Rotor wear degree analysis and evaluation method

Abstract

The invention discloses a rotor wear degree analysis and evaluation method, which comprises the steps of obtaining historical monitoring logs of a rotor, and determining an optimal monitoring mode and optimal wear data according to an association relationship between historical wear data of a monitoring mode corresponding to each historical monitoring log and a preset evaluation index and historical monitoring information of the monitoring mode; according to the optimal wear data and historical evaluation values of the corresponding preset evaluation indexes, generating the wear degree of the rotor, and according to the optimal wear data, the plurality of historical evaluation values and the corresponding wear degree, constructing a wear degree prediction model; the method comprises the following steps: acquiring real-time optimal wear data of a rotor, inputting the real-time optimal wear data into a wear degree prediction model to obtain a predicted wear degree, and judging whether to send an alarm instruction or not according to the predicted wear degree. By selecting a reasonable monitoring mode, the evaluation precision of the wear degree of the rotor is improved, and corresponding effective measures are timely found and taken. And the motor performance and efficiency are improved.

Description

Technical Field

[0001] The present application relates to the technical field of rotor monitoring, and in particular to a method for analyzing and evaluating the degree of rotor wear. Background Art

[0002] As the core component of steam turbines, generators and fans, rotors need to withstand harsh working conditions such as high temperature, high pressure, high speed, and high stress. Due to other unfavorable operating factors such as lubricating oil quality, impurities, and frequent start-up and shutdown, the rotor shaft neck will experience varying degrees of wear and roughening. The damage depth is generally within 2mm. Minor rotor wear may not immediately affect the normal operation of the motor, but over time, the wear will gradually increase, which may eventually lead to a decline in motor performance and even cause more serious problems.

[0003] In the prior art, there are multiple ways to monitor the degree of rotor wear. Each monitoring method has differences in monitoring accuracy and timeliness, resulting in different analysis and evaluation results of the rotor wear degree. Therefore, there is an urgent need for a rotor wear degree analysis and evaluation method to select a reasonable monitoring method, accurately evaluate the degree of rotor wear, and promptly discover and take corresponding effective measures to improve motor performance and efficiency. Summary of the invention

[0004] In order to solve the above technical problems, the present application provides a method for analyzing and evaluating the degree of rotor wear. By analyzing the correlation between historical wear data of multiple monitoring methods and preset evaluation indicators, the correlation coefficient is determined and a comprehensive analysis is performed based on the monitoring accuracy, monitoring timeliness and monitoring cost of the historical wear data to determine the preferred monitoring method and preferred wear data. A wear degree prediction model is constructed based on the preferred wear data to obtain the predicted degree of wear. Whether to send an alarm signal is determined based on the predicted degree of wear. By selecting a reasonable monitoring method, the accuracy of rotor wear degree assessment is improved, and corresponding effective measures are discovered and taken in time to improve motor performance and efficiency.

[0005] In some embodiments of the present application, a method for analyzing and evaluating the degree of rotor wear is provided, comprising:

[0006] Obtain historical monitoring logs of the rotor, and determine the monitoring method corresponding to each historical monitoring log and the historical wear data of the monitoring method;

[0007] Determine the preferred monitoring method and preferred wear data according to the correlation between the historical wear data of the monitoring method and the preset evaluation index and the historical monitoring information of the monitoring method;

[0008] Generate the wear degree of the rotor according to the preferred wear data and the historical evaluation values ​​of the corresponding preset evaluation indicators, and build a wear degree prediction model according to the preferred wear data, multiple historical evaluation values ​​and the corresponding wear degrees;

[0009] The real-time optimal wear data of the rotor is obtained, and the real-time optimal wear data is input into the wear degree prediction model to obtain the predicted wear degree, and it is determined whether to send an alarm instruction according to the predicted wear degree.

[0010] In some embodiments of the present application, determining a preferred monitoring method and preferred wear data includes:

[0011] Determine the monitoring method involved in each historical monitoring log according to the historical monitoring log of the rotor, classify the historical wear data obtained by the same monitoring method into the same category, and construct historical wear data sets of multiple categories;

[0012] Preset a plurality of preset evaluation indicators, and determine a historical evaluation value of the preset evaluation indicator corresponding to a historical node of each historical wear data according to the historical operation data in the historical operation log of the motor, wherein the preset evaluation indicator includes a performance evaluation indicator, an efficiency evaluation indicator, and a state evaluation indicator;

[0013] A focus period is set according to the mutation period of the historical evaluation value of each preset evaluation indicator, and a preset collection node is set at a preset time interval with the focus period as a reference time line;

[0014] Randomly select S1 pieces of historical wear data from each category of historical wear data set, and collect corresponding historical wear data according to multiple preset collection nodes of each reference timeline, and map them to the corresponding reference timeline to generate a data change curve of the selected historical wear data;

[0015] Analyze the data change curve to obtain the change characteristics of the historical wear data selected from the historical wear data sets of multiple categories, wherein the change characteristics include the change time node and the change value;

[0016] Generate the correlation degree between the selected historical wear data and the corresponding preset evaluation index according to the change time node and the change value;

[0017] The historical wear data with a correlation degree greater than a preset correlation degree threshold is set as the concerned wear data corresponding to the preset evaluation index;

[0018] The number of concerned wear data of the same category in each preset evaluation indicator is compared with the number of S1 screened out to obtain the concern ratio of the corresponding category, and the monitoring method of the corresponding category whose concern ratio is greater than the preset concern ratio threshold is set as the concerned monitoring method of the corresponding preset evaluation indicator;

[0019] Comprehensively analyze the monitoring methods of each preset evaluation indicator to determine the preferred monitoring method and preferred wear data.

