Vibration stress monitoring method of wireless telemetering rotary transmission system

Abstract

The invention discloses a vibration stress monitoring method for a wireless telemetering rotary transmission system. The method comprises the following steps: S1, collecting multi-dimensional vibration acceleration data of the rotary transmission system in real time; s2, performing signal conditioning on the multi-dimensional vibration acceleration data output by the vibration acceleration sensor through a signal conditioning circuit; s3, generating a vibration signal data packet with a timestamp and a sensor identifier; s4, dynamically adjusting a data transmission frequency band according to the environment interference condition; s5, automatically generating a normal working state reference model of the rotary transmission system, wherein the normal working state reference model covers feature distribution of the multi-dimensional vibration acceleration data under different working conditions; s6, generating abnormal vibration characteristic data; and S7, storing the abnormal vibration characteristic data in a vibration data storage module of the mobile terminal, and sending an alarm signal to an operator through an alarm module. According to the invention, the defects of vibration signal acquisition, transmission, processing and environmental adaptability in the prior art are overcome.

Description

Technical Field

[0001] The invention relates to the technical field of vibration stress monitoring, and in particular to a vibration stress monitoring method for a wireless telemetry rotation transmission system. Background Art

[0002] In the prior art, vibration stress monitoring of rotating transmission systems mainly relies on wired sensors and traditional signal acquisition devices. Wired sensors and traditional signal acquisition devices realize signal transmission and power supply through physical connection and transmit the vibration data collected by the sensor to the stationary end for analysis. However, under high-speed rotation and complex working conditions, traditional wired monitoring technology has many limitations and it is difficult to meet the requirements of industrial equipment for real-time, high reliability and environmental adaptability.

[0003] Signal acquisition and transmission technology in rotating environments has long been limited by physical connections. Traditional wired sensors transmit vibration signals through slip rings or torsional cables. Traditional methods are prone to long-term use, which can cause cable wear, breakage or poor contact, leading to signal loss. In addition, the slip ring contact method is difficult to maintain a stable electrical connection during high-speed rotation, which may lead to a decrease in signal transmission quality and the inability to guarantee the accuracy and continuity of vibration monitoring data.

[0004] On the other hand, rotating transmission systems are usually in harsh environments with high temperature, high humidity, high dust or strong oil mist. The sealing and heat dissipation design of existing devices can hardly cope with these complex working conditions. The equipment is prone to failure due to overheating or sealing failure during long-term operation, which not only affects the service life of the monitoring equipment but also reduces the stability and reliability of vibration monitoring.

[0005] In addition, the existing technology also has deficiencies in data processing and model analysis. Traditional methods usually use a preset fixed threshold to determine whether the vibration is abnormal. However, due to the complex and dynamic working conditions of the rotating transmission system, the fixed threshold method is prone to false alarms or missed alarms and is difficult to adapt to the changing working environment. At the same time, the existing technology lacks the ability to perform self-learning analysis on vibration signals and cannot update the discrimination model in real time according to the actual working conditions, resulting in the limitation of the effect of predictive maintenance. Summary of the invention

[0006] An object of the present invention is to provide a vibration stress monitoring method for a wireless telemetry rotary transmission system, which solves the deficiencies of the prior art in vibration signal acquisition, transmission, processing and environmental adaptability.

[0007] A vibration stress monitoring method for a wireless telemetry rotating transmission system according to an embodiment of the present invention comprises the following steps: S1. Vibration acceleration sensors are evenly arranged in the circumferential direction of the flange of the rotating transmission system and fixed on the surface of the transmission shaft to collect multi-dimensional vibration acceleration data of the rotating transmission system in real time. The multi-dimensional vibration acceleration data is transmitted to the signal conditioning circuit of the vibration signal monitoring device through a radio frequency connection line; S2, performing signal conditioning on the multi-dimensional vibration acceleration data output by the vibration acceleration sensor through a signal conditioning circuit, and performing noise reduction processing on the multi-dimensional vibration acceleration data through a high-order low-pass filter; S3, using the embedded processor in the vibration signal monitoring device to encode the conditioned multi-dimensional vibration acceleration data to generate a vibration signal data packet with a timestamp and a sensor identifier; S4. The encoded vibration signal data packet is transmitted to the mobile terminal connected thereto through the wireless communication module of the vibration signal monitoring device. The wireless communication module adopts an adaptive frequency band switching mechanism to dynamically adjust the data transmission frequency band according to the environmental interference situation. S5. The mobile terminal decodes the received vibration signal data packet, extracts the vibration acceleration data, analyzes the long-term accumulated vibration acceleration data through an artificial intelligence self-learning algorithm, and automatically generates a normal working state benchmark model of the rotary transmission system. The normal working state benchmark model covers the characteristic distribution of multi-dimensional vibration acceleration data under different working conditions; S6. Compare and analyze the real-time transmitted vibration signal data one by one based on the normal working state benchmark model, update the model parameters through the self-learning algorithm, and adapt to the changes in the operating conditions of the rotation transmission system. When the comparison result shows that the vibration signal data exceeds the setting range of the benchmark model, it is determined to be abnormal vibration, and abnormal vibration characteristic data is generated; S7. The abnormal vibration characteristic data is stored in the vibration data storage module of the mobile terminal, and an alarm signal is sent to the operator through the alarm module.

