Industrial equipment self-diagnosis system based on algorithm model prediction

Abstract

The invention relates to the technical field of industrial automation, in particular to an industrial equipment self-diagnosis system based on algorithm model prediction, and the system comprises a model building module which comprises a historical fault model; the data acquisition module divides a plurality of to-be-detected point locations and acquires and converts vibration signals; the period judgment module performs periodic judgment on the real-time vibration graph, determines that the vibration period is abnormal according to the judgment result and gives an alarm, or analyzes the abnormal condition of the point location; the data analysis module adjusts a preset frequency based on the point location abnormal conditions, or determines an overall abnormal condition based on a plurality of point location abnormal conditions so as to determine a real-time fault degree, and performs fault early warning on a comparison result of the real-time fault degree and a preset fault degree, or adjusts a preset interval; the model comparison module inputs a plurality of pieces of real-time vibration data into a historical fault model and calculates a fault degree difference value; the model calibration module calibrates the historical fault model based on the fault degree difference. According to the invention, the accuracy of the model fault prediction result is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial automation, and in particular to an industrial equipment self-diagnosis system based on algorithm model prediction. Background Art

[0002] In the industrial production process, the stable operation of equipment is crucial. Traditional equipment fault diagnosis relies on manual inspection and experience judgment, which has problems such as slow response speed and low diagnostic accuracy. With the development of the Internet of Things, big data and machine learning technology, automated and intelligent equipment fault diagnosis systems have become a research hotspot. However, most existing systems are based on only a single data source or simple models, making it difficult to achieve accurate prediction and real-time diagnosis of equipment failures.

[0003] The patent document with Chinese patent application publication number CN117540315A discloses an online diagnosis method for the health of process industrial equipment system, including obtaining historical data of industrial equipment, wherein the historical data includes equipment operation data; establishing a data recognition and prediction model based on the historical data, performing horizontal correlation analysis and vertical deep learning on the historical data based on the data recognition and prediction, and obtaining a predicted value for fault diagnosis; establishing an evaluation model, and performing evaluation and scoring for fault diagnosis based on the predicted value and the acquired real-time operation parameters; determining the operating status of the equipment based on the evaluation score, and displaying the abnormal operating status.

[0004] The models built based on historical data in the prior art rely on specific data sets, which results in insufficient generalization ability of the models when faced with new operating conditions or different equipment, leading to inaccurate fault prediction results. Summary of the invention

[0005] To this end, the present invention provides an industrial equipment self-diagnosis system based on algorithm model prediction. By selecting vibration parameters as the basis for fault judgment, vibration signals are comprehensively collected based on accurate division of points to be detected and setting appropriate preset frequencies, so as to analyze real-time vibration data to judge the real-time fault degree, and compare it with the prediction results of historical fault models. Then, real-time calibration of the model can solve the problem of inaccurate fault prediction results of industrial equipment.

[0006] To achieve the above object, the present invention provides an industrial equipment self-diagnosis system based on algorithm model prediction, the system comprising: A model building module, used to determine historical vibration abnormality data and its corresponding historical fault degree according to historical vibration data, and to build a historical fault model based on a number of historical vibration abnormality data and its corresponding historical fault degree; A data acquisition module, connected to the model building module, is used to divide a number of points to be detected on the output shaft of the industrial equipment based on a preset interval, collect a number of real-time vibration signals corresponding to the number of points to be detected in real time based on a preset frequency, convert the number of real-time vibration signals corresponding to any of the points to be detected into a number of real-time vibration data and draw a real-time vibration graph; A period judgment module, connected to the data acquisition module, for performing periodic judgment on any of the real-time vibration patterns, so as to determine the abnormality of the vibration period and issue a warning according to the periodic judgment result, or to analyze the abnormality of the point; A data analysis module connected to the period determination module, configured to adjust the preset frequency based on the abnormal situation of the point, or determine the overall abnormal situation based on the abnormal situations of several points, determine the real-time fault degree based on the overall abnormal situation, perform fault warning based on the comparison result between the real-time fault degree and the preset fault degree, or adjust the preset interval; A model comparison module, connected to the data analysis module, for inputting a number of the real-time vibration data into the historical fault model to obtain a predicted fault degree, and calculating a fault degree difference between the predicted fault degree and the real-time fault degree; A model calibration module is connected to the model comparison module and is used to calibrate the historical fault model based on the fault degree difference to obtain a calibrated fault model.

