Intelligent detection method and device for lubrication equipment fault based on vibration analysis

Abstract

The present invention provides a method and device for intelligently detecting lubrication equipment faults based on vibration analysis, which relates to the technical field of fault detection. The method and device perform real-time vibration signal monitoring on the lubrication pump of the lubrication equipment, identify time-frequency features based on the vibration signal monitoring set, detect faults on the lubrication pump, determine a first fault risk coefficient, activate a lubrication pump fault detection auxiliary factor, perform fault detection on the lubrication pump in combination with an Internet of Things sensor array, and correct the first fault risk coefficient based on a second fault risk coefficient. If the lubrication pump fault risk coefficient is greater than / equal to the lubrication pump fault risk threshold, a lubrication pump fault warning signal is generated. The method solves the technical problem in the prior art that the lubrication pump fault of the lubrication equipment relies on experience and manual judgment, resulting in low accuracy and reliability in lubrication pump fault identification. The method achieves the technical effect of improving the intelligent level of lubrication pump fault identification and significantly improving the accuracy and reliability of lubrication pump fault detection.

Description

Technical Field

[0001] The present invention relates to the technical field of fault detection, and in particular to a method and a device for intelligently detecting faults of lubrication equipment based on vibration analysis. Background Art

[0002] As a key component in mechanical equipment to ensure efficient operation and extend service life, the state of the lubrication pump has an important impact on the normal operation and production efficiency of mechanical equipment. In order to ensure the normal operation of the lubrication pump, traditional lubrication pump maintenance methods mostly rely on experience and manual judgment. Although this method can prevent failures to a certain extent, it is often unable to accurately capture the actual operating status of the lubrication pump due to the inflexible setting of the lubrication pump maintenance cycle and the subjectivity of manual labor, resulting in low accuracy and reliability of lubrication pump fault identification, which in turn increases unnecessary downtime and maintenance costs of the lubrication pump.

[0003] In the prior art, there is a technical problem that the failure of a lubrication pump of a lubrication equipment depends on experience and manual judgment, resulting in low accuracy and reliability in the identification of the lubrication pump failure. Summary of the invention

[0004] The present application provides a method and device for intelligently detecting lubrication equipment faults based on vibration analysis, which is used to solve the technical problem in the prior art that the lubrication pump faults of lubrication equipment rely on experience and manual judgment, resulting in low accuracy and reliability in lubrication pump fault identification.

[0005] In view of the above problems, the present application provides a method and device for intelligent detection of lubrication equipment faults based on vibration analysis.

[0006] The first aspect of the present application provides an intelligent detection method for lubrication equipment faults based on vibration analysis, the method comprising: performing real-time vibration signal monitoring of a lubrication pump of the lubrication equipment according to an Internet of Things sensor array to obtain a vibration signal monitoring set; performing time-frequency feature recognition based on the vibration signal monitoring set to obtain a vibration signal time-frequency feature; performing fault detection on the lubrication pump based on the vibration signal time-frequency feature to determine a first fault risk coefficient; activating a lubrication pump fault detection auxiliary factor, performing fault detection on the lubrication pump in combination with the Internet of Things sensor array to determine a second fault risk coefficient; correcting the first fault risk coefficient based on the second fault risk coefficient to generate a lubrication pump fault risk coefficient; determining whether the lubrication pump fault risk coefficient is greater than / equal to a lubrication pump fault risk threshold; if the lubrication pump fault risk coefficient is greater than / equal to the lubrication pump fault risk threshold, generating a lubrication pump fault warning signal.

[0007] The second aspect of the present application provides an intelligent detection device for lubrication equipment faults based on vibration analysis, the device comprising: a signal monitoring module, the signal monitoring module is used to perform real-time vibration signal monitoring of a lubrication pump of the lubrication equipment according to an Internet of Things sensor array to obtain a vibration signal monitoring set; a time-frequency feature recognition module, the time-frequency feature recognition module is used to perform time-frequency feature recognition based on the vibration signal monitoring set to obtain the vibration signal time-frequency feature; a first fault risk coefficient determination module, the first fault risk coefficient determination module is used to perform fault detection on the lubrication pump based on the vibration signal time-frequency feature to determine a first fault risk coefficient; a second fault risk coefficient determination module, the second fault risk coefficient determination module is used to activate the lubrication A pump fault detection auxiliary factor is used to perform fault detection on the lubrication pump in combination with the Internet of Things sensor array to determine a second fault risk coefficient; a lubrication pump fault risk coefficient generation module is used to correct the first fault risk coefficient based on the second fault risk coefficient to generate a lubrication pump fault risk coefficient; a lubrication pump fault risk coefficient judgment module is used to judge whether the lubrication pump fault risk coefficient is greater than / equal to the lubrication pump fault risk threshold; a lubrication pump fault warning signal generation module is used to generate a lubrication pump fault warning signal if the lubrication pump fault risk coefficient is greater than / equal to the lubrication pump fault risk threshold.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] The method provided in the embodiment of the present application monitors the real-time vibration signal of the lubrication pump of the lubrication equipment according to the Internet of Things sensor array to obtain a vibration signal monitoring set; performs time-frequency feature recognition based on the vibration signal monitoring set to obtain the vibration signal time-frequency feature; performs fault detection on the lubrication pump based on the vibration signal time-frequency feature to determine the first fault risk coefficient; activates the lubrication pump fault detection auxiliary factor, performs fault detection on the lubrication pump in combination with the Internet of Things sensor array to determine the second fault risk coefficient; based on the second fault risk coefficient, corrects the first fault risk coefficient to generate a lubrication pump fault risk coefficient; determines whether the lubrication pump fault risk coefficient is greater than / equal to the lubrication pump fault risk threshold; if the lubrication pump fault risk coefficient is greater than / equal to the lubrication pump fault risk threshold, generates a lubrication pump fault warning signal. The technical effect of improving the intelligent level of lubrication pump fault identification, significantly improving the accuracy and reliability of lubrication pump fault detection, and reducing lubrication pump downtime and maintenance costs is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A schematic diagram of the flow of the intelligent detection method for lubrication equipment faults based on vibration analysis provided in this application;

[0011] Figure 2 A schematic diagram of the structure of the intelligent detection device for lubrication equipment faults based on vibration analysis provided in this application.