[0008] Optionally, the S1 includes the following steps: S11. Arrange multiple vibration acceleration sensors evenly in the circumferential direction of the flange of the rotating transmission system, and fix them to the surface of the transmission shaft by bolts, so that the vibration acceleration sensors rotate synchronously with the transmission shaft; S12, each vibration acceleration sensor collects vibration acceleration data of the rotating transmission system in real time during operation, wherein the vibration acceleration data includes lateral vibration acceleration , longitudinal vibration acceleration and circumferential vibration acceleration : Lateral vibration acceleration , represents the vibration acceleration in the radial direction along the flange plane; Longitudinal vibration acceleration , represents the vibration acceleration along the axial direction of the transmission shaft; Circumferential vibration acceleration , represents the vibration acceleration along the circumferential direction of the transmission shaft; in, , , are the vibration forces in the corresponding directions respectively, and m is the local mass of the transmission shaft; S13, each vibration acceleration sensor collects the lateral vibration acceleration , longitudinal vibration acceleration and circumferential vibration acceleration Converted into electrical signals and transmitted in real time to the signal conditioning circuit of the vibration signal monitoring device through a radio frequency connection line; S14. The installation position of the vibration acceleration sensor must meet the following symmetry conditions: If n vibration acceleration sensors are arranged in the circumferential direction of the flange, the angular interval of the sensor positions needs to be ; The sensors need to be arranged symmetrically in pairs in the semicircular direction of the flange to eliminate the coupling effect of gravity acceleration g on the vibration acceleration data; S15. Combined with the installation error of the vibration acceleration sensor, the following error compensation model is established: ; in, is the vibration acceleration data after compensation, is the uncompensated vibration acceleration data, is the installation error compensation value; S16. Verify the stability of the sensor output signal. Regularly verify the signal amplitude variation range to ensure that the vibration acceleration data collected by the sensor meets the following range under normal working conditions: ; in, and They are the minimum and maximum values ​​of the vibration acceleration data respectively.

[0009] Optionally, S2 includes the following steps: S21, preprocessing the multi-dimensional vibration acceleration data output by the vibration acceleration sensor through a signal conditioning circuit, and amplifying the multi-dimensional vibration acceleration data in the form of an electrical signal through a preamplifier circuit to an input range that matches the signal conditioning circuit; S22, using a high-order low-pass filter to perform noise reduction processing on the amplified multi-dimensional vibration acceleration data to remove high-frequency noise and obtain filtered vibration acceleration data; S23, adjusting the amplitude of the filtered multi-dimensional vibration acceleration data respectively, and adjusting the amplitude of each dimension of data to the input range of the analog-to-digital conversion module ,in: ; in, is the voltage value of the electrical signal in the corresponding direction, is the gain coefficient corresponding to the direction, which is adjusted according to the signal strength and the dynamic range of the analog-to-digital conversion module; S24, performing signal verification on the amplitude-adjusted multidimensional vibration acceleration data, the verification content includes signal integrity verification, judging whether there is data loss by calculating the verification of the data packet, and outputting the verified multidimensional vibration acceleration data to the data encoding module.

[0010] Optionally, S3 includes the following steps: S31, receiving the multi-dimensional vibration acceleration data after signal conditioning through an embedded processor, and the processor performs time synchronization marking on the received vibration acceleration data; S32. Assigning a sensor identifier to each set of vibration acceleration data , sensor identifier The vibration acceleration sensor used to uniquely identify the multi-dimensional vibration acceleration data satisfies the following relationship: ; in, is the sensor layout position on the flange, is the corresponding signal processing channel number; S34, vibration acceleration data, timestamp and sensor identifier Combination generation data recording unit , represents the kth sampling moment of sensor data collection; S35, multiple data recording units through embedded processor Perform data encoding to generate vibration signal data packets , the encoding rules are: ; in, is the preset encoding function; S36, check the integrity of the generated vibration signal data packet, through the check code Determine whether the data packet is lost or erroneous during the encoding process.

