Equipment defect prediction method and system based on machine learning
By collecting and using machine learning to train and predict equipment defect data, the problem of low accuracy in equipment defect prediction in the prior art is solved, and accurate positioning and prediction of equipment defects is achieved, and production risks are reduced.
Patent Information
- Application Number
- CN202510148371.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-10
- Publication Date
- 2025-05-16
AI Technical Summary
When the existing machine learning equipment defect prediction method deals with the defect type and feature data of diverse equipment, the training data error leads to low accuracy of defect type analysis results and poor model interpretation.
By collecting equipment defect data, using machine learning to train iterative prediction models, and evaluate the prediction model and alert the defect based on real-time data to accurately locate defect locations and improve the accuracy of equipment defect prediction.
It improves the accuracy of equipment defect prediction, can accurately locate defect locations, discover problems in advance, and reduce production risks.
Smart Images
Figure CN120011887A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of machine learning technology, and in particular to a method and system for predicting equipment defects based on machine learning. Background Art
[0002] At present, the equipment defect prediction method based on machine learning has made progress in technological development. However, due to the data diversity of equipment defect types and characteristics, errors in training data caused by machine learning training usually make the final defect type analysis results less accurate and the model less interpretable.
[0003] Therefore, the present invention proposes a method and system for predicting equipment defects based on machine learning. Summary of the invention
[0004] The present invention provides a method and system for predicting equipment defects based on machine learning, which collects equipment defect data, trains an iterative prediction model using machine learning, and finally evaluates the prediction model and issues defect warnings based on real-time data. The defect location is accurately located, improving the accuracy of equipment defect prediction.
[0005] In one aspect, the present invention provides a method for predicting equipment defects based on machine learning, comprising: Step 1: Collect sample defect data of equipment in operation of chemical water treatment, and classify and label the sample defect data based on the defect type of the equipment to obtain label parameter data; Step 2: preprocess the tag parameter data, and extract features of the parameter data under different tags according to the characteristics of equipment defects