[0020] In some embodiments of the present application, determining the preferred monitoring method and the preferred wear data further includes:

[0021] Comprehensively analyze the monitoring methods of each preset evaluation indicator, eliminate the monitoring methods that do not appear in all the preset evaluation indicators at the same time, and determine the correlation coefficient between the corresponding monitoring method and each preset evaluation indicator according to the correlation degree between the historical wear data screened by the remaining monitoring methods and each preset evaluation indicator;

[0022] The correlation coefficients of all preset evaluation indicators corresponding to the same monitoring method are averaged to obtain the average correlation coefficient of the corresponding monitoring method;

[0023] The historical monitoring information of the monitoring method includes historical monitoring cost, historical monitoring accuracy and historical monitoring timeliness;

[0024] Quantify the historical monitoring cost, historical monitoring accuracy, and historical monitoring timeliness of the remaining monitoring methods of concern to obtain a first quantized value, a second quantized value, and a third quantized value, and compare them with the corresponding standard quantized values, and set a correction coefficient according to the comparison result;

[0025] The correlation coefficient mean of the corresponding monitoring method of interest is corrected according to the correction coefficient to obtain the characteristic correlation coefficient, and the remaining monitoring methods of interest are ranked according to the characteristic correlation coefficient. The first-ranked monitoring method of interest is set as the preferred monitoring method, and the historical wear data in the historical wear data set corresponding to the preferred monitoring method is set as the preferred wear data.

[0026] In some embodiments of the present application, the calculation formula of the degree of association is:

[0027] g=e -∣αb1-αb0∣ ×p;

[0028] Among them, g is the degree of association, b1 is the change time node, b0 is the mutation time node of the preset evaluation index, and p is the change value;

[0029] The calculation formula of the correlation coefficient is:

[0030]

[0031] Wherein, t0 is the correlation coefficient, S0 is the number of historical wear data with a correlation degree greater than a preset correlation degree threshold, and gi is the i-th historical wear data with a correlation degree greater than the preset correlation degree threshold;

[0032] The calculation formula of the characteristic correlation coefficient is:

[0033]

[0034] Where T is the characteristic correlation coefficient, is the mean of the correlation coefficient, n is the total number of preset evaluation indicators, t0j is the correlation coefficient of the jth preset evaluation indicator, z×[(a1-a0)×c1+(m1-m0)×c2+(k1-k0)×c3] is the correction coefficient, z is the correction conversion coefficient, a1 is the first quantitative value, a0 is the standard quantitative value corresponding to the first quantitative value, c1 is the weight coefficient corresponding to the monitoring cost, m1 is the second quantitative value, m0 is the standard quantitative value corresponding to the second quantitative value, c2 is the weight coefficient corresponding to the monitoring accuracy, k1 is the third quantitative value, k0 is the standard quantitative value corresponding to the third quantitative value, and c3 is the weight coefficient corresponding to the monitoring timeliness.

[0035] In some embodiments of the present application, generating the wear degree of the rotor according to the preferred wear data and the historical evaluation value of the corresponding preset evaluation index includes:

[0036] Generate a comprehensive evaluation value according to the preferred monitoring method and the historical evaluation value of each preset evaluation indicator of the historical monitoring log corresponding to the preferred wear data and the weight coefficient of the corresponding preset evaluation indicator;

[0037] The wear degree of the rotor is set according to the relationship between the comprehensive evaluation value and the preset evaluation value interval;

[0038] Presetting a first preset evaluation value interval, a second preset evaluation value interval, a third preset evaluation value interval and a fourth preset evaluation value interval;

[0039] When the comprehensive evaluation value is within the first preset evaluation value interval, the wear degree of the rotor is set to a fourth preset wear degree f4;

[0040] When the comprehensive evaluation value is within the second preset evaluation value interval, the wear degree of the rotor is set to a third preset wear degree f3;

[0041] When the comprehensive evaluation value is within the third preset evaluation value interval, the wear degree of the rotor is set to the second preset wear degree f2;

[0042] When the comprehensive evaluation value is within the fourth preset evaluation value interval, the wear degree of the rotor is set to the first preset wear degree f1;

[0043] Among them, f1<f2<f3<f4.

