Method for predicting wear of friction plate of fan coupler

Abstract

The invention discloses a fan coupling friction plate wear prediction method, which comprises the following steps: carrying out a loading test on a coupling, and obtaining a temperature reference value when the coupling does not slip, and a rotation degree and a temperature when the coupling slips; establishing a temperature curve graph based on the temperature data; after the temperature curve graph is corrected and fitted, feature extraction is carried out to form a temperature feature value set; acquiring rotating speeds corresponding to the temperature characteristic value set to form a rotating speed characteristic value set; processing the temperature characteristic value set to obtain a critical temperature early warning value; inputting the temperature characteristic value set, the rotating speed characteristic value set and the critical temperature early warning value into a neural network for training; obtaining a target temperature characteristic value set and a target rotating speed characteristic value set; inputting the target temperature characteristic value set and the target rotating speed characteristic value set into a neural network model for prediction; comparing the target temperature characteristic value set with the target critical temperature early warning value; if the target critical temperature early warning value is larger than or equal to the data in the target temperature characteristic value set, a worker is informed to maintain or replace the friction plate.

Description

Technical Field

[0001] The invention relates to the technical field of mechanical equipment fault monitoring, in particular to a method for predicting wear of a friction plate of a fan coupling. Background Art

[0002] The wind power coupling is used to connect the rotor of the wind turbine and the rotor of the generator, and is used to transmit the rotational torque generated by the wind turbine to the generator. It is a key component for converting wind energy into electrical energy. When the wind turbine load is too large, the slippage of the coupling can prevent excessive torque from being transmitted to the generator, thereby avoiding damage to the generator. As the use time increases, the wear of the friction plate increases, and eventually the friction plate will be worn to the point of being unusable and require maintenance or replacement. However, since the coupling is sealed in the fan cover and squeezed against the friction flange, the existing technical means cannot monitor the wear of the friction plate in real time, and staff can only be dispatched regularly to disassemble the fan cover and coupling to detect the wear of the friction plate. This method not only wastes manpower costs, but also may cause damage to the coupling or other components during the disassembly and reinstallation of the cover and coupling, further affecting the normal operation of the wind turbine equipment.

[0003] Therefore, designing a method that can monitor the degree of friction plate wear in real time has become an urgent problem to be solved. Summary of the invention

[0004] In order to at least overcome the above-mentioned deficiencies in the prior art, the purpose of the present application is to provide a method for predicting the wear of the friction plate of a fan coupling.

[0005] The present application provides a method for predicting wear of a fan coupling friction plate, comprising: S1: Perform a rotation loading test at different speeds on the coupling to be tested under different test environment temperature conditions, obtain the normal working temperature reference value of the friction plate and the friction flange when there is no slippage, and obtain the rotation speed and temperature data of the friction plate and the friction flange in different slipping periods, and establish a mapping coordinate set of the rotation speed and temperature data; S2: Establish a temperature curve graph based on the temperature data; S3: performing fitting and correction processing on the temperature curve according to the normal working temperature reference value, and performing feature extraction on the fitted and corrected temperature curve to form a temperature feature value set; S4: updating the temperature characteristic value set to the mapping coordinate set, and obtaining the rotation speed corresponding to the temperature characteristic value to form a rotation speed characteristic value set; S5: extracting the temperature characteristic values ​​under the working condition that the wear degree of the friction plate reaches the condition that maintenance or replacement is required from the temperature characteristic value set to form a wear temperature characteristic value set, and calculating the wear temperature characteristic value set to obtain a critical temperature warning value, which is the critical temperature warning value when the wear degree of the friction plate reaches the critical temperature that requires maintenance or replacement; S6: assigning the critical temperature warning value, the rotation speed characteristic value set and the temperature characteristic value set as sample data to the neural network model for prediction training, wherein the output data of the neural network model is the critical temperature warning value, and the input data is the rotation speed characteristic value set and the temperature characteristic value set; S7: Establish an early warning notification model; S8: Obtain a target temperature characteristic value set of the friction plate and the friction flange of the coupling in different slipping periods under actual working conditions, and a target speed characteristic value set corresponding to the target temperature characteristic value set; S9: assigning the target speed characteristic value set and the target temperature characteristic value set to the neural network model for prediction, and the neural network model predicts the corresponding target critical temperature warning value; S10: assigning the target critical temperature warning value and the target temperature characteristic value set to the warning notification model for comparison; S11: When at least one of the data in the target temperature characteristic value set is greater than or equal to the critical temperature warning value, the warning switch state is triggered to be on, and the staff is notified to repair or replace the friction plate.

