Speed reducer efficiency testing method based on intelligent sensor

Abstract

The invention relates to the technical field of speed reducer testing, and discloses a speed reducer efficiency testing method based on an intelligent sensor, which comprises the steps of judging whether vibration data of a speed reducer is normal or not, judging whether a preliminary judgment result has misjudgment or not, and analyzing and determining the cause of vibration data abnormity. According to the method, multiple factors possibly causing vibration data abnormity are comprehensively considered, the running state of the speed reducer is analyzed more comprehensively, the accuracy of abnormal reason judgment is improved through combined analysis of multi-dimensional data and comparison with a fault mode database, potential problems can be found and solved in time, and the efficiency is improved from data collection to final abnormity analysis. All the steps are connected with one another, operation and implementation are convenient, the running state of the speed reducer is monitored and analyzed in real time, problems can be found in time at the initial stage, measures can be taken, further development of faults is avoided, and maintenance cost and downtime are reduced.

Description

Technical Field

[0001] This invention relates to the field of speed reducer testing technology, specifically a speed reducer efficiency testing method based on intelligent sensors. Background Technology

[0002] Speed ​​reducers are a crucial component of mechanical transmission systems, and their efficiency directly affects the energy consumption and performance of the entire mechanical system. Accurate and reliable speed reducer efficiency testing is of great significance for product design optimization, quality control, fault diagnosis, and energy consumption assessment. The core of the speed reducer efficiency testing method based on intelligent sensors lies in utilizing "intelligent sensors" that integrate multi-functional sensing modules, microprocessors, and communication interfaces. These sensors are directly installed on key parts of the speed reducer's input and output shafts. The intelligent sensors can collect key operating parameters (such as torque, speed, and temperature) in real time, and perform preliminary processing and fusion of the raw data through the built-in microprocessor. Then, the processed data is transmitted to the central control / data acquisition unit via wireless or bus, and finally, the speed reducer efficiency is calculated and evaluated. Integrating multiple sensing functions into the intelligent sensor node simplifies the testing device, improves integration and intelligence, reduces costs, and enables more accurate and flexible speed reducer efficiency testing. Existing methods for testing reducer efficiency based on smart sensors cannot determine whether abnormal vibration data is affecting the reducer's vibration data by considering the reducer's own temperature, ambient temperature, and the operation and temperature of other equipment in the environment. They also cannot use diverse data to determine if preliminary conclusions are misjudged, nor can they use diverse data combined with a database to determine the cause of abnormal vibration data. This can easily lead to incorrect conclusions due to misjudgment, increasing production and usage costs, thus limiting their practicality. Summary of the Invention

[0003] This invention provides a speed reducer efficiency testing method based on intelligent sensors, which helps to solve the problems mentioned in the background art.

[0004] This invention provides the following technical solution: a method for testing the efficiency of a speed reducer based on intelligent sensors, comprising: Obtain vibration data from the reducer; Parameter determination function To determine whether the vibration data is normal; like If the vibration data is normal, then a normal misjudgment analysis will be performed. If a normal misjudgment exists, determine the cause of the misjudgment and correct it; If there are no normal misjudgments, record the vibration data; like If so, the vibration data is determined to be abnormal, and an anomaly misjudgment analysis is performed; If any abnormal misjudgments are found, a comprehensive analysis will be conducted. If no abnormalities are detected, then perform anomaly analysis.

[0005] As an optional solution to the speed reducer efficiency testing method based on intelligent sensors described in this invention, the following is included: Performing normal misjudgment analysis, specifically: Vibration signals were collected under different operating conditions of the reducer, and time-domain and frequency-domain features were extracted to form a normal vibration feature library matrix, denoted as Z. Normalize each feature component in the feature library; Extract the currently acquired vibration data into a feature vector. Calculate its Euclidean distance with each feature vector in the feature library, denoted as . ; Through normal misjudgment function To determine whether there is a normal misjudgment; like If so, it is determined that there is a normal misjudgment; like If so, it is determined that there is no normal misjudgment.

[0006] As an optional solution to the speed reducer efficiency testing method based on intelligent sensors described in this invention, the following is included: Anomaly misjudgment analysis is performed, specifically: During the operation of the reducer, temperature data is collected periodically to form a temperature time series dataset, denoted as . ; Calculate the standard deviation of the temperature data, denoted as . ; By setting a temperature fluctuation determination function To determine whether the temperature data of the reducer fluctuates significantly during operation; like If the temperature fluctuation of the reducer is small during operation, it can be determined that the reducer temperature fluctuation is small. Therefore, it is determined that there is no abnormal misjudgment; like If the temperature fluctuates significantly during operation, it can be determined that the reducer experiences large temperature fluctuations. Then, within the same timeframe as the reducer's operation period, ambient temperature data is collected to form an ambient temperature time series dataset, denoted as... ; Calculate the standard deviation of the ambient temperature data, denoted as . ; Determination function based on temperature influence To determine whether ambient temperature affects the speed reducer during operation; like If the ambient temperature has a significant impact on the reducer, it indicates that the ambient temperature is abnormal. Then, perform a temperature diffusion analysis; like If the ambient temperature has little effect on the reducer, then the ambient temperature is considered normal. Then, perform load analysis.

[0007] As an optional solution to the speed reducer efficiency testing method based on intelligent sensors described in this invention, the following is included: performing temperature diffusion analysis, specifically: The number of devices in the environment where the speed reducer is located is denoted as . ; Through device analysis functions To determine whether the abnormal ambient temperature is affected by other equipment; like If so, it is determined that there are no other devices. Then, it is determined that there is an abnormal misjudgment; like If so, it is determined that other devices exist; Then, obtain the airflow conditions in the environment where the reducer is located; To obtain the operating status of other equipment in the environment where the reducer is located; Through environmental comprehensive analysis function To determine whether the abnormal ambient temperature is affected by other equipment; like If so, it is determined that the ambient temperature was not affected by other equipment; Then, it is determined that there is an abnormal misjudgment; like If so, it is determined that the ambient temperature is affected by other equipment; Then, stop other operating equipment, and after the temperature of the stopped equipment has cooled and stabilized, collect the ambient temperature again. The other operating equipment refers to the equipment that is currently running among the other equipment. Calculate the new standard deviation of ambient temperature, denoted as . ; If the ambient temperature is normal, perform load analysis; If the ambient temperature is abnormal, it is determined that there is an abnormal misjudgment.

