Reducer maintenance and inspection system based on multi-source sensing fusion

Abstract

The invention relates to a speed reducer maintenance and inspection system based on multi-source sensing fusion, which comprises a gear speed reducer, data detection equipment, a data acquisition subsystem, a communication subsystem, a data analysis subsystem, an intelligent diagnosis subsystem, a three-axis vibration sensor and microphones arranged in an acoustic array form a Cartesian coordinate system, fault information can be accurately captured from two dimensions of vibration and acoustics, the detection sensitivity of early faults is greatly improved, and the detection delay is effectively reduced.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and specifically relates to a reducer maintenance inspection system based on multi-source sensing fusion. Background Art

[0002] Existing online detection systems mostly rely on sensors such as vibration, oil, and temperature. Through weighted average, Kalman filtering, or neural networks, preliminary fusion of multi-source sensors is performed. For example, joint detection of vibration and temperature, mechanical faults are analyzed through vibration spectrum analysis, and temperature sensors assist in warning of overheating risks. This solution has a high delay and is not sensitive enough to early fault detection; there are also those that use inductive or optical sensors to detect oil abrasive particles and judge the degree of wear to achieve partial condition prediction; there are also those that use Hall speed sensors to assist vibration sensors to detect through time-domain interpolation algorithms, but there are problems with cumulative phase errors.

[0003] At the same time, maintaining high-frequency data acquisition by sensors results in too high power consumption, and the data volume is huge, putting great pressure on the bandwidth; existing faults and defects, the fault sources are fuzzy, and traditional vibration sensors are difficult to distinguish faults of bearings, gears, and shafts. Especially in the case of multi-component coupled vibration scenarios, the misjudgment rate is high; the calculation of fault characteristic frequencies depends on the rotational speed accuracy, and rotational speed sensors are not integrated, resulting in spectrum analysis errors; there is a lack of acoustic and vibration joint positioning capabilities, and the physical positions of faulty components cannot be determined. Oil, temperature, and vibration data are analyzed independently, and a cross-parameter correlation model has not been established. Therefore, there is still room for improvement in the existing technology. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a reducer maintenance inspection system based on multi-source sensing fusion in view of the deficiencies of the above-mentioned existing technologies.

[0005] To achieve the above object, the present invention provides the following technical solution. A reducer maintenance inspection system based on multi-source sensing fusion includes a gear reducer, a data detection device, a data acquisition subsystem, a communication subsystem, a data analysis subsystem, and an intelligent diagnosis subsystem. It is characterized in that: the gear reducer includes a box body, and an input shaft, an output shaft, a high-speed shaft, and a low-speed shaft are rotatably arranged in the box body. The low-speed shaft, the high-speed shaft, the input shaft, and the output shaft are connected by a gear set; The data detection device includes a three-axis vibration sensor, a microphone, a temperature sensor, a Hall speed sensor, a multi-parameter oil fluid sensor, an adsorption abrasive particle sensor, and a moisture content sensor. The three-axis vibration sensors are respectively installed at the connections between the box body and the input shaft and the output shaft. The microphone is arranged on the outer periphery of the box body through an acoustic array to detect the internal noise of the gearbox, facilitating the positioning of faulty components and the analysis of fault problems. The installation positions of the three-axis vibration sensors and the microphone form a Cartesian coordinate system, and the Cartesian coordinate system is coplanar with the low-speed shaft, the high-speed shaft, the input shaft, and the output shaft. The temperature sensors are distributed at the connections between the high-speed shaft, the low-speed shaft, the input shaft, and the output shaft and the box body. The Hall speed sensor is fixedly installed on the input shaft. The multi-parameter oil fluid sensor, the adsorption abrasive particle sensor, and the moisture content sensor are arranged on one side of the box body away from the output shaft; The data acquisition subsystem is used to collect and integrate the data from the data detection device; The communication subsystem uses a wireless communication method to transmit the collected data to the data analysis system and the intelligent diagnosis system; The data analysis subsystem includes receiving the data from the data acquisition subsystem and calculating the fault coordinates; The intelligent diagnosis subsystem includes a sub-database. The intelligent diagnosis subsystem uses the fault coordinates calculated by the data analysis subsystem to determine the faults of the components.