[0011] Optionally, S5 includes the following steps: S51, receiving vibration signal data packet via mobile terminal , using the preset decoding function Decode the data packet to extract multi-dimensional vibration acceleration data, timestamp and sensor identifier; S52, according to the timestamp and sensor identifier The extracted vibration acceleration data Classify and store in the corresponding time series database; S53, perform discrete Fourier transform on each dimension of vibration acceleration data to calculate its main frequency And spectrum characteristic distribution : ; in, is the vibration acceleration value corresponding to the nth sampling point, N represents the total number of sampling points, T is the sampling period, and f is the frequency component; The amplitude fluctuation range of the vibration acceleration data is statistically analyzed, and the segment probability is calculated as follows: ; in, Amplitude In the interval The probability distribution within, p(a) is the probability density function of the amplitude; Directional coupling of vibration acceleration data in different dimensions through mutual information calculate: ; in, for and The joint probability density of and are their respective marginal probability density functions; S54. Automatically generate a normal working state benchmark model of the rotating transmission system based on the extracted feature data : .

[0012] Optionally, the S6 comprises the following steps: S61, through the decoding function The vibration signal data packet transmitted in real time Decode to extract vibration acceleration data, timestamp and sensor identifier; S62. By analyzing the frequency domain of real-time data, the difference between the main frequency deviation and the power spectrum density is defined as: ; ; in, and are the main frequencies of real-time data and benchmark model respectively, and are the power spectral densities of real-time data and benchmark model, respectively. and are the minimum analysis frequency range and the maximum analysis frequency range respectively. When , it is judged as abnormal; S63. Calculate the amplitude deviation of the real-time vibration acceleration data and compare it with the normal range: ; in, is the amplitude of real-time vibration acceleration data, is the amplitude in the baseline model, when When , it is judged as abnormal; S64, by real-time calculation of the directional coupling mutual information of multi-dimensional vibration acceleration data, the difference is defined as follows: ; in, and are the joint probability densities of real-time data and the benchmark model, respectively. , it is judged as abnormal.

[0013] S65. When any comparison result in step S62, S63 or S64 exceeds the normal range, comprehensively record abnormal characteristic data; S66. Use the real-time abnormal feature data to update the baseline model through a self-learning algorithm.

[0014] The beneficial effects of the present invention are: (1) The present invention introduces an artificial intelligence self-learning algorithm, which combines the feature extraction of long-term vibration acceleration data to automatically generate and dynamically update the normal working state benchmark model. Through the real-time iterative update of the self-learning algorithm, the system can adapt to the dynamic changes in the operating conditions of the rotating transmission system, avoiding the defects of false alarms or missed alarms of the traditional fixed threshold method under complex working conditions. When the main frequency, power spectrum characteristics and directional coupling characteristics of the multi-dimensional vibration signal are slightly offset, the algorithm can quickly capture the anomaly and accurately determine the fault type. Compared with the traditional method, the accuracy of vibration anomaly detection is improved by more than 15%, and the model update time is reduced from minutes to seconds.

[0015] (2) The present invention realizes real-time processing and encoding of multi-dimensional vibration acceleration data through an embedded processor in the vibration signal monitoring device. By utilizing optimized encoding rules, the data packet has high traceability and compactness, which greatly reduces the redundancy in the data transmission process. It has been verified that the size of the generated vibration signal data packet is reduced by about 30%, but still retains complete vibration characteristic information, thereby improving wireless transmission efficiency and data reliability. In addition, combined with a multi-level verification mechanism, the data packet transmission error rate is reduced to below 0.1%, which significantly improves the stability of the system.

[0016] (3) The present invention adopts a sealed design and an optimized heat dissipation structure, which significantly enhances the operating stability of the vibration monitoring device in high temperature, high humidity, and strong oil mist environments. The wireless power supply and signal transmission technology completely eliminates the mechanical wear problem of the traditional slip ring power supply method, ensuring that the equipment can operate stably for a long time under high-speed rotation conditions. After testing, the present invention has been continuously operated for 120 hours under high temperature of 70°C and humidity of 95% without performance degradation; the failure rate in an oil mist environment is reduced to 0.05%. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 The present invention provides a flow chart of a vibration stress monitoring method for a wireless telemetry rotating transmission system. DETAILED DESCRIPTION