[0008] As an optional solution to the speed reducer efficiency testing method based on intelligent sensors described in this invention, the load analysis specifically includes: During the operation of the reducer, speed data is collected periodically to form a speed time series dataset, denoted as R; Calculate the difference between adjacent speed data points to obtain the speed change sequence, denoted as . ; Calculate the absolute value of the speed change sequence to obtain the absolute speed change sequence, denoted as . ; The number of points in the absolute speed change sequence that show drastic speed changes is denoted as . ; Speed ​​determination function To determine whether there are drastic changes in speed during the operation of the reducer; like If so, it is determined that there is a situation of drastic change in rotational speed; Then, it is determined that there is an abnormal misjudgment; like If so, it is determined that there is no situation where the rotational speed changes drastically; Therefore, during the operation of the reducer, load data is collected periodically to form a load time series dataset, denoted as L; Calculate the difference between adjacent load data points to obtain the load change sequence, denoted as . ; Calculate the absolute value of the load change sequence to obtain the absolute load change sequence, denoted as . ; The number of load change points in the absolute load change sequence is denoted as . ; Through load determination function To determine whether there is a dynamic load increase in the reducer during operation; like If so, it is determined that there is an increase in dynamic load; Then, it is determined that there is an abnormal misjudgment; like If so, it is determined that there is no increase in dynamic load; Therefore, it is determined that there is no abnormal misjudgment.

[0009] As an optional solution to the speed reducer efficiency testing method based on intelligent sensors described in this invention, a comprehensive analysis is performed, specifically: Perform a fast Fourier transform on the time-domain vibration signal in the vibration data to obtain the frequency-domain signal, denoted as X(f); Extracting gear meshing frequency Rotation frequency and the amplitude value at its harmonic frequency and ; The extracted key frequency features are compared with the spectrum baseline under normal conditions, and the deviation of amplitude at each frequency is calculated and denoted as ΔA(f). Set an amplitude deviation threshold, denoted as ; like If so, it can be determined that there is a problem with poor gear meshing; like If not, no judgment conclusion will be formed; like If so, it can be determined that there is a problem with the shaft system or bearings; like If not, no judgment conclusion will be formed; Acquire the temperature data of the reducer, denoted as TEM; Through local analysis functions Analyze whether there is localized overheating. like If so, it can be determined that there is a local overheating phenomenon; like If so, it can be determined that there is no local overheating phenomenon; Obtain the load data of the reducer and record it as LOAD; Based on the load vibration determination function To determine if there is any abnormal vibration under high load; like If the vibration data is abnormal, then it is determined that the vibration is abnormal under high load. like If the vibration data is normal, then the vibration is considered normal under high load. Acquire vibration signals from different sensor locations and calculate their phase difference PD(f); Phase difference analysis function Analyze whether there is local asymmetric vibration; like If so, it can be determined that there is local asymmetric vibration; like If so, it can be determined that there is no local asymmetric vibration; If any of the conditions for judging poor gear meshing or judging shaft or bearing problems are met, then the abnormal vibration data is determined to be caused by mechanical structure problems or poor meshing, and anomaly analysis is performed. If all conditions for judging poor gear meshing or shaft or bearing problems are not met, then an anomaly cause analysis should be performed.

[0010] As an optional solution to the speed reducer efficiency testing method based on intelligent sensors described in this invention, the following is included: Anomaly cause analysis is performed, specifically: Key features are extracted from the collected vibration signals to form a feature vector. ; Read the database containing characteristics of various failure modes and their corresponding root causes; Calculate eigenvectors With each fault mode feature vector in the database The similarity score is denoted as ; Find the minimum distance Corresponding Fault Mode Index ; Using distance comparison function Determine the cause of the malfunction; like The cause of the fault is confirmed to be an index in the database. The root cause corresponding to the failure mode; like If it is, then it is marked as unknown; The abnormal information, the root cause description of the best-matching failure mode, and the similarity score are compiled into a report format for output.

[0011] As an optional solution to the speed reducer efficiency testing method based on intelligent sensors described in this invention, the following is included: Anomaly analysis is performed, specifically: Extracting gear meshing frequency Rotation frequency and the amplitude value at its harmonic frequency and ; Extract bearing failure frequency ; Calculate bearing failure frequency The amplitude value at the harmonic frequency is denoted as ; Compare the extracted key frequency features with the baseline amplitude of the spectrum under normal conditions. , as well as Calculate the amplitude deviation; Through type analysis functions Determine the type of abnormality in the reducer and establish a preliminary conclusion on the cause. like If so, it can be preliminarily determined that the gears are not meshing properly; like If so, it is preliminarily determined to be a mechanical structure problem; Obtain gearbox tooth surface temperature data and bearing area temperature data ; Through temperature analysis function The preliminary cause determination conclusion is then re-evaluated to obtain the secondary cause determination conclusion. like If so, the second determination is that the gear meshing is faulty; like If so, the secondary determination is a mechanical structure problem; If the preliminary cause determination conclusion is consistent with the secondary cause determination conclusion, and both are determined to be poor gear meshing, then the cause of the abnormal vibration data is determined to be poor gear meshing. If the preliminary cause determination conclusion is consistent with the secondary cause determination conclusion, and both are determined to be mechanical structure problems, then the cause of the abnormal vibration data is determined to be a mechanical structure problem. If the preliminary cause determination conclusion is inconsistent with the secondary cause determination conclusion, then a manual determination shall be made.