[0044] In some embodiments of the present application, a wear degree prediction model is constructed according to the preferred wear data, multiple historical evaluation values ​​and corresponding wear degrees, including:

[0045] Using the preferred wear data as first input data and using the historical evaluation value corresponding to each preset evaluation index as first output data;

[0046] Generate a comprehensive evaluation value according to the historical evaluation value corresponding to each preset evaluation indicator, use the comprehensive evaluation value as the second input data, and use the wear degree corresponding to the rotor as the second output data;

[0047] Performing a first neural network training according to a plurality of first input data and corresponding first output data, and performing a second neural network training according to a plurality of second input data and corresponding second output data;

[0048] The trained neural network model is set as the wear degree prediction model.

[0049] In some embodiments of the present application, real-time preferred wear data of the rotor is obtained, and the real-time preferred wear data is input into a wear degree prediction model to obtain a predicted wear degree, including:

[0050] Acquire real-time wear data of the rotor according to the preferred monitoring method, input the real-time wear data into a wear degree prediction model, and obtain predicted evaluation values ​​of multiple preset evaluation indicators;

[0051] Generate a predicted comprehensive evaluation value based on the predicted evaluation values ​​of multiple preset evaluation indicators and the weight coefficients of the corresponding preset evaluation indicators;

[0052] The predicted comprehensive evaluation value is input into the wear degree prediction model to obtain the predicted wear degree.

[0053] In some embodiments of the present application, before determining whether to send an alarm instruction according to the predicted wear degree, the method further includes:

[0054] Acquire real-time preferred wear data for a preset period, and construct a real-time preferred wear data curve for the preset period based on the real-time preferred wear data for the preset period;

[0055] Constructing a plurality of historical preferred wear data curve segments according to the historical preferred wear data of the preferred monitoring method, and selecting the historical wear data curve segments that match the curve characteristics of the real-time preferred wear data curve;

[0056] According to the time interval between the preset time period and the future time period, historical wear data of the time period corresponding to the historical wear data curve segment matching the curve characteristics of the real-time preferred wear data curve is obtained, and set as the first predicted wear data of the real-time preferred wear data in the future time period;

[0057] Inputting the first predicted wear data of multiple preset moments in the future period into the wear degree prediction model, obtaining the predicted wear degree at each preset moment in the future period, and constructing a predicted wear degree change curve;

[0058] A compensation coefficient is generated based on the change value at each preset moment in the predicted wear degree change curve and the degree of fluctuation in the future time period, the predicted wear degree is corrected based on the compensation coefficient, and whether to send an alarm signal is determined based on the corrected predicted wear degree.

[0059] In some embodiments of the present application, the calculation formula of the compensation coefficient is:

[0060]

[0061] Among them, r is the compensation coefficient, r0 is the compensation conversion coefficient, σ 2 is the variance of the predicted wear degree at multiple preset moments in the future period, z is the total number of preset moments in the future period, and F c is the predicted wear degree at the cth preset moment, F c+1 is the predicted wear degree at the c+1th preset moment.

[0062] In some embodiments of the present application, determining whether to send an alarm signal according to the corrected predicted wear degree includes:

[0063] Pre-set wear level threshold;

[0064] The corrected predicted wear degree is compared with the wear degree threshold. If the corrected predicted wear degree is less than the wear degree threshold, no alarm instruction is sent. If the corrected predicted wear degree is greater than the wear degree threshold, an alarm instruction is sent.

[0065] Compared with the prior art, the rotor wear degree analysis and evaluation method of the embodiment of the present application has the following beneficial effects:

[0066] By analyzing the correlation between historical wear data of multiple monitoring methods and preset evaluation indicators, the correlation coefficient is determined and a comprehensive analysis is conducted based on the monitoring accuracy, monitoring timeliness and monitoring cost of historical wear data to determine the preferred monitoring method and preferred wear data. A wear degree prediction model is constructed based on the preferred wear data to obtain the predicted wear degree. Based on the predicted wear degree, it is determined whether to send an alarm signal. By selecting a reasonable monitoring method, the accuracy of rotor wear degree assessment is improved, and corresponding effective measures are discovered in time and taken to improve motor performance and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 It is a flow chart of a rotor wear degree analysis and evaluation method in a preferred embodiment of the present application. DETAILED DESCRIPTION

[0068] The specific implementation methods of the present application are further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present application but are not intended to limit the scope of the present application.

[0069] In the description of the present application, it should be understood that the terms "center", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.

[0070] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0071] In the description of this application, it should be noted that, 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 it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0072] like Figure 1 As shown, a rotor wear degree analysis and evaluation method according to a preferred embodiment of the present application includes:

[0073] Step S101: Obtain historical monitoring logs of the rotor, and determine the monitoring method corresponding to each historical monitoring log and the historical wear data of the monitoring method;

[0074] Step S102: determining a preferred monitoring method and preferred wear data according to the correlation between the historical wear data of the monitoring method and the preset evaluation index and the historical monitoring information of the monitoring method;

[0075] Step S103: generating the wear degree of the rotor according to the preferred wear data and the historical evaluation values ​​of the corresponding preset evaluation indicators, and constructing a wear degree prediction model according to the preferred wear data, the multiple historical evaluation values ​​and the corresponding wear degrees;

[0076] Step S104: Acquire the real-time optimal wear data of the rotor, input the real-time optimal wear data into the wear degree prediction model, obtain the predicted wear degree, and determine whether to send an alarm instruction based on the predicted wear degree.