[0018] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0019] refer to Figure 1 , a vibration stress monitoring method for a wireless telemetry rotating transmission system comprises the following steps: S1. Vibration acceleration sensors are evenly arranged in the circumferential direction of the flange of the rotating transmission system and fixed on the surface of the transmission shaft to collect multi-dimensional vibration acceleration data of the rotating transmission system in real time. The multi-dimensional vibration acceleration data is transmitted to the signal conditioning circuit of the vibration signal monitoring device through a radio frequency connection line; S2, performing signal conditioning on the multi-dimensional vibration acceleration data output by the vibration acceleration sensor through a signal conditioning circuit, and performing noise reduction processing on the multi-dimensional vibration acceleration data through a high-order low-pass filter; S3, using the embedded processor in the vibration signal monitoring device to encode the conditioned multi-dimensional vibration acceleration data to generate a vibration signal data packet with a timestamp and a sensor identifier; S4. The encoded vibration signal data packet is transmitted to the mobile terminal connected thereto through the wireless communication module of the vibration signal monitoring device. The wireless communication module adopts an adaptive frequency band switching mechanism to dynamically adjust the data transmission frequency band according to the environmental interference situation. S5. The mobile terminal decodes the received vibration signal data packet, extracts the vibration acceleration data, analyzes the long-term accumulated vibration acceleration data through an artificial intelligence self-learning algorithm, and automatically generates a normal working state benchmark model of the rotary transmission system. The normal working state benchmark model covers the characteristic distribution of multi-dimensional vibration acceleration data under different working conditions; S6. Compare and analyze the real-time transmitted vibration signal data one by one based on the normal working state benchmark model, update the model parameters through the self-learning algorithm, and adapt to the changes in the operating conditions of the rotation transmission system. When the comparison result shows that the vibration signal data exceeds the setting range of the benchmark model, it is determined to be abnormal vibration, and abnormal vibration characteristic data is generated; S7. The abnormal vibration characteristic data is stored in the vibration data storage module of the mobile terminal, and an alarm signal is sent to the operator through the alarm module.

[0020] In this implementation, S1 includes the following steps: S11. Arrange multiple vibration acceleration sensors evenly in the circumferential direction of the flange of the rotating transmission system, and fix them to the surface of the transmission shaft by bolts, so that the vibration acceleration sensors rotate synchronously with the transmission shaft; S12. Each vibration acceleration sensor collects the vibration acceleration data of the rotating transmission system in real time during operation. The vibration acceleration data includes the lateral vibration acceleration. , longitudinal vibration acceleration and circumferential vibration acceleration : Lateral vibration acceleration , represents the vibration acceleration in the radial direction along the flange plane; Longitudinal vibration acceleration , represents the vibration acceleration along the axial direction of the transmission shaft; Circumferential vibration acceleration , represents the vibration acceleration along the circumferential direction of the transmission shaft; in, , , are the vibration forces in the corresponding directions respectively, and m is the local mass of the transmission shaft; S13, each vibration acceleration sensor collects the lateral vibration acceleration , longitudinal vibration acceleration and circumferential vibration acceleration Converted into electrical signals and transmitted in real time to the signal conditioning circuit of the vibration signal monitoring device through a radio frequency connection line; S14. The installation position of the vibration acceleration sensor must meet the following symmetry conditions: If n vibration acceleration sensors are arranged in the circumferential direction of the flange, the angular interval of the sensor positions needs to be ; The sensors need to be arranged symmetrically in pairs in the semicircular direction of the flange to eliminate the coupling effect of gravity acceleration g on the vibration acceleration data; S15. Combined with the installation error of the vibration acceleration sensor, the following error compensation model is established: ; in, is the vibration acceleration data after compensation, is the uncompensated vibration acceleration data, is the installation error compensation value; S16. Verify the stability of the sensor output signal. Regularly verify the signal amplitude variation range to ensure that the vibration acceleration data collected by the sensor meets the following range under normal working conditions: ; in, and They are the minimum and maximum values ​​of the vibration acceleration data respectively.

[0021] In this implementation, S2 includes the following steps: S21, preprocessing the multi-dimensional vibration acceleration data output by the vibration acceleration sensor through a signal conditioning circuit, and amplifying the multi-dimensional vibration acceleration data in the form of an electrical signal through a preamplifier circuit to an input range that matches the signal conditioning circuit; S22, using a high-order low-pass filter to perform noise reduction processing on the amplified multi-dimensional vibration acceleration data to remove high-frequency noise and obtain filtered vibration acceleration data; S23, adjusting the amplitude of the filtered multi-dimensional vibration acceleration data respectively, and adjusting the amplitude of each dimension of data to the input range of the analog-to-digital conversion module ,in: ; in, is the voltage value of the electrical signal in the corresponding direction, is the gain coefficient corresponding to the direction, which is adjusted according to the signal strength and the dynamic range of the analog-to-digital conversion module; S24, performing signal verification on the amplitude-adjusted multidimensional vibration acceleration data, the verification content includes signal integrity verification, judging whether there is data loss by calculating the verification of the data packet, and outputting the verified multidimensional vibration acceleration data to the data encoding module.