[0006] Adopting the above technical solution, through multi-source sensing fusion, multiple sensors such as a triaxial vibration sensor, a microphone, a temperature sensor, a Hall speed sensor, a multi-parameter oil sensor, an adsorption abrasive particle sensor, and a moisture content sensor work together. The triaxial vibration sensor and the microphones arranged in an acoustic array form a Cartesian coordinate system, which can accurately capture fault information from two dimensions of vibration and acoustics, greatly improving the detection sensitivity to early faults and effectively reducing the detection delay. The multi-parameter oil sensor serves as a lubrication environment detection and auxiliary judgment function. The multi-parameter oil sensor includes detecting temperature, density, and viscosity. When the oil is agitated inside the box, the oil sensor will detect the corresponding state of the oil. If the oil viscosity drops to a certain value and the box body temperature rises, corresponding alarm signals will be sent. In the previous steps, if faults such as bearing or gear wear are judged, a wear condition assessment will be carried out. Based on the change of the metal particle density in the oil over a period of time and the changes in the noise and vibration of the gearbox, it is easier to judge the wear condition. When the density changes suddenly, it may be caused by problems such as external water ingress, and corresponding judgments are made. The oil sensor accurately evaluates the lubricating oil state and the wear trend of each part of the gearbox through multi-physical quantity collaborative detection. The intelligent diagnosis subsystem of this system combines the fault coordinates calculated by the data analysis subsystem and uses the sub-database for fault determination, which can effectively avoid these errors and improve the accuracy of fault diagnosis. In the existing system, the vibration sensor has a high misjudgment rate when distinguishing bearing, gear, and shaft faults, especially in the case of multi-component coupled vibration scenarios. However, in this system, multiple sensors work together and acoustic and vibration joint positioning can accurately determine the physical position of the faulty component, greatly reducing the misjudgment rate. The system conducts comprehensive analysis on the data of different types of sensors through multi-source sensing fusion, establishes a more comprehensive and in-depth correlation model, can more accurately reflect the overall operating state of the reducer, realizes precise diagnosis and predictive maintenance of reducer faults, reduces equipment failure rate and maintenance costs, and improves equipment reliability and operating efficiency.

[0007] The above-mentioned reducer maintenance inspection system based on multi-source sensing fusion can be further set as follows: The corresponding relationship between the triaxial vibration sensor and the microphone to form a Cartesian coordinate system is that the microphones are respectively installed on the X-axis, Y-axis, and the coordinate origin, and the triaxial vibration sensors are respectively installed on the X-axis and Y-axis to make the coordinate plane coplanar with the gear shaft axis, and all fault sources are inside the coordinate system.

[0008] With the above technical solution, in the axis plane, the axis perpendicular to the output shaft is the x-axis, and the axis perpendicular to the input shaft is the y-axis. The x-axis and the y-axis intersect at the microphone installation position, which is the coordinate origin. The microphones are respectively installed on the x-axis, the y-axis and the coordinate origin. The vibration sensors are installed on the x-axis and the y-axis, and the coordinate plane is coplanar with the axis of the gear shaft. And all the fault sources are inside the coordinate system. The microphones distributed on the coordinate axes and the origin can capture the noise information inside the gearbox from multiple directions and positions, and can effectively collect the noise generated at different positions, and can more accurately locate the fault source. When a fault occurs, by analyzing the time difference of the sound signals received by different microphones, the approximate area where the fault occurs can be quickly determined. The vibration sensors installed on the x-axis and the y-axis can effectively detect the vibration conditions of the gears and bearings. Since the coordinate plane is coplanar with the axis of the gear shaft, the vibration information related to the operating state of the gear shaft can be obtained to the greatest extent, and the sensitivity and accuracy of the vibration signal can be improved. The method of combining acoustic and vibration monitoring in a Cartesian coordinate system can realize the synchronous analysis of acoustic and vibration data.

[0009] The above reducer maintenance inspection system based on multi-source sensing fusion can be further set as follows: The algorithm of the data analysis subsystem specifically includes the following steps: Step S1: Data preparation, receiving data from the data acquisition subsystem; Step S2: Determine the position coordinates (x, y) of the fault source through the time difference of sound propagation to different positions. The formula used is as follows: : represents the time for sound to propagate from the fault source S(x,y) to the origin (or reference point); d1 refers to the distance between the microphone on the x-axis and the origin coordinates; : d2 refers to the distance between the microphone on the y-axis and the origin coordinates; Among them, the sound propagation speed in the air vair = 343 m / s, and the fault source S(x,y); Step S3: Estimate the time delay difference through the generalized cross-correlation phase transformation algorithm, establish and solve the hyperbolic equation using the time difference to obtain the signal source; Step S4: Obtain the fault source coordinates through the Newton iteration method, and further determine the fault source position through particle filter fusion optimization. Associate the data of multiple databases with the fault source state, and the subsequent change law of the fault source position over time. In the process of particle filter iteration, first perform initialization, and the generated particles obey , and then perform weight update; Step S5: Retain the high-weight particles through resampling and perform weighted averaging to obtain the final coordinates: in, refers to the weight of the i-th particle at time k, is the actual observed value, refers to the predicted observation value of the ith particle.

[0010] By adopting the above technical scheme, in the data preparation stage, data from the data acquisition subsystem is received, which provides a basis for subsequent analysis and ensures the integrity and accuracy of the data. The coordinates of the fault source position are determined by the time difference of sound propagation to different positions. The time of sound propagation to the origin, the first sensor and the second sensor is calculated by the formula. Combined with the known sound propagation speed in air and solid, the fault source can be initially located. The generalized cross-correlation phase transformation algorithm estimates the delay difference and establishes and solves the hyperbolic equation, which further narrows the scope of the fault source and improves the positioning accuracy. The Newton iteration method obtains the coordinates of the fault source, and then the particle filter fusion optimization is used to further determine the fault source position. The advantages of the Newton iteration method such as fast convergence speed are utilized. The particle filter is also used to fuse the association of multiple database data with the fault source state and the change law of the fault source position over time, making the positioning more accurate and dynamic. Resampling retains high-weight particles and performs weighted averaging to obtain the final coordinates, which can effectively remove the interference caused by low-weight particles, making the final fault source coordinates more reliable and stable.