[0006] When this embodiment is implemented, as the use time of the fan coupling friction plate increases, the abrasive particles evenly distributed on its surface will gradually decrease, and even cracks will appear on the friction plate surface in some specific cases, causing the temperature of the coupling to increase when it slips. Because of the reduction of abrasive particles, the contact area between the friction plate and the friction flange will increase relatively, resulting in an increase in pressure per unit area, which in turn increases the friction force, and the heat generated thereby also increases. Secondly, in some specific cases, the surface of the worn friction plate may form irregular shapes and tiny cracks. These defects will cause local stress concentration, resulting in uneven contact of the friction surface, resulting in fluctuations in friction force and concentration of heat, further increasing the heat generated by friction. Based on the above phenomenon, the inventor of the present application applied a rotational loading test to the coupling. When the rotational speed increased, the coupling would slip. Based on a large number of rotational loading experiments, the inventor of the present application found that the heat generated by the friction plate during the slipping process can be used to characterize the degree of wear of the friction plate. When the degree of wear of the friction plate requires maintenance or replacement, the wear temperature of the friction plate will increase to the critical temperature warning value as the rotational speed increases. Therefore, the purpose of this application is to find the critical temperature warning value when the degree of wear of the friction plate reaches the critical temperature warning value that requires maintenance or replacement. When the highest temperature generated by friction reaches or exceeds the critical temperature warning value, the staff is notified to maintain or replace the friction plate.

[0007] In the implementation of this embodiment, since the fan under actual working conditions is under different ambient temperature conditions, different critical temperature warning values ​​correspond to different ambient temperatures. Based on this, the inventor of the present application sets different test ambient temperatures in the laboratory to simulate different ambient temperatures, and performs rotation loading tests at different speeds on the coupling to be tested under different ambient temperature conditions. During the test, the normal working temperature reference value of the friction plate and the friction flange when there is no slip is obtained. In this embodiment, the working temperature reference value is the average value within the normal temperature range of the coupling when there is no slip, and the rotation speed and temperature data of the friction plate and the friction flange in different slip periods are obtained, and a mapping coordinate set of the rotation speed and temperature data is established. In this embodiment, one coordinate in the mapping coordinate set includes the rotation speed of the friction plate at the current moment, the rotation speed of the friction flange at the current moment, and the temperature data generated when the friction flange and the friction plate slip at the current moment.

[0008] When this embodiment is implemented, the temperature curve graph is fitted and corrected, and then feature extraction is performed to form a temperature feature value set. The temperature feature values ​​in the temperature feature value set are the peak values ​​of each curve segment in the temperature curve graph. Then, the speed values ​​corresponding to each temperature feature value are extracted to form a speed feature value set. The temperature feature values ​​under the working conditions where the wear degree of the friction plate in the temperature feature value set reaches the working conditions requiring maintenance or replacement are extracted to form a wear temperature feature value set. The temperature feature values ​​in the wear temperature feature value set are calculated to obtain the critical temperature warning value.

[0009] When this embodiment is implemented, the critical temperature warning value, the speed characteristic value set and the temperature characteristic value set are assigned as sample data to the neural network model for prediction training. After the neural network model training is completed, the target speed characteristic value set and the target temperature characteristic value set in the actual working conditions are assigned to the neural network model for prediction. The neural network model predicts the target critical temperature warning value, and compares the target critical temperature warning value with the data in the target temperature characteristic value set. When there is data greater than or equal to the target critical temperature warning value in the target temperature characteristic value set, the switch of the warning notification model is triggered to turn on, and a warning text message is sent to notify the staff to repair or replace the friction plate.

[0010] In a possible implementation, in step S2, the step of establishing a temperature curve graph includes: Use Kalman filter to process temperature data and filter out noise in temperature data; Use a smoothing algorithm to smooth the temperature data and extract the data trend of the temperature data within the slip period; Based on the temperature data and data trends, use the Matplotlib tool to draw a temperature curve graph.