[0022] In this implementation, S3 includes the following steps: S31, receiving the multi-dimensional vibration acceleration data after signal conditioning through an embedded processor, and the processor performs time synchronization marking on the received vibration acceleration data; S32. Assigning a sensor identifier to each set of vibration acceleration data , sensor identifier The vibration acceleration sensor used to uniquely identify the multi-dimensional vibration acceleration data satisfies the following relationship: ; in, is the sensor layout position on the flange, is the corresponding signal processing channel number; S34, vibration acceleration data, timestamp and sensor identifier Combination generation data recording unit , represents the kth sampling moment of sensor data collection; S35, multiple data recording units through embedded processor Perform data encoding to generate vibration signal data packets , the encoding rules are: ; in, is the preset encoding function; S36, check the integrity of the generated vibration signal data packet, through the check code Determine whether the data packet is lost or erroneous during the encoding process.

[0023] In this implementation, S5 includes the following steps: S51, receiving vibration signal data packet via mobile terminal , using the preset decoding function Decode the data packet to extract multi-dimensional vibration acceleration data, timestamp and sensor identifier; S52, according to the timestamp and sensor identifier The extracted vibration acceleration data Classify and store in the corresponding time series database; S53, perform discrete Fourier transform on each dimension of vibration acceleration data to calculate its main frequency And spectrum characteristic distribution : ; in, is the vibration acceleration value corresponding to the nth sampling point, N represents the total number of sampling points, T is the sampling period, and f is the frequency component; The amplitude fluctuation range of the vibration acceleration data is statistically analyzed, and the segment probability is calculated as follows: ; in, Amplitude In the interval The probability distribution within, p(a) is the probability density function of the amplitude; Directional coupling of vibration acceleration data in different dimensions through mutual information calculate: ; in, for and The joint probability density of and are their respective marginal probability density functions; S54. Automatically generate a normal working state benchmark model of the rotating transmission system based on the extracted feature data : .

[0024] In this implementation, S6 includes the following steps: S61, through the decoding function The vibration signal data packet transmitted in real time Decode to extract vibration acceleration data, timestamp and sensor identifier; S62. By analyzing the frequency domain of real-time data, the difference between the main frequency deviation and the power spectrum density is defined as: ; ; in, and are the main frequencies of real-time data and benchmark model, and are the power spectral densities of real-time data and benchmark model, respectively. and are the minimum analysis frequency range and the maximum analysis frequency range respectively. When , it is judged as abnormal; S63. Calculate the amplitude deviation of the real-time vibration acceleration data and compare it with the normal range: ; in, is the amplitude of real-time vibration acceleration data, is the amplitude in the baseline model, when When , it is judged as abnormal; S64, by real-time calculation of the directional coupling mutual information of multi-dimensional vibration acceleration data, the difference is defined as follows: ; in, and are the joint probability densities of real-time data and the benchmark model, respectively. , it is judged as abnormal.

[0025] S65. When any comparison result in step S62, S63 or S64 exceeds the normal range, comprehensively record abnormal characteristic data; S66. Use the real-time abnormal feature data to update the baseline model through a self-learning algorithm.

[0026] Example 1: In a certain industrial equipment manufacturing plant, a key rotating transmission system in the mechanical production workshop is used to drive the large rollers on the production line. The transmission system has been in a high-speed rotation, high temperature, high humidity and strong oil mist environment for a long time. The production line has been shut down many times due to abnormal vibration, causing serious economic losses. The traditional vibration monitoring method uses wired sensors and fixed thresholds to judge abnormalities, but in high-speed rotation and complex environments, the wired slip ring has poor contact, signal loss and false alarms frequently occur, which cannot meet the needs of high reliability and real-time fault monitoring of the workshop production line.

[0027] To solve the above problems, this embodiment uses the wireless telemetry rotation transmission system vibration stress monitoring method of the present invention to perform comprehensive vibration monitoring and fault detection on the rotation transmission system of the workshop.

[0028] Eight vibration acceleration sensors are evenly installed on the circumference of the flange of the transmission shaft. Each sensor is fixed by bolts and rotates synchronously with the transmission shaft. To eliminate the gravity acceleration coupling effect that may be introduced by the asymmetric sensor arrangement, each pair of sensors is symmetrically arranged along the semicircle direction of the flange and the installation error is corrected by the built-in compensation model. The sensor collects the vibration acceleration signals of the transmission shaft in the lateral, longitudinal and circumferential directions in real time, and the sensor output signal is transmitted to the vibration signal monitoring device of the rotating part through the radio frequency connecting line.