[0011] The above-mentioned reducer maintenance and inspection system based on multi-source sensor fusion can be further configured as follows: the step S3 includes step S3.1: estimating the delay difference by using a generalized cross-correlation phase transformation algorithm, and correcting the delay difference, and the formula used is as follows: ; Among them, Δtcomp is the corrected delay difference, dwall is the box wall thickness, vwall=5000m / s is the speed of sound in the solid (box wall), dair is the distance of the air part in the box, vair is the speed of sound in air; Step S3.2 uses the time difference to establish and solve the hyperbola equation to obtain the signal source. The hyperbola equation is: ; ; Where, t10=t1-t0, t1 is the time required for the signal to propagate from the target point to the reference point (d1,0), t0 is the time required for the signal to propagate from the target point to the origin (0,0), and Δt10 is the difference between these two propagation times; Δt20 = t2 - t0, where t2 is the time required for the signal to propagate from the target point to the reference point (0, d2), t0 is the time required for the signal to propagate from the target point to the origin (0, 0), and Δt20 is the difference between these two propagation times.

[0012] With the above technical solution, in step S3.1, the time delay difference is estimated by the generalized cross-correlation phase transformation algorithm, and the formula is used to correct the time delay difference. Considering that the sound will pass through two different media, the box wall and the air inside the box, during the propagation process, and the sound speeds in different media are different, this correction formula fully considers factors such as the box wall thickness, the sound speed in the solid (box wall), the distance of the air part inside the box, and the sound speed in the air, making the calculation of the time delay difference more accurate and eliminating the error caused by the medium difference. In step S3.2, a hyperbola equation is established and solved using the time difference. Based on the time difference of sound propagation, the position of the signal source can be associated with the position of the sensor, as well as the sound propagation speed and time difference. By solving these two hyperbola equations, the position of the signal source can be determined, realizing the accurate positioning of the fault source.

[0013] The above reducer maintenance inspection system based on multi-source sensing fusion can be further set as follows: Step S4 includes step S4.1: Obtain the fault source coordinates through the Newton iteration method, and further determine the fault source position through particle filter fusion optimization. Associate multiple databases with the data fault source status. The formula used is as follows: ; where is the state vector, including the fault source coordinates at time and its change rate , is the state transition matrix, is the process noise; Step S4.2: For the subsequent change law of the fault source position over time, list the observation equation: ; where, is the observation vector, including the RMS which is the root mean square value of the vibration signal, is the bearing temperature rise rate (unit: °C / min), is the acoustic initial value coordinate, H is the observation matrix, which maps the state vector to the observation space, is the observation noise, and the observation noise follows the Gaussian distribution Ν(0, R), where R is the noise covariance matrix; Step S4.3: In the particle filter iteration process, first perform initialization, and the generated particles follow , and then perform weight update. The weight update formula is: 。

[0014] With the above technical solution, in step S4.1, the Newton iteration method is used to obtain the coordinates of the fault source. This method can quickly converge to an approximate value of the fault source coordinates, laying a foundation for subsequent precise analysis. The position of the fault source is further determined through particle filter fusion optimization, associating various database data with the state of the fault source and making full use of the state vector to describe the coordinates of the fault source and its rate of change, combined with the state transition matrix F and process noise , making the determination of the fault source position more in line with the actual physical process, considering the dynamic change characteristics of the fault source, enhancing the accuracy and reliability of positioning. In step S4.2, the observation equation is listed, associating the root mean square value of the vibration signal, the bearing temperature rise rate, the acoustic initial value coordinates, and other information in the observation vector with the state vector through the observation matrix H, and considering the observation noise at the same time, enabling the system to analyze the position of the fault source by integrating multi-source data and improving the perception ability of the fault source state; In step S4.3, during the particle filter iteration process, particles obeying are initialized and generated, and the particle weights are updated through the weight update formula. The particle weights can be dynamically adjusted according to the difference between the actual observation value and the predicted observation value, retaining the particles that are more in line with the actual situation and removing interference information, so that the finally obtained fault source position is more accurate.

[0015] The above reducer maintenance inspection system based on multi-source sensing fusion can be further set as follows: The algorithm of the intelligent diagnosis subsystem specifically includes the following steps: Step Q1: After obtaining the coordinates from the data analysis subsystem, the components are judged. First, through a rough sample comparison, the sub-database is loaded according to the coordinates; Step Q2: Use the dynamic time warping method to align the real-time features with the samples and calculate the comprehensive similarity. The calculation formula is: ; Among them, is the real-time vibration feature vector, is the database sample feature vector, is the temperature feature similarity (normalized difference), is the acoustic positioning coordinate matching degree. If S≥0.95, the fault type is directly judged. If 0.7≤S≤0.9, it is uploaded to the cloud to trigger secondary verification.