[0011] In a possible implementation, the specific steps of step S3 include: S301: Using a fitting equation to fit the temperature curves in different slipping periods under the same ambient temperature condition, a plurality of fitting temperature curves in each slipping period are obtained, and the starting temperature and the ending temperature of each fitting temperature curve are both normal working temperature reference values. The fitting equation is:

[0012] In the formula, y is the temperature generated during the slip period, x is the time corresponding to the temperature, and is the fitting parameter value; S302: if there is an overlapping area between the current fitting temperature curve and the previous fitting temperature curve, the overlapping area is divided into a front overlapping area and a rear overlapping area by using the temperature value corresponding to the intersection of the current fitting temperature curve and the previous fitting temperature curve as a dividing line; S303: Eliminate the overlapping area before, and use the correction equation to correct the current fitting temperature curve to obtain the corrected and accurate current temperature curve; S304: performing a sliding process on the current temperature curve so that the starting temperature position of the current temperature curve slides to the ending temperature position of the previous fitting temperature curve, thereby obtaining a complete and continuous temperature curve graph; S305: Establishing a working temperature comparison table, and storing the ambient temperature and the normal working temperature reference value in the working temperature comparison table; S306: Extracting features from the fitted and corrected temperature curve to form a set of temperature feature values.

[0013] In one possible implementation, the correction equation is:

[0014] In the formula, is the corrected current fitting temperature, is the normal operating temperature reference value, is the time corresponding to the intersection of the current fitting temperature curve and the previous fitting temperature curve, is the time when the temperature of the previous fitting temperature curve drops to the temperature reference value, is the time when the temperature of the current fitting temperature curve drops to the temperature reference value. is the temperature function of the current fitting temperature curve, is the temperature function of the previous fitting temperature curve.

[0015] In a possible implementation, the step of obtaining a target temperature characteristic value set in step S8 includes: S801: Continuously record target temperature data of the friction plate and the friction flange by setting a temperature sensor in a wrapping cover wrapping the coupling; S802: Perform frequency statistical analysis on the target temperature data to count the frequency of occurrence of each target temperature data; S803: taking the target temperature data with the highest frequency as the target ambient temperature; S804: extracting a normal operating temperature reference value corresponding to the target ambient temperature from the operating temperature comparison table as a target operating temperature reference value; S805: Establishing a target temperature curve graph based on the target temperature data; S806: performing fitting correction processing on the target temperature curve according to the target operating temperature reference value, and performing feature extraction on the fitted and corrected target temperature curve to form a target temperature feature value set.

[0016] In a possible implementation, the formula for calculating the critical temperature warning value in step S5 is:

[0017] In the formula, is the critical temperature warning value, is the mean of the wear temperature characteristic value, K is the temperature warning safety factor, is the standard deviation of the wear temperature characteristic value.

[0018] In a possible implementation, the neural network model in step S6 is a temperature DNN neural network model.

[0019] In one possible implementation, the temperature DNN neural network model training steps include: The speed characteristic value set and the temperature characteristic value set are used as input feature X, and the critical temperature warning value is used as label Y, and are input into the input layer of the temperature DNN neural network model for normalization and standardization. The hidden layer uses an activation function to obtain the output value sent from the input layer to each hidden layer neuron; The output layer calculates the predicted critical temperature warning value through the activation function; Use the loss function to evaluate the predicted critical temperature warning value, and use the back propagation algorithm to update the weight parameters of the temperature DNN neural network model after the evaluation is qualified; The output layer outputs the qualified critical temperature warning value as the final result.

[0020] In a possible implementation, the specific steps of step S11 include: The warning notification model traverses all values ​​in the target temperature characteristic value set to detect whether at least one value is greater than or equal to the critical temperature warning value; When the detection conditions are met, the traversal operation is stopped, and the warning notification model triggers the warning notification switch state to be turned on; The warning notification model calls the SMS service interface to send warning information to the staff; After the warning information is sent, the warning notification model writes the warning time, content, target temperature feature value set and critical temperature warning value into the log.

[0021] The present invention provides a method for predicting the wear of a friction plate of a fan coupling. Through the above technical scheme, the rotational speed and temperature data of the friction plate and the friction flange of the fan coupling in actual working conditions can be monitored in real time. A target temperature characteristic value set is formed by processing the temperature data. After the target temperature characteristic value set and the expected corresponding target rotational speed characteristic value set are assigned to a neural network model, the neural network model predicts a target critical temperature warning value of the friction plate. By comparing the target critical temperature warning value with the target temperature characteristic value set, the current degree of wear of the friction plate can be determined, thereby avoiding the need to waste human resources to frequently and regularly dispatch staff to disassemble the fan cover and coupling to detect the degree of wear of the friction plate. At the same time, it also avoids the possibility of damage to the coupling or other components due to frequent disassembly and installation of the cover and coupling. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings: Figure 1 is a flow chart of the prediction method of the present invention; Figure 2 This is a flow chart of the present invention for fitting and correcting the temperature curve and performing feature extraction; Figure 3 A flow chart of obtaining a target temperature characteristic value set for the present invention; Figure 4 This is a temperature curve diagram when the present invention is not fitted and corrected; Figure 5 It is the temperature curve diagram after fitting of the present invention; Figure 6 A temperature curve graph after dividing the temperature curve graph fitted by the present invention into regions; Figure 7 This is the temperature curve diagram after correction of the present invention; Figure 8 It is a temperature curve diagram after the corrected temperature curve of the present invention is subjected to slip processing. DETAILED DESCRIPTION

[0023] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0024] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0025] In addition, 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. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0026] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like 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 an indirect connection 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 the present invention can be understood according to specific circumstances.