[0077] In this embodiment, the monitoring methods include vibration monitoring, acoustic emission technology monitoring, magnetic particle flaw detection monitoring, ultrasonic flaw detection monitoring and other monitoring methods. For example, the historical wear data of vibration monitoring includes historical vibration signals, abnormal frequency components, peak values ​​and waveforms of historical vibration signals, etc.

[0078] In some embodiments of the present application, determining a preferred monitoring method and preferred wear data includes:

[0079] Determine the monitoring method involved in each historical monitoring log according to the historical monitoring log of the rotor, classify the historical wear data obtained by the same monitoring method into the same category, and construct historical wear data sets of multiple categories;

[0080] Preset a plurality of preset evaluation indicators, and determine a historical evaluation value of the preset evaluation indicator corresponding to a historical node of each historical wear data according to the historical operation data in the historical operation log of the motor, wherein the preset evaluation indicator includes a performance evaluation indicator, an efficiency evaluation indicator, and a state evaluation indicator;

[0081] A focus period is set according to the mutation period of the historical evaluation value of each preset evaluation indicator, and a preset collection node is set at a preset time interval with the focus period as a reference time line;

[0082] Randomly select S1 pieces of historical wear data from each category of historical wear data set, and collect corresponding historical wear data according to multiple preset collection nodes of each reference timeline, and map them to the corresponding reference timeline to generate a data change curve of the selected historical wear data;

[0083] Analyze the data change curve to obtain the change characteristics of the historical wear data selected from the historical wear data sets of multiple categories, wherein the change characteristics include the change time node and the change value;

[0084] Generate the correlation degree between the selected historical wear data and the corresponding preset evaluation index according to the change time node and the change value;

[0085] The historical wear data with a correlation degree greater than a preset correlation degree threshold is set as the concerned wear data corresponding to the preset evaluation index;

[0086] The number of concerned wear data of the same category in each preset evaluation indicator is compared with the number of S1 screened out to obtain the concern ratio of the corresponding category, and the monitoring method of the corresponding category whose concern ratio is greater than the preset concern ratio threshold is set as the concerned monitoring method of the corresponding preset evaluation indicator;

[0087] Comprehensively analyze the monitoring methods of each preset evaluation indicator to determine the preferred monitoring method and preferred wear data.

[0088] In this embodiment, the performance evaluation index refers to the evaluation of the realization, stability and durability of the motor function, the efficiency evaluation index refers to the evaluation of the motor's usage efficiency and energy consumption, and the status evaluation index refers to the evaluation of the motor's operating status, including vibration, noise and temperature. The historical operating data associated with each preset evaluation indicator in the motor's historical operation log is compared with the preset standard operating data interval to obtain multiple historical operating data differences, and generate historical evaluation values ​​corresponding to the preset evaluation indicators.

[0089] In this embodiment, the mutation period refers to the 20 minutes before the time point when the historical evaluation value fluctuates and the fluctuation value is greater than the preset fluctuation value. The preset time interval is set in advance based on the length of the mutation period and the monitoring accuracy.

[0090] In this embodiment, the preferred monitoring method is screened out according to the degree of correlation, correlation coefficient and characteristic correlation coefficient between the historical wear data of different monitoring methods and the preset evaluation indicators, so as to select a monitoring method that improves the monitoring accuracy of the rotor wear degree, the monitoring timeliness and reduces the monitoring cost, so as to timely discover the rotor wear problem and take effective measures, laying the foundation for improving the performance, efficiency and operating status of the motor.

[0091] In some embodiments of the present application, determining the preferred monitoring method and the preferred wear data further includes:

[0092] Comprehensively analyze the monitoring methods of each preset evaluation indicator, eliminate the monitoring methods that do not appear in all the preset evaluation indicators at the same time, and determine the correlation coefficient between the corresponding monitoring method and each preset evaluation indicator according to the correlation degree between the historical wear data screened by the remaining monitoring methods and each preset evaluation indicator;

[0093] The correlation coefficients of all preset evaluation indicators corresponding to the same monitoring method are averaged to obtain the average correlation coefficient of the corresponding monitoring method;

[0094] The historical monitoring information of the monitoring method includes historical monitoring cost, historical monitoring accuracy and historical monitoring timeliness;

[0095] Quantify the historical monitoring cost, historical monitoring accuracy, and historical monitoring timeliness of the remaining monitoring methods of concern to obtain a first quantized value, a second quantized value, and a third quantized value, and compare them with the corresponding standard quantized values, and set a correction coefficient according to the comparison result;

[0096] The correlation coefficient mean of the corresponding monitoring method of interest is corrected according to the correction coefficient to obtain the characteristic correlation coefficient, and the remaining monitoring methods of interest are ranked according to the characteristic correlation coefficient. The first-ranked monitoring method of interest is set as the preferred monitoring method, and the historical wear data in the historical wear data set corresponding to the preferred monitoring method is set as the preferred wear data.