[0016] With the above technical solution, in step Q1, according to the coordinates obtained by the data analysis subsystem, first conduct a rough sample comparison, and then load the sub-database based on the coordinates to quickly narrow down the search range and accurately locate the data related to the components that may have faults; in step Q2, the dynamic time warping method is used to align the real-time features with the samples, and the comprehensive similarity is calculated. The formula comprehensively considers the real-time vibration feature vector, the database sample feature vector, the temperature feature similarity, and the acoustic positioning coordinate matching degree, evaluates the fault features from multiple dimensions, avoids the limitations of single-feature judgment, and improves the accuracy of fault diagnosis; different diagnostic strategies are formulated according to the value of the comprehensive similarity S. When S≥0.95, the fault type is directly judged, which can quickly handle faults with high similarity and obvious features, improve the diagnostic efficiency, and take corresponding maintenance measures in a timely manner. When 0.7≤S≤0.9, it is uploaded to the cloud to trigger secondary verification. For the situation where the similarity is in the medium range and the fault features are not very clear, the powerful computing power and rich data resources of the cloud are used for further analysis to ensure the reliability of the diagnostic results.

[0017] The above reducer maintenance inspection system based on multi-source sensing fusion can be further set as follows: Step Q1 includes step Q1.1: By using PCA dimensionality reduction technology, reduce the 32-dimensional vibration feature to 8 dimensions, and the dimensionality reduction formula is: ; where V is the original 32-dimensional feature vector and P is the principal component matrix; Step Q1.2: By using INT8 quantization technology, convert 32-bit floating-point numbers (FP32) to 8-bit integers (INT8) to reduce the amount of calculation and speed up the inference speed. The quantization formula is: where μ is the data mean and σ is the data standard deviation. Compress the FP32 floating-point number to an INT8 integer, so that the model memory occupancy is reduced from 32MB to 8MB and the inference time is reduced.

[0018] With the above technical solution, in step Q1.1, the PCA dimensionality reduction technology is used to reduce the 32-dimensional vibration features to 8 dimensions, which can effectively remove redundant information in the original features. Through the dimensionality reduction formula, while reducing the data dimension, the key information of the data is retained to the greatest extent, reducing the complexity of subsequent calculations and improving the calculation efficiency. In step Q1.2, the INT8 quantization technology is adopted to convert 32-bit floating-point numbers (FP32) into 8-bit integers (INT8), significantly reducing the amount of calculation and accelerating the inference speed because the operation of 8-bit integers is more efficient on hardware. At the same time, the model memory occupancy is reduced from 32MB to 8MB, greatly saving memory resources and enabling the system to run more smoothly in an environment with limited resources.

[0019] The above reducer maintenance inspection system based on multi-source sensing fusion can be further set as follows: step Q2 includes step Q2.1: in the interval of secondary verification (0.7 ≤ S ≤ 0.9), the system will immediately upload relevant data to the cloud, and through the cloud, further determine the type, severity, and possible development trend of the fault; Step Q2.1: The cloud will feedback the judgment result to the local device and provide corresponding maintenance suggestions according to the fault situation.

[0020] With the above technical solution, in the secondary verification interval (0.7 ≤ S ≤ 0.9), the system uploads relevant data to the cloud. This measure makes full use of the powerful computing power and rich data resources of the cloud. Based on a large amount of historical data and advanced algorithm models, the cloud can conduct a more in-depth and comprehensive analysis of faults, further determining the type, severity, and possible development trend of faults, making up for the deficiencies in data processing and analysis capabilities of local devices, improving the accuracy and reliability of fault diagnosis. Subsequently, the cloud feeds back the judgment results to the local device and provides corresponding maintenance suggestions according to the fault situation, enabling the operators of local devices to handle faults in a timely manner. During the entire monitoring and fault handling process, the system will continuously learn and optimize itself. By accumulating and analyzing a large amount of device operation data, the system can automatically adjust the sampling frequency and working status of sensors. For devices with stable operation and good conditions, some sensors can reduce the sampling frequency or enter the sleep state at appropriate times, thereby reducing energy consumption and extending the service life of sensors; while for devices with abnormal conditions, the system will automatically increase the sampling frequency of sensors to more closely monitor the device status and timely capture detailed information on the development of faults. The monitoring data of different devices will be aggregated and analyzed in the cloud. Through technologies such as federated learning, each device can share fault knowledge and handling experience. When a new fault mode is detected by a certain device, the relevant information will be sorted out and analyzed in the cloud, and a more complete judgment algorithm will be shared via OTA with other devices, continuously improving the fault prediction and diagnosis capabilities of the entire system and achieving more efficient and intelligent predictive maintenance.

[0021] The above-mentioned reducer maintenance inspection system based on multi-source sensing fusion can be further set as: the input shaft is vertically arranged on the high-speed shaft, and the high-speed shaft, low-speed shaft, and output shaft are arranged in parallel.

[0022] With the above technical solution, the structure where the input shaft is perpendicular to the high-speed shaft can change the power direction through appropriate transmission components, enabling the power to be smoothly transmitted from the input shaft to the high-speed shaft. The parallel arrangement of the high-speed shaft, low-speed shaft, and output shaft is conducive to adopting the common and efficient parallel-axis gear transmission method. This transmission method has high transmission efficiency and strong load-bearing capacity, and can effectively transmit the power of the high-speed shaft to the low-speed shaft and output shaft step by step, reducing the energy loss during the power transmission process.