[0012] As an optional solution to the speed reducer efficiency testing method based on intelligent sensors described in this invention, the following steps are taken: determining the cause of misjudgment and correcting it: Extract the minimum distance between the current feature and the feature library. The corresponding eigenvector is defined as the smallest eigenvector, denoted as . ; Calculate the current feature vector The feature vector with the smallest distance The pointwise difference is denoted as ; Extracting satisfaction The feature dimension h; Adjust the feature extraction algorithm or parameters based on the significantly different feature dimensions; New feature vector Adding normal vibration feature library to form a new feature library matrix. ; The adjusted feature extraction method is used to re-extract features from the currently acquired vibration data, resulting in new feature vectors. ; Calculate the new feature vector Euclidean distance to each feature vector in the updated normal vibration feature library Find the feature vector with the smallest distance. ; By verification function To verify whether the misjudgment has been corrected; like If so, the misjudgment has been corrected; like If the error is not corrected, it will be determined that a misjudgment has not been corrected and manual correction will be performed.

[0013] The present invention has the following beneficial effects:

[0014] 1. This method for testing the efficiency of a reducer based on intelligent sensors acquires vibration data from the reducer and determines whether the vibration data is normal. If the vibration data is normal, a normal misjudgment analysis is performed. This involves collecting vibration signals under different operating conditions, generating frequency and time domain feature maps to form a "normal vibration feature library," and comparing the currently acquired vibration data with the data in the "normal vibration feature library." If they match, it is determined that there is no normal misjudgment, and the vibration data is recorded. If they do not match, it is determined that there is a normal misjudgment, and the cause of the misjudgment is identified and corrected. This method can quickly filter vibration data, promptly identify potential anomalies, and save time and resources for subsequent detailed analysis. By establishing a normal vibration feature library and comparing it, anomalies in normal data can be identified, misjudgments can be avoided, and the accuracy of monitoring can be improved.

[0015] 2. This intelligent sensor-based reducer efficiency testing method, if the vibration data is abnormal, performs anomaly misjudgment analysis. Specifically, it acquires reducer temperature data and determines if the temperature fluctuates significantly. If the temperature fluctuates significantly, it acquires the ambient temperature and analyzes its impact on the reducer (e.g., a high ambient temperature causing the reducer temperature to rise rapidly at a certain time). If the ambient temperature is normal, it performs load analysis. If the ambient temperature is abnormal, it acquires the number of devices in the reducer's environment. If other devices are present in the reducer's environment, it uses airflow patterns and the operating status of other devices (including whether they are running and their temperatures) to determine if the abnormal ambient temperature is due to the influence of other devices (e.g., the ambient temperature itself is not high, but other devices...). (The equipment has many high-temperature devices, and the airflow blows the high temperature towards the vicinity of the reducer.) If the ambient temperature is abnormal due to the influence of other equipment, the other operating equipment is stopped. After the temperature of the stopped equipment cools down and stabilizes, the ambient temperature is collected again. If the ambient temperature is normal, load analysis is performed. If the ambient temperature is abnormal, it is determined that there is an abnormality and a misjudgment. If the ambient temperature is abnormal and is not affected by other equipment, it is determined that there is an abnormality and a misjudgment. If there is no other equipment in the environment where the reducer is located (i.e., the ambient temperature itself is in a high-temperature environment), it is determined that there is an abnormality and a misjudgment. If the temperature data fluctuation is small, it is determined that there is no abnormality and a misjudgment. The influence of ambient temperature and load changes on vibration data is considered, avoiding misjudgments caused by these external factors and improving the reliability of abnormal situation judgment.

[0016] 3. This intelligent sensor-based reducer efficiency testing method acquires the reducer's speed data during load analysis to determine if there are drastic speed changes during operation. If drastic speed changes are found, an anomaly is identified. If no drastic speed changes are found, the method acquires the reducer's load data to determine if there is a dynamic load increase during operation. If a dynamic load increase is found, an anomaly is identified. If no dynamic load increase is found, no anomaly is identified. This further analyzes the reducer's operating status under different load conditions, which helps to identify anomalies caused by load changes and provides a basis for subsequent maintenance and adjustment.

[0017] 4. This intelligent sensor-based reducer efficiency testing method combines vibration signal analysis with multi-dimensional data for comprehensive analysis. If the abnormal vibration data is determined to be due to mechanical structure problems or poor meshing, anomaly analysis is performed. This involves using frequency domain analysis and temperature monitoring to determine whether the abnormal vibration data is caused by mechanical structure problems or poor meshing. If the abnormal vibration data is determined not to be due to mechanical structure problems or poor meshing, the cause of the anomaly is determined by comparing with a fault mode database, and an anomaly report is generated. By combining vibration signal analysis and multi-dimensional data, the method can more comprehensively analyze the causes of abnormal situations, improving the diagnostic capability for complex problems. Through frequency domain analysis and temperature monitoring, the cause of the anomaly can be accurately determined, providing a clear direction for subsequent maintenance and improvement. Attached Figure Description