[0027] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may include that the first and second features are in direct contact, or may include that the first and second features are not in direct contact but are in contact through another feature between them. Moreover, a first feature being "above", "above" and "above" a second feature includes that the first feature is directly above and obliquely above the second feature, or simply indicates that the first feature is higher in level than the second feature. A first feature being "below", "below" and "below" a second feature includes that the first feature is directly below and obliquely below the second feature, or simply indicates that the first feature is lower in level than the second feature.

[0028] In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the fact that ordinary technicians in the field can implement it. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of the present invention.

[0029] In the following description, suffixes such as "module", "component", "assembly" or "unit" are used only to facilitate the description of the present invention and have no specific meanings. Therefore, they can be used in combination.

[0030] The present invention will be further described in detail below through specific implementation modes in conjunction with the accompanying drawings.

[0031] Please refer to Figure 1 , which is a flow chart of a method for predicting wear of a friction plate of a fan coupling provided in an embodiment of the present invention. Furthermore, the method for predicting wear of a friction plate of a fan coupling may specifically include the contents described in the following steps S1 to S11.

[0032] S1: performing a rotation loading test at different speeds on the coupling to be tested under different test environment temperature conditions, obtaining a normal working temperature reference value of the friction plate and the friction flange when there is no slippage, and obtaining the rotation speed and temperature data of the friction plate and the friction flange in different slippage periods, and establishing a mapping coordinate set of the rotation speed and the temperature data; S2: Establishing a temperature curve graph based on the temperature data; S3: performing fitting correction processing on the temperature curve according to the normal working temperature reference value, and performing feature extraction on the fitted and corrected temperature curve to form a temperature feature value set; S4: updating the temperature characteristic value set to the mapping coordinate set, and obtaining the rotation speed corresponding to the temperature characteristic value to form a rotation speed characteristic value set; S5: extracting the temperature characteristic values ​​under the working condition that the wear degree of the friction plate reaches the condition that maintenance or replacement is required from the temperature characteristic value set to form a wear temperature characteristic value set, and calculating the wear temperature characteristic value set to obtain a critical temperature warning value, wherein the critical temperature warning value is the critical temperature when the wear degree of the friction plate reaches the condition that maintenance or replacement is required; S6: assigning the critical temperature warning value, the rotation speed characteristic value set and the temperature characteristic value set as sample data to a neural network model for prediction training, wherein the output data of the neural network model is the critical temperature warning value, and the input data is the rotation speed characteristic value set and the temperature characteristic value set; S7: Establish an early warning notification model; S8: Obtaining a target temperature characteristic value set of the friction plate and the friction flange of the coupling in different slipping periods under actual working conditions, and a target speed characteristic value set corresponding to the target temperature characteristic value set; S9: assigning the target speed characteristic value set and the target temperature characteristic value set to the neural network model for prediction, and the neural network model predicts a corresponding target critical temperature warning value; S10: assigning the target critical temperature warning value and the target temperature characteristic value set to the warning notification model for comparison; S11: When at least one of the data in the target temperature characteristic value set is greater than or equal to the critical temperature warning value, the warning switch state is triggered to be on, and the staff is notified to repair or replace the friction plate.

[0033] When implementing this embodiment, the inventor prepared three groups of friction plates, each group of friction plates has several states, the first group of friction plates is brand new and unused friction plates, the second group of friction plates is that the friction plates have experienced a certain degree of wear but have not reached the conditions for maintenance or replacement, and the third group is that the friction plates have reached the conditions for maintenance or replacement; the friction plates in the three states are respectively installed on the coupling, and the coupling to be tested is subjected to rotation loading tests at different speeds under different ambient temperature conditions; during the test, the normal operating temperature reference value of each group of friction plates and friction flanges when not slipping is obtained, and the rotation speed and temperature data of each group of friction plates and friction flanges in different slipping periods are obtained, and the rotation speed and temperature data are assigned to the speed-temperature mapping coordinate system to form a mapping coordinate set, and then the temperature curve is drawn based on the temperature data using the Matplotlib tool.