[0097] In this embodiment, the standard quantitative values ​​of the historical monitoring cost, the historical monitoring accuracy and the historical monitoring timeliness refer to the values ​​obtained by quantifying the monitoring cost, the monitoring accuracy and the monitoring timeliness of the corresponding monitoring method under normal monitoring conditions.

[0098] In this embodiment, the timeliness and accuracy of the historical wear data in evaluating the preset evaluation indicators are determined by the correlation between the historical wear data and the preset evaluation indicators, and the monitoring accuracy, timeliness and cost of the corresponding monitoring method in evaluating the degree of rotor wear are comprehensively evaluated in combination with the monitoring accuracy, timeliness and cost of the corresponding monitoring method in evaluating the degree of rotor wear, and the preferred monitoring method is selected to lay the foundation for the subsequent evaluation of the degree of rotor wear.

[0099] In some embodiments of the present application, the calculation formula of the degree of association is:

[0100] g=e -∣b1-b0∣ ×p;

[0101] Among them, g is the degree of association, b1 is the change time node, b0 is the mutation time node of the preset evaluation index, and p is the change value;

[0102] The calculation formula of the correlation coefficient is:

[0103]

[0104] Wherein, t0 is the correlation coefficient, S0 is the number of historical wear data with a correlation degree greater than a preset correlation degree threshold, and gi is the i-th historical wear data with a correlation degree greater than the preset correlation degree threshold;

[0105] The calculation formula of the characteristic correlation coefficient is:

[0106]

[0107] Where T is the characteristic correlation coefficient, is the mean of the correlation coefficient, n is the total number of preset evaluation indicators, t0j is the correlation coefficient of the jth preset evaluation indicator, z×[(a1-a0)×c1+(m1-m0)×c2+(k1-k0)×c3] is the correction coefficient, z is the correction conversion coefficient, a1 is the first quantitative value, a0 is the standard quantitative value corresponding to the first quantitative value, c1 is the weight coefficient corresponding to the monitoring cost, m1 is the second quantitative value, m0 is the standard quantitative value corresponding to the second quantitative value, c2 is the weight coefficient corresponding to the monitoring accuracy, k1 is the third quantitative value, k0 is the standard quantitative value corresponding to the third quantitative value, and c3 is the weight coefficient corresponding to the monitoring timeliness.

[0108] In this embodiment, the correction conversion coefficient refers to converting the comparison result of the first quantization value and the second quantization value with the corresponding standard quantization value into a corresponding preset correction coefficient, and the preset correction coefficient is set in advance and is determined based on multiple preset comparison results.

[0109] In some embodiments of the present application, generating the wear degree of the rotor according to the preferred wear data and the historical evaluation value of the corresponding preset evaluation index includes:

[0110] Generate a comprehensive evaluation value according to the preferred monitoring method and the historical evaluation value of each preset evaluation indicator of the historical monitoring log corresponding to the preferred wear data and the weight coefficient of the corresponding preset evaluation indicator;

[0111] The wear degree of the rotor is set according to the relationship between the comprehensive evaluation value and the preset evaluation value interval;

[0112] Presetting a first preset evaluation value interval, a second preset evaluation value interval, a third preset evaluation value interval and a fourth preset evaluation value interval;

[0113] When the comprehensive evaluation value is within the first preset evaluation value interval, the wear degree of the rotor is set to a fourth preset wear degree f4;

[0114] When the comprehensive evaluation value is within the second preset evaluation value interval, the wear degree of the rotor is set to a third preset wear degree f3;

[0115] When the comprehensive evaluation value is within the third preset evaluation value interval, the wear degree of the rotor is set to the second preset wear degree f2;

[0116] When the comprehensive evaluation value is within the fourth preset evaluation value interval, the wear degree of the rotor is set to the first preset wear degree f1;

[0117] Among them, f1<f2<f3<f4.

[0118] In this embodiment, the first preset evaluation value interval < the second preset evaluation value interval < the third preset evaluation value interval < the fourth preset evaluation value interval.

[0119] In this embodiment, the historical evaluation values ​​of the preset evaluation indicators corresponding to the preferred wear data in each historical monitoring log of the priority monitoring method are used to generate a comprehensive evaluation value, which is calculated based on the historical evaluation values ​​of multiple preset evaluation indicators and the weight coefficients of the corresponding preset evaluation indicators.

[0120] In this embodiment, the larger the comprehensive evaluation value is, the better the comprehensive evaluation of the preset evaluation indicators is, that is, the better the comprehensive evaluation of the operating status, performance and efficiency of the motor is, the smaller the degree of wear of the rotor is; and the smaller the comprehensive evaluation value is, the greater the degree of wear of the rotor is.

