Permanent magnet synchronous host bearing state detection method and system

Abstract

The invention relates to a permanent magnet synchronous host bearing state detection method and system, and the method comprises the steps: 1, enabling a sharp end of a special measurement rod to directly contact with the surface of a host bearing end cover, and collecting an audio signal and a sound wave signal in the operation of a bearing in real time; secondly, the elevator is made to operate under the three different load working conditions of full load, 50% load and no load, and bearing operation signals under all the working conditions are collected; and step 3, transmitting the collected signals to an intelligent AI analysis platform for preprocessing and feature extraction. According to the invention, the intelligent decision-making system can generate a gradient maintenance scheme considering economy and reliability based on an optimization algorithm of reinforcement learning, effectively prolongs the service life of the bearing, reduces the maintenance cost, guarantees the data safety and traceability through the application of the block chain technology, achieves the collaborative diagnosis and knowledge sharing of cross-brand equipment, and improves the reliability of the bearing. The method has remarkable advantages in the aspects of improving the equipment reliability, optimizing the maintenance strategy, reducing the operation and maintenance cost and the like.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of permanent magnet synchronous host bearing detection, in particular to a permanent magnet synchronous host bearing state detection method and system. BACKGROUND

[0002] In the operation process of an elevator permanent magnet synchronous host, the bearing as a key transmission component directly affects the safety and service life of the equipment, the traditional bearing detection method mainly relies on regular manual inspection or offline vibration analysis, and has defects such as long detection cycle, poor real-time performance, and easy to be affected by subjective factors, although the existing online monitoring system based on a vibration sensor improves the detection frequency, due to the limited installation position of the sensor, it is difficult to accurately capture the internal fault characteristics of the bearing, and is easily affected by electromagnetic interference and environmental noise, in addition, the conventional signal processing method has insufficient recognition ability for early weak fault characteristics, and the detection data under a single working condition is difficult to comprehensively reflect the actual wear state of the bearing. SUMMARY

[0003] In order to solve the above problems, the application provides a permanent magnet synchronous host bearing state detection method and system.

[0004] The permanent magnet synchronous host bearing state detection method and system provided by the application adopt the following technical scheme: A permanent magnet synchronous host bearing state detection method, comprising the following steps: Step one, directly contact the surface of the host bearing end cover through the sharp end of the special measuring rod, and collect audio signals and sound wave signals in real time when the bearing is running; Step two, make the elevator run under three different load working conditions of full load, 50% load and no load, and collect the bearing running signals under each working condition; Step three, transmit the collected signals to an intelligent AI analysis platform for preprocessing and feature extraction; Step four, analyze the extracted signal features by using a pre-established bearing wear degree evaluation model, and judge the wear state of the current bearing; Step five, combine the service life of the elevator and the historical detection data to predict the remaining service life of the bearing and generate maintenance suggestions.

[0005] As a preferred technical scheme of the application, the special measuring rod is made of iron rod material, one end of the special measuring rod is processed into a sharp shape for contacting the bearing end cover, and the other end of the special measuring rod is provided with a high-power information collector, the information collector is used for collecting sound wave signals and audio signals and has anti-electromagnetic interference capability.

[0006] As a preferred technical solution of the present application, in step three, more specifically, the collected original signal is subjected to band-pass filtering processing to filter out 50Hz power frequency interference and environmental noise, the effective signal is subjected to amplification processing with gain controlled within 20-40dB range, the signal is subjected to time domain and frequency domain analysis to extract at least twelve characteristic parameters including peak frequency, harmonic component and envelope characteristics, and the wavelet transform is used to eliminate transient noise in the signal to improve the accuracy of feature extraction.

[0007] As a preferred technical solution of the present application, the bearing wear degree evaluation model is constructed by using a deep convolutional neural network, the deep convolutional neural network comprises five convolutional layers, three pooling layers and two fully connected layers, the input layer receives the preprocessed signal characteristic parameters, and the output layer provides a five-level classification result of the bearing wear degree, and a data set comprising more than one thousand bearing samples with different wear degrees is used for model training, and the adaptability of the model to a specific elevator model is optimized by using a transfer learning technology.