[0034] In this embodiment, after fitting and correcting the temperature curve graph, an accurate temperature curve graph is obtained, and then feature extraction is performed on the accurate temperature curve graph to form a temperature characteristic value set, where the temperature characteristic values ​​in the temperature characteristic value set are the peak values ​​of each curve segment in the temperature curve graph; each temperature characteristic value is updated to a mapping coordinate set, and then the speed value corresponding to each temperature characteristic value in the mapping coordinate set is extracted to form a speed characteristic value set; the temperature characteristic value under the working condition that the wear degree of the friction plate in the temperature characteristic value set reaches the working condition that requires maintenance or replacement is extracted to form a wear temperature characteristic value set, and the temperature characteristic values ​​in the wear temperature characteristic value set are calculated to obtain a critical temperature warning value; the temperature characteristic value set, the speed characteristic value set and the critical temperature warning value are assigned to the neural network model for training.

[0035] When this embodiment is implemented, the target speed characteristic value set and the target temperature characteristic value set in the actual working conditions are assigned to the neural network model for prediction. The neural network model predicts the target critical temperature warning value, and the target critical temperature warning value is compared with the data in the target temperature characteristic value set. When there is data greater than or equal to the target critical temperature warning value in the target temperature characteristic value set, the switch of the early warning notification model is triggered to turn on, and a warning text message is sent to notify the staff to repair or replace the friction plate.

[0036] In a possible implementation, the step of establishing a temperature curve graph includes: Processing the temperature data using a Kalman filter to filter out noise in the temperature data; Using a smoothing algorithm to smooth the temperature data and extract the data trend of the temperature data within the slip period; A temperature curve graph is drawn using Matplotlib tool according to the temperature data and the data trend.

[0037] When this embodiment is implemented, the use of Kalman filter to process temperature data can effectively filter out the noise in the sensor data and provide a more accurate temperature reading. Secondly, the smoothing algorithm helps to extract the temperature data trend during the slip period, making the temperature change at the critical moment clearer, thus facilitating analysis and decision-making. Finally, using the Matplotlib tool to draw a temperature curve graph can not only intuitively display the change of temperature over time, but also help users better understand the laws and anomalies of temperature changes through visualization.

[0038] In one possible implementation, please refer to Figure 2 , Figure 4~Figure 8 The specific steps of step S3 include: S301: Using a fitting equation to fit the temperature curves in different slipping periods under the same ambient temperature condition, a plurality of fitting temperature curves in each slipping period are obtained, and the starting temperature and the ending temperature of each fitting temperature curve are both the normal operating temperature reference value, and the fitting equation is:

[0039] In the formula, y is the temperature generated during the slip period, x is the time corresponding to the temperature, and is the fitting parameter value; S302: if there is an overlapping area between the current fitting temperature curve and the previous fitting temperature curve, the overlapping area is divided into a front overlapping area and a rear overlapping area by using the temperature value corresponding to the intersection of the current fitting temperature curve and the previous fitting temperature curve as a dividing line; S303: Eliminate the front overlapping area, and use a correction equation to correct the current fitting temperature curve to obtain a corrected and accurate current temperature curve; S304: performing a sliding process on the current temperature curve so that the starting temperature position of the current temperature curve slides to the ending temperature position of the previous fitting temperature curve, thereby obtaining a complete and continuous temperature curve graph; S305: Establishing a working temperature comparison table, and storing the ambient temperature and the normal working temperature reference value in the working temperature comparison table; S306: extracting features from the temperature curve after fitting correction to form a set of temperature feature values.

[0040] When implementing this embodiment, the inventor of this application found that when the coupling slips continuously for many times, the temperature data generated by the friction plate is distorted and is not real temperature data. The temperature data will directly affect the critical temperature warning value obtained by subsequent calculations. The critical temperature warning value is a key indicator for judging whether the friction plate is worn to the point where it needs to be replaced or maintained. If the critical temperature warning value is inaccurate, the result data predicted by the neural network model will be inaccurate. Therefore, how to "eliminate the false and retain the true" for the temperature data generated by continuous slipping becomes a key issue. In this embodiment, if Figure 4 As shown in the figure, the horizontal axis is time and the vertical axis is temperature. In this figure, two consecutive slips occurred, that is, the previous temperature curve went from the temperature to the normal working temperature reference value ( ) starts to slip. At this time, the temperature starts to rise with time. When the temperature rises to the peak, the slip stops. Then the temperature starts to drop, but before it drops to the normal working temperature reference value ( ), at time At , the second slippage occurred continuously to form the current temperature curve. The current temperature curve stopped slipping after rising to the peak temperature, and then the temperature dropped to the normal working temperature reference value ( ).