[0023] The above-mentioned reducer maintenance inspection system based on multi-source sensing fusion can be further set as follows: The gear set includes a first rotating gear with one end of the output shaft facing the high-speed shaft. The high-speed shaft is provided with a second transmission gear meshing with the first transmission gear. The high-speed shaft is provided with a third transmission gear on one side of the second transmission gear. The low-speed shaft is provided with a fourth transmission gear meshing with the third transmission gear. The low-speed shaft is provided with a fifth transmission gear on one side of the fourth transmission gear. The output shaft is provided with a sixth transmission gear meshing with the fifth transmission gear.

[0024] Adopting the above technical solution, through the meshing of the first rotating gear on the output shaft and the second transmission gear on the high-speed shaft, the power of the output shaft can be smoothly and efficiently transmitted to the high-speed shaft, realizing the initial conversion and transmission of power. The third transmission gear on the high-speed shaft meshes with the fourth transmission gear on the low-speed shaft, and the fifth transmission gear on the low-speed shaft meshes with the sixth transmission gear on the output shaft, forming a multi-stage transmission structure, meeting the requirements for speed and torque under different working conditions, and enabling the reducer to adapt to a wider range of working scenarios.

[0025] The beneficial effects of the present invention: 1. By forming a Cartesian coordinate system with a triaxial vibration sensor and microphones arranged in an acoustic array, fault information can be accurately captured from two dimensions of vibration and acoustics, greatly improving the detection sensitivity to early faults and effectively reducing detection latency; 2. The multi-parameter oil condition detection sensor analyzes the changes in the oil condition caused by internal and external conditions, improves the recognition rate of compound faults, and thus establishes a vibration-acoustics-temperature-oil joint degradation model; 3. It can dynamically adjust the sensor frequency according to the algorithm, reducing the power consumption of the sensor and the amount of data generated. Description of the Drawings

[0026] Figure 1 It is a schematic structural diagram of the gear reducer of the present invention; Figure 2 It is a side view of the gear reducer structure of the present invention; Figure 3 It is a structural diagram of the gear set of the present invention; Figure 4 It is a schematic diagram of the system flow of the present invention; Figure 5 It is a diagram of the establishment of the Cartesian coordinate of the present invention; Label Annotations: 1 - Gear Reducer, 2 - Housing, 3 - Input Shaft, 4 - Output Shaft, 5 - High-Speed Shaft, 6 - Low-Speed Shaft, 7 - Triaxial Vibration Sensor, 8 - Microphone, 9 - Temperature Sensor, 10 - Hall Speed Sensor, 11 - Multi-Parameter Oil Sensor, 12 - Adsorptive Abrasive Sensor, 13 - Moisture Content Sensor, 14 - First Rotating Gear, 15 - Second Transmission Gear, 16 - Third Transmission Gear, 17 - Fourth Transmission Gear, 18 - Fifth Transmission Gear, 19 - Sixth Transmission Gear. Specific Embodiment

[0027] This specific embodiment is only an interpretation of the present invention and is not a limitation thereof. After reading this specification, those skilled in the art can make modifications to this embodiment without creative contributions as needed, but as long as it is within the scope of the claims of the present invention, it is protected by the patent law.

[0028] Embodiment 1: A reducer maintenance inspection system based on multi-source sensing fusion, including a gear reducer 1, a data detection device, a data acquisition subsystem, a communication subsystem, a data analysis subsystem, and an intelligent diagnosis subsystem. It is characterized in that: the gear reducer includes a housing 2, and an input shaft 3, an output shaft 4, a high-speed shaft 5, and a low-speed shaft 6 are rotatably arranged inside the housing. The low-speed shaft 6, the high-speed shaft 5, the input shaft 3, and the output shaft 4 are connected by a gear set. The data detection device includes a triaxial vibration sensor 7, a microphone 8, a temperature sensor 9, a Hall speed sensor 10, a multi-parameter oil sensor 11, an adsorptive abrasive sensor 12, and a moisture content sensor 13. The triaxial vibration sensor 7 is respectively installed at the connections between the housing 2 and the input shaft 3 and the output shaft 4. The microphone 8 is arranged on the outer periphery of the housing 2 through an acoustic array for detecting the internal noise of the gearbox to facilitate the positioning of faulty components and the analysis of fault problems. The installation positions of the triaxial vibration sensor 7 and the microphone 8 form a Cartesian coordinate system, and the Cartesian coordinate system is coplanar with the low-speed shaft 6, the high-speed shaft 5, the input shaft 3, and the output shaft 4. The temperature sensor 9 is distributed at the connections between the high-speed shaft 5, the low-speed shaft 6, the input shaft 3, and the output shaft 4 and the housing 2. The Hall speed sensor 10 is fixedly installed on the input shaft 3. The multi-parameter oil sensor 11, the adsorptive abrasive sensor 12, and the moisture content sensor 13 are arranged on the side of the housing 2 away from the output shaft. The data acquisition subsystem is used to collect and integrate the data from the data detection device. The communication subsystem uses a wireless communication method to transmit the collected data to the data analysis system and the intelligent diagnosis system. The data analysis subsystem includes receiving the data from the data acquisition subsystem and calculating the fault coordinates. The intelligent diagnosis subsystem includes a sub-database. The intelligent diagnosis subsystem determines the faults of components based on the fault coordinates calculated by the data analysis subsystem. The corresponding relationship formed by the three-axis vibration sensor 7 and the microphone 8 in the Cartesian coordinate system is as follows: The microphone 8 is respectively installed on the X-axis, the Y-axis, and the coordinate origin. The three-axis vibration sensor 7 is respectively installed on the X-axis and the Y-axis to make the coordinate plane coplanar with the axis of the gear shaft. All fault sources are inside the coordinate system.