[0008] As a preferred technical solution of the present application, a permanent magnet synchronous host bearing state detection system based on the permanent magnet synchronous host bearing state detection method comprises a signal acquisition module, a signal transmission module electrically connected to the output end of the signal acquisition module, an intelligent analysis module electrically connected to the output end of the signal transmission module, a result display module electrically connected to the output end of the intelligent analysis module, and a database module electrically connected to the output end of the result display module.

[0009] As a preferred technical solution of the present application, the signal acquisition module is composed of a special measuring rod and a high-power information collector, the signal acquisition module is used for collecting sound waves and audio signals of the bearing operation, the signal transmission module adopts a Wi-Fi wireless transmission mode, the transmission rate of the signal transmission module is not less than 100Mbps, the signal transmission module supports real-time data uploading, the intelligent analysis module internally loads a signal preprocessing unit, a feature extraction unit, an adaptive learning unit and an AI analysis unit, the result display module is used for displaying the bearing wear state evaluation result and the remaining life prediction, and the database module is used for storing historical detection data, standard signal characteristics and bearing wear evaluation models.

[0010] As a preferred technical solution of the present application, the adaptive learning unit dynamically optimizes the parameter weight of the bearing wear evaluation model by continuously collecting actual bearing replacement case data, the adaptive learning unit internally integrates a correction coefficient library based on the device operating environment, the correction coefficient library internally stores a machine room temperature influence coefficient, an operation frequency compensation coefficient and a device aging attenuation coefficient, and the adaptive learning unit applies the correction coefficient to the detection result calibration in real time through a multi-dimensional data fusion technology.

[0011] As a preferred technical solution of the present application, the output end of the intelligent analysis module is further electrically connected with an intelligent decision module, the intelligent decision module is integrated with a maintenance strategy optimization algorithm based on reinforcement learning, the maintenance strategy optimization algorithm can automatically generate an optimal maintenance scheme according to multi-dimensional data such as bearing wear degree prediction results, elevator use frequency, and maintenance resource distribution, the maintenance scheme includes recommending a non-contact lubrication maintenance scheme for a slight wear state, formulating a periodic vibration monitoring plan for a moderate wear state, and automatically booking spare parts and planning a shortest downtime maintenance path for a severe wear state, and the intelligent decision module simultaneously supports maintenance cost prediction and risk assessment functions and provides a quantitative basis for decision-making.

[0012] As a preferred technical solution of the present application, the output end of the database module is electrically connected with a cloud collaborative detection network, the cloud collaborative detection network realizes distributed storage and tamper-proof verification of detection data through blockchain technology, the cloud collaborative detection network supports standardized access and cross-platform analysis of multi-elevator brand data, and the cloud collaborative detection network is internally loaded with an expert consultation function module, and the cloud collaborative detection network can automatically initiate multi-party video consultation and synchronously share detection waveforms and characteristic maps when detecting abnormal data.

[0013] In summary, the present application includes at least one of the following beneficial technical effects of the permanent magnet synchronous host bearing state detection method and system: In terms of detection accuracy, the present application realizes high-sensitivity acquisition of micron-level vibration characteristics through the optimized design of a specially designed measuring rod and multi-modal signal fusion technology, significantly improves the early fault recognition capability, the multi-working-condition excitation strategy and the adaptive feature extraction algorithm used by the system can effectively separate mechanical and electromagnetic mixed fault features, significantly reducing the false positive rate and the missed detection rate, the intelligent analysis module combines deep convolutional neural networks with transfer learning, enabling the model to have cross-model generalization capability, and the environment parameter correction function of the adaptive learning unit ensures the detection stability under different operating conditions, the intelligent decision system based on the optimization algorithm of reinforcement learning can generate a gradient maintenance scheme that takes into account economy and reliability, effectively extending the service life of the bearing and reducing maintenance costs, the application of blockchain technology not only ensures data security and traceability, but also enables collaborative diagnosis and knowledge sharing across brand devices, the present application has significant advantages in improving equipment reliability, optimizing maintenance strategies, and reducing operation and maintenance costs. BRIEF DESCRIPTION OF DRAWINGS

[0014] Figure 1 is a flow chart of the permanent magnet synchronous host bearing state detection method of the present application; Figure 2 is a distribution diagram of the permanent magnet synchronous host bearing state detection system of the present application. DETAILED DESCRIPTION

[0015] The following description will be made in conjunction with the accompanying drawings Figures 1-2 The application is further described in detail.