[0041] During the two consecutive skidding processes mentioned above, the peak value of the current temperature curve has exceeded the actual critical temperature warning value, that is, during the second skidding process, the temperature generated by the previous skidding has not dropped to the normal operating temperature reference value ( ) starts to slip, there is a temperature superposition, so the peak value of the second slip is higher than its true peak value. Based on this, the present invention calculates the true temperature curve of the second slip through a fitting correction method.

[0042] In the implementation of this embodiment, Figure 5 As shown in the figure, the horizontal axis is time and the vertical axis is temperature. The previous temperature curve and the current temperature curve are fitted by the fitting equation so that the starting temperature and the ending temperature of the previous temperature curve and the current temperature curve are both the normal working temperature reference value ( ), forming the previous fitting temperature curve and the current fitting temperature curve; in this embodiment, Figure 6 As shown in the figure, the horizontal axis is time, the vertical axis is temperature, and the temperature value corresponding to the intersection of the current fitting temperature curve and the previous fitting temperature curve is the dividing line (that is, the time is The temperature value at that time is the dividing line), and the overlapping area of ​​the current fitting temperature curve and the previous fitting temperature curve is divided into the front overlapping area (i.e. Figure 6 The area composed of the underscores in the Figure 6 After removing the overlapping area, the current fitting temperature curve is corrected using the correction equation to obtain the corrected current temperature curve. The corrected current temperature curve is as follows: Figure 7 As shown in the figure, the horizontal axis is time and the vertical axis is temperature. Figure 7 In the embodiment, the corrected actual peak value of the current temperature curve is less than the critical temperature warning value; Figure 8 As shown in the figure, the horizontal axis is time and the vertical axis is temperature. The corrected current temperature curve is slid so that the starting temperature position of the current temperature curve slides to the ending temperature position of the previous fitting temperature curve to obtain a complete and continuous temperature curve graph; by performing feature extraction on the complete and continuous temperature curve graph, a set of temperature characteristic values ​​can be obtained, and the set of temperature characteristic values ​​is each peak value in the complete and continuous temperature curve graph.

[0043] In a possible implementation, the correction equation is:

[0044] In the formula, is the corrected current fitting temperature, is the normal operating temperature reference value, is the time corresponding to the intersection of the current fitting temperature curve and the previous fitting temperature curve, is the time when the temperature of the previous fitting temperature curve drops to the temperature reference value, is the time when the temperature of the current fitting temperature curve drops to the temperature reference value. is the temperature function of the current fitting temperature curve, is the temperature function of the previous fitting temperature curve.

[0045] In one possible implementation, please refer to Figure 3 In step S8, the step of obtaining the target temperature characteristic value set includes: S801: Continuously record target temperature data of the friction plate and the friction flange by setting a temperature sensor in a wrapping cover wrapping the coupling; S802: Perform frequency statistical analysis on the target temperature data to count the frequency of occurrence of each target temperature data; S803: taking the target temperature data with the highest frequency as the target ambient temperature; S804: extracting, from the operating temperature comparison table, a normal operating temperature reference value corresponding to the target ambient temperature as a target operating temperature reference value; S805: Establishing a target temperature curve graph based on the target temperature data; S806: performing fitting correction processing on the target temperature curve according to the target operating temperature reference value, and performing feature extraction on the fitted and corrected target temperature curve to form a target temperature feature value set.

[0046] During the implementation of this embodiment, when extracting the temperature data during the slipping period, the temperature data within the first 10 seconds before the start of slipping and the temperature data within the last 10 seconds after the slipping stops and reaches the lowest temperature are extracted according to the time series during the entire slipping period, and frequency statistical analysis is performed. This processing method makes the target ambient temperature we analyze and count more in line with the ambient temperature of the fan coupling under actual working conditions.

[0047] In a possible implementation, the formula for calculating the critical temperature warning value in step S5 is:

[0048] In the formula, is the critical temperature warning value, is the mean of the wear temperature characteristic value, K is the temperature warning safety factor, is the standard deviation of the wear temperature characteristic value.

[0049] In a possible implementation, in step S6, the neural network model is a temperature DNN neural network model.