[0029] Embodiment 2: The algorithm of the data analysis subsystem specifically includes the following steps: Step S1: Data preparation, receiving data from the data acquisition subsystem; Step S2: Determine the position coordinates (x, y) of the fault source through the time difference of sound propagation to different positions. The formula used is as follows: : represents the time for sound to propagate from the fault source S(x, y) to the origin (or reference point); d1 refers to the distance between the microphone on the X-axis and the origin coordinates; : d2 refers to the distance between the microphone on the Y-axis and the origin coordinates; Among them, the sound propagation speed vair = 343 m / s in the air, and the fault source S(x, y); Step S3: Estimate the time delay difference through the generalized cross-correlation phase transformation algorithm, establish and solve the hyperbolic equation using the time difference to obtain the signal source; Step S4: Obtain the fault source coordinates through the Newton iteration method, and further determine the fault source position through particle filter fusion optimization. Associate the fault source states of multiple official data. The subsequent change law of the fault source position over time. During the particle filter iteration process, first perform initialization, and the generated particles obey , and then perform weight update; Step S5: Retain high-weight particles through resampling and perform weighted averaging to obtain the final coordinates: Among them, refers to the weight of the i-th particle at the k-th moment, is the actual observed value, refers to the predicted observed value of the i-th particle; Step S3 includes Step S3.1: Estimate the time delay difference through the generalized cross-correlation phase transformation algorithm and correct the time delay difference. The formula used is as follows: ; Among them, Δtcomp is the corrected time delay difference, dwal l is the wall thickness of the box, and vwall = 5000 m / s is the sound speed in the solid (box wall), dairis the distance of the air part inside the box, vair is the speed of sound in air; Step S3.2 uses the time difference to establish and solve the hyperbolic equation to obtain the signal source. The hyperbolic equation is: ; ; where t10 = t1 - t0, t1 is the time required for the signal to propagate from the target point to the reference point (d1,0), t0 is the time required for the signal to propagate from the target point to the origin (0,0), and Δt10 is the difference between these two propagation times; Δt20 = t2 - t0, t2 is the time required for the signal to propagate from the target point to the reference point (0, d2), t0 is the time required for the signal to propagate from the target point to the origin (0,0), and Δt20 is the difference between these two propagation times; Step S4 includes Step S4.1: Obtain the coordinates of the fault source through the Newton iteration method, and further determine the position of the fault source through particle filter fusion optimization. Associate various database data with the state of the fault source. The formula used is as follows: ; where is the state vector, including the coordinates of the fault source at time and its rate of change , is the state transition matrix, is the process noise; Step S4.2: List the observation equation for the subsequent variation law of the fault source position with time: ; where, is the observation vector, including the RMS which is the root mean square value of the vibration signal, is the bearing temperature rise rate (unit: °C / min), is the initial acoustic value coordinate, H is the observation matrix that maps the state vector to the observation space, is the observation noise, and the observation noise follows the Gaussian distribution Ν(0,R), where R is the noise covariance matrix; Step S4.3: In the particle filter iteration process, first perform initialization, and the generated particles follow , and then perform weight update. The weight update formula is: .

[0030] Example 3: The algorithm of the intelligent diagnosis subsystem specifically includes the following steps: Step Q1: After obtaining the coordinates from the data analysis subsystem, determine the component. First, perform a rough sample comparison and load the sub-database according to the coordinates. Step Q2: Use the dynamic time warping method to align the real-time features with the samples and calculate the comprehensive similarity. The calculation formula is: ; where is the real-time vibration feature vector, is the database sample feature vector, is the temperature feature similarity (normalized difference), is the acoustic positioning coordinate matching degree. If S≥0.95, directly determine the fault type. If 0.7≤S≤0.9, upload to the cloud to trigger secondary verification. Step Q1 includes Step Q1.1: Through the PCA dimensionality reduction technique, reduce the 32-dimensional vibration feature to 8 dimensions. The dimensionality reduction formula is: ; where V is the original 32-dimensional feature vector and P is the principal component matrix. Step Q1.2: Through the INT8 quantization technique, convert 32-bit floating-point numbers (FP32) to 8-bit integers (INT8) to reduce the calculation amount and speed up the inference speed. The quantization formula is: where μ is the data mean and σ is the data standard deviation. Compress the FP32 floating-point number to an INT8 integer, reducing the model memory occupancy from 32MB to 8MB and reducing the inference time. Step Q2 includes Step Q2.1: In the interval of secondary verification (0.7≤S≤0.9), the system will immediately upload the relevant data to the cloud. Through the cloud, further determine the type, severity, and possible development trend of the fault. Step Q2.1: The cloud will feedback the judgment result to the local device and provide corresponding maintenance suggestions according to the fault situation.