[0007] Furthermore, the cycle determination module includes: A period identification unit, used to identify whether there is a repeated waveform in any of the real-time vibration graphs, and to identify the time interval between two adjacent repeated waveforms as a real-time period; a period determination unit connected to the period identification unit, for determining that the real-time vibration graph is a non-periodic vibration when there is no repeated waveform in the real-time vibration graph, or when there is a repeated waveform in the real-time vibration graph and several real-time periods are inconsistent, and for determining that the real-time vibration graph is a periodic vibration when there is a repeated waveform in the real-time vibration graph and several real-time periods are consistent; The abnormality judgment unit is connected to the period judgment unit, and is used to determine that the vibration period is abnormal and issue a warning when the real-time vibration graph shows non-periodic vibration, and to analyze the abnormal situation of the point when the real-time vibration graph shows periodic vibration.

[0008] Furthermore, the cycle identification unit includes: The pattern recognition subunit is used to determine a preset time window, extract the waveform profile within the preset time window, calculate the similarity between several waveform profiles and compare it with a similarity threshold, and determine whether there is a repeated waveform based on the similarity comparison result; The period recognition subunit is connected to the pattern recognition subunit and is used to calculate the time interval between any two adjacent repeated waveforms to obtain the real-time period.

[0009] Furthermore, the graphic recognition subunit includes: A window setting block is used to identify a number of rising waveform segments and falling waveform segments in the real-time vibration graph, and extract a time period corresponding to any connected rising waveform segment and falling waveform segment as a time period corresponding to the preset time window; A similarity calculation block is used to calculate the similarity between any two waveform profiles within a preset time window based on a similarity algorithm to obtain a number of similarity values; The similarity comparison block is connected to the similarity calculation block, and is used to determine that the corresponding waveform contours are similar when the similarity value is greater than the similarity threshold, that is, to determine that there is a repeated waveform.

[0010] Furthermore, the abnormality judgment unit includes: A point analysis block is used to mark the area in the real-time vibration map that does not belong to the preset vibration range as a marked area; The abnormality judgment block is connected to the point analysis block to determine the abnormal distribution and abnormal proportion of several marked areas. When the abnormal distribution is dense, or the abnormal proportion is greater than the preset proportion, the point is determined to be abnormal.

[0011] Furthermore, the data analysis module includes: A frequency adjustment unit, configured to adjust the preset frequency when the abnormal distribution is dispersed and the abnormal proportion is less than the preset proportion; The overall abnormality determination unit is used to determine the number of abnormal points among the points to be detected when the point is abnormal, determine the real-time fault degree based on the number of abnormalities, and issue a fault warning based on the comparison result between the real-time fault degree and the preset fault degree, or adjust the preset interval.

[0012] Furthermore, the overall abnormality determination unit includes: A degree determination subunit, used to determine the real-time fault degree based on the number of abnormalities and the total number of the plurality of points to be detected; A fault warning subunit, connected to the degree determination subunit, for performing a fault warning when the real-time fault degree is greater than a preset fault degree; The interval adjustment subunit is connected to the degree determination subunit, and is used to determine the abnormal point area when the real-time fault degree is less than the preset fault degree, and adjust the preset interval according to the abnormal point area.

[0013] Furthermore, the model building module includes: A data acquisition unit, used to determine the historical vibration abnormality data corresponding to the abnormality of the output shaft of the industrial equipment, and to determine the corresponding historical fault degree; The model building unit is connected to the data acquisition unit and is used to divide the historical vibration abnormality data and the corresponding historical fault degrees into a training set and a test set, build a model based on the training set, and test the built model based on the test set to obtain the historical fault model.

[0014] Furthermore, the model calibration module includes: A coefficient determination unit, configured to determine an adjustment coefficient based on the fault degree difference; A model calibration unit is connected to the coefficient determination unit and is used to calibrate the model coefficients in the historical fault model based on the adjustment coefficients, so as to calibrate the historical fault model and obtain a calibrated fault model.

[0015] Furthermore, the model calibration unit includes: A coefficient determination subunit, used to determine a plurality of correlation coefficients based on a plurality of the historical vibration abnormality data and their corresponding historical fault degrees, and to use the average of the plurality of correlation coefficients as the model coefficient; The coefficient calibration subunit is connected to the coefficient determination subunit and is used to calibrate the model coefficient in the historical fault model based on the adjustment coefficient.

