stabilizing for 5 minutes, the operating parameters were collected, with a vibration peak of 0.10 mm and a peak frequency of 23.1 Hz accounting for 38%. The second adjustment increased the damping to 0.05 kN. m, parameters collected after stabilization for 5 minutes, vibration peak value 0.11mm, 23.1Hz peak value accounted for 40%; The third adjustment increased the damping to 0.06 kN. After stabilizing for 5 minutes, the vibration peak was 0.13 mm, with the 23.1 Hz peak accounting for 42%. By making multiple adjustments, the impact pulse becomes sharper than its initial state, and its characteristics are significantly amplified, thus forming enhanced abnormal waveform features. The simulation module is used to reproduce the problem. The input parameters for the simulation are the load, speed and other parameters corresponding to the third damping adjustment. The simulation reference waveform is generated as the theoretical waveform and compared with the actual waveform after adjustment and enhancement. If the second similarity calculation is 94%, which is greater than the second high threshold of 80%, then the abnormal type of bearing outer ring wear initially determined is accurate.

[0046] If the dominant anomaly initially identified in the selected strategy is anomaly type A with a similarity of 48%, which is less than the first low threshold of 60%, it is judged as low similarity, and the second analysis scheme is executed; or if the dominant anomaly is initially identified as anomaly type A with a similarity of 91%, which is greater than the first high threshold of 85%, the first analysis scheme is executed, but after the first analysis scheme is strengthened and verified, the secondary similarity is lower than the second low threshold, and the second analysis scheme is also triggered.

[0047] The second analysis plan is detailed below: Establish a baseline pattern for parameter fluctuations in a pre-built database. Specifically, the baseline pattern for parameter fluctuations is the synergistic pattern between parameter fluctuations and related influencing parameters. The fluctuation pattern of the waveform of the parameter to be analyzed is compared with the parameter fluctuation benchmark pattern to determine whether it conforms to the parameter fluctuation benchmark pattern. If it conforms, it is determined to be a normal fluctuation; if it does not conform, it is determined to be an anomaly, and the anomaly confirmation sub-strategy is executed. The simulation module is used to reproduce the waveform of the simulated parameter of the parameter to be analyzed, which is determined to be fluctuating normally. The deviation is compared with the waveform of the parameter to be analyzed. If the waveform deviation is determined to be lower than the deviation threshold, the passive monitoring mode is switched; otherwise, the third analysis scheme is executed.

[0048] The specific content of the exception confirmation sub-strategy is as follows: Feature extraction is performed on the waveforms of the parameters to be analyzed that do not conform to the fluctuation benchmark law to generate anomaly feature vectors; The associated influencing parameters are adjusted based on the active monitoring mode to obtain the waveform difference before and after adjustment, and the abnormal response intensity is calculated. If the abnormal response intensity is higher than the response threshold, the third analysis scheme will be executed based on the abnormal response intensity.