[0029] The signal conditioning circuit integrated in the monitoring device amplifies the vibration signal and performs high-order low-pass filtering. After removing high-frequency noise, the amplitude range of the signal is adjusted to the input range of the analog-to-digital conversion module. The data after signal conditioning is received by the embedded processor, which encodes the vibration signal according to the timestamp and sensor identifier to generate a vibration signal data packet with a time series. The data packet is transmitted to the mobile terminal in the workshop control room through the wireless communication module.

[0030] The vibration signal data decoded on the mobile terminal is stored in the time series database. The system conducts offline training on the vibration signal data accumulated within three months, and uses artificial intelligence self-learning algorithm to extract signal features, including main frequency distribution, power spectrum density characteristics, amplitude fluctuation range and directional coupling, to generate a normal working state benchmark model of the rotating transmission system. The model shows that the main frequency range of the transmission system under normal working conditions is 23.5Hz to 25Hz, and the power spectrum characteristics show that the lateral vibration is mainly concentrated at 23.8Hz, the amplitude fluctuation range is ±0.15g (unit of gravity acceleration), and the directional coupling mutual information index is 0.85 to 0.92.

[0031] As the production line runs, the system enters real-time monitoring mode. The system compares the received real-time vibration data with the benchmark model one by one. One morning, the monitoring system found that the main frequency of the vibration signal shifted from the normal range of 23.8Hz to 26.3Hz, with a shift rate of 10.5%. The power spectrum density was obviously abnormal, and the vibration amplitude exceeded the normal fluctuation range by about 35% in the longitudinal direction. The directional coupling index dropped to 0.76, which was lower than the lower limit of the benchmark range. The system immediately determined that the transmission system was abnormal.

[0032] After analyzing the abnormal data through a self-learning algorithm, it was found that the significant increase in the amplitude of longitudinal vibration was mainly due to bearing wear. The system generated an abnormal feature report, recording in detail the time, location, feature deviation and judgment results of the abnormality, and pushed the information to the workshop operation and maintenance personnel through the alarm module. The maintenance personnel quickly located the problem and replaced the bearing, avoiding shutdown of the production line.

[0033] Compared with traditional methods, the performance of the present invention in this case is significantly superior. Traditional monitoring methods have experienced signal loss due to poor contact of the slip ring, resulting in 4 false alarms in the past 3 months. The method of the present invention achieves high reliability of data transmission under the guarantee of wireless transmission technology, with 0 false alarms. In addition, the fixed threshold judgment method of the traditional method cannot identify early fault signals, while the self-learning algorithm based on artificial intelligence of the present invention can dynamically update the benchmark model, and abnormal signals have been detected within 5 hours before the fault occurs, thus achieving early warning.

[0034] Experiments show that the monitoring system based on the present invention effectively solves the problems of traditional methods in vibration signal collection, processing and transmission through optimized algorithms and hardware designs, and significantly improves the reliability and accuracy of vibration monitoring in industrial environments.

[0035] The present invention introduces an artificial intelligence self-learning algorithm, which combines the feature extraction of long-term vibration acceleration data to automatically generate and dynamically update the normal working state benchmark model. Through the real-time iterative update of the self-learning algorithm, the system can adapt to the dynamic changes of the operating conditions of the rotating transmission system, and avoid the defects of false alarms or missed alarms of the traditional fixed threshold method under complex working conditions. When the main frequency, power spectrum characteristics and directional coupling characteristics of the multi-dimensional vibration signal are slightly offset, the algorithm can quickly capture the anomaly and accurately determine the fault type. Compared with traditional methods, the accuracy of vibration anomaly detection is improved by more than 15%, and the model update time is reduced from minutes to seconds.

[0036] The present invention realizes real-time processing and encoding of multi-dimensional vibration acceleration data through an embedded processor in a vibration signal monitoring device. By utilizing optimized encoding rules, the data packet has high traceability and compactness, which greatly reduces redundancy in the data transmission process. It has been verified that the size of the generated vibration signal data packet is reduced by about 30%, but still retains complete vibration characteristic information, thereby improving wireless transmission efficiency and data reliability. In addition, combined with a multi-level verification mechanism, the data packet transmission error rate is reduced to below 0.1%, which significantly improves the stability of the system.