[0016] Compared with the prior art, the beneficial effect of the present invention is that, by setting the data acquisition module to accurately divide the points to be detected and setting a suitable preset frequency, the vibration signal on the output shaft of the industrial equipment can be collected in real time and comprehensively, and converted into an intuitive real-time vibration graph, which provides a reliable data basis for subsequent analysis and judgment. By setting the period judgment module to perform periodic judgment on the real-time vibration graph, the vibration period anomaly can be quickly identified and a warning can be issued. At the same time, the abnormal situation of the point can be analyzed, which provides an accurate data basis for subsequent data analysis and model comparison. By setting the data analysis module to adjust the preset frequency based on the abnormal situation of the point, the key monitoring of key areas can be achieved. At the same time, by determining the overall abnormal situation and the real-time fault degree, it is possible to provide a fault warning in time to avoid further development of the fault, adjust the preset interval to optimize the sensor layout, and improve the efficiency and accuracy of data collection. By setting the model comparison module in combination with the real-time vibration data and the historical fault model, the fault degree of the equipment is predicted and compared with the real-time fault degree, so as to calculate the fault degree deviation, which provides an important basis and accurate data foundation for the calibration of the model. By setting the model calibration module to calibrate the historical fault model based on the fault degree difference, the prediction accuracy of the model is continuously improved, which is helpful to achieve accurate prediction and timely warning of equipment failures and improve the reliability and safety of the equipment.

[0017] In particular, by setting the period identification unit to accurately identify whether there are repeated waveforms in the real-time vibration graph and measuring the time interval between adjacent repeated waveforms, an accurate data basis is provided for subsequent period judgment. By setting the period judgment unit to determine whether the real-time vibration graph shows periodic vibration, it is helpful to quickly identify the vibration state of the equipment. By setting the abnormality judgment unit, when the real-time vibration graph shows non-periodic vibration, it can immediately determine that the vibration period is abnormal and issue a warning, which helps to avoid further damage or failure of the equipment caused by abnormal vibration period. When the real-time vibration graph shows periodic vibration, further analysis of point abnormalities helps to accurately locate the faults of each monitoring point of the equipment and improve the accuracy of equipment fault judgment.

[0018] In particular, by setting a graphic recognition subunit to calculate the similarity between several waveform contours and comparing it with a preset similarity threshold, it is possible to accurately determine whether there are repeated waveforms, thereby providing a reliable data basis for subsequent period identification. By setting the period identification subunit to calculate the time interval between any two adjacent repeated waveforms, the real-time period is obtained, which helps to quickly identify the vibration periodic characteristics of the device, provides a secondary judgment basis for determining whether the device vibration is periodic, and improves the accuracy of the periodic judgment results. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A structural block diagram of an industrial equipment self-diagnosis system based on algorithm model prediction provided by an embodiment of the present invention; Figure 2 A structural block diagram of a cycle determination module of an industrial equipment self-diagnosis system based on algorithm model prediction provided by an embodiment of the present invention; Figure 3 A structural block diagram of a data analysis module of an industrial equipment self-diagnosis system based on algorithm model prediction provided by an embodiment of the present invention; Figure 4 A structural block diagram of a model comparison module of an industrial equipment self-diagnosis system based on algorithm model prediction provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0021] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0022] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be understood as a limitation on the present invention.

[0023] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0024] See also Figure 1 As shown, an embodiment of the present invention provides an industrial equipment self-diagnosis system based on algorithm model prediction, the system comprising: A model building module 10 is used to determine historical vibration abnormality data and its corresponding historical fault degree according to the historical vibration data, and to build a historical fault model based on a number of historical vibration abnormality data and its corresponding historical fault degree; The data acquisition module 20 is connected to the model building module 10, and is used to divide a plurality of points to be detected on the output shaft of the industrial equipment based on a preset interval, collect a plurality of real-time vibration signals corresponding to the plurality of points to be detected in real time based on a preset frequency, convert the plurality of real-time vibration signals corresponding to any of the points to be detected into a plurality of real-time vibration data and draw a real-time vibration graph; The period determination module 30 is connected to the data acquisition module 20 and is used to perform period determination on any of the real-time vibration patterns, so as to determine the abnormality of the vibration period and issue a warning according to the period determination result, or analyze the abnormality of the point; The data analysis module 40 is connected to the period determination module 30, and is used to adjust the preset frequency based on the abnormal situation of the point, or determine the overall abnormal situation based on the abnormal situations of several points, determine the real-time fault degree based on the overall abnormal situation, and perform fault warning based on the comparison result between the real-time fault degree and the preset fault degree, or adjust the preset interval; A model comparison module 50, connected to the data analysis module 40, is used to input a number of the real-time vibration data into the historical fault model to obtain a predicted fault degree, and calculate a fault degree difference between the predicted fault degree and the real-time fault degree; The model calibration module 60 is connected to the model comparison module 50 and is used to calibrate the historical fault model based on the fault degree difference to obtain a calibrated fault model.