[0016] Referring to Figure 1 A permanent magnet synchronous main bearing state detection method, comprising the following steps: Step one, through the sharp end of the special measuring rod directly contacting the surface of the main bearing end cover, real-time acquisition of the audio signal and sound wave signal of the bearing running; the special measuring rod is made of iron rod material, one end of the special measuring rod is processed into a sharp shape for contacting the bearing end cover, and the other end of the special measuring rod is installed with a high-power information collector, which is used for collecting sound wave signals and audio signals and has anti-electromagnetic interference capability; The measuring rod of iron rod material utilizes its high elastic modulus characteristics to efficiently conduct the broadband mechanical vibration wave generated on the surface of the bearing end cover to the information collector, the sharp end design improves the local pressure by reducing the contact area, ensures the contact sensitivity of micron-level vibration, and avoids sliding friction noise interference at the same time, the length of the measuring rod is optimized through acoustic resonance to suppress the standing wave effect of a specific frequency band, ensuring the linear transmission of 20Hz-20kHz sound wave signals, and the high-power information collector adopts an electromagnetic shielding cavity design, internally integrating a MEMS acceleration sensor and a capacitive microphone array, respectively capturing structure-conducted sound and air noise, eliminating common-mode electromagnetic interference through a differential amplification circuit to realize the improvement of signal-to-noise ratio; Step two, make the elevator run under three different load conditions of full load, 50% load and no load respectively, and collect the bearing running signals under each condition; Through the dynamic switching of the elevator load condition to excite the potential fault characteristics of the bearing, the radial force of the bearing increases under the full load condition, causing the characteristic harmonics of the raceway fatigue crack, and the high-speed running exposes the high-frequency resonance caused by the cage looseness under no load, and the 50% load excites the non-integer frequency component caused by the slip of the steel ball, the system automatically synchronously collects the motor current signal and vibration waveform, uses the current ripple to back-propagate the electromagnetic excitation force, separates the mechanical and electromagnetic mixed fault components, and uses a load adaptive weighting algorithm in the feature extraction stage to normalize and fuse the time domain peak-to-peak value, frequency domain sideband energy and other parameters of the three conditions, enhancing the saliency of weak fault characteristics; Step three, the collected signals are transmitted to an intelligent AI analysis platform for preprocessing and feature extraction; the collected original signals are subjected to band-pass filtering processing and filtering of 50Hz power frequency interference and environmental noise, the effective signals are subjected to amplification processing and the gain is controlled within the range of 20-40dB, the signals are subjected to time domain and frequency domain analysis and at least twelve characteristic parameters including peak frequency, harmonic component and envelope feature are extracted, the transient noise in the signal is eliminated through wavelet transform, and the accuracy of feature extraction is improved; Step four, the extracted signal features are analyzed by using a pre-established bearing wear degree evaluation model to judge the wear state of the current bearing; the bearing wear degree evaluation model is constructed by using a deep convolutional neural network, the deep convolutional neural network comprises five convolutional layers, three pooling layers and two fully connected layers, the input layer receives the preprocessed signal feature parameters, and the output layer provides a five-level classification result of the bearing wear degree, and a data set comprising more than one thousand bearing samples with different wear degrees is used for model training, and a migration learning technology is used to optimize the adaptability of the model to a specific elevator model; The wavelet packet analysis and the convolution denoising network are cascaded, the first stage adopts Db8 wavelet basis for 6-layer decomposition, the transient impact noise is eliminated by an adaptive threshold algorithm, and the sudden waveform containing fault information is retained, the second stage inputs the pretreated signal into a 1D convolution autoencoder, the encoder captures long-period characteristics through dilated convolution, and the decoder reconstructs the pure signal by using a skip connection, and the joint architecture can effectively separate the modulation sideband caused by millimeter-level defects; Step five, the remaining service life of the bearing is predicted in combination with the service life of the elevator and the historical detection data, and a maintenance suggestion is generated.