[0012] Explanation of the accompanying drawings: signal monitoring module 11, time-frequency feature recognition module 12, first fault risk coefficient determination module 13, second fault risk coefficient determination module 14, lubrication pump fault risk coefficient generation module 15, lubrication pump fault risk coefficient judgment module 16, lubrication pump fault warning signal generation module 17. DETAILED DESCRIPTION

[0013] The present application provides a method and device for intelligent detection of lubrication equipment faults based on vibration analysis, which is used to solve the technical problem that the prior art relies on experience and manual judgment to identify the faults of lubrication pumps of lubrication equipment, resulting in low accuracy and reliability in the identification of lubrication pump faults. The method achieves the technical effect of improving the intelligent level of lubrication pump fault identification, significantly improving the accuracy and reliability of lubrication pump fault detection, and reducing the downtime and maintenance cost of lubrication pumps.

[0014] Below, the technical solutions in the present invention will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments described herein. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. It should also be noted that, for the convenience of description, only the parts related to the present invention are shown in the accompanying drawings, rather than all of them.

[0015] Embodiment 1, as Figure 1 As shown, the present application provides a method for intelligent detection of lubrication equipment faults based on vibration analysis, the method comprising:

[0016] The lubrication pump of the lubrication equipment is monitored for real-time vibration signals based on the IoT sensor array to obtain a vibration signal monitoring set.

[0017] Specifically, the lubrication pump is the core component of the lubrication equipment, which is used to transport lubricating oil or grease from the storage container to the parts that need lubrication. Through multiple sensors distributed at different positions of the lubrication pump, that is, the Internet of Things sensor array, the lubrication pump of the lubrication equipment is monitored for real-time vibration signals. Each sensor in the sensor array is equipped with a highly sensitive vibration detection element, which can capture the tiny vibration changes generated by the lubrication pump during operation. In the vibration signal monitoring process, each sensor in the Internet of Things sensor array collects vibration data in real time, converts these data into digital signals, and transmits them to the data processing center to form a vibration signal monitoring set. The vibration signal monitoring set includes the vibration amplitude and frequency of the lubrication pump at different time points, which can reflect the operating status of the lubrication equipment. Real-time vibration signal monitoring of the lubrication pump of the lubrication equipment through the Internet of Things sensor array can ensure the real-time and accuracy of the monitoring data, and provide reliable data support for fault detection.

[0018] Time-frequency features are identified based on the vibration signal monitoring set to obtain the time-frequency features of the vibration signal.

[0019] Furthermore, time-frequency feature identification is performed based on the vibration signal monitoring set to obtain the time-frequency features of the vibration signal, including: performing EMD filtering according to the vibration signal monitoring set to obtain a vibration signal set; based on the vibration signal set, obtaining a spectrum analysis result of the vibration signal according to a spectrum analyzer; performing time domain feature analysis according to the vibration signal set to generate a time domain feature analysis result of the vibration signal; and collating the spectrum analysis result of the vibration signal and the time domain feature analysis result of the vibration signal to generate the time-frequency features of the vibration signal.

[0020] Specifically, first, the vibration signal monitoring set is processed using EMD filtering (empirical mode decomposition) to identify useful vibration signal components from complex vibration signals. EMD filtering is an adaptive signal processing technology that can decompose complex signals into several intrinsic mode functions (IMFs), which represent different frequency components in the signal. Through comprehensive analysis of IMFs, noise and interference in the signal can be removed to obtain a purer vibration signal set. Then, based on the vibration signal set processed by EMD filtering, a spectrum analyzer is used to analyze the vibration signal set signal in the frequency domain to determine the distribution of each frequency component in the signal. Vibration signal spectrum analysis using a spectrum analyzer can identify the main vibration frequencies and corresponding amplitudes in the vibration signal, and obtain vibration signal spectrum analysis results. The vibration signal spectrum analysis results can intuitively display the vibration characteristics of the lubrication pump, such as whether there is abnormal vibration of a specific frequency, indicating the possible fault type of the equipment.

[0021] After completing the frequency domain analysis, the vibration signal set is subjected to time domain feature extraction to generate a vibration signal time domain feature analysis result. Time domain feature analysis refers to processing the signal in the time domain to extract the timing characteristics of the signal, such as the peak value, mean value, variance, etc. of the signal. The vibration signal time domain feature analysis result can reflect the temporal variation pattern of the vibration signal, such as the fluctuation of the vibration amplitude, the occurrence of sudden vibration, etc. The time domain analysis result and the frequency domain analysis result complement each other and together provide a comprehensive feature description of the vibration signal.

[0022] Finally, the vibration signal spectrum analysis results and the vibration signal time domain feature analysis results are sorted out to generate comprehensive vibration signal time-frequency features. The vibration signal time-frequency features combine the performance of the vibration signal in the time domain and frequency domain, and provide a comprehensive description of the vibration state of the equipment, which helps to accurately identify the normal or abnormal state of the lubrication pump and provide a reliable data basis for subsequent fault prediction and risk assessment.