[0050] In a possible implementation, the temperature DNN neural network model training step includes: The speed characteristic value set and the temperature characteristic value set are used as input feature X, and the critical temperature warning value is used as label Y, and are input into the input layer of the temperature DNN neural network model for normalization and standardization processing; The hidden layer uses an activation function to obtain the output value sent from the input layer to each hidden layer neuron; The output layer calculates the predicted critical temperature warning value through the activation function; The predicted critical temperature warning value is evaluated using a loss function, and the weight parameters of the temperature DNN neural network model are updated using a back propagation algorithm after the evaluation is qualified; The output layer outputs the qualified critical temperature warning value as the final result.

[0051] When this embodiment is implemented, the DNN neural network model has powerful nonlinear modeling capabilities and can effectively capture the complex relationship between input features and output labels, which is better than the traditional linear model; secondly, the DNN neural network model can automatically extract features, reduce manual intervention, improve the adaptability and prediction accuracy of the model, and make the predicted critical temperature warning value more accurate.

[0052] In a possible implementation, the warning notification model performs a traversal operation on all values ​​in the target temperature characteristic value set to detect whether at least one value is greater than or equal to the critical temperature warning value; When the detection condition is met, the traversal operation is stopped, and the warning notification model triggers the warning notification switch state to be turned on; The warning notification model calls the SMS service interface to send the warning information to the staff; After the warning information is sent, the warning notification model writes the time, content, target temperature feature value set and critical temperature warning value of the warning into the log.

[0053] When this embodiment is implemented, there is a log document specifically used to record warning events. The role of the log document is to provide detailed traceability of warning events, which is convenient for subsequent analysis and auditing; secondly, the target temperature characteristic value set, target speed characteristic value set and target critical temperature warning value in the log will be regularly pushed to the DNN neural network model for training and learning via timestamps, which provides valuable historical data support for subsequent DNN neural network model prediction training and data analysis, and helps to optimize and improve the DNN neural network model.

[0054] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for predicting wear of friction plates of fan couplings, characterized in that: The steps of the prediction method include: S1: performing a rotation loading test at different speeds on the coupling to be tested under different test environment temperature conditions, obtaining a normal working temperature reference value of the friction plate and the friction flange when there is no slippage, and obtaining the rotation speed and temperature data of the friction plate and the friction flange in different slippage periods, and establishing a mapping coordinate set of the rotation speed and the temperature data; S2: Establishing a temperature curve graph based on the temperature data; S3: performing fitting correction processing on the temperature curve according to the normal working temperature reference value, and performing feature extraction on the fitted and corrected temperature curve to form a temperature feature value set; S4: updating the temperature characteristic value set to the mapping coordinate set, and obtaining the rotation speed corresponding to the temperature characteristic value to form a rotation speed characteristic value set; S5: extracting the temperature characteristic values ​​under the working condition that the wear degree of the friction plate reaches the condition that maintenance or replacement is required from the temperature characteristic value set to form a wear temperature characteristic value set, and calculating the wear temperature characteristic value set to obtain a critical temperature warning value, wherein the critical temperature warning value is the critical temperature when the wear degree of the friction plate reaches the condition that maintenance or replacement is required; S6: assigning the critical temperature warning value, the rotation speed characteristic value set and the temperature characteristic value set as sample data to a neural network model for prediction training, wherein the output data of the neural network model is the critical temperature warning value, and the input data is the rotation speed characteristic value set and the temperature characteristic value set; S7: Establish an early warning notification model; S8: Obtaining a target temperature characteristic value set of the friction plate and the friction flange of the coupling in different slipping periods under actual working conditions, and a target speed characteristic value set corresponding to the target temperature characteristic value set; S9: assigning the target speed characteristic value set and the target temperature characteristic value set to the neural network model for prediction, and the neural network model predicts a corresponding target critical temperature warning value; S10: assigning the target critical temperature warning value and the target temperature characteristic value set to the warning notification model for comparison; S11: When at least one of the data in the target temperature characteristic value set is greater than or equal to the critical temperature warning value, the warning switch state is triggered to be on, and the staff is notified to repair or replace the friction plate.

2. A method for predicting wear of a fan coupling friction plate according to claim 1, characterized in that: In step S2, the step of establishing a temperature curve diagram includes: Processing the temperature data using a Kalman filter to filter out noise in the temperature data; Using a smoothing algorithm to smooth the temperature data and extract the data trend of the temperature data within the slip period; A temperature curve graph is drawn using Matplotlib tool according to the temperature data and the data trend.