[0031] Embodiment 4: The input shaft 3 of the gear speed reducer 1 is vertically arranged on the high-speed shaft 5, and the high-speed shaft 5, the low-speed shaft 6, and the output shaft 4 are arranged in parallel.

[0032] The gear set includes a first rotating gear 14 provided at one end of the output shaft 4 facing the high-speed shaft 5. The high-speed shaft 5 is provided with a second transmission gear 15 meshing with the first transmission gear. The high-speed shaft 5 is provided with a third transmission gear 16 on one side of the second transmission gear 15. The low-speed shaft 6 is provided with a fourth transmission gear 17 meshing with the third transmission gear 16. The low-speed shaft 6 is provided with a fifth transmission gear 18 on one side of the fourth transmission gear 17. The output shaft 4 is provided with a sixth transmission gear 19 meshing with the fifth transmission gear 18. By meshing the first rotating gear 14 on the output shaft 4 with the second transmission gear 15 on the high-speed shaft 5, the power of the output shaft 4 can be smoothly and efficiently transmitted to the high-speed shaft 5, realizing the preliminary conversion and transmission of power. The third transmission gear 16 on the high-speed shaft 5 meshes with the fourth transmission gear 17 on the low-speed shaft 6, and the fifth transmission gear 18 on the low-speed shaft 6 meshes with the sixth transmission gear 19 on the output shaft 4, forming a multi-stage transmission structure, meeting the requirements for speed and torque under different working conditions, and enabling the reducer to adapt to a wider range of working scenarios.

Claims

1. A reducer maintenance inspection system based on multi-source sensing fusion, comprising a gear reducer, a data detection device, a data acquisition subsystem, a communication subsystem, a data analysis subsystem, and an intelligent diagnosis subsystem, characterized in that: The gear reducer includes a housing, in which an input shaft, an output shaft, a high-speed shaft and a low-speed shaft are rotatably arranged, and the low-speed shaft, the high-speed shaft, the input shaft and the output shaft are connected by a gear set; The data detection device includes a three-axis vibration sensor, a microphone, a temperature sensor, a Hall speed sensor, a multi-parameter oil fluid sensor, an adsorption abrasive particle sensor, and a moisture content sensor. The three-axis vibration sensor is respectively installed at the joints of the housing with the input shaft and the output shaft. The microphone is arranged on the periphery of the housing through an acoustic array for detecting the internal noise of the gearbox to facilitate the positioning of faulty components and the analysis of fault problems. The installation positions of the three-axis vibration sensor and the microphone form a Cartesian coordinate system, and the Cartesian coordinate system is coplanar with the low-speed shaft, the high-speed shaft, the input shaft and the output shaft. The temperature sensors are distributed at the joints of the high-speed shaft, the low-speed shaft, the input shaft and the output shaft with the housing. The Hall speed sensor is fixedly installed on the input shaft. The multi-parameter oil fluid sensor, the adsorption abrasive particle sensor, and the moisture content sensor are arranged on one side of the housing away from the output shaft; The data acquisition subsystem is used to acquire and integrate the data from the data detection device; The communication subsystem uses a wireless communication method to transmit the acquired data to the data analysis system and the intelligent diagnosis system; The data analysis subsystem includes receiving the data from the data acquisition subsystem and calculating the fault coordinates; The intelligent diagnosis subsystem includes a sub-database, and the intelligent diagnosis subsystem determines the faults of components based on the fault coordinates calculated by the data analysis subsystem.

2. The reducer maintenance inspection system based on multi-source sensing fusion according to claim 1, characterized in that: The corresponding relationship formed by the three-axis vibration sensor and the microphone in the Cartesian coordinate system is as follows: The microphones are respectively installed on the X-axis, the Y-axis and the coordinate origin. The three-axis vibration sensors are respectively installed on the X-axis and the Y-axis to make the coordinate plane coplanar with the axis of the gear shaft, and all fault sources are inside the coordinate system.

3. The reducer maintenance inspection system based on multi-source sensing fusion according to claim 1, characterized in that: The algorithm of the data analysis subsystem specifically includes the following steps: Step S1: Data preparation, receiving the data from the data acquisition subsystem; Step S2: Determine the position coordinates (x, y) of the fault source through the time difference of sound propagation to different positions. The formula used is as follows: : represents the time for sound to propagate from the fault source S(x,y) to the origin (or reference point); d1 refers to the distance between the microphone and the origin coordinates on the X-axis; : d2 refers to the distance between the microphone and the origin coordinate on the Y-axis; where, the propagation speed of sound in the air vair = 343 m / s, and the fault source S(x, y); Step S3: Estimate the time delay difference through the generalized cross-correlation phase transformation algorithm, and establish and solve the hyperbolic equation using the time difference to obtain the signal source; Step S4: Obtain the fault source coordinates by the Newton iteration method, and further determine the fault source location through particle filter fusion optimization. Associate various database data with the fault source state, and the subsequent variation law of the fault source location over time. During the particle filter iteration process, first perform initialization, and the generated particles obey , and then update the weights; Step S5: Retain the high-weight particles through resampling and perform weighted averaging to obtain the final coordinates: ； wherein, refers to the weight of the i-th particle at the k-th moment, is the actual observed value, refers to the predicted observed value of the i-th particle.