[0017] Referring to Figure 2 A permanent magnet synchronous host bearing state detection system based on a permanent magnet synchronous host bearing state detection method, comprising a signal acquisition module, an output end of the signal acquisition module being electrically connected with a signal transmission module, an output end of the signal transmission module being electrically connected with an intelligent analysis module, an output end of the intelligent analysis module being electrically connected with a result display module, and an output end of the result display module being electrically connected with a database module; the signal acquisition module is composed of a special measuring rod and a high-power information collector, the signal acquisition module is used for collecting sound waves and audio signals of bearing operation, the signal transmission module adopts a Wi-Fi wireless transmission mode, the transmission rate of the signal transmission module is not less than 100 Mbps, the signal transmission module supports real-time data uploading, the intelligent analysis module is internally loaded with a signal preprocessing unit, a feature extraction unit, an adaptive learning unit and an AI analysis unit, the result display module is used for displaying bearing wear state evaluation results and remaining life prediction, and the database module is used for storing historical detection data, standard signal features and bearing wear evaluation models; The special measuring rod is made of iron rod material, the high mechanical impedance characteristic of which ensures that the vibration signal of the bearing end cover can be efficiently conducted to the high-power information collector, and the sharp end design effectively suppresses the contact noise interference; the information collector integrates a piezoelectric acceleration sensor and a wideband microphone, respectively captures the bearing structure vibration and air noise, and eliminates common-mode electromagnetic interference through differential amplification technology; the signal transmission module adopts low-delay Wi-Fi communication to ensure that the high-frequency vibration waveform is not distorted during transmission, and supports multi-channel synchronous uploading to provide a high-fidelity data basis for subsequent intelligent analysis; the adaptive learning unit dynamically adjusts the filtering parameters by continuously optimizing the signal preprocessing algorithm to adapt to the differences in bearing vibration characteristics of different elevator models, thereby improving the detection rate of weak fault signals.

[0018] The adaptive learning unit dynamically optimizes the parameter weights of the bearing wear evaluation model by continuously collecting actual bearing replacement case data; the adaptive learning unit is internally integrated with a correction coefficient library based on the device operating environment, which internally stores the machine room temperature influence coefficient, the running frequency compensation coefficient and the device aging attenuation coefficient; the adaptive learning unit applies the correction coefficients to the detection result calibration in real time through multi-dimensional data fusion technology; The bearing wear degree evaluation model is constructed by using a deep convolutional neural network, the architecture of which is optimized for typical features of bearing faults; the input layer receives preprocessed multi-dimensional signal features, including time domain peak value, frequency domain harmonic component and envelope spectrum features; the convolution layer adapts to signal distortion caused by different installation methods through deformable convolution kernels; the adaptive learning unit dynamically corrects the model weights in combination with device operating environment parameters such as machine room temperature and running frequency, thereby improving the evaluation accuracy; the intelligent analysis module uses transfer learning technology to optimize the generalization ability of the model on new models using historical detection data, thereby ensuring the consistency of bearing state evaluation for different brands of elevators; when abnormal features are detected, the system automatically triggers expert knowledge base comparison and generates confidence evaluation in combination with the historical case library, thereby providing a reliable basis for maintenance decision-making.

[0019] The output end of the intelligent analysis module is also electrically connected to an intelligent decision-making module, which integrates a maintenance strategy optimization algorithm based on reinforcement learning; the maintenance strategy optimization algorithm can automatically generate an optimal maintenance plan based on multi-dimensional data such as bearing wear degree prediction results, elevator usage frequency and maintenance resource distribution; the maintenance plan includes recommending a non-contact lubrication maintenance plan for slight wear state, formulating a periodic vibration monitoring plan for moderate wear state, and automatically reserving spare parts and planning the shortest downtime maintenance path for severe wear state; the intelligent decision-making module also supports maintenance cost prediction and risk assessment functions and provides quantitative basis for decision-making; The intelligent decision-making module uses reinforcement learning algorithms to construct a dynamic maintenance strategy optimization model. The system maps parameters such as bearing wear, elevator operating load, and maintenance resource distribution into a multi-dimensional state space. Through the maintenance strategy optimization algorithm, the optimal decision strategy is trained in a simulated environment. For minor wear, the system recommends a non-contact magnetohydrodynamic lubrication solution to reduce downtime. For moderate wear, the system automatically adjusts the vibration monitoring cycle and dynamically optimizes the detection frequency based on real-time data. In the case of severe wear, the system links with the spare parts inventory database to intelligently plan the shortest repair path and coordinate the maintenance team simultaneously. The decision-making module also integrates a risk assessment engine to quantitatively analyze the cost-effectiveness ratio of different maintenance solutions, generate solutions, and assist managers in formulating scientific maintenance plans.