[0121] In some embodiments of the present application, a wear degree prediction model is constructed according to the preferred wear data, multiple historical evaluation values ​​and corresponding wear degrees, including:

[0122] Using the preferred wear data as first input data and using the historical evaluation value corresponding to each preset evaluation index as first output data;

[0123] Generate a comprehensive evaluation value according to the historical evaluation value corresponding to each preset evaluation indicator, use the comprehensive evaluation value as the second input data, and use the wear degree corresponding to the rotor as the second output data;

[0124] Performing a first neural network training according to a plurality of first input data and corresponding first output data, and performing a second neural network training according to a plurality of second input data and corresponding second output data;

[0125] The trained neural network model is set as the wear degree prediction model.

[0126] In this embodiment, the first neural network training refers to the process of training the neural network with data through preferred wear data and corresponding historical evaluation values ​​of different preset evaluation indicators, so that the neural network model can obtain the evaluation value of each preset evaluation indicator for different preferred wear data. The second neural network training refers to the process of training data through the comprehensive evaluation value of multiple preset evaluation indicators and the corresponding wear degree of the rotor. The evaluation values ​​of multiple preset evaluation indicators and the wear degree of the rotor are obtained through the first neural network training and the second neural network training, so as to improve the analysis accuracy of the rotor wear degree, timely discover and take corresponding effective measures, and improve the motor performance and efficiency.

[0127] In some embodiments of the present application, real-time preferred wear data of the rotor is obtained, and the real-time preferred wear data is input into a wear degree prediction model to obtain a predicted wear degree, including:

[0128] Acquire real-time wear data of the rotor according to the preferred monitoring method, input the real-time wear data into a wear degree prediction model, and obtain predicted evaluation values ​​of multiple preset evaluation indicators;

[0129] Generate a predicted comprehensive evaluation value based on the predicted evaluation values ​​of multiple preset evaluation indicators and the weight coefficients of the corresponding preset evaluation indicators;

[0130] The predicted comprehensive evaluation value is input into the wear degree prediction model to obtain the predicted wear degree.

[0131] In some embodiments of the present application, before determining whether to send an alarm instruction according to the predicted wear degree, the method further includes:

[0132] Acquire real-time preferred wear data for a preset period, and construct a real-time preferred wear data curve for the preset period based on the real-time preferred wear data for the preset period;

[0133] Constructing a plurality of historical preferred wear data curve segments according to the historical preferred wear data of the preferred monitoring method, and selecting the historical wear data curve segments that match the curve characteristics of the real-time preferred wear data curve;

[0134] According to the time interval between the preset time period and the future time period, historical wear data of the time period corresponding to the historical wear data curve segment matching the curve characteristics of the real-time preferred wear data curve is obtained, and set as the first predicted wear data of the real-time preferred wear data in the future time period;

[0135] Inputting the first predicted wear data of multiple preset moments in the future period into the wear degree prediction model, obtaining the predicted wear degree at each preset moment in the future period, and constructing a predicted wear degree change curve;

[0136] A compensation coefficient is generated based on the change value at each preset moment in the predicted wear degree change curve and the degree of fluctuation in the future time period, the predicted wear degree is corrected based on the compensation coefficient, and whether to send an alarm signal is determined based on the corrected predicted wear degree.

[0137] In this embodiment, the curve characteristics include the curve change trend and the curve fluctuation value. The curve characteristics of the real-time preferred wear data curve are compared with the curve characteristics of the historical preferred wear data curve segment, and the historical preferred wear data curve segment with the greatest similarity is set as the historical wear data curve segment that matches the curve characteristics of the real-time preferred wear data curve.

[0138] In some embodiments of the present application, the calculation formula of the compensation coefficient is:

[0139]

[0140] Among them, r is the compensation coefficient, r0 is the compensation conversion coefficient, σ 2 is the variance of the predicted wear degree at multiple preset moments in the future period, z is the total number of preset moments in the future period, and F c is the predicted wear degree at the cth preset moment, F c+1 is the predicted wear degree at the c+1th preset moment.

[0141] In this embodiment, the corrected predicted wear degree is fi*r, i=1,2,3,4.

[0142] In this embodiment, the compensation conversion coefficient is set in advance. The change value calculated at each preset moment and the degree of fluctuation in the future time period are converted into a preset compensation coefficient through the compensation conversion coefficient. The preset compensation coefficient is set in advance, and the value range of the compensation coefficient is (0.75, 1.25).

[0143] In this embodiment, a predicted wear degree change curve is constructed to obtain the change trend and fluctuation size of the predicted wear degree in future time periods, and a corresponding compensation coefficient is set to correct the predicted wear degree obtained from real-time preferred wear data, thereby improving the accuracy of the predicted wear degree, timely discovering and taking corresponding effective measures, and improving motor performance and efficiency.

[0144] In some embodiments of the present application, determining whether to send an alarm signal according to the corrected predicted wear degree includes:

[0145] Pre-set wear level threshold;

[0146] The corrected predicted wear degree is compared with the wear degree threshold. If the corrected predicted wear degree is less than the wear degree threshold, no alarm instruction is sent. If the corrected predicted wear degree is greater than the wear degree threshold, an alarm instruction is sent.

[0147] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and substitutions can be made without departing from the technical principles of the present application. These improvements and substitutions should also be regarded as the scope of protection of the present application.