[0018] Figure 1 This is a flowchart of the speed reducer efficiency testing method based on intelligent sensors according to the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Example 1: A method for testing the efficiency of a speed reducer based on intelligent sensors, see [link / reference]. Figure 1 ,include: Obtain vibration data from the reducer; Parameter determination function To determine if the vibration data is normal: ; in, This indicates that the parameter is normal. This indicates that the parameter is abnormal. This means that all conditions must be met simultaneously. It means "or", meaning that either condition must be met. The determination result representing the time-domain amplitude is as follows: ; in, This is the time-domain amplitude threshold, used to determine whether the time-domain amplitude is normal. If the RMS value is less than... If the time domain condition is met, the signal is considered normal; otherwise, it is considered abnormal. Here, x represents the vibration signal, and RMS(x) is the root mean square value of the vibration signal. Specifically: ; in, Let be the vibration signal value at the i-th sampling point, and N be the total number of sampling points. The specific steps for calculating the root mean square (RMS) value of the vibration signal are as follows: obtain discrete data points of the vibration signal through sensor sampling; if there is a DC offset in the signal, first subtract the average value of the signal to eliminate the influence of the DC component; square each data point and add all the squared values; divide the sum of squares by the total number of data points to obtain the average value of the squares; take the square root of the average value to obtain the root mean square (RMS) value of the vibration signal. Parameter determination function middle, The determination result of the critical frequency amplitude is as follows: ; in, This is the frequency domain amplitude threshold for the gear meshing frequency, used to determine whether the frequency domain amplitude of the gear meshing frequency is normal. This is the frequency domain amplitude threshold for the rotation frequency, used to determine whether the frequency domain amplitude of the rotation frequency is normal. The amplitude of the frequency domain signal at the gear meshing frequency is specifically: ; in, The gear meshing frequency is given by the following formula: ; Where z is the number of gear teeth and n is the rotational speed; In the formula for the amplitude of a frequency domain signal at the gear meshing frequency, To represent the frequency domain signal corresponding to the vibration signal in the time domain obtained by Fast Fourier Transform (FFT), In this context, k is the frequency index. Gear meshing frequency The corresponding frequency index, the specific formula is: ; Where N is the number of sampling points, which are the signal values ​​at discrete time points selected when converting a continuous signal into a discrete signal. It is the sampling frequency. This indicates rounding down, which means rounding the value to the nearest integer. In determining the critical frequency amplitude, The amplitude of the frequency domain signal at the rotation frequency is specifically: ; in, The rotation frequency is given by the following formula: ; Where n is the rotational speed; In the formula for the amplitude of a frequency domain signal at the rotation frequency, To represent the frequency domain signal corresponding to the vibration signal in the time domain obtained by Fast Fourier Transform (FFT), In this context, k is the frequency index. Rotation frequency The corresponding frequency index, the specific formula is: ; Where N is the number of sampling points, It is the sampling frequency. This indicates rounding down, which means rounding the value to the nearest integer. like If the vibration data is normal, then a normal misjudgment analysis will be performed. If a normal misjudgment exists, determine the cause of the misjudgment and correct it; If there are no normal misjudgments, record the vibration data; like If so, the vibration data is determined to be abnormal, and an anomaly misjudgment analysis is performed; If any abnormal misjudgments are found, a comprehensive analysis will be conducted. If no abnormalities are detected, then perform anomaly analysis.

[0021] The above method comprehensively considers various factors that may cause abnormal vibration data, including mechanical structure problems, poor meshing, temperature effects, and load changes. It enables a more comprehensive analysis of the reducer's operating status. By combining and analyzing multi-dimensional data and comparing it with a fault mode database, the accuracy of identifying the cause of abnormalities is improved, which helps to discover and resolve potential problems in a timely manner. It provides a systematic method and process, from data acquisition to final anomaly analysis, with each step interconnected, facilitating operation and implementation. Through real-time monitoring and analysis of the reducer's operating status, problems can be detected and addressed in their early stages, preventing further development of the fault and reducing maintenance costs and downtime.

[0022] Example 2 is an improvement on Example 1. This method for testing the efficiency of a reducer based on intelligent sensors performs anomaly misjudgment analysis, specifically as follows: During the operation of the reducer, temperature data is collected periodically to form a temperature time series dataset, denoted as . ; in, This represents the temperature value of the reducer at time rf, where rf is the index, and its value ranges from 1 to tn, where tn is the number of sampling points; Calculate the standard deviation of the temperature data, denoted as . ; in, The average value of the temperature data is as follows: ; By setting a temperature fluctuation determination function To determine whether the temperature data of the reducer fluctuates significantly during operation: ; in, This is the temperature fluctuation threshold, used to determine whether the temperature data of the reducer fluctuates significantly during operation. like If the temperature fluctuation of the reducer is small during operation, it can be determined that the reducer temperature fluctuation is small. Therefore, it is determined that there is no abnormal misjudgment; like If the temperature fluctuates significantly during operation, it can be determined that the reducer experiences large temperature fluctuations. Then, within the same timeframe as the reducer's operation period, ambient temperature data is collected to form an ambient temperature time series dataset, denoted as... : ; in, This represents the ambient temperature value at time ef, where ef is the index, and its value ranges from 1 to tn. Calculate the standard deviation of the ambient temperature data, denoted as . : ; in, The average value of the ambient temperature data is as follows: ; Determination function based on temperature influence To determine whether ambient temperature affects the speed reducer during operation: ; in, The ambient temperature fluctuation threshold is used to determine whether the ambient temperature affects the reducer during operation, that is, to determine whether the large temperature fluctuation of the reducer during operation is due to the high ambient temperature. like If the ambient temperature has a significant impact on the reducer, it indicates that the ambient temperature is abnormal. Then, perform a temperature diffusion analysis; like If the ambient temperature has little effect on the reducer, then the ambient temperature is considered normal. Then, perform load analysis.

[0023] Specifically, temperature diffusion analysis was performed as follows: The number of devices in the environment where the speed reducer is located is denoted as . The number of devices refers to the total number of devices in the environment excluding the speed reducer; Through device analysis functions To determine whether the abnormal ambient temperature is affected by other equipment, the other equipment refers to other equipment in the environment where the reducer is located, excluding the reducer itself. ; like If so, it is determined that there are no other devices. Then, it is determined that there is an abnormal misjudgment; like If so, it is determined that other devices exist; Then, obtain the airflow conditions in the environment where the reducer is located, including the wind speed and wind direction in the environment; The operating status of other equipment in the environment where the reducer is located is obtained, including whether the equipment is running and the temperature of the equipment. Through environmental comprehensive analysis function Determine whether the abnormal ambient temperature is affected by other equipment: ; in, The number of high-temperature equipment refers to the total number of other equipment that is in operation and has a high temperature. The specific formula for counting the number of high-temperature equipment is as follows: ; in, It is the total number of other devices in the environment. It is the temperature of the eith device. This refers to the operating status of the eith device. This is the high temperature threshold, used to determine whether the temperature of the equipment is too high. "Running" indicates that the equipment is in operation, and "stop" indicates that the equipment is in shutdown. If the equipment temperature is high and the equipment is in operation, it is considered a high temperature equipment, and the output is 1. At this time, the number of high temperature equipment is incremented by 1. If the equipment temperature is low or the equipment is in shutdown, it is considered that the equipment is not a high temperature equipment, and the output is 0. At this time, the number of high temperature equipment is not increased. Environmental comprehensive analysis function In this context, AIR is the high-temperature diffusion analysis function, specifically: ; in, The wind speed threshold, The wind direction thresholds, using the reducer location as a reference point, define the range of wind directions that might blow high temperatures towards the reducer. These two thresholds are used to determine whether, under current ambient airflow conditions, the heat generated by the high-temperature equipment might be blown towards the reducer, causing an increase in the ambient temperature around the reducer and thus affecting its operation and temperature. Wind speed is the speed at which air flows within an environment. AIR represents the direction of airflow in the environment, i.e., wind direction. If AIR=1, it is assumed that the airflow may blow the high temperature towards the reducer. If AIR=0, it is assumed that the airflow cannot blow the high temperature towards the reducer. like If so, it is determined that the ambient temperature was not affected by other equipment; Then, it is determined that there is an abnormal misjudgment; like If so, it is determined that the ambient temperature is affected by other equipment; Then, stop other operating equipment, and after the temperature of the stopped equipment has cooled and stabilized, collect the ambient temperature again. The other operating equipment refers to the equipment that is currently running among the other equipment. Calculate the new standard deviation of ambient temperature, denoted as . : ; in, The average value of the new ambient temperature data is as follows: ; If the ambient temperature is normal, that is Then perform load analysis; If the ambient temperature is abnormal, that is If so, it is determined that there is an abnormal misjudgment.