[0025] Specifically, the embodiment of the present invention can collect the vibration signal on the output shaft of the industrial equipment in real time and comprehensively by setting the data acquisition module to accurately divide the points to be detected and set the appropriate preset frequency, and convert it into an intuitive real-time vibration graph, thereby providing a reliable data basis for subsequent analysis and judgment. By setting the period judgment module to perform periodic judgment on the real-time vibration graph, it can quickly identify vibration period anomalies and issue warnings. At the same time, it can also analyze the abnormal situation of the point, thereby providing an accurate data basis for subsequent data analysis and model comparison. By setting the data analysis module to adjust the preset frequency based on the abnormal situation of the point, it is possible to achieve key monitoring of key areas. By determining the overall abnormal situation and the real-time fault degree, it is possible to provide timely fault warning to avoid further development of the fault, adjust the preset interval to optimize the sensor layout, and improve the efficiency and accuracy of data collection. By setting the model comparison module to combine the real-time vibration data and the historical fault model, the fault degree of the equipment is predicted and compared with the real-time fault degree, so as to calculate the fault degree deviation, which provides an important basis and accurate data foundation for the calibration of the model. By setting the model calibration module to calibrate the historical fault model based on the fault degree difference, the prediction accuracy of the model is continuously improved, which helps to achieve accurate prediction and timely warning of equipment failures and improve the reliability and safety of the equipment.

[0026] Industrial equipment refers to industrial production equipment and various machine tools, such as lathes, milling machines, grinders, planers and other machines. Industrial equipment includes many types, mainly including general equipment, special equipment, process equipment, testing equipment and instruments, etc. Among them, general equipment is suitable for the production process of various industries, mainly including machine tools, lifting equipment, complete sets of equipment, etc.; special equipment is suitable for the production process of specific industries, such as steel smelting equipment, petrochemical equipment, food processing equipment, etc.; process equipment is equipment used in production processes, such as processing equipment and pressure vessels, etc.; testing equipment is equipment used to detect product quality and performance, including Various testing instruments and test benches, etc.; Instruments and meters are equipment used to measure, control and record various physical quantities, such as thermometers, pressure gauges, flow meters, etc.; in addition, industrial equipment can be further classified according to their uses, among which production equipment is various mechanical equipment used in the production process, and power equipment includes steam generating equipment (such as steam boilers), power generating equipment (such as steam engines, steam turbines), power conversion equipment (such as generators, motors), power adjustment or regulation equipment (such as transformers, rectifier equipment) and power transmission equipment (such as power lines, steam pipes), etc. The industrial equipment described in the embodiments of the present invention may be a generator.

[0027] Specifically, the preset interval in the embodiment of the present invention needs to be small enough to accurately capture the vibration conditions at different positions on the shaft, but large enough to avoid mutual interference between sensors. The specific value will be determined based on the length of the shaft and the performance of the sensor. For example, if the length of the shaft is 5 meters, the preset interval can be selected as 10 centimeters, so that approximately 50 detection points can be evenly distributed on the shaft. The preset frequency is determined based on the historical failure frequency of the equipment. If the historical failure frequency of the equipment is once a month, then the preset frequency may be set to collect data once a week to promptly detect signs of failure.

[0028] Specifically, the real-time vibration diagram can be obtained by performing Fourier series expansion on a number of sampled real-time vibration signals, and obtaining a spectrum diagram of the real-time vibration signal, i.e., a representation of the real-time vibration signal in the frequency domain, through Fourier transformation. The spectrum diagram shows the intensity (or amplitude) of each frequency component in the signal, so that key features such as the main frequency and secondary frequency in the signal can be intuitively identified.