Claims

1. A rotor wear degree analysis and evaluation method, characterized in that: include: Obtain historical monitoring logs of the rotor, and determine the monitoring method corresponding to each historical monitoring log and the historical wear data of the monitoring method; Determine the preferred monitoring method and preferred wear data according to the correlation between the historical wear data of the monitoring method and the preset evaluation index and the historical monitoring information of the monitoring method; Generate the wear degree of the rotor according to the preferred wear data and the historical evaluation values ​​of the corresponding preset evaluation indicators, and build a wear degree prediction model according to the preferred wear data, multiple historical evaluation values ​​and the corresponding wear degrees; The real-time optimal wear data of the rotor is obtained, and the real-time optimal wear data is input into the wear degree prediction model to obtain the predicted wear degree, and it is determined whether to send an alarm instruction according to the predicted wear degree.

2. The rotor wear degree analysis and evaluation method according to claim 1, characterized in that: Determine the preferred monitoring method and preferred wear data, including: Determine the monitoring method involved in each historical monitoring log according to the historical monitoring log of the rotor, classify the historical wear data obtained by the same monitoring method into the same category, and construct historical wear data sets of multiple categories; Preset a plurality of preset evaluation indicators, and determine a historical evaluation value of the preset evaluation indicator corresponding to a historical node of each historical wear data according to the historical operation data in the historical operation log of the motor, wherein the preset evaluation indicator includes a performance evaluation indicator, an efficiency evaluation indicator, and a state evaluation indicator; A focus period is set according to the mutation period of the historical evaluation value of each preset evaluation indicator, and a preset collection node is set at a preset time interval with the focus period as a reference time line; Randomly select S1 pieces of historical wear data from each category of historical wear data set, and collect corresponding historical wear data according to multiple preset collection nodes of each reference timeline, and map them to the corresponding reference timeline to generate a data change curve of the selected historical wear data; Analyze the data change curve to obtain the change characteristics of the historical wear data selected from the historical wear data sets of multiple categories, wherein the change characteristics include the change time node and the change value; Generate the correlation degree between the selected historical wear data and the corresponding preset evaluation index according to the change time node and the change value; The historical wear data with a correlation degree greater than a preset correlation degree threshold is set as the concerned wear data corresponding to the preset evaluation index; The number of concerned wear data of the same category in each preset evaluation indicator is compared with the number of S1 screened out to obtain the concern ratio of the corresponding category, and the monitoring method of the corresponding category whose concern ratio is greater than the preset concern ratio threshold is set as the concerned monitoring method of the corresponding preset evaluation indicator; Comprehensively analyze the monitoring methods of each preset evaluation indicator to determine the preferred monitoring method and preferred wear data.

3. The rotor wear degree analysis and evaluation method according to claim 2, characterized in that: Determine the preferred monitoring method and preferred wear data, including: Comprehensively analyze the monitoring methods of each preset evaluation indicator, eliminate the monitoring methods that do not appear in all the preset evaluation indicators at the same time, and determine the correlation coefficient between the corresponding monitoring method and each preset evaluation indicator according to the correlation degree between the historical wear data screened by the remaining monitoring methods and each preset evaluation indicator; The correlation coefficients of all preset evaluation indicators corresponding to the same monitoring method are averaged to obtain the average correlation coefficient of the corresponding monitoring method; The historical monitoring information of the monitoring method includes historical monitoring cost, historical monitoring accuracy and historical monitoring timeliness; Quantify the historical monitoring cost, historical monitoring accuracy, and historical monitoring timeliness of the remaining monitoring methods of concern to obtain a first quantized value, a second quantized value, and a third quantized value, and compare them with the corresponding standard quantized values, and set a correction coefficient according to the comparison result; The correlation coefficient mean of the corresponding monitoring method of interest is corrected according to the correction coefficient to obtain the characteristic correlation coefficient, and the remaining monitoring methods of interest are ranked according to the characteristic correlation coefficient. The first-ranked monitoring method of interest is set as the preferred monitoring method, and the historical wear data in the historical wear data set corresponding to the preferred monitoring method is set as the preferred wear data.

4. The rotor wear degree analysis and evaluation method according to claim 3, characterized in that: The calculation formula of the degree of association is: g=e -∣b1-b0∣ ×p; Among them, g is the degree of association, b1 is the change time node, b0 is the mutation time node of the preset evaluation index, and p is the change value; The calculation formula of the correlation coefficient is: Wherein, t0 is the correlation coefficient, S0 is the number of historical wear data with a correlation degree greater than a preset correlation degree threshold, and gi is the i-th historical wear data with a correlation degree greater than the preset correlation degree threshold; The calculation formula of the characteristic correlation coefficient is: Where T is the characteristic correlation coefficient, is the mean of the correlation coefficient, n is the total number of preset evaluation indicators, t0j is the correlation coefficient of the jth preset evaluation indicator, z×[(a1-a0)×c1+(m1-m0)×c2+(k1-k0)×c3] is the correction coefficient, z is the correction conversion coefficient, a1 is the first quantitative value, a0 is the standard quantitative value corresponding to the first quantitative value, c1 is the weight coefficient corresponding to the monitoring cost, m1 is the second quantitative value, m0 is the standard quantitative value corresponding to the second quantitative value, c2 is the weight coefficient corresponding to the monitoring accuracy, k1 is the third quantitative value, k0 is the standard quantitative value corresponding to the third quantitative value, and c3 is the weight coefficient corresponding to the monitoring timeliness.