[0049] Case 1: The main anomaly identified in the preliminary analysis was the wear of the bearing outer ring, with a similarity of 48%, which is lower than the first low threshold and is judged as low similarity, triggering the second analysis scheme; First, the baseline pattern of bearing outer ring wear in the database is that the 23Hz peak value increases by ≥5% with the increase of load rate. The load rate is adjusted by adjusting the damper damping to adjust the load rate from 40% to 60%. The fluctuation pattern of the 23Hz peak value in the parameter to be analyzed is obtained. The 23Hz peak value increases from 15% to 18%. It is compared with the baseline pattern and it is determined that it does not conform to the baseline pattern. Therefore, it is determined that there is an anomaly and enters the anomaly confirmation sub-strategy. Features were extracted from the waveform of the parameter to be analyzed. The vibration parameter features are: 23Hz peak value 15%-18%, intermittent pulse interval 10-15ms, waveform baseline drift 0.01mm, and the generated abnormal feature vector is [15-18%, 10-15ms, 0.01mm]. Switching to active monitoring mode, the damper damping associated with abnormal wear of the bearing outer ring is adjusted based on the active monitoring mode, reducing the damper damping from 0.03kN. m increased to 0.06kN m, obtain the waveform difference before and after adjustment: the 23Hz peak value increased from 18% to 25%, the pulse interval stabilized at 10ms, and the baseline drift increased to 0.02mm. Based on the collected data, and using the existing response intensity calculation formula, the abnormal response intensity is calculated by using the ratio of the relative change rate of abnormal signs to the adjustment amplitude of associated parameters. The response intensity corresponding to the 23Hz peak is found to be 12.96% / (kN). m), higher than the response threshold of 10% / (kN) m), which is determined to be a significant abnormal response, triggers the third analysis scheme to perform abnormal combination analysis; Case 2: Initial assessment indicated bearing outer ring wear, but the similarity score in the first analysis scheme was below the second low threshold, triggering the second analysis scheme. Similarly, if the parameter waveform does not conform to the baseline pattern, the anomaly confirmation strategy is entered. The intensity of the abnormal response is calculated to determine whether it is a significant abnormal response. If it is higher than the response threshold, the third analysis scheme is entered. Case 3: Initially, a slight fluctuation in the torque parameter was detected, with a similarity of 53% to the abnormal waveform, which is below the first low threshold of 60%, triggering the second analysis scheme. The database stores the baseline law for torque parameter fluctuation. When the damper damping fluctuation is ±5%, the motor output torque fluctuation is ≤3%. By damping the damper from 0.04kN m rises to 0.042 kN m, the fluctuation of the torque waveform to be analyzed ranges from 120N. m rises to 123N The fluctuation of m, at +2.5%, is consistent with the baseline pattern and is preliminarily judged as normal fluctuation. The simulation module was used to input a damping of 0.042 kN. Using the parameters m and normal operating conditions, a simulated torque waveform is generated with an average torque of 122.8 N. The fluctuation range of m is ±2.2%, and the deviation from the actual torque waveform is 0.3%, which is less than the deviation threshold of 1%. This confirms that the slight fluctuation of the torque parameter is a normal mechanical response caused by the natural fluctuation of the damper damping, and is not abnormal.

[0050] If the waveform of the parameter to be analyzed is determined to be between the first high threshold and the first low threshold in the selected strategy, it will enter the third analysis scheme. Or, if the waveform similarity of the abnormal parameter is higher than the first high threshold, and the second similarity is between the second high threshold and the second low threshold after the first analysis scheme, the third analysis scheme will be triggered. Or, if the parameter does not conform to the benchmark pattern after the second analysis scheme and the abnormal response intensity is significant, it will enter the third analysis scheme.

[0051] The third analysis plan is as follows: The pre-built database stores combinations of anomaly types and corresponding combined collaborative waveform features, as well as combined association influence parameters associated with each combination of anomaly types. The combined collaborative waveform feature refers to the composite waveform feature formed by the superposition and influence of the single waveform features corresponding to each anomaly when two or more anomaly types exist simultaneously. The characteristics of the waveform to be analyzed are compared with the characteristics of the combined and coordinated waveforms, and the combination of anomaly types corresponding to the maximum similarity is used as the criterion for judgment. If the similarity is greater than the third highest threshold, then the corresponding combination of anomaly types is determined; If the similarity is below the third lowest threshold, it is treated as a new anomaly type and stored. If the similarity is between the third high threshold and the third low threshold, the corresponding combined correlation influence parameters are actively adjusted through intervention in active monitoring mode to obtain the characteristics of the amplified waveform to be analyzed. These characteristics are then compared again with the combined collaborative waveform features corresponding to the abnormal type combinations in the database. If the similarity of all abnormal type combinations is lower than the third low threshold, it is determined to be a new abnormal type and written into the database.

[0052] Case 1: Based on the second analysis plan, the parameters that do not conform to the baseline pattern and have a significant abnormal response intensity are identified as true anomalies, triggering the third analysis plan. First, the database stores multiple abnormal combinations, including combination C and combination D. Assuming combination C is slight wear on the bearing outer ring combined with shaft misalignment, its combined waveform characteristics show a 20%-25% peak value at 23Hz and a 30%-35% peak value at the first harmonic. The waveform exhibits periodic pulses with a tilted baseline, pulse intervals of 8-12ms, and a baseline tilt of 0.01-0.02mm / 10s. The corresponding associated parameters are damper damping and load rate. Combination D is wear on the bearing inner ring combined with a slight short circuit between winding turns. Its combined waveform characteristics show a 25%-30% peak value at 35Hz, an impact pulse rise slope ≥0.025mm / ms, and a 6%-8% third harmonic current. The current waveform has spikes in the positive half-cycle. The corresponding associated parameters are rotational speed and damper damping. The waveform features to be analyzed are compared with the waveform features of combinations of anomaly types in the database, and the similarity is calculated. The waveform features to be analyzed are: 23Hz peak percentage 22%, 24Hz peak percentage 32%, pulse interval 9.8ms, and baseline tilt 0.014mm / 10s. After similarity calculation, and taking the anomaly combination with the highest similarity as the dominant anomaly, the similarity with anomaly type C is found to be 78%. A similarity score of 78% is determined to be between the third highest threshold of 80% and the third lowest threshold of 65%. Therefore, active monitoring mode is initiated for adjustment. This involves gradually adjusting the damper damping in the combined correlation parameters using a small, incremental approach, reducing the damper damping from 0.03 kN. m increased to 0.05kN The amplified waveform features a 24% peak value at 23Hz, a 34% peak value at 24Hz, and a sharper pulse. Its rise slope is 0.028 mm / ms, and its baseline tilt is 0.016 mm / 10s. The amplified waveform features are compared again with the combined waveform features of the anomaly type combination C. The similarity is 83%, which is higher than the third highest threshold of 80%. Therefore, the anomaly type is determined to be a combination of slight wear on the outer ring of the bearing and misalignment of the shaft system.