[0029] See also Figure 2 As shown, the period determination module 30 includes: A period identification unit 31 is used to identify whether there is a repeated waveform in any of the real-time vibration graphs, and to identify the time interval between two adjacent repeated waveforms as a real-time period; The period determination unit 32 is connected to the period identification unit 31, and is used to determine that the real-time vibration graph is a non-periodic vibration when there is no repeated waveform in the real-time vibration graph, or when there is a repeated waveform in the real-time vibration graph and several real-time periods are inconsistent, and to determine that the real-time vibration graph is a periodic vibration when there is a repeated waveform in the real-time vibration graph and several real-time periods are consistent; The abnormality judgment unit 33 is connected to the period judgment unit 32, and is used to determine that the vibration period is abnormal and issue a warning when the real-time vibration graph shows non-periodic vibration, and to analyze the abnormal situation of the point when the real-time vibration graph shows periodic vibration.

[0030] Specifically, the embodiment of the present invention provides an accurate data basis for subsequent period judgment by setting the period identification unit to accurately identify whether there are repeated waveforms in the real-time vibration graph, and measures the time interval between adjacent repeated waveforms. By setting the period judgment unit to determine whether the real-time vibration graph shows periodic vibration, it is helpful to quickly identify the vibration state of the equipment. By setting the abnormality judgment unit, when the real-time vibration graph shows non-periodic vibration, it can immediately determine that the vibration period is abnormal and issue a warning, which helps to avoid further damage or failure of the equipment caused by abnormal vibration period. When the real-time vibration graph shows periodic vibration, further analyzing the abnormal situation of the point helps to accurately locate the failure of each monitoring point of the equipment and improve the accuracy of equipment failure judgment.

[0031] It can be understood that the embodiment of the present invention determines that several real-time cycles are consistent by judging the proportion of real-time cycles corresponding to the same numerical value. When the proportion is greater than the preset proportion, it is determined that the several real-time cycles are consistent. When the proportion is less than the preset proportion, it is determined that the several real-time cycles are inconsistent. The preset proportion is set to 90%. Setting a proportion of 90% ensures that only when the vibration periods are the same in most cases will they be judged as periodic vibrations, so as to reduce misjudgment and improve the reliability of judgment.

[0032] Specifically, the cycle identification unit includes: The pattern recognition subunit is used to determine a preset time window, extract the waveform profile within the preset time window, calculate the similarity between several waveform profiles and compare it with a similarity threshold, and determine whether there is a repeated waveform based on the similarity comparison result; The period recognition subunit is connected to the pattern recognition subunit and is used to calculate the time interval between any two adjacent repeated waveforms to obtain the real-time period.

[0033] Specifically, the embodiment of the present invention calculates the similarity between several waveform contours by setting a graphic recognition subunit, and compares it with a preset similarity threshold, so as to accurately determine whether there is a repeated waveform, thereby providing a reliable data basis for subsequent period identification. By setting the period identification subunit to calculate the time interval between any two adjacent repeated waveforms, a real-time period is obtained, which helps to quickly identify the vibration periodic characteristics of the device, provides a secondary judgment basis for determining whether the device vibration is periodic, and improves the accuracy of the periodic judgment results.

[0034] It can be understood that, in the period identification subunit described in the embodiment of the present invention, “adjacent” refers to two repeated waveforms that are immediately adjacent in time series.

[0035] It can be understood that the waveform contour extracted in the embodiment of the present invention can be extracted by an edge detection algorithm. The process of extraction based on the edge detection algorithm is a prior art and will not be described in detail here.

[0036] Specifically, the graphic recognition subunit includes: A window setting block is used to identify a number of rising waveform segments and falling waveform segments in the real-time vibration graph, and extract a time period corresponding to any connected rising waveform segment and falling waveform segment as a time period corresponding to the preset time window; A similarity calculation block is used to calculate the similarity between any two waveform profiles within a preset time window based on a similarity algorithm to obtain a number of similarity values; The similarity comparison block is connected to the similarity calculation block, and is used to determine that the corresponding waveform contours are similar when the similarity value is greater than the similarity threshold, that is, to determine that there is a repeated waveform.

[0037] Specifically, the embodiments of the present invention timely discover fault signs by identifying and comparing vibration waveforms, and similarity calculation provides quantitative indicators to reduce the possibility of misjudgment. By identifying repeated waveforms, a basis is provided for determining whether the vibration is a periodic change.

[0038] It can be understood that, in the embodiment of the present invention, the rising band and the falling band in the real-time vibration graph can be determined by calculating the derivatives of several vibration values ​​corresponding to the real-time vibration graph, identifying the point where the derivative changes from positive to negative as the inflection point, and defining the part from the inflection point to the part before the next inflection point where the derivatives are all positive as the rising band, and defining the part from the end point of the rising band to the part before the beginning of the next rising band as the falling band.