5. The rotor wear degree analysis and evaluation method according to claim 4, characterized in that: The wear degree of the rotor is generated according to the preferred wear data and the historical evaluation value of the corresponding preset evaluation index, including: Generate a comprehensive evaluation value according to the preferred monitoring method and the historical evaluation value of each preset evaluation indicator of the historical monitoring log corresponding to the preferred wear data and the weight coefficient of the corresponding preset evaluation indicator; The wear degree of the rotor is set according to the relationship between the comprehensive evaluation value and the preset evaluation value interval; Presetting a first preset evaluation value interval, a second preset evaluation value interval, a third preset evaluation value interval and a fourth preset evaluation value interval; When the comprehensive evaluation value is within the first preset evaluation value interval, the wear degree of the rotor is set to a fourth preset wear degree f4; When the comprehensive evaluation value is within the second preset evaluation value interval, the wear degree of the rotor is set to a third preset wear degree f3; When the comprehensive evaluation value is within the third preset evaluation value interval, the wear degree of the rotor is set to the second preset wear degree f2; When the comprehensive evaluation value is within the fourth preset evaluation value interval, the wear degree of the rotor is set to the first preset wear degree f1; Among them, f1<f2<f3<f4.

6. The rotor wear degree analysis and evaluation method according to claim 5, characterized in that: A wear degree prediction model is constructed based on the preferred wear data, multiple historical evaluation values ​​and the corresponding wear degrees, including: Using the preferred wear data as first input data and using the historical evaluation value corresponding to each preset evaluation index as first output data; Generate a comprehensive evaluation value according to the historical evaluation value corresponding to each preset evaluation indicator, use the comprehensive evaluation value as the second input data, and use the wear degree corresponding to the rotor as the second output data; Performing a first neural network training according to a plurality of first input data and corresponding first output data, and performing a second neural network training according to a plurality of second input data and corresponding second output data; The trained neural network model is set as the wear degree prediction model.

7. The rotor wear degree analysis and evaluation method according to claim 6, characterized in that: Obtain the real-time optimal wear data of the rotor, input the real-time optimal wear data into the wear degree prediction model, and obtain the predicted wear degree, including: Acquire real-time wear data of the rotor according to the preferred monitoring method, input the real-time wear data into a wear degree prediction model, and obtain predicted evaluation values ​​of multiple preset evaluation indicators; Generate a predicted comprehensive evaluation value based on the predicted evaluation values ​​of multiple preset evaluation indicators and the weight coefficients of the corresponding preset evaluation indicators; The predicted comprehensive evaluation value is input into the wear degree prediction model to obtain the predicted wear degree.

8. The rotor wear degree analysis and evaluation method according to claim 7, characterized in that: Before determining whether to send an alarm instruction based on the predicted wear degree, the following is also included: Acquire real-time preferred wear data for a preset period, and construct a real-time preferred wear data curve for the preset period based on the real-time preferred wear data for the preset period; Constructing a plurality of historical preferred wear data curve segments according to the historical preferred wear data of the preferred monitoring method, and selecting the historical wear data curve segments that match the curve characteristics of the real-time preferred wear data curve; According to the time interval between the preset time period and the future time period, historical wear data of the time period corresponding to the historical wear data curve segment matching the curve characteristics of the real-time preferred wear data curve is obtained, and set as the first predicted wear data of the real-time preferred wear data in the future time period; Inputting the first predicted wear data of multiple preset moments in the future period into the wear degree prediction model, obtaining the predicted wear degree at each preset moment in the future period, and constructing a predicted wear degree change curve; A compensation coefficient is generated based on the change value at each preset moment in the predicted wear degree change curve and the degree of fluctuation in the future time period, the predicted wear degree is corrected based on the compensation coefficient, and whether to send an alarm signal is determined based on the corrected predicted wear degree.

9. The rotor wear degree analysis and evaluation method according to claim 8, characterized in that: The calculation formula of the compensation coefficient is: Among them, r is the compensation coefficient, r0 is the compensation conversion coefficient, σ 2 is the variance of the predicted wear degree at multiple preset moments in the future period, z is the total number of preset moments in the future period, and F c is the predicted wear degree at the cth preset moment, F c+1 is the predicted wear degree at the c+1th preset moment.

10. The rotor wear degree analysis and evaluation method according to claim 9, characterized in that: Determine whether to send an alarm signal based on the corrected predicted wear level, including: Pre-set wear level threshold; The corrected predicted wear degree is compared with the wear degree threshold. If the corrected predicted wear degree is less than the wear degree threshold, no alarm instruction is sent. If the corrected predicted wear degree is greater than the wear degree threshold, an alarm instruction is sent.

Images