3. A method for predicting wear of a fan coupling friction plate according to claim 2, characterized in that: The specific steps of step S3 include: S301: Using a fitting equation to fit the temperature curves in different slipping periods under the same ambient temperature condition, a plurality of fitting temperature curves in each slipping period are obtained, and the starting temperature and the ending temperature of each fitting temperature curve are both the normal operating temperature reference value, and the fitting equation is: In the formula, y is the temperature generated during the slip period, x is the time corresponding to the temperature, and K and b are the fitting parameter values; S302: if there is an overlapping area between the current fitting temperature curve and the previous fitting temperature curve, the overlapping area is divided into a front overlapping area and a rear overlapping area by using the temperature value corresponding to the intersection of the current fitting temperature curve and the previous fitting temperature curve as a dividing line; S303: Eliminate the front overlapping area, and use a correction equation to correct the current fitting temperature curve to obtain a corrected and accurate current temperature curve; S304: performing a sliding process on the current temperature curve so that the starting temperature position of the current temperature curve slides to the ending temperature position of the previous fitting temperature curve, thereby obtaining a complete and continuous temperature curve graph; S305: Establishing a working temperature comparison table, and storing the ambient temperature and the normal working temperature reference value in the working temperature comparison table; S306: performing feature extraction on the temperature curve after fitting correction to form a temperature feature value set.

4. A method for predicting wear of a fan coupling friction plate according to claim 3, characterized in that: The correction equation is: Wherein, T is the corrected current fitting temperature, T0 is the normal working temperature reference value, t1 is the time corresponding to the intersection of the current fitting temperature curve and the previous fitting temperature curve, t2 is the time corresponding to when the temperature of the previous fitting temperature curve drops to the temperature reference value, t3 is the time corresponding to when the temperature of the current fitting temperature curve drops to the temperature reference value, T1(t) is the temperature function of the current fitting temperature curve, and T2(t) is the temperature function of the previous fitting temperature curve.

5. The method for predicting wear of friction plates of fan coupling according to claim 3, characterized in that: The step of obtaining the target temperature characteristic value set in step S8 includes: S801: Continuously record target temperature data of the friction plate and the friction flange by setting a temperature sensor in a wrapping cover wrapping the coupling; S802: Perform frequency statistical analysis on the target temperature data to count the frequency of occurrence of each target temperature data; S803: taking the target temperature data with the highest frequency as the target ambient temperature; S804: extracting, from the operating temperature comparison table, a normal operating temperature reference value corresponding to the target ambient temperature as a target operating temperature reference value; S805: Establishing a target temperature curve graph based on the target temperature data; S806: performing fitting correction processing on the target temperature curve according to the target operating temperature reference value, and performing feature extraction on the fitted and corrected target temperature curve to form a target temperature feature value set.

6. The method for predicting wear of a fan coupling friction plate according to claim 1, characterized in that: The formula for calculating the critical temperature warning value in step S5 is: Where, T critical is the critical temperature warning value, μ T is the mean of the wear temperature characteristic value, K is the temperature warning safety factor, σ t is the standard deviation of the wear temperature characteristic value.

7. The method for predicting wear of friction plates of fan coupling according to claim 1, characterized in that: In step S6, the neural network model is a temperature DNN neural network model.

8. The method for predicting wear of friction plates of fan coupling according to claim 5, characterized in that: The temperature DNN neural network model training steps include: The speed characteristic value set and the temperature characteristic value set are used as input feature X, and the critical temperature warning value is used as label Y, and are input into the input layer of the temperature DNN neural network model for normalization and standardization processing; The hidden layer uses an activation function to obtain the output value sent from the input layer to each hidden layer neuron; The output layer calculates the predicted critical temperature warning value through the activation function; The predicted critical temperature warning value is evaluated using a loss function, and the weight parameters of the temperature DNN neural network model are updated using a back propagation algorithm after the evaluation is qualified; The output layer outputs the qualified critical temperature warning value as the final result.

9. The method for predicting wear of friction plates of fan coupling according to claim 1, characterized in that: The specific steps of step S11 include: The early warning notification model performs a traversal operation on all values ​​in the target temperature characteristic value set to detect whether at least one value is greater than or equal to the critical temperature early warning value; When the detection condition is met, the traversal operation is stopped, and the warning notification model triggers the warning notification switch state to be turned on; The warning notification model calls the SMS service interface to send the warning information to the staff; After the warning information is sent, the warning notification model writes the time, content, target temperature feature value set and critical temperature warning value of the warning into the log.

Images