[0024] The load analysis includes, specifically: During the operation of the reducer, speed data is collected periodically to form a speed time series dataset, denoted as R: ; in, This represents the rotational speed at time q, where rm is the number of sampling points; Calculate the difference between adjacent speed data points to obtain the speed change sequence, denoted as ΔR: ; in, ; Calculate the absolute value of the speed change sequence to obtain the absolute speed change sequence, denoted as . : ; The number of points in the absolute speed change sequence that show drastic speed changes is denoted as . : ; in, This is the threshold for drastic speed changes, used to extract data points with drastic speed changes in the absolute speed change sequence; Speed ​​determination function To determine whether there are drastic changes in speed during the operation of the reducer: ; in, The threshold for the number of points with drastic speed changes is used to determine whether there are drastic speed changes in the reducer during operation; like If so, it is determined that there is a situation of drastic change in rotational speed; Then, it is determined that there is an abnormal misjudgment; like If so, it is determined that there is no situation where the rotational speed changes drastically; Therefore, during the operation of the reducer, load data is collected periodically to form a load time series dataset, denoted as L: ; in, This represents the load value at time q, where rm is the number of sampling points; Calculate the difference between adjacent load data points to obtain the load change sequence, denoted as ΔL: ; in, ; Calculate the absolute value of the load change sequence to obtain the absolute load change sequence, denoted as . : ; The number of load change points in the absolute load change sequence is denoted as . : ; in, This is the load change threshold, used to extract data points of load change in the absolute load change sequence; Through load determination function To determine whether there is an increase in dynamic load during the operation of the reducer: ; in, This is a threshold for the number of load change points, used to determine whether there is a dynamic increase in load during the operation of the reducer; like If so, it is determined that there is an increase in dynamic load; Then, it is determined that there is an abnormal misjudgment; like If so, it is determined that there is no increase in dynamic load; Therefore, it is determined that there is no abnormal misjudgment.

[0025] This embodiment also provides a comprehensive analysis, specifically as follows: Perform a Fast Fourier Transform (FFT) on the time-domain vibration signal in the vibration data to obtain the frequency-domain signal, denoted as . : ; Where N is the number of sampling points, f is the frequency, x(t) represents the signal value of the time-domain signal at time point t. In the vibration analysis of the reducer, it represents the intensity of the vibration signal collected by the vibration sensor at time point t. t is the time index, ranging from 0 to N-1, corresponding to the time position of the sampling point. Each t value represents a discrete time point used to identify the position of x(t) in the time domain. i is the imaginary unit, satisfying... In the formula, is used to represent a complex exponential function in order to convert a time-domain signal into a frequency-domain signal; Extracting gear meshing frequency Rotation frequency and the amplitude value at its harmonic frequency and ; The extracted key frequency features are compared with the spectral baseline under normal conditions, and the amplitude deviation at each frequency is calculated and denoted as . : ; in, This represents the baseline amplitude of the spectrum under normal conditions. Set an amplitude deviation threshold, denoted as This is used to determine the cause of abnormalities in the speed reducer; like If so, it can be determined that there may be a problem with poor gear meshing; like If not, no judgment conclusion will be formed; in, This represents the deviation between the amplitude value at the currently extracted gear meshing frequency and the baseline amplitude value at the gear meshing frequency under normal conditions. It is used to determine whether the current gear meshing state is normal. Specifically: , It is the baseline amplitude value of the spectrum at the gear meshing frequency under normal conditions; like If so, it can be determined that there may be a problem with the shaft system or bearings; like If not, no judgment conclusion will be formed; Here, represents the deviation between the amplitude value at the currently extracted rotational frequency and the baseline amplitude value of the spectrum at the rotational frequency under normal conditions, used to determine whether the current shaft system or bearing condition is normal, specifically: , It is the baseline amplitude value of the spectrum at the rotation frequency under normal conditions; Acquire the temperature data of the reducer, denoted as TEM; Through local analysis functions Analyze whether there is localized overheating: ; in, This is the local temperature threshold, used to analyze whether local overheating exists; like If so, it can be determined that there may be local overheating. like If so, it can be determined that there is likely no local overheating phenomenon; Obtain the load data of the reducer and record it as LOAD; Based on the load vibration determination function To determine if there is any abnormal vibration under high load: ; in, This is the load threshold, used to determine whether there is abnormal vibration under high load. like If the vibration data is abnormal, then it is determined that the vibration is abnormal under high load. like If the vibration data is normal, then the vibration is considered normal under high load. Obtain vibration signals from different sensor locations and calculate their phase difference PD(f): ; in, and They represent and phase angle, and These are vibration signals from different sensor locations. and The frequency domain signal obtained after Fast Fourier Transform (FFT); Phase difference analysis function Analyze whether local asymmetric vibrations exist: ; in, This is the phase difference threshold, used to analyze whether local asymmetric vibrations exist; like If so, it can be determined that there may be local asymmetric vibration; like If so, it can be determined that there may be no local asymmetric vibration; If any of the conditions for judging poor gear meshing or judging shaft or bearing problems are met, then the abnormal vibration data is determined to be caused by mechanical structure problems or poor meshing, and anomaly analysis is performed. If all conditions for judging poor gear meshing or shaft or bearing problems are not met, then an anomaly cause analysis should be performed.