[0020] The output of the database module is electrically connected to a cloud-based collaborative detection network. The cloud-based collaborative detection network uses blockchain technology to achieve distributed storage and tamper-proof verification of detection data. The cloud-based collaborative detection network supports standardized access and cross-platform analysis of data from multiple elevator brands. The cloud-based collaborative detection network is equipped with an expert consultation function module. When the cloud-based collaborative detection network detects abnormal data, it can automatically initiate multi-party video consultations and synchronously share detection waveforms and feature maps. By leveraging blockchain technology, distributed storage and trusted verification of test data are achieved. All collected bearing vibration data are encrypted and uploaded to the distributed storage system, and data integrity is automatically verified through smart contracts. When abnormal waveforms are detected, the system triggers a multi-party expert consultation mechanism. Based on the expert consultation function module, real-time waveform sharing and collaborative diagnosis are realized. After the diagnostic opinions are confirmed by the consensus mechanism, they are written into the immutable blockchain log. The network supports standardized access to elevator data from different brands and achieves secure fusion and analysis of multi-source data through federated learning technology to avoid the leakage of sensitive data. Historical test data and maintenance records form a traceable and complete lifecycle archive, providing high-quality data support for bearing reliability research.

[0021] In this application, at the signal acquisition level, a specially designed measuring rod made of iron rod material achieves high-fidelity transmission of broadband vibration signals through an optimized sharp contact end. Combined with the multimodal sensing capabilities of an integrated information acquisition unit, it effectively integrates structural vibration and air noise characteristics. Electromagnetic shielding and differential amplification technologies ensure the reliability of signal acquisition. The system employs a multi-condition excitation strategy, using dynamic changes in elevator load to excite bearing fault characteristics under different stress states. Combined with synchronous analysis of motor current signals, it achieves accurate separation of mechanical and electromagnetic faults. The signal processing stage innovatively adopts a cascaded architecture of wavelet packet analysis and convolutional noise reduction networks, preserving the transient characteristics of fault signals while enhancing the identification of weak defects through deep learning algorithms. The bearing wear assessment model uses a deep convolutional neural network architecture, employing deformable convolution... The system adapts to signal distortion caused by different installation methods and introduces transfer learning technology to enhance the model's generalization performance across different models. The adaptive learning unit continuously collects actual maintenance case data, dynamically optimizes model parameters, and integrates environmental correction coefficients, significantly improving diagnostic accuracy under different operating conditions. The intelligent decision-making module constructs a multi-dimensional state space based on reinforcement learning algorithms, which can automatically generate gradient solutions from preventive lubrication to emergency repair based on variables such as bearing health status, usage intensity, and maintenance resources, and quantitatively evaluate the cost-effectiveness ratio of each solution. The cloud-based collaborative detection network uses blockchain technology to build a distributed trust mechanism, which not only ensures the immutability of detection data but also enables collaborative analysis of cross-brand data through federated learning. The real-time interactive capability of the expert consultation function greatly improves the diagnostic efficiency of complex faults.

[0022] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A method for detecting the condition of a permanent magnet synchronous motor bearing, characterized in that: Includes the following steps: Step 1: The sharp end of a specially designed measuring rod is used to directly contact the surface of the bearing end cover of the main unit to collect audio and sound wave signals during bearing operation in real time. Step 2: Run the elevator under three different load conditions: full load, 50% load, and no load, and collect the bearing operation signals under each condition. Step 3: The collected signals are transmitted to the intelligent AI analysis platform for preprocessing and feature extraction; Step 4: Analyze the extracted signal features using a pre-established bearing wear assessment model to determine the current wear state of the bearing. Step 5: Based on the elevator's service life and historical inspection data, predict the remaining service life of the bearings and generate maintenance recommendations.