[0039] Specifically, the abnormality judgment unit includes: A point analysis block is used to mark the area in the real-time vibration map that does not belong to the preset vibration range as a marked area; The abnormality judgment block is connected to the point analysis block to determine the abnormal distribution and abnormal proportion of several marked areas. When the abnormal distribution is dense, or the abnormal proportion is greater than the preset proportion, the point is determined to be abnormal.

[0040] Specifically, the embodiment of the present invention provides a basis for subsequent abnormality judgment by setting the point analysis block to accurately mark the area in the real-time vibration map that does not belong to the preset vibration range. By setting the abnormal judgment block based on the marked area provided by the point analysis block, the abnormal distribution and the abnormal proportion are further determined, and whether the vibration state of the equipment is abnormal, as well as the specific location and degree of the abnormality are more accurately judged. This helps to improve the accuracy of fault diagnosis and provide an important basis for subsequent equipment maintenance and fault handling. When the abnormal distribution is dense or the abnormal proportion is greater than the preset proportion, the abnormal judgment block can immediately determine the point abnormality and trigger an early warning mechanism, which helps to avoid further damage or failure of the equipment due to abnormal vibration cycle and improve the reliability and safety of the equipment.

[0041] It can be understood that the embodiment of the present invention can determine the abnormal distribution of the marked area by counting the number of times each marked area appears in a preset time window, averaging the number of occurrences of each marked area, and obtaining the average occurrence frequency. The calculation formula can be: average occurrence frequency = (total number of occurrences) / (number of statistical time windows). For example, if the marked area appears 3 times, 4 times, 5 times, 2 times and 3 times in 5 statistical time windows respectively, then its average occurrence frequency is (3+4+5+2+3) / 5=3.4 times / time window. The average occurrence frequency is compared with the preset frequency. When the average occurrence frequency is greater than the preset frequency, the abnormal distribution is dense, wherein the preset frequency is a certain percentage of the total frequency of the marked area, such as 70%, to ensure that the average occurrence frequency is greater than the vast majority of the total frequency to further determine the dense distribution.

[0042] It can be understood that, in the embodiment of the present invention, the abnormal proportion can be determined by dividing the sum of the time periods corresponding to the plurality of marked areas by the total time period of the waveform in the real-time vibration diagram.

[0043] See also Figure 3 As shown, the data analysis module 40 includes: A frequency adjustment unit 41, configured to adjust the preset frequency when the abnormal distribution is dispersed and the abnormal proportion is less than the preset proportion; The overall abnormality determination unit 42 is used to determine the number of abnormal points among the points to be detected when the point is abnormal, determine the real-time fault degree based on the number of abnormalities, and issue a fault warning based on the comparison result between the real-time fault degree and the preset fault degree, or adjust the preset interval.

[0044] Specifically, the overall abnormality determination unit includes: A degree determination subunit, used to determine the real-time fault degree based on the number of abnormalities and the total number of the plurality of points to be detected; A fault warning subunit, connected to the degree determination subunit, for performing a fault warning when the real-time fault degree is greater than a preset fault degree; The interval adjustment subunit is connected to the degree determination subunit, and is used to determine the abnormal point area when the real-time fault degree is less than the preset fault degree, and adjust the preset interval according to the abnormal point area.

[0045] It can be understood that the real-time fault degree described in the embodiment of the present invention is determined by calculating the ratio of the number of abnormalities to the total number of points.

[0046] It can be understood that the preset degree in the embodiment of the present invention is 1 / 3, that is, when the number of abnormal points exceeds 1 / 3 of the total number of points to be detected, it proves that the output shaft of the industrial equipment is faulty, and a fault warning is issued at this time.

[0047] It can be understood that the abnormal point area in the embodiment of the present invention is a formation area centered on the abnormal point, with the nearest non-abnormal points on the left and right as the starting point and the end point, and adjusting the preset interval includes reducing the preset interval according to the severity of the corresponding point abnormality, thereby increasing the monitoring equipment to achieve more intensive monitoring, wherein the severity of the point abnormality = (preset frequency / average occurrence frequency) × (preset proportion / abnormal proportion), wherein the severity of the point abnormality is used as a reduction factor to reduce the preset interval.

[0048] See also Figure 4 As shown, the model building module 10 includes: The data acquisition unit 11 is used to determine the historical vibration abnormality data corresponding to the abnormality of the output shaft of the industrial equipment, and determine the corresponding historical fault degree; The model building unit 12 is connected to the data acquisition unit 11, and is used to divide the historical vibration abnormality data and the corresponding historical fault degrees into a training set and a test set, build a model based on the training set, and test the built model based on the test set to obtain the historical fault model.