[0037] The present invention adopts a sealing design and an optimized heat dissipation structure, which significantly enhances the operating stability of the vibration monitoring device in high temperature, high humidity, and strong oil mist environments. At the same time, the wireless power supply and signal transmission technology are used to completely eliminate the mechanical wear problem of the traditional slip ring power supply method, ensuring that the equipment can operate stably for a long time under high-speed rotation conditions. After testing, the present invention has been continuously operated for 120 hours under high temperature of 70°C and humidity of 95% without performance degradation; the failure rate in an oil mist environment is reduced to 0.05%.

[0038] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A vibration stress monitoring method for a wireless telemetry rotating transmission system, characterized in that: The steps include: S1. Vibration acceleration sensors are evenly arranged in the circumferential direction of the flange of the rotating transmission system and fixed on the surface of the transmission shaft to collect multi-dimensional vibration acceleration data of the rotating transmission system in real time. The multi-dimensional vibration acceleration data is transmitted to the signal conditioning circuit of the vibration signal monitoring device through a radio frequency connection line; S2, performing signal conditioning on the multi-dimensional vibration acceleration data output by the vibration acceleration sensor through a signal conditioning circuit, and performing noise reduction processing on the multi-dimensional vibration acceleration data through a high-order low-pass filter; S3, using the embedded processor in the vibration signal monitoring device to encode the conditioned multi-dimensional vibration acceleration data to generate a vibration signal data packet with a timestamp and a sensor identifier; S4. The encoded vibration signal data packet is transmitted to the mobile terminal connected thereto through the wireless communication module of the vibration signal monitoring device. The wireless communication module adopts an adaptive frequency band switching mechanism to dynamically adjust the data transmission frequency band according to the environmental interference situation. S5. The mobile terminal decodes the received vibration signal data packet, extracts the vibration acceleration data, analyzes the long-term accumulated vibration acceleration data through an artificial intelligence self-learning algorithm, and automatically generates a normal working state benchmark model of the rotary transmission system. The normal working state benchmark model covers the characteristic distribution of multi-dimensional vibration acceleration data under different working conditions; S6. Compare and analyze the real-time transmitted vibration signal data one by one based on the normal working state benchmark model, update the model parameters through the self-learning algorithm, and adapt to the changes in the operating conditions of the rotation transmission system. When the comparison result shows that the vibration signal data exceeds the setting range of the benchmark model, it is determined to be abnormal vibration, and abnormal vibration characteristic data is generated; S7. The abnormal vibration characteristic data is stored in the vibration data storage module of the mobile terminal, and an alarm signal is sent to the operator through the alarm module.

2. A vibration stress monitoring method for a wireless telemetry rotating transmission system according to claim 1, characterized in that: The S1 comprises the following steps: S11. Arrange multiple vibration acceleration sensors evenly in the circumferential direction of the flange of the rotating transmission system, and fix them to the surface of the transmission shaft by bolts, so that the vibration acceleration sensors rotate synchronously with the transmission shaft; S12, each vibration acceleration sensor collects vibration acceleration data of the rotating transmission system in real time during operation, wherein the vibration acceleration data includes lateral vibration acceleration , longitudinal vibration acceleration and circumferential vibration acceleration : Lateral vibration acceleration , represents the vibration acceleration in the radial direction along the flange plane; Longitudinal vibration acceleration , represents the vibration acceleration along the axial direction of the transmission shaft; Circumferential vibration acceleration , represents the vibration acceleration along the circumferential direction of the transmission shaft; in, , , are the vibration forces in the corresponding directions respectively, and m is the local mass of the transmission shaft; S13, each vibration acceleration sensor collects the lateral vibration acceleration , longitudinal vibration acceleration and circumferential vibration acceleration Converted into electrical signals and transmitted in real time to the signal conditioning circuit of the vibration signal monitoring device through a radio frequency connection line; S14. The installation position of the vibration acceleration sensor must meet the following symmetry conditions: If n vibration acceleration sensors are arranged in the circumferential direction of the flange, the angular interval of the sensor positions needs to be ; The sensors need to be arranged symmetrically in pairs in the semicircular direction of the flange to eliminate the coupling effect of gravity acceleration g on the vibration acceleration data; S15. Combined with the installation error of the vibration acceleration sensor, the following error compensation model is established: ； in, is the vibration acceleration data after compensation, is the uncompensated vibration acceleration data, is the installation error compensation value; S16. Verify the stability of the sensor output signal. Regularly verify the signal amplitude variation range to ensure that the vibration acceleration data collected by the sensor meets the following range under normal working conditions: ； in, and They are the minimum and maximum values ​​of the vibration acceleration data respectively.