[0053] In Case 2, the waveform features to be analyzed are compared with the combined waveform features of each combination of anomaly types in the database. The calculated similarity is 62%, which is still less than the third highest threshold of 65%. Therefore, the anomaly is determined to be a new anomaly type. The system automatically records the waveform features of the anomaly and writes them into the database to provide a basis for the identification of similar anomalies in the future.

[0054] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principle of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A multi-mode switching monitoring system based on level-position analysis, characterized in that, include: A damper is mounted on the output end of the motor and located between the motor and the load, the damper outputs damping to change the load; The data monitoring module acquires parameters of the motor during operation according to either a passive monitoring mode or an active monitoring mode. The analysis module extracts features from the motor operating parameters and performs anomaly comparison. It matches and executes the corresponding analysis scheme in the chain analysis strategy according to the preset selected strategy. The analysis module matches the next analysis scheme based on the comparison result of the previous analysis scheme. The simulation module is used to analyze the scheme to reproduce and obtain the benchmark parameters, and obtain the benchmark parameter waveform as the comparison basis for anomaly judgment.

2. The multi-mode switching monitoring system based on level position analysis according to claim 1, characterized in that, The analysis module is configured with a selected strategy, specifically: The pre-built database stores multiple anomaly types, and each anomaly type is pre-matched with corresponding abnormal waveform features and associated influence parameters. The abnormal waveform features reflect the parameter waveforms of the motor when the corresponding anomaly type is running, and the associated influence parameters reflect the parameters affected by the anomaly type. The analysis module obtains the motor operating parameters as real-time waveforms of regular parameters, extracts the features of the real-time waveforms of regular parameters and compares them with the abnormal waveform features of the corresponding parameters in the database to obtain the similarity. The abnormal type with the highest similarity is taken as the dominant abnormality, and the highest similarity is judged. The module also switches from passive monitoring mode to active monitoring mode. If the similarity is higher than the first high threshold, then execute the first analysis scheme; If the similarity is lower than the first low threshold, then the second analysis scheme is executed; If the similarity is between the first high threshold and the first low threshold, then the third analysis scheme is executed.

3. The multi-mode switching monitoring system based on level position analysis according to claim 2, characterized in that, The first analysis scheme specifically includes: The abnormal waveforms are filtered by matching the characteristics of real-time waveforms with the characteristics of abnormal waveforms in the database based on the similarity. The abnormal waveforms with the highest similarity are used to preliminarily determine the corresponding abnormal types. Based on the determined anomaly type, the associated impact parameters are retrieved from the pre-built database, and in the active monitoring mode, the associated impact parameters are gradually adjusted to obtain enhanced anomaly waveform features. Specifically, the enhanced anomaly waveform features are the actual waveforms after amplification of the real-time waveforms of regular parameters that are highly similar to the anomaly waveforms in the database. The simulation module is used to reproduce the simulation baseline waveform after adjusting the associated influence parameters, and the similarity is calculated by comparing it with the enhanced abnormal waveform features. If the similarity is higher than the second high threshold when comparing similarity in the selected strategy, the preliminary anomaly type is accurate; if the similarity is lower than the second low threshold, the second analysis scheme is executed; if the similarity is between the second high threshold and the second low threshold, the third analysis scheme is executed.