4. The reducer maintenance inspection system based on multi-source sensing fusion according to claim 3, characterized in that: Step S3 includes Step S3.1: Estimate the time delay difference through the generalized cross-correlation phase transformation algorithm and correct the time delay difference. The formula used is as follows: ； where Δtcomp is the corrected time delay difference, dwal l is the wall thickness of the box, vwall = 5000 m / s is the sound speed in the solid (box wall), dair is the distance of the air part inside the box, vair is the sound speed in air; Step S3.2 Use the time difference to establish and solve the hyperbolic equation to obtain the signal source. The hyperbolic equation is: ； ； where, t10 = t1 - t0, t1 is the time required for the signal to propagate from the target point to the reference point (d1, 0), t0 is the time required for the signal to propagate from the target point to the origin (0, 0), and Δt10 is the difference between these two propagation times; Δt20 = t2 - t0, where t2 is the time required for the signal to propagate from the target point to the reference point (0, d2), t0 is the time required for the signal to propagate from the target point to the origin (0, 0), and Δt20 is the difference between these two propagation times.

5. The reducer maintenance inspection system based on multi-source sensing fusion according to claim 3, characterized in that: Step S4 includes Step S4.1: Obtain the coordinates of the fault source through the Newton iteration method, and further determine the position of the fault source through particle filter fusion optimization. Associate various database data with the state of the fault source. The formula used is as follows: ； where is the state vector, including the coordinates of the fault source at the moment and its rate of change , is the state transition matrix, is the process noise; Step S4.2: For the subsequent variation law of the fault source position over time, list the observation equation: ； Among them, is the observation vector, where RMS is the root mean square value of the vibration signal, is the bearing temperature rise rate (unit: °C / min), is the acoustic initial value coordinate, and H is the observation matrix that maps the state vector to the observation space, is the observation noise, and the observation noise follows a Gaussian distribution Ν(0, R), where R is the noise covariance matrix; Step S4.3: During the particle filter iteration process, first initialize so that the generated particles follow , and then update the weights. The weight update formula is: 。 6. The reducer maintenance inspection system based on multi-source sensing fusion according to claim 1, characterized in that: The algorithm of the intelligent diagnosis subsystem specifically includes the following steps: Step Q1: After obtaining the coordinates from the data analysis subsystem, determine the component. First, perform a rough sample comparison, and load the sub-database according to the coordinates. Step Q2: Use the dynamic time warping method to align the real-time features with the samples, and calculate the comprehensive similarity. The calculation formula is: ； Among them, is the real-time vibration feature vector, is the database sample feature vector, is the temperature feature similarity (normalized difference), is the acoustic positioning coordinate matching degree. If S≥0.95, the fault type is directly judged. If 0.7≤S≤0.9, it is uploaded to the cloud to trigger secondary verification.

7. The reducer maintenance inspection system based on multi-source sensing fusion according to claim 6, characterized in that: Step Q1 includes Step Q1.1: Through the PCA dimensionality reduction technique, reduce the 32-dimensional vibration feature to 8 dimensions. The dimensionality reduction formula is: ； Among them, V is the original 32-dimensional feature vector, and P is the principal component matrix. Step Q1.2: Through the INT8 quantization technique, convert 32-bit floating-point numbers (FP32) to 8-bit integers (INT8) to reduce the calculation amount and speed up the inference speed. The quantization formula is: Among them, μ is the data mean, σ is the data standard deviation. Compress the FP32 floating-point numbers to INT8 integers, so that the model memory occupancy is reduced from 32MB to 8MB, and the inference time is reduced.

8. The reducer maintenance inspection system based on multi-source sensing fusion according to claim 6, wherein: Step Q2 includes Step Q2.1: In the interval of secondary verification (0.7 ≤ S ≤ 0.9), the system will immediately upload the relevant data to the cloud. Through the cloud, further determine the type, severity, and possible development trend of the fault. Step Q2.1: The cloud feeds back the judgment result to the local device and provides corresponding maintenance suggestions according to the fault situation.

9. The reducer maintenance inspection system based on multi-source sensing fusion according to any one of claims 1-2, characterized in that: The input shaft is vertically arranged on the high-speed shaft, and the high-speed shaft, low-speed shaft, and output shaft are arranged in parallel.

10. The reducer maintenance inspection system based on multi-source sensing fusion according to any one of claims 1-2, characterized in that: The gear set includes a first rotating gear arranged at one end of the output shaft facing the high-speed shaft. The high-speed shaft is provided with a second transmission gear meshing with the first transmission gear. The high-speed shaft is provided with a third transmission gear on one side of the second transmission gear. The low-speed shaft is provided with a fourth transmission gear meshing with the third transmission gear. The low-speed shaft is provided with a fifth transmission gear on one side of the fourth transmission gear. The output shaft is provided with a sixth transmission gear meshing with the fifth transmission gear.

Images