[0023] Fault detection is performed on the lubrication pump based on the time-frequency characteristics of the vibration signal to determine a first fault risk coefficient.

[0024] Furthermore, fault detection is performed on the lubrication pump based on the time-frequency characteristics of the vibration signal to determine a first fault risk coefficient, including: performing a local search based on the lubrication pump to obtain a vibration signal time-frequency feature sample set and a vibration feature fault risk sample set; performing a global search based on the lubrication pump to obtain a global vibration signal time-frequency feature sample set and a global vibration feature fault risk sample set; constructing a vibration feature fault risk prediction model using the global vibration signal time-frequency feature sample set as input data and the global vibration feature fault risk sample set as output data; performing sensitive optimization learning on the vibration feature fault risk prediction model based on the vibration signal time-frequency feature sample set and the vibration feature fault risk sample set to obtain a vibration feature fault risk prediction channel; and obtaining the first fault risk coefficient based on the vibration signal time-frequency features and the vibration feature fault risk prediction channel.

[0025] Specifically, first, a local search is performed on the lubrication pump. The local search refers to analyzing only the vibration signal of the current lubrication pump to be tested. The time-frequency feature sample set of the lubrication pump vibration signal is obtained through analysis, and the corresponding vibration feature fault risk sample set is formed by analyzing historical fault data or existing fault labels. The vibration signal time-frequency feature sample set and the vibration feature fault risk sample set reflect the vibration performance of the lubrication pump under different operating conditions and its potential fault risk.

[0026] Then, a global search is performed on the lubrication pump, and the object is a plurality of lubrication pumps of the same model as the lubrication pump. By expanding the scope of the search to a plurality of lubrication pumps of the same model through global search, vibration signal data in a larger range can be collected to form a global vibration signal time-frequency feature sample set. By collecting and comparing data between a plurality of lubrication pumps of the same model, the comprehensiveness and representativeness of the global vibration signal time-frequency feature sample set are ensured. In addition, by analyzing the fault records of these lubrication pumps under different operating conditions, a global vibration feature fault risk sample set is generated. The global vibration signal time-frequency feature sample set is used as input data, and the global vibration feature fault risk sample set is used as output data to construct a vibration feature fault risk prediction model. The vibration feature fault risk prediction model utilizes machine learning or deep learning algorithms, and can identify and learn the relationship between vibration features and fault risks.

[0027] After obtaining the vibration feature fault risk prediction model, the model is subjected to sensitive optimization learning. Sensitive optimization learning means that the sensitivity of the model to specific vibration features is optimized by further training the model, thereby improving the fault detection capability of the model in practical applications. Based on the vibration signal time-frequency feature sample set and the vibration feature fault risk sample set obtained by local retrieval, the vibration feature fault risk prediction model is subjected to sensitive optimization learning. By adjusting the model parameters, optimizing the algorithm structure, etc., the vibration feature fault risk prediction channel is obtained, so that the vibration feature fault risk prediction model is more suitable for the actual operation of the lubrication pump, and the accuracy and reliability of fault detection are improved.

[0028] Finally, the first fault risk coefficient of the lubrication pump is calculated by using the optimized vibration characteristic fault risk prediction channel in combination with the time-frequency characteristics of the vibration signal through the vibration fault risk prediction channel. The first fault risk coefficient is a quantitative expression of the fault risk, which intuitively reflects the potential degree of fault risk under the current operating state of the lubrication pump. The higher the value, the greater the possibility of fault. By combining global and local data and continuously optimizing the model, the model can maintain accurate and reliable fault risk prediction capabilities under different operating conditions.

[0029] The lubrication pump fault detection auxiliary factor is activated, and the lubrication pump is subjected to fault detection in combination with the Internet of Things sensor array to determine a second fault risk coefficient.

[0030] Furthermore, a lubrication pump fault detection auxiliary factor is activated, and the lubrication pump is detected for fault in combination with the Internet of Things sensor array to determine a second fault risk coefficient, including: the lubrication pump fault detection auxiliary factor includes the lubrication pump operation noise, the lubrication pump temperature, the lubrication pump pressure signal and the lubrication pump speed, wherein the lubrication pump pressure signal includes the pump outlet pressure and the pump inlet pressure; based on the lubrication pump fault detection auxiliary factor, the lubrication pump is monitored in real time according to the Internet of Things sensor array to obtain a lubrication pump operation noise monitoring set, a lubrication pump temperature monitoring set, a lubrication pump pressure signal monitoring set and a lubrication pump speed monitoring set; based on the lubrication pump Perform fault detection on the lubrication pump based on the operating noise monitoring set to obtain a noise characteristic fault risk coefficient; perform fault detection on the lubrication pump based on the lubrication pump temperature monitoring set to obtain a temperature characteristic fault risk coefficient; perform fault detection on the lubrication pump based on the lubrication pump pressure signal monitoring set to obtain a pressure characteristic fault risk coefficient; perform fault detection on the lubrication pump based on the lubrication pump speed monitoring set to obtain a speed characteristic fault risk coefficient; add the noise characteristic fault risk coefficient, the temperature characteristic fault risk coefficient, the pressure characteristic fault risk coefficient and the speed characteristic fault risk coefficient to the second fault risk coefficient.

[0031] Specifically, the lubrication pump fault detection auxiliary factor is activated, and the lubrication pump fault detection auxiliary factor includes multiple key parameters during the operation of the lubrication pump, and the multiple key parameters are the lubrication pump operation noise, the lubrication pump temperature, the lubrication pump pressure signal, and the lubrication pump speed. The pressure signal of the lubrication pump is further subdivided into pump outlet pressure and pump inlet pressure, and the two signals of the pump outlet pressure and the pump inlet pressure respectively reflect the working pressure state inside the lubrication pump. Each key parameter in the lubrication pump fault detection auxiliary factor provides equipment status information of different dimensions, which is helpful to comprehensively judge the health status of the lubrication pump.