3. A vibration stress monitoring method for a wireless telemetry rotating transmission system according to claim 1, characterized in that: The S2 comprises the following steps: S21, preprocessing the multi-dimensional vibration acceleration data output by the vibration acceleration sensor through a signal conditioning circuit, and amplifying the multi-dimensional vibration acceleration data in the form of an electrical signal through a preamplifier circuit to an input range that matches the signal conditioning circuit; S22, using a high-order low-pass filter to perform noise reduction processing on the amplified multi-dimensional vibration acceleration data to remove high-frequency noise and obtain filtered vibration acceleration data; S23, adjusting the amplitude of the filtered multi-dimensional vibration acceleration data respectively, and adjusting the amplitude of each dimension of data to the input range of the analog-to-digital conversion module ,in: ； in, is the voltage value of the electrical signal in the corresponding direction, is the gain coefficient corresponding to the direction, which is adjusted according to the signal strength and the dynamic range of the analog-to-digital conversion module; S24, performing signal verification on the amplitude-adjusted multidimensional vibration acceleration data, the verification content includes signal integrity verification, judging whether there is data loss by calculating the verification of the data packet, and outputting the verified multidimensional vibration acceleration data to the data encoding module.

4. A vibration stress monitoring method for a wireless telemetry rotating transmission system according to claim 1, characterized in that: The S3 comprises the following steps: S31, receiving the multi-dimensional vibration acceleration data after signal conditioning through an embedded processor, and the processor performs time synchronization marking on the received vibration acceleration data; S32. Assigning a sensor identifier to each set of vibration acceleration data , sensor identifier The vibration acceleration sensor used to uniquely identify the multi-dimensional vibration acceleration data satisfies the following relationship: ； in, is the sensor layout position on the flange, is the corresponding signal processing channel number; S34, vibration acceleration data, timestamp and sensor identifier Combination generation data recording unit , represents the kth sampling moment of sensor data collection; S35, multiple data recording units through embedded processor Perform data encoding to generate vibration signal data packets , the encoding rules are: ； in, is the preset encoding function; S36, check the integrity of the generated vibration signal data packet, through the check code Determine whether the data packet is lost or erroneous during the encoding process.

5. The vibration stress monitoring method of a wireless telemetry rotating transmission system according to claim 1, characterized in that: The S5 comprises the following steps: S51, receiving vibration signal data packet via mobile terminal , using the preset decoding function Decode the data packet to extract multi-dimensional vibration acceleration data, timestamp and sensor identifier; S52, according to the timestamp and sensor identifier The extracted vibration acceleration data Classify and store in the corresponding time series database; S53, perform discrete Fourier transform on each dimension of vibration acceleration data to calculate its main frequency And spectrum characteristic distribution : ； in, is the vibration acceleration value corresponding to the nth sampling point, N represents the total number of sampling points, T is the sampling period, and f is the frequency component; The amplitude fluctuation range of the vibration acceleration data is statistically analyzed, and the segment probability is calculated as follows: ； in, Amplitude In the interval The probability distribution within, p(a) is the probability density function of the amplitude; Directional coupling of vibration acceleration data in different dimensions through mutual information calculate: ； in, for and The joint probability density of and are their respective marginal probability density functions; S54. Automatically generate a normal working state benchmark model of the rotating transmission system based on the extracted feature data : 。 6. A vibration stress monitoring method for a wireless telemetry rotating transmission system according to claim 5, characterized in that: The S6 comprises the following steps: S61, through the decoding function The vibration signal data packet transmitted in real time Decode to extract vibration acceleration data, timestamp and sensor identifier; S62. By analyzing the frequency domain of real-time data, the difference between the main frequency deviation and the power spectrum density is defined as: ； ； in, and are the main frequencies of real-time data and benchmark model respectively, and are the power spectral densities of real-time data and benchmark model, respectively. and are the minimum analysis frequency range and the maximum analysis frequency range respectively. When , it is judged as abnormal; S63. Calculate the amplitude deviation of the real-time vibration acceleration data and compare it with the normal range: ； in, is the amplitude of real-time vibration acceleration data, is the amplitude in the baseline model, when When , it is judged as abnormal; S64, by real-time calculation of the directional coupling mutual information of multi-dimensional vibration acceleration data, the difference is defined as follows: ； in, and are the joint probability densities of real-time data and the benchmark model, respectively. When , it is judged as abnormal; S65. When any comparison result in step S62, S63 or S64 exceeds the normal range, comprehensively record abnormal characteristic data; S66. Use the real-time abnormal feature data to update the baseline model through a self-learning algorithm.

Images