[0026] The analysis of the causes of the anomalies includes: Key features are extracted from the collected vibration signals to form a feature vector. ; Read the database containing characteristics of various failure modes and their corresponding root causes; Calculate the eigenvectors using the weighted Euclidean distance formula. With each fault mode feature vector in the database The similarity score is denoted as : ; in, Let g be the weight of the g-th feature. and Let be the g-th component of the current anomaly feature and the k-th fault mode feature, respectively, and dn represent the dimension of the feature vector, i.e., the number of features, which represents the current anomaly feature vector. and fault mode feature vectors in the database The number of features contained in each feature, in practical applications, may include features such as the root mean square value, peak value, and amplitude at different frequencies of the vibration signal. Each feature describes the operating state of the reducer from different perspectives. By comparing these features, different fault modes can be identified and classified. Find the minimum distance Corresponding Fault Mode Index ; Using distance comparison function Determine the cause of the fault: ; in, This is the distance matching threshold, used to determine the cause of the fault; like The cause of the fault is confirmed to be an index in the database. The root cause corresponding to the failure mode; like If it is, then it is marked as unknown; The abnormal information, the root cause description of the best-matching failure mode, and the similarity score are compiled into a report format for output.

[0027] This embodiment also provides anomaly analysis, specifically: Extracting gear meshing frequency Rotation frequency and the amplitude value at its harmonic frequency and ; Extract bearing failure frequency ; Calculate bearing failure frequency The amplitude value at the harmonic frequency is denoted as : ; Compare the extracted key frequency features with the baseline amplitude of the spectrum under normal conditions. , as well as Calculate the amplitude deviation: ; ; ; Through type analysis functions Determine the type of gearbox malfunction and establish a preliminary cause assessment conclusion: ; in, The amplitude deviation threshold is used to determine the type of abnormality in the reducer and to make a preliminary conclusion on the cause. like If so, it can be preliminarily determined that the gears are not meshing properly; like If so, it is initially determined to be a mechanical structure problem, such as bearing wear or shaft bending; Obtain gearbox tooth surface temperature data and bearing area temperature data ; Through temperature analysis function A second determination was made based on the preliminary causal assessment, resulting in a second causal assessment conclusion: ; in, The temperature threshold of the part is used to make a secondary determination on the preliminary cause determination conclusion to obtain the secondary cause determination conclusion. like If so, the second determination is that the gear meshing is faulty; like If so, the secondary determination is a mechanical structure problem, such as bearing wear or shaft bending; If the preliminary cause determination conclusion is consistent with the secondary cause determination conclusion, and both are determined to be poor gear meshing, then the cause of the abnormal vibration data is determined to be poor gear meshing. If the preliminary cause determination conclusion is consistent with the secondary cause determination conclusion, and both are determined to be mechanical structure problems, such as bearing wear or shaft bending, then the cause of the abnormal vibration data is determined to be a mechanical structure problem, such as bearing wear or shaft bending. If the preliminary cause determination conclusion is inconsistent with the secondary cause determination conclusion, then a manual determination shall be made.

[0028] Example 3 is an improvement on Example 2. In this example, normal misjudgment analysis is performed, specifically as follows: Vibration signals were collected under different operating conditions of the reducer, and time-domain and frequency-domain features were extracted to form a normal vibration feature library matrix, denoted as Z: ; in, This represents the feature vector of the m-th operating condition, where m is the total number of operating conditions in the normal vibration feature library; Normalize each feature component in the feature library to eliminate dimensional differences: ; in, It is the b-th characteristic component of the a-th working condition. and These are the minimum and maximum values ​​of feature b under all operating conditions. This represents the normalized value; Extract the currently acquired vibration data into a feature vector. Calculate its Euclidean distance with each feature vector in the feature library, denoted as . : ; in, It is the b-th component of the current feature, where n is the total dimension of the feature. This represents the result after normalization; Through normal misjudgment function Determine if there are any normal misjudgments: ; in, The minimum distance between the current feature and the feature library. This is the normal misjudgment threshold, used to determine whether there is a misjudgment when the vibration data is currently judged to be normal, that is, whether the currently collected vibration data is indeed normal. The specific formula is: ; Where k is an empirical value, such as k = 1.5. The average distance between all features in the feature library is as follows: ; Where m is the number of feature vectors in the feature library. and This represents two feature vectors in the feature library. It is the Euclidean distance between them. In the formula for the normal false alarm threshold, The standard deviation of all features in the feature library is: ; like If so, it is determined that there is a normal misjudgment; like If so, it is determined that there is no normal misjudgment; The specific steps for constructing the normal vibration feature library matrix are as follows: Select several typical operating conditions of the reducer, such as low speed and high load, high speed and low load, and medium speed and medium load, to ensure that its main working states are covered. Under each operating condition, vibration sensors are used to collect vibration signals from the reducer to obtain a sufficient number of data samples. The acquired vibration signals are preprocessed, including removing DC components and filtering, to reduce noise interference and improve signal quality. Calculate statistical characteristics such as mean, variance, peak value, root mean square (RMS), and kurtosis to reflect the fluctuation characteristics of the signal on the time axis; Perform a Fast Fourier Transform (FFT) on the signal to extract the amplitude spectrum, energy distribution, and amplitude at key frequencies (such as gear meshing frequency and rotation frequency), thus revealing the frequency components of the signal. By drawing the correspondence between different time-domain and frequency-domain features and operating conditions, or by using simple clustering algorithms (such as K-Means) to perform cluster analysis on the features, we can verify whether the selected features can effectively distinguish different operating conditions and ensure the discriminative ability of the feature library. The time-domain and frequency-domain features under different operating conditions are arranged in a certain order to form a "normal vibration feature library" matrix. The rows of the matrix represent different operating conditions, and the columns represent each feature value, which facilitates the subsequent judgment and analysis of the reducer's operating status.