[0032] In the process of monitoring the vibration of the lubrication pump, the lubrication pump fault detection auxiliary factors are monitored and collected in real time according to the IoT sensor array, and a monitoring data set including a lubrication pump operation noise monitoring set, a lubrication pump temperature monitoring set, a lubrication pump pressure signal monitoring set and a lubrication pump speed monitoring set is generated. Then, for the lubrication pump operation noise, fault detection is performed based on the lubrication pump operation noise monitoring set. By analyzing the characteristics of the noise signal, such as the intensity, frequency distribution and sudden changes of the noise, the abnormal noise characteristics can be identified, thereby calculating the noise characteristic fault risk coefficient. The noise characteristic fault risk coefficient reflects the correlation between the noise anomaly and the equipment failure. The higher the value, the more serious the noise anomaly and the greater the possibility of equipment failure. Similarly, the lubrication pump temperature fault is analyzed based on the lubrication pump temperature monitoring set, and the temperature characteristic fault risk coefficient is generated by comparing the real-time monitoring data of the temperature change with the historical data, which is used to evaluate the impact of the temperature anomaly on the equipment operation. The lubrication pump pressure signal monitoring set is used to analyze the pressure state at the pump outlet and inlet. By comparing the difference between the normal operating pressure of the pump and the real-time monitoring pressure, it is identified whether the lubrication pump has pressure anomaly, and the pressure characteristic fault risk coefficient is calculated. The pressure characteristic failure risk factor can help identify problems such as pipe blockage, leakage or reduced pump efficiency. Speed ​​is an important indicator of equipment operating efficiency and load conditions. Abnormal speed fluctuations may indicate mechanical failure or electrical problems of the equipment. By analyzing the speed data of the lubrication pump through the speed monitoring of the lubrication pump, the speed characteristic failure risk factor can be calculated to assess the risk level of abnormal speed.

[0033] Finally, the calculated noise characteristic fault risk coefficient, the temperature characteristic fault risk coefficient, the pressure characteristic fault risk coefficient and the speed characteristic fault risk coefficient are integrated and added to the second fault risk coefficient to form a comprehensive fault risk coefficient. The second fault risk coefficient reflects the overall fault risk level of the lubrication pump and provides a basis for real-time monitoring and maintenance decisions of the equipment. By activating the fault detection auxiliary factor and combining the real-time monitoring and analysis of various operating parameters of the lubrication pump with the IoT sensor array, the operating status of the lubrication pump can be comprehensively evaluated, which can further accurately and comprehensively identify potential fault risks.

[0034] Based on the second failure risk coefficient, the first failure risk coefficient is corrected to generate a lubrication pump failure risk coefficient.

[0035] Furthermore, based on the second fault risk coefficient, the first fault risk coefficient is corrected to generate a lubrication pump failure risk coefficient, including: identifying fault risk dimensions based on the first fault risk coefficient and the second fault risk coefficient to obtain a fault risk dimension set; initializing weight distribution based on the fault risk dimension set to obtain a first result of fault risk weight distribution; performing proportion calculation based on the first fault risk coefficient and the second fault risk coefficient to obtain a second result of fault risk weight distribution; performing concentrated value calculation of weights of each fault risk dimension based on the first result of fault risk weight distribution and the second result of fault risk weight distribution to obtain a third result of fault risk weight distribution; performing weighted calculation on the first fault risk coefficient and the second fault risk coefficient based on the third result of fault risk weight distribution to output the lubrication pump failure risk coefficient.

[0036] Specifically, the first fault risk coefficient and the second fault risk coefficient determined are analyzed in terms of fault risk dimensions, different characteristics or factors included in the first fault risk coefficient and the second fault risk coefficient are obtained, and fault risk dimensions related to the fault are determined, such as mechanical wear, excessive oil temperature, oil pollution, etc. By identifying these dimensions, a comprehensive set of fault risk dimensions is obtained. By identifying the fault risk dimensions, it is ensured that all factors that may affect the lubrication pump failure can be comprehensively and accurately considered in the subsequent correction process.

[0037] Based on the obtained fault risk dimension set, a weight distribution is initialized for each dimension. The initialized weight distribution refers to assigning an initial weight to each fault risk dimension in the absence of further information. The weight represents the proportion or importance of each fault risk dimension in the overall fault risk. The initial weight allocation can be set based on expert experience, historical data or industry standards to obtain a first result of the fault risk weight distribution.

[0038] After obtaining the first result of the fault risk weight distribution of the initial weight distribution, a proportion calculation is performed based on the relative size of the first fault risk coefficient and the second fault risk coefficient. The proportion calculation refers to analyzing the proportion of each fault risk dimension in the first fault risk coefficient and the second fault risk coefficient, and quantifying the size of each dimension in the current fault risk. Through the proportion calculation, the second result of the fault risk weight distribution is obtained, and the second result of the fault risk weight distribution is the weight distribution result adjusted according to the actual risk level.