[0049] Specifically, when the model building module of the embodiment of the present invention builds the historical fault model, it first collects historical vibration data, which includes vibration information of the equipment under different working conditions and corresponding historical fault records. By analyzing these data, the correlation between vibration characteristics and faults can be identified. Then, the data is trained using statistical methods or machine learning algorithms, such as support vector machines (SVM), random forests, neural networks, etc., to obtain model coefficients that can predict faults. These model coefficients are the core of the historical fault model, and they can reflect the strength of the relationship between different vibration characteristics and faults. The above-mentioned model building technology is a prior art and will not be repeated here.

[0050] Specifically, the model calibration module includes: A coefficient determination unit, configured to determine an adjustment coefficient based on the fault degree difference; A model calibration unit is connected to the coefficient determination unit and is used to calibrate the model coefficients in the historical fault model based on the adjustment coefficients, so as to calibrate the historical fault model and obtain a calibrated fault model.

[0051] It can be understood that, in the embodiment of the present invention, when the fault degree difference is greater than 0 and less than the set degree difference threshold, the model coefficients in the historical fault model are calibrated to calibrate the historical fault model, and when the fault degree difference is greater than the set degree difference threshold, the vibration data is re-collected to analyze the fault degree to construct a new fault prediction model, wherein the set degree difference threshold is 10% of the predicted fault degree.

[0052] Specifically, the model calibration unit includes: A coefficient determination subunit, used to determine a plurality of correlation coefficients based on a plurality of the historical vibration abnormality data and their corresponding historical fault degrees, and to use the average of the plurality of correlation coefficients as the model coefficient; The coefficient calibration subunit is connected to the coefficient determination subunit and is used to calibrate the model coefficient in the historical fault model based on the adjustment coefficient.

[0053] It can be understood that the correlation coefficients described in the embodiment of the present invention can be calculated by using statistical methods or machine learning algorithms, such as linear regression, nonlinear fitting or neural network, to analyze the correlation between abnormal vibration data and the degree of fault, thereby calculating a number of correlation coefficients.

[0054] It can be understood that the adjustment coefficient in the embodiment of the present invention = fault degree difference / set degree difference threshold.

[0055] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0056] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An industrial equipment self-diagnosis system based on algorithm model prediction, characterized in that: include: A model building module, used to determine historical vibration abnormality data and its corresponding historical fault degree according to historical vibration data, and to build a historical fault model based on a number of historical vibration abnormality data and its corresponding historical fault degree; A data acquisition module, connected to the model building module, is used to divide a number of points to be detected on the output shaft of the industrial equipment based on a preset interval, collect a number of real-time vibration signals corresponding to the number of points to be detected in real time based on a preset frequency, convert the number of real-time vibration signals corresponding to any of the points to be detected into a number of real-time vibration data and draw a real-time vibration graph; A period judgment module, connected to the data acquisition module, for performing periodic judgment on any of the real-time vibration patterns, so as to determine the abnormality of the vibration period and issue a warning according to the periodic judgment result, or to analyze the abnormality of the point; A data analysis module connected to the period determination module, configured to adjust the preset frequency based on the abnormal situation of the point, or determine the overall abnormal situation based on the abnormal situations of several points, determine the real-time fault degree based on the overall abnormal situation, perform fault warning based on the comparison result between the real-time fault degree and the preset fault degree, or adjust the preset interval; A model comparison module, connected to the data analysis module, for inputting a number of the real-time vibration data into the historical fault model to obtain a predicted fault degree, and calculating a fault degree difference between the predicted fault degree and the real-time fault degree; A model calibration module is connected to the model comparison module and is used to calibrate the historical fault model based on the fault degree difference to obtain a calibrated fault model.

2. The industrial equipment self-diagnosis system based on algorithm model prediction according to claim 1 is characterized in that: The cycle determination module comprises: A period identification unit, used to identify whether there is a repeated waveform in any of the real-time vibration graphs, and to identify the time interval between two adjacent repeated waveforms as a real-time period; a period determination unit connected to the period identification unit, for determining that the real-time vibration graph is a non-periodic vibration when there is no repeated waveform in the real-time vibration graph, or when there is a repeated waveform in the real-time vibration graph and several real-time periods are inconsistent, and for determining that the real-time vibration graph is a periodic vibration when there is a repeated waveform in the real-time vibration graph and several real-time periods are consistent; The abnormality judgment unit is connected to the period judgment unit, and is used to determine that the vibration period is abnormal and issue a warning when the real-time vibration graph shows non-periodic vibration, and to analyze the abnormal situation of the point when the real-time vibration graph shows periodic vibration.