2. The method for detecting the condition of a permanent magnet synchronous motor bearing according to claim 1, characterized in that: The specially designed measuring rod is made of iron. One end of the rod is machined into a sharp shape to contact the bearing end cap, and the other end is equipped with a high-power information acquisition device. The information acquisition device is used to collect sound wave signals and audio signals and has anti-electromagnetic interference capabilities.

3. The method for detecting the condition of a permanent magnet synchronous motor bearing according to claim 1, characterized in that: In step three, more specifically, the acquired raw signal is bandpass filtered to remove 50Hz power frequency interference and environmental noise, the effective signal is amplified and the gain is controlled within the range of 20-40dB, the signal is analyzed in the time and frequency domains and at least twelve feature parameters, including peak frequency, harmonic components, and envelope features, are extracted, and transient noise in the signal is eliminated by wavelet transform, thereby improving the accuracy of feature extraction.

4. The method for detecting the condition of a permanent magnet synchronous motor bearing according to claim 1, characterized in that: The bearing wear assessment model is constructed using a deep convolutional neural network, which contains five convolutional layers, three pooling layers, and two fully connected layers. The input layer receives preprocessed signal feature parameters, and the output layer provides a five-level classification result of the bearing wear degree. During model training, a dataset containing more than 1,000 bearing samples with different wear degrees is used, and the adaptability of the model to specific elevator models is optimized through transfer learning techniques.

5. A permanent magnet synchronous motor bearing condition detection system, based on the permanent magnet synchronous motor bearing condition detection method according to any one of claims 1-4, characterized in that: It includes a signal acquisition module, the output of which is electrically connected to a signal transmission module, the output of which is electrically connected to an intelligent analysis module, the output of which is electrically connected to a result display module, and the output of which is electrically connected to a database module.

6. The permanent magnet synchronous motor bearing condition detection system according to claim 5, characterized in that: The signal acquisition module consists of a specially designed measuring rod and a high-power information acquisition device. The signal acquisition module is used to acquire the sound waves and audio signals of the bearing operation. The signal transmission module adopts Wi-Fi wireless transmission and has a transmission rate of no less than 100Mbps. The signal transmission module supports real-time data upload. The intelligent analysis module is equipped with a signal preprocessing unit, a feature extraction unit, an adaptive learning unit, and an AI analysis unit. The result display module is used to display the bearing wear condition assessment results and remaining life prediction. The database module is used to store historical detection data, standard signal features, and bearing wear assessment models.

7. The permanent magnet synchronous motor bearing condition detection system according to claim 6, characterized in that: The adaptive learning unit dynamically optimizes the parameter weights of the bearing wear assessment model by continuously collecting actual bearing replacement case data. The adaptive learning unit integrates a correction coefficient library based on the equipment operating environment. The correction coefficient library stores the influence coefficient of machine room temperature, the compensation coefficient of operating frequency, and the equipment aging attenuation coefficient. The adaptive learning unit applies the correction coefficients to the test results calibration in real time through multi-dimensional data fusion technology.

8. The permanent magnet synchronous motor bearing condition detection system according to claim 6, characterized in that: The output of the intelligent analysis module is also electrically connected to an intelligent decision-making module. The intelligent decision-making module integrates a maintenance strategy optimization algorithm based on reinforcement learning. The maintenance strategy optimization algorithm can automatically generate the optimal maintenance plan based on multi-dimensional data such as bearing wear prediction results, elevator usage frequency, and maintenance resource distribution. The maintenance plan includes recommending a non-contact lubrication maintenance plan for minor wear, formulating a periodic vibration monitoring plan for moderate wear, and automatically reserving spare parts and planning the shortest downtime maintenance path for severe wear. The intelligent decision-making module also supports maintenance cost prediction and risk assessment functions and provides quantitative basis for decision-making.

9. The permanent magnet synchronous motor bearing condition detection system according to claim 6, characterized in that: The output of the database module is electrically connected to a cloud-based collaborative detection network. The cloud-based collaborative detection network uses blockchain technology to achieve distributed storage and tamper-proof verification of detection data. The cloud-based collaborative detection network supports standardized access and cross-platform analysis of data from multiple elevator brands. The cloud-based collaborative detection network is equipped with an expert consultation function module. When the cloud-based collaborative detection network detects abnormal data, it can automatically initiate multi-party video consultations and synchronously share detection waveforms and feature maps.

Images