[0039] Then, the first result of the fault risk weight distribution and the second result of the fault risk weight distribution are combined, and the concentrated value of the weight of each fault risk dimension is calculated through comprehensive analysis. The concentrated value calculation includes weighted average, median calculation or other mathematical methods, and finally determines the relative importance of each dimension to obtain a third result of the fault risk weight distribution that takes into account both expert experience and actual conditions. Finally, according to the third result of the fault risk weight distribution, the first fault risk coefficient and the second fault risk coefficient are weighted. Weighted calculation refers to linearly combining the first and second fault risk coefficients according to the weight of each risk dimension, so as to obtain a more accurate and comprehensive fault risk assessment value, namely the lubrication pump fault risk coefficient, which represents the current overall fault risk level of the equipment. Through the method of step-by-step weighting and correction, the relative importance of different fault risk dimensions can be fully considered to ensure that the final fault risk assessment result is both accurate and comprehensive, providing reliable decision support for the maintenance and management of the lubrication pump.

[0040] Determine whether the lubrication pump failure risk coefficient is greater than / equal to a lubrication pump failure risk threshold.

[0041] If the lubrication pump failure risk coefficient is greater than / equal to the lubrication pump failure risk threshold, a lubrication pump failure warning signal is generated.

[0042] Specifically, after the lubrication pump failure risk coefficient is obtained through calibration, the lubrication pump failure risk coefficient is compared with a preset lubrication pump failure risk threshold to determine whether the current failure risk exceeds the safety range, that is, whether the lubrication pump failure risk coefficient is greater than / equal to the lubrication pump failure risk threshold. The lubrication pump failure risk threshold is set based on multiple factors such as the historical operation data of the equipment, expert experience, and equipment standards, and represents the highest risk level that the lubrication equipment can accept.

[0043] If the lubrication pump failure risk coefficient is less than the lubrication pump failure risk threshold, it means that the equipment is within the safe range and continues to monitor without issuing an alarm. If the lubrication pump failure risk coefficient is greater than or equal to the lubrication pump failure risk threshold, it means that the lubrication pump has a high risk of failure and a lubrication pump failure warning signal is immediately generated. The lubrication pump failure warning signal is a reminder to equipment management personnel, indicating that the current equipment has a high risk of failure and requires timely inspection and maintenance. The warning signal can be issued in a variety of forms, such as alarm sounds, warning lights, SMS notifications, etc., to ensure that relevant personnel can know the equipment status at the first time and take corresponding measures. Through real-time risk assessment and early warning mechanisms, sudden failures of lubrication pumps can be effectively prevented, reducing equipment downtime and maintenance costs.

[0044] Furthermore, EMD filtering is performed according to the vibration signal monitoring set to obtain a vibration signal set, including: adding white noise according to the vibration signal monitoring set to obtain a characteristic vibration signal monitoring set; performing EMD decomposition according to the characteristic vibration signal monitoring set to obtain a vibration signal intrinsic mode function feature sequence set; integrating and averaging the vibration signal intrinsic mode function feature sequence set to obtain a CEEMD decomposition result set; and performing denoising and reconstruction according to the CEEMD decomposition result set to generate the vibration signal set.

[0045] Specifically, the vibration signal monitoring set is preliminarily processed and white noise is added. White noise is a random signal whose spectral density is uniform over the entire frequency range and the energy of all frequency components is equal. By adding an appropriate amount of white noise to the vibration signal in the vibration signal monitoring set, the characteristic components in the signal can be enhanced, so that the intrinsic mode function (IMF) can be more accurately extracted in the subsequent signal decomposition. By adding white noise, a characteristic vibration signal monitoring set is obtained. The characteristic vibration signal monitoring set obtained by adding white noise can prevent the signal from experiencing mode aliasing during the EMD decomposition process, thereby improving the stability and accuracy of the decomposition results.

[0046] The characteristic vibration signal monitoring set is subjected to EMD decomposition. Through EMD decomposition, different features of the vibration signal in the characteristic vibration signal monitoring set can be separated, so that subsequent analysis can focus more on key fault features. Through EMD decomposition, a vibration signal intrinsic mode function feature sequence set is obtained, and the vibration signal intrinsic mode function feature sequence set contains the main information components in the original signal. Then, the vibration signal intrinsic mode function feature sequence set is integrated and averaged, multiple vibration signal intrinsic mode function feature sequences are integrated together, and they are averaged to eliminate random fluctuations introduced by white noise, and obtain the CEEMD (collective empirical mode decomposition) decomposition result set. CEEMD decomposition is an improvement of the EMD method. By performing EMD decomposition on multiple signals with different white noises added, and then integrating and averaging the results, the influence of mode aliasing is effectively reduced. The result of the integration and averaging processing is a CEEMD decomposition result set, which contains the average signal after multiple decompositions and represents the core characteristics of the vibration signal. Finally, the CEEMD decomposition result set is denoised and reconstructed. Denoising reconstruction means reconstructing a purer vibration signal set by removing the noise components in the decomposition results. The multiple IMFs obtained by CEEMD decomposition are used to remove high-frequency noise and low-energy components, retaining only the key signals related to equipment failure, and generating a vibration signal set. The vibration signal set has a high signal-to-noise ratio and can more accurately reflect the operating status of the lubrication pump, ensuring the reliability and accuracy of the vibration signal and improving the accuracy of fault detection and risk assessment.

[0047] Furthermore, fault detection is performed on the lubrication pump based on the lubrication pump operation noise monitoring set to obtain a noise characteristic fault risk coefficient, including: obtaining an operating task scenario of the lubrication pump; performing noise health prediction on the lubrication pump based on the operating task scenario to obtain a predicted healthy noise matrix; organizing the lubrication pump operation noise monitoring set to establish a noise monitoring matrix; performing deviation detection on the noise monitoring matrix based on the predicted healthy noise matrix to obtain a noise deviation detection matrix; based on the noise deviation detection matrix, according to P noise characteristic fault risk prediction models, P noise characteristic fault risk prediction coefficients are obtained, wherein P is a positive integer greater than 1; performing centralized value calculation based on the P noise characteristic fault risk prediction coefficients to output the noise characteristic fault risk coefficient.