4. The multi-mode switching monitoring system based on level position analysis according to claim 3, characterized in that, The second analysis scheme specifically includes: Establish a baseline pattern for parameter fluctuations in a pre-built database. Specifically, the baseline pattern for parameter fluctuations is the synergistic pattern between parameter fluctuations and related influencing parameters. The fluctuation pattern of the waveform of the parameter to be analyzed is compared with the parameter fluctuation benchmark pattern to determine whether it conforms to the parameter fluctuation benchmark pattern. If it conforms, it is determined to be a normal fluctuation; if it does not conform, it is determined to be an anomaly, and the anomaly confirmation sub-strategy is executed. The simulation module is used to reproduce the waveform of the simulated parameter of the parameter to be analyzed, which is determined to be fluctuating normally. The deviation is compared with the waveform of the parameter to be analyzed. If the waveform deviation is determined to be lower than the deviation threshold, the passive monitoring mode is switched; otherwise, the third analysis scheme is executed.

5. The multi-mode switching monitoring system based on level position analysis according to claim 4, characterized in that, The specific details of the anomaly confirmation sub-strategy are as follows: Feature extraction is performed on the waveforms of the parameters to be analyzed that do not conform to the fluctuation benchmark law to generate anomaly feature vectors; The associated influencing parameters are adjusted based on the active monitoring mode to obtain the waveform difference before and after adjustment, and the abnormal response intensity is calculated. If the abnormal response intensity is higher than the response threshold, the third analysis scheme will be executed based on the abnormal response intensity.

6. The multi-mode switching monitoring system based on level position analysis according to claim 5, characterized in that, The third analysis scheme is as follows: The pre-built database pre-stores combinations of anomaly types and corresponding combined collaborative waveform features, as well as combined association influence parameters associated with each group of anomaly types. The characteristics of the waveform to be analyzed are compared with the characteristics of the combined and coordinated waveforms, and the combination of anomaly types corresponding to the maximum similarity is used as the criterion for judgment. If the similarity is greater than the third highest threshold, then the corresponding combination of anomaly types is determined; If the similarity is below the third lowest threshold, it is treated as a new anomaly type and stored. If the similarity is between the third high threshold and the third low threshold, the characteristics of the amplified waveform to be analyzed are obtained by adjusting the corresponding combined correlation influence parameters. The characteristics are then compared again with the combined collaborative waveform features corresponding to the abnormal type combinations in the database. If the similarity of all abnormal type combinations is lower than the third low threshold, it is determined to be a new abnormal type and written into the database.

7. The multi-mode switching monitoring system based on level position analysis according to claim 6, characterized in that, The passive monitoring mode is specifically as follows: The data monitoring module acquires actual parameters according to a preset benchmark sampling frequency, summarizes and records the actual parameters at intervals, collects real-time waveforms of regular parameters, and generates log reports periodically based on historical records.

8. The multi-mode switching monitoring system based on level position analysis according to claim 6, characterized in that, The active monitoring mode is specifically as follows: The active monitoring mode specifically includes scenarios that only collect data without intervention and scenarios that collect data and intervene. The scenario of collecting data without intervention specifically refers to the following: the data monitoring module collects parameters at a sampling frequency higher than the benchmark, records the parameters point by point, and records the real-time waveform of the parameters; The specific acquisition and intervention scenario is as follows: when the triggering condition is met, the intervention operation is initiated. The intervention operation is carried out by adjusting the associated influencing parameters and simultaneously acquiring parameters at a sampling frequency higher than the benchmark, recording the parameters point by point, and recording the real-time waveform of the parameters. The triggering conditions include the preliminary determination of the anomaly type in the first analysis scheme, the execution of the anomaly confirmation sub-strategy, and the analysis parameters whose similarity is between the third high threshold and the third low threshold in the third analysis scheme.

9. The multi-mode switching monitoring system based on level position analysis according to claim 1, characterized in that, The data monitoring module also includes a torque and speed sensor, a piezoelectric acceleration sensor, a current sensor, a voltage sensor, and a temperature sensor, specifically: The input terminal of the torque and speed sensor is connected to the output terminal of the damper, and the output terminal is connected to the load to obtain the torque and speed parameters of the motor operation; the piezoelectric acceleration sensor is installed on the motor to obtain vibration parameters; the current sensor and the voltage sensor are installed in the power supply circuit of the motor to obtain voltage and current parameters; the temperature sensor is built into the motor to obtain the internal temperature parameters of the motor. And the power parameters of the motor are calculated using the voltage and current parameters of the motor.

Images