[0029] This embodiment also provides a method for determining the cause of misjudgment and correcting it, specifically as follows: Extract the minimum distance between the current feature and the feature library. The corresponding eigenvector is defined as the smallest eigenvector, denoted as . ; Calculate the current feature vector The feature vector with the smallest distance The pointwise difference is denoted as : ; Extracting satisfaction The feature dimension h, where, This is the difference threshold, used to identify significantly different feature dimensions; Based on the significantly different feature dimensions, the feature extraction algorithm or parameters can be adjusted. For example, if the amplitude difference at a certain frequency is significant, the parameters of the bandpass filter can be optimized to extract the amplitude at that frequency more accurately. New feature vector Adding normal vibration feature library to form a new feature library matrix. To better reflect the actual operating status of the reducer; The adjusted feature extraction method is used to re-extract features from the currently acquired vibration data, resulting in new feature vectors. ; Calculate the new feature vector Euclidean distance to each feature vector in the updated normal vibration feature library Find the feature vector with the smallest distance. : ; By verification function Verify whether the misjudgment has been corrected: ; like If so, the misjudgment has been corrected; like If the error is not corrected, it will be determined that a misjudgment has not been corrected and manual correction will be performed.

[0030] This embodiment comprehensively considers various factors that may lead to abnormal vibration data, including mechanical structure problems, poor meshing, temperature effects, and load changes. It can more comprehensively analyze the operating status of the reducer. By combining and analyzing multi-dimensional data and comparing it with a fault mode database, it improves the accuracy of judging the cause of abnormalities, helps to discover and solve potential problems in a timely manner, and provides a systematic method and process. From data acquisition to the final anomaly analysis, each step is interconnected, which is convenient for operation and implementation. Through real-time monitoring and analysis of the reducer's operating status, problems can be discovered in the early stages and measures can be taken in time to avoid further development of the fault and reduce maintenance costs and downtime.

Claims

1. A method for testing the efficiency of a speed reducer based on intelligent sensors, characterized in that: include: Obtain vibration data from the reducer; Parameter determination function To determine whether the vibration data is normal; like If the vibration data is normal, then a normal misjudgment analysis will be performed. If a normal misjudgment exists, determine the cause of the misjudgment and correct it; If there are no normal misjudgments, record the vibration data; like If so, the vibration data is determined to be abnormal, and an anomaly misjudgment analysis is performed; If any abnormal misjudgment is found, a comprehensive analysis will be conducted; If no abnormalities are detected, then perform anomaly analysis.

2. The method for testing the efficiency of a reducer based on a smart sensor according to claim 1, characterized in that: Perform normal misjudgment analysis, specifically as follows: Vibration signals were collected under different operating conditions of the reducer, and time-domain and frequency-domain features were extracted to form a normal vibration feature library matrix, denoted as Z. Normalize each feature component in the feature library; Extract the currently acquired vibration data into a feature vector. Calculate its Euclidean distance with each feature vector in the feature library, denoted as . ; Through normal misjudgment function To determine whether there is a normal misjudgment; like If so, it is determined that there is a normal misjudgment; like If so, it is determined that there is no normal misjudgment.

3. The method for testing the efficiency of a reducer based on a smart sensor according to claim 1, characterized in that: Anomaly and false positive analysis are performed, specifically as follows: During the operation of the reducer, temperature data is collected periodically to form a temperature time series dataset, denoted as . ; Calculate the standard deviation of the temperature data, denoted as . ; By setting a temperature fluctuation determination function To determine whether the temperature data of the reducer fluctuates significantly during operation; like If the temperature fluctuation of the reducer is small during operation, it can be determined that the reducer temperature fluctuation is small. Therefore, it is determined that there is no abnormal misjudgment; like If the temperature fluctuates significantly during operation, it can be determined that the reducer experiences large temperature fluctuations. Then, within the same timeframe as the reducer's operation period, ambient temperature data is collected to form an ambient temperature time series dataset, denoted as... ; Calculate the standard deviation of the ambient temperature data, denoted as . ; Determination function based on temperature influence To determine whether ambient temperature affects the speed reducer during operation; like If the ambient temperature has a significant impact on the reducer, it indicates that the ambient temperature is abnormal. Then, perform a temperature diffusion analysis; like If the ambient temperature has little effect on the reducer, then the ambient temperature is considered normal. Then, perform load analysis.

4. The method for testing the efficiency of a reducer based on a smart sensor according to claim 3, characterized in that: Perform temperature diffusion analysis, specifically: The number of devices in the environment where the speed reducer is located is denoted as . ; Through device analysis functions To determine whether the abnormal ambient temperature is affected by other equipment; like If so, it is determined that there are no other devices; Then, it is determined that there is an abnormal misjudgment; like If so, it is determined that other devices exist; Then, obtain the airflow conditions in the environment where the reducer is located; To obtain the operating status of other equipment in the environment where the reducer is located; Through environmental comprehensive analysis function To determine whether the abnormal ambient temperature is affected by other equipment; like If so, it is determined that the ambient temperature was not affected by other equipment; Then, it is determined that there is an abnormal misjudgment; like If so, it is determined that the ambient temperature is affected by other equipment; Then, stop other operating equipment, and after the temperature of the stopped equipment has cooled and stabilized, collect the ambient temperature again. The other operating equipment refers to the equipment that is currently running among the other equipment. Calculate the new standard deviation of ambient temperature, denoted as . ; If the ambient temperature is normal, perform load analysis; If the ambient temperature is abnormal, it is determined that there is an abnormal misjudgment.