[0048] Specifically, the operation task scenario of the lubrication pump is obtained. The operation task scenario refers to the working environment and task requirements of the lubrication pump under specific operating conditions, including the working state of the lubrication pump under different load, temperature and pressure conditions. Then, the noise health prediction of the lubrication pump is performed based on the operation task scenario, that is, the normal noise of the lubrication pump under the operation task scenario is predicted by analyzing the historical data and the characteristics of the lubrication pump, and a predicted healthy noise matrix is ​​generated. The predicted healthy noise matrix contains the expected noise characteristics of the lubrication pump under the operation task scenario, and the expected noise characteristics are the baseline noise of the lubrication pump in a healthy state. Then, the real-time collected lubrication pump operation noise monitoring set is sorted to obtain a noise monitoring matrix, which is the actual operation noise data of the lubrication pump under the current operation task scenario. The noise deviation detection matrix is ​​generated by calculating the difference between the noise monitoring matrix and the predicted healthy noise matrix at each frequency and time point. The noise deviation detection matrix reflects the degree of noise deviation under the current operating state. The higher the value, the greater the deviation.

[0049] Furthermore, according to the noise deviation detection matrix, multiple (P, P is a positive integer greater than 1) noise feature fault risk prediction models are used to evaluate the fault risk. Each of the P models analyzes different noise features, such as frequency features, amplitude features, periodic features, etc. Through the P models, a corresponding noise feature fault risk prediction coefficient can be calculated for each noise feature to obtain P noise feature fault risk prediction coefficients. The P noise feature fault risk prediction coefficients reflect the risk level of different noise features in the current state. Finally, the centralized value calculation is performed based on the P noise feature fault risk prediction coefficients to ensure that the risk assessment results of each noise feature can be effectively integrated to avoid misjudgment caused by a single feature abnormality. Through the centralized value calculation, the P noise feature risk coefficients are comprehensively calculated into a noise feature fault risk coefficient. The noise feature fault risk coefficient is a comprehensive comparison between the current noise state and the fault state of the lubrication pump, which accurately and comprehensively reflects the overall noise risk level of the lubrication pump, and provides an important basis for fault warning and maintenance of the lubrication pump, ensuring that the lubrication equipment can be detected and processed in time before problems occur, reducing equipment downtime and maintenance costs.

[0050] Embodiment 2 is based on the same inventive concept as the intelligent detection method for lubrication equipment fault based on vibration analysis in the above embodiment. Figure 2 As shown, the present application provides an intelligent detection device for lubrication equipment faults based on vibration analysis, wherein the device comprises:

[0051] A signal monitoring module 11, the signal monitoring module 11 is used to perform real-time vibration signal monitoring of the lubrication pump of the lubrication equipment according to the Internet of Things sensor array to obtain a vibration signal monitoring set; a time-frequency feature recognition module 12, the time-frequency feature recognition module 12 is used to perform time-frequency feature recognition based on the vibration signal monitoring set to obtain the vibration signal time-frequency feature; a first fault risk coefficient determination module 13, the first fault risk coefficient determination module 13 is used to perform fault detection on the lubrication pump based on the vibration signal time-frequency feature to determine a first fault risk coefficient; a second fault risk coefficient determination module 14, the second fault risk coefficient determination module 14 is used to activate the lubrication pump fault detection auxiliary factor, combined with the Internet of Things sensor The sensing array performs fault detection on the lubrication pump to determine a second fault risk coefficient; a lubrication pump fault risk coefficient generation module 15, the lubrication pump fault risk coefficient generation module 15 is used to correct the first fault risk coefficient based on the second fault risk coefficient to generate a lubrication pump fault risk coefficient; a lubrication pump fault risk coefficient judgment module 16, the lubrication pump fault risk coefficient judgment module 16 is used to judge whether the lubrication pump fault risk coefficient is greater than / equal to the lubrication pump fault risk threshold; a lubrication pump fault warning signal generation module 17, the lubrication pump fault warning signal generation module 17 is used to generate a lubrication pump fault warning signal if the lubrication pump fault risk coefficient is greater than / equal to the lubrication pump fault risk threshold.

[0052] Furthermore, the time-frequency feature identification module 12 is used to perform the following steps: perform EMD filtering according to the vibration signal monitoring set to obtain a vibration signal set; based on the vibration signal set, obtain a vibration signal spectrum analysis result according to a spectrum analyzer; perform time domain feature analysis according to the vibration signal set to generate a vibration signal time domain feature analysis result; organize the vibration signal spectrum analysis result and the vibration signal time domain feature analysis result to generate the vibration signal time-frequency feature.

[0053] Furthermore, the time-frequency feature recognition module 12 is also used to perform the following steps: add white noise according to the vibration signal monitoring set to obtain a characteristic vibration signal monitoring set; perform EMD decomposition according to the characteristic vibration signal monitoring set to obtain a vibration signal intrinsic mode function feature sequence set; perform integration and averaging according to the vibration signal intrinsic mode function feature sequence set to obtain a CEEMD decomposition result set; perform denoising and reconstruction according to the CEEMD decomposition result set to generate the vibration signal set.

[0054] Furthermore, the first fault risk coefficient determination module 13 is used to perform the following steps: perform local retrieval based on the lubrication pump to obtain a vibration signal time-frequency feature sample set and a vibration feature fault risk sample set; perform global retrieval based on the lubrication pump to obtain a global vibration signal time-frequency feature sample set and a global vibration feature fault risk sample set; use the global vibration signal time-frequency feature sample set as input data and the global vibration feature fault risk sample set as output data to construct a vibration feature fault risk prediction model; perform sensitive optimization learning on the vibration feature fault risk prediction model based on the vibration signal time-frequency feature sample set and the vibration feature fault risk sample set to obtain a vibration feature fault risk prediction channel; based on the vibration signal time-frequency features, obtain the first fault risk coefficient according to the vibration feature fault risk prediction channel.