3. The industrial equipment self-diagnosis system based on algorithm model prediction according to claim 2 is characterized in that: The cycle identification unit comprises: The pattern recognition subunit is used to determine a preset time window, extract the waveform profile within the preset time window, calculate the similarity between several waveform profiles and compare it with a similarity threshold, and determine whether there is a repeated waveform based on the similarity comparison result; The period recognition subunit is connected to the pattern recognition subunit and is used to calculate the time interval between any two adjacent repeated waveforms to obtain the real-time period.

4. The industrial equipment self-diagnosis system based on algorithm model prediction according to claim 3 is characterized in that: The graphic recognition subunit comprises: A window setting block is used to identify a number of rising waveform segments and falling waveform segments in the real-time vibration graph, and extract a time period corresponding to any connected rising waveform segment and falling waveform segment as a time period corresponding to the preset time window; A similarity calculation block is used to calculate the similarity between any two waveform profiles within a preset time window based on a similarity algorithm to obtain a number of similarity values; The similarity comparison block is connected to the similarity calculation block, and is used to determine that the corresponding waveform contours are similar when the similarity value is greater than the similarity threshold, that is, to determine that there is a repeated waveform.

5. The industrial equipment self-diagnosis system based on algorithm model prediction according to claim 4 is characterized in that: The abnormality judgment unit comprises: A point analysis block is used to mark the area in the real-time vibration map that does not belong to the preset vibration range as a marked area; The abnormality judgment block is connected to the point analysis block to determine the abnormal distribution and abnormal proportion of several marked areas. When the abnormal distribution is dense, or the abnormal proportion is greater than the preset proportion, the point is determined to be abnormal.

6. The industrial equipment self-diagnosis system based on algorithm model prediction according to claim 5 is characterized in that: The data analysis module includes: A frequency adjustment unit, configured to adjust the preset frequency when the abnormal distribution is dispersed and the abnormal proportion is less than the preset proportion; The overall abnormality determination unit is used to determine the number of abnormal points among the points to be detected when the point is abnormal, determine the real-time fault degree based on the number of abnormalities, and issue a fault warning based on the comparison result between the real-time fault degree and the preset fault degree, or adjust the preset interval.

7. The industrial equipment self-diagnosis system based on algorithm model prediction according to claim 6 is characterized in that: The overall abnormality determination unit includes: A degree determination subunit, used to determine the real-time fault degree based on the number of abnormalities and the total number of the plurality of points to be detected; A fault warning subunit, connected to the degree determination subunit, for performing a fault warning when the real-time fault degree is greater than a preset fault degree; The interval adjustment subunit is connected to the degree determination subunit, and is used to determine the abnormal point area when the real-time fault degree is less than the preset fault degree, and adjust the preset interval according to the abnormal point area.

8. The industrial equipment self-diagnosis system based on algorithm model prediction according to claim 7 is characterized in that: The model building module includes: A data acquisition unit, used to determine the historical vibration abnormality data corresponding to the abnormality of the output shaft of the industrial equipment, and to determine the corresponding historical fault degree; The model building unit is connected to the data acquisition unit and is used to divide the historical vibration abnormality data and the corresponding historical fault degrees into a training set and a test set, build a model based on the training set, and test the built model based on the test set to obtain the historical fault model.

9. The industrial equipment self-diagnosis system based on algorithm model prediction according to claim 8, characterized in that: The model calibration module comprises: A coefficient determination unit, configured to determine an adjustment coefficient based on the fault degree difference; A model calibration unit is connected to the coefficient determination unit and is used to calibrate the model coefficients in the historical fault model based on the adjustment coefficients, so as to calibrate the historical fault model and obtain a calibrated fault model.

10. The industrial equipment self-diagnosis system based on algorithm model prediction according to claim 9, characterized in that: The model calibration unit comprises: A coefficient determination subunit, used to determine a plurality of correlation coefficients based on a plurality of the historical vibration abnormality data and their corresponding historical fault degrees, and to use the average of the plurality of correlation coefficients as the model coefficient; The coefficient calibration subunit is connected to the coefficient determination subunit and is used to calibrate the model coefficient in the historical fault model based on the adjustment coefficient.

Images