5. The method for testing the efficiency of a reducer based on a smart sensor according to claim 4, characterized in that: Perform load analysis, specifically: During the operation of the reducer, speed data is collected periodically to form a speed time series dataset, denoted as R; Calculate the difference between adjacent speed data points to obtain the speed change sequence, denoted as . ; Calculate the absolute value of the speed change sequence to obtain the absolute speed change sequence, denoted as . ; The number of points in the absolute speed change sequence that show drastic speed changes is denoted as . ; Speed ​​determination function To determine whether there are drastic changes in speed during the operation of the reducer; like If so, it is determined that there is a situation of drastic change in rotational speed; Then, it is determined that there is an abnormal misjudgment; like If so, it is determined that there is no situation where the rotational speed changes drastically; Therefore, during the operation of the reducer, load data is collected periodically to form a load time series dataset, denoted as L; Calculate the difference between adjacent load data points to obtain the load change sequence, denoted as . ; Calculate the absolute value of the load change sequence to obtain the absolute load change sequence, denoted as . ; The number of load change points in the absolute load change sequence is denoted as . ; Through load determination function To determine whether there is an increase in dynamic load during the operation of the reducer; like If so, it is determined that there is an increase in dynamic load; Then, it is determined that there is an abnormal misjudgment; like If so, it is determined that there is no increase in dynamic load; Therefore, it is determined that there is no abnormal misjudgment.

6. The method for testing the efficiency of a reducer based on a smart sensor according to claim 1, characterized in that: A comprehensive analysis was conducted, specifically as follows: Perform a fast Fourier transform on the time-domain vibration signal in the vibration data to obtain the frequency-domain signal, denoted as X(f); Extracting gear meshing frequency Rotation frequency and the amplitude value at its harmonic frequency and ; The extracted key frequency features are compared with the spectrum baseline under normal conditions, and the deviation of amplitude at each frequency is calculated and denoted as ΔA(f). Set an amplitude deviation threshold, denoted as ; like If so, it can be determined that there is a problem with poor gear meshing; like If not, no judgment conclusion will be formed; like If so, it can be determined that there is a problem with the shaft system or bearings; like If not, no judgment conclusion will be formed; Acquire the temperature data of the reducer, denoted as TEM; Through local analysis functions Analyze whether there is localized overheating. like If so, it can be determined that there is a local overheating phenomenon; like If so, it can be determined that there is no local overheating phenomenon; Obtain the load data of the reducer and record it as LOAD; Based on the load vibration determination function To determine if there is any abnormal vibration under high load; like If the vibration data is abnormal, then it is determined that the vibration is abnormal under high load. like If the vibration data is normal, then the vibration is considered normal under high load. Acquire vibration signals from different sensor locations and calculate their phase difference PD(f); Phase difference analysis function Analyze whether there is local asymmetric vibration; like If so, it can be determined that there is local asymmetric vibration; like If so, it can be determined that there is no local asymmetric vibration; If any of the conditions for judging poor gear meshing or judging shaft or bearing problems are met, then the abnormal vibration data is determined to be caused by mechanical structure problems or poor meshing, and anomaly analysis is performed. If all conditions for judging poor gear meshing or shaft or bearing problems are not met, then an anomaly cause analysis should be performed.

7. The method for testing the efficiency of a reducer based on a smart sensor according to claim 6, characterized in that: An anomaly cause analysis was conducted, specifically as follows: Key features are extracted from the collected vibration signals to form a feature vector. ; Read the database containing characteristics of various failure modes and their corresponding root causes; Calculate eigenvectors With each fault mode feature vector in the database The similarity score is denoted as ; Find the minimum distance Corresponding Fault Mode Index ; Using distance comparison function Determine the cause of the malfunction; like The cause of the fault is confirmed to be an index in the database. The root cause corresponding to the failure mode; like If it is, then it is marked as unknown; The abnormal information, the root cause description of the best-matching failure mode, and the similarity score are compiled into a report format for output.

8. The method for testing the efficiency of a reducer based on a smart sensor according to claim 7, characterized in that: Anomaly analysis is performed, specifically as follows: Extracting gear meshing frequency Rotation frequency and the amplitude value at its harmonic frequency and ; Extract bearing failure frequency ; Calculate bearing failure frequency The amplitude value at the harmonic frequency is denoted as ; Compare the currently extracted key frequency features with the baseline amplitude of the spectrum under normal conditions. , as well as Calculate the amplitude deviation; Through type analysis functions Determine the type of abnormality in the reducer and establish a preliminary conclusion on the cause. like If so, it can be preliminarily determined that the gears are not meshing properly; like If so, it is preliminarily determined to be a mechanical structure problem; Obtain gearbox tooth surface temperature data and bearing area temperature data ; Through temperature analysis function The preliminary cause determination conclusion is then re-evaluated to obtain the secondary cause determination conclusion. like If so, the second determination is that the gears are not meshing properly; like If so, the secondary determination is a mechanical structure problem; If the preliminary cause determination conclusion is consistent with the secondary cause determination conclusion, and both are determined to be poor gear meshing, then the cause of the abnormal vibration data is determined to be poor gear meshing. If the preliminary cause determination conclusion is consistent with the secondary cause determination conclusion, and both are determined to be mechanical structure problems, then the cause of the abnormal vibration data is determined to be a mechanical structure problem. If the preliminary cause determination conclusion is inconsistent with the secondary cause determination conclusion, then a manual determination shall be made.

9. The method for testing the efficiency of a reducer based on a smart sensor according to claim 2, characterized in that: The causes of misjudgments were identified and corrected, specifically as follows: Extract the minimum distance between the current feature and the feature library. The corresponding eigenvector is defined as the smallest eigenvector, denoted as . ; Calculate the current feature vector The feature vector with the smallest distance The pointwise difference is denoted as ; Extracting satisfaction The feature dimension h; Adjust the feature extraction algorithm or parameters based on the significantly different feature dimensions; New feature vector Adding normal vibration feature library to form a new feature library matrix. ; The adjusted feature extraction method is used to re-extract features from the currently acquired vibration data, resulting in new feature vectors. ; Calculate the new feature vector Euclidean distance to each feature vector in the updated normal vibration feature library Find the feature vector with the smallest distance. ; By verification function Verify whether the misjudgment has been corrected; like If so, the misjudgment has been corrected; like If the error is not corrected, it will be determined that a misjudgment has not been corrected and manual correction will be performed.

Images