[0055] Further, the second fault risk coefficient determination module 14 is used to perform the following steps: the lubrication pump fault detection auxiliary factor includes the lubrication pump operation noise, the lubrication pump temperature, the lubrication pump pressure signal and the lubrication pump speed, wherein the lubrication pump pressure signal includes the pump outlet pressure and the pump inlet pressure; based on the lubrication pump fault detection auxiliary factor, the lubrication pump is monitored in real time according to the Internet of Things sensor array to obtain the lubrication pump operation noise monitoring set, the lubrication pump temperature monitoring set, the lubrication pump pressure signal monitoring set and the lubrication pump speed monitoring set; based on the lubrication pump operation noise monitoring set, the lubrication pump is fault detected on the lubrication pump to obtain the noise characteristic fault risk coefficient; based on the lubrication pump temperature monitoring set, the lubrication pump is fault detected to obtain the temperature characteristic fault risk coefficient; based on the lubrication pump pressure signal monitoring set, the lubrication pump is fault detected to obtain the pressure characteristic fault risk coefficient; based on the lubrication pump speed monitoring set, the lubrication pump is fault detected to obtain the speed characteristic fault risk coefficient; the noise characteristic fault risk coefficient, the temperature characteristic fault risk coefficient, the pressure characteristic fault risk coefficient and the speed characteristic fault risk coefficient are added to the second fault risk coefficient.

[0056] Furthermore, the second fault risk coefficient determination module 14 is also used to perform the following steps: obtain an operating task scenario of the lubrication pump; perform noise health prediction on the lubrication pump based on the operating task scenario to obtain a predicted healthy noise matrix; organize the lubrication pump operation noise monitoring set to establish a noise monitoring matrix; perform deviation detection on the noise monitoring matrix based on the predicted healthy noise matrix to obtain a noise deviation detection matrix; based on the noise deviation detection matrix, according to P noise feature fault risk prediction models, obtain P noise feature fault risk prediction coefficients, where P is a positive integer greater than 1; perform centralized value calculation based on the P noise feature fault risk prediction coefficients, and output the noise feature fault risk coefficient.

[0057] Furthermore, the lubrication pump failure risk coefficient generation module 15 is used to perform the following steps: identify the failure risk dimension based on the first failure risk coefficient and the second failure risk coefficient to obtain a failure risk dimension set; initialize the weight distribution based on the failure risk dimension set to obtain a first result of the failure risk weight distribution; perform a proportion calculation based on the first failure risk coefficient and the second failure risk coefficient to obtain a second result of the failure risk weight distribution; calculate the concentrated value of the weights of each failure risk dimension based on the first result of the failure risk weight distribution and the second result of the failure risk weight distribution to obtain a third result of the failure risk weight distribution; perform a weighted calculation on the first failure risk coefficient and the second failure risk coefficient based on the third result of the failure risk weight distribution, and output the lubrication pump failure risk coefficient.

[0058] Any of the methods or steps described above may be stored as computer instructions or programs in various types of computer memories, and the computer instructions or programs may be recognized by various types of computer processors to implement any of the methods or steps described above.

[0059] Based on the above specific embodiments of the present invention, any improvements and modifications made to the present invention by those skilled in the art without departing from the principles of the present invention shall fall within the patent protection scope of the present invention.

Claims

1. An intelligent detection method for lubrication equipment faults based on vibration analysis, characterized in that: The method comprises: The lubrication pump of the lubrication equipment is monitored for real-time vibration signals based on the IoT sensor array to obtain a vibration signal monitoring set; Performing time-frequency feature recognition based on the vibration signal monitoring set to obtain the time-frequency features of the vibration signal; Performing fault detection on the lubrication pump based on the time-frequency characteristics of the vibration signal to determine a first fault risk coefficient; Activate the lubrication pump fault detection auxiliary factor, perform fault detection on the lubrication pump in combination with the Internet of Things sensor array, and determine a second fault risk coefficient; Based on the second failure risk coefficient, correcting the first failure risk coefficient to generate a lubrication pump failure risk coefficient; Determining whether the lubrication pump failure risk coefficient is greater than / equal to a lubrication pump failure risk threshold; If the lubrication pump failure risk coefficient is greater than / equal to the lubrication pump failure risk threshold, a lubrication pump failure warning signal is generated; Performing fault detection on the lubrication pump based on the time-frequency characteristics of the vibration signal to determine a first fault risk coefficient includes: Performing local retrieval based on the lubrication pump to obtain a vibration signal time-frequency feature sample set and a vibration feature fault risk sample set; Performing a global search based on the lubrication pump to obtain a global vibration signal time-frequency feature sample set and a global vibration feature fault risk sample set; Taking the global vibration signal time-frequency feature sample set as input data and the global vibration feature fault risk sample set as output data, a vibration feature fault risk prediction model is constructed; Based on the vibration signal time-frequency feature sample set and the vibration feature fault risk sample set, the vibration feature fault risk prediction model is subjected to sensitive optimization learning to obtain a vibration feature fault risk prediction channel; Based on the time-frequency characteristics of the vibration signal and according to the vibration characteristic fault risk prediction channel, obtaining the first fault risk coefficient; Performing fault detection on the lubrication pump in combination with the IoT sensor array to determine a second fault risk coefficient includes: The lubrication pump fault detection auxiliary factors include the lubrication pump operation noise, the lubrication pump temperature, the lubrication pump pressure signal and the lubrication pump speed, wherein the lubrication pump pressure signal includes the pump outlet pressure and the pump inlet pressure; Based on the lubrication pump fault detection auxiliary factor, the lubrication pump is monitored in real time according to the Internet of Things sensor array to obtain a lubrication pump operation noise monitoring set, a lubrication pump temperature monitoring set, a lubrication pump pressure signal monitoring set and a lubrication pump speed monitoring set; Performing fault detection on the lubrication pump based on the lubrication pump operation noise monitoring set to obtain a noise characteristic fault risk coefficient; Performing fault detection on the lubrication pump based on the lubrication pump temperature monitoring set to obtain a temperature characteristic fault risk coefficient; Performing fault detection on the lubrication pump based on the lubrication pump pressure signal monitoring set to obtain a pressure characteristic fault risk coefficient; Performing fault detection on the lubrication pump based on the lubrication pump speed monitoring set to obtain a speed characteristic fault risk coefficient; adding the noise characteristic failure risk coefficient, the temperature characteristic failure risk coefficient, the pressure characteristic failure risk coefficient, and the rotation speed characteristic failure risk coefficient to the second failure risk coefficient; Based on the second fault risk coefficient, correcting the first fault risk coefficient to generate a lubrication pump fault risk coefficient includes: Performing fault risk dimension identification based on the first fault risk coefficient and the second fault risk coefficient to obtain a fault risk dimension set; Initialize weight distribution based on the fault risk dimension set to obtain a first result of fault risk weight distribution; Performing a proportion calculation based on the first failure risk coefficient and the second failure risk coefficient to obtain a second failure risk weight distribution result; Calculate the concentrated value of each fault risk dimension weight based on the first fault risk weight distribution result and the second fault risk weight distribution result to obtain a third fault risk weight distribution result; Based on the third result of the failure risk weight distribution, a weighted calculation is performed on the first failure risk coefficient and the second failure risk coefficient, and the lubrication pump failure risk coefficient is output.

2. The method according to claim 1, characterized in that Performing time-frequency feature recognition based on the vibration signal monitoring set to obtain the time-frequency features of the vibration signal includes: Perform EMD filtering according to the vibration signal monitoring set to obtain a vibration signal set; Based on the vibration signal set, obtaining a vibration signal spectrum analysis result using a spectrum analyzer; Performing time domain feature analysis on the vibration signal set to generate a vibration signal time domain feature analysis result; Arrange the vibration signal spectrum analysis result and the vibration signal time domain feature analysis result to generate the vibration signal time-frequency feature.

3. The method according to claim 2, characterized in that Performing EMD filtering according to the vibration signal monitoring set to obtain a vibration signal set includes: Adding white noise according to the vibration signal monitoring set to obtain a characteristic vibration signal monitoring set; Perform EMD decomposition according to the characteristic vibration signal monitoring set to obtain a vibration signal intrinsic mode function characteristic sequence set; Integrate and average the characteristic sequence set of the intrinsic mode function of the vibration signal to obtain a CEEMD decomposition result set; De-noising and reconstruction are performed according to the CEEMD decomposition result set to generate the vibration signal set.

4. The method according to claim 1, characterized in that Performing fault detection on the lubrication pump based on the lubrication pump operation noise monitoring set to obtain a noise characteristic fault risk coefficient includes: Obtain the operating task scenario of the lubrication pump; Performing noise health prediction on the lubrication pump based on the operation task scenario to obtain a predicted healthy noise matrix; Arrange the lubrication pump operation noise monitoring set and establish a noise monitoring matrix; Perform deviation detection on the noise monitoring matrix based on the predicted healthy noise matrix to obtain a noise deviation detection matrix; Based on the noise deviation detection matrix, and according to P noise feature fault risk prediction models, P noise feature fault risk prediction coefficients are obtained, where P is a positive integer greater than 1; A centralized value calculation is performed based on the P noise characteristic fault risk prediction coefficients, and the noise characteristic fault risk coefficient is output.

5. Intelligent detection device for lubrication equipment fault based on vibration analysis, characterized in that: The steps for implementing the method according to any one of claims 1 to 4 include: A signal monitoring module, wherein the signal monitoring module is used to perform real-time vibration signal monitoring on a lubrication pump of a lubrication equipment according to an IoT sensor array to obtain a vibration signal monitoring set; A time-frequency feature recognition module, the time-frequency feature recognition module is used to perform time-frequency feature recognition based on the vibration signal monitoring set to obtain the time-frequency features of the vibration signal; a first fault risk coefficient determination module, the first fault risk coefficient determination module being used to perform fault detection on the lubrication pump based on the time-frequency characteristics of the vibration signal and determine a first fault risk coefficient; A second fault risk coefficient determination module, the second fault risk coefficient determination module is used to activate the lubrication pump fault detection auxiliary factor, perform fault detection on the lubrication pump in combination with the Internet of Things sensor array, and determine a second fault risk coefficient; a lubrication pump failure risk coefficient generating module, the lubrication pump failure risk coefficient generating module being used to correct the first failure risk coefficient based on the second failure risk coefficient to generate a lubrication pump failure risk coefficient; A lubrication pump failure risk coefficient judgment module, the lubrication pump failure risk coefficient judgment module is used to judge whether the lubrication pump failure risk coefficient is greater than / equal to a lubrication pump failure risk threshold; A lubrication pump failure warning signal generation module is used to generate a lubrication pump failure warning signal if the lubrication pump failure risk coefficient is greater than / equal to the lubrication pump failure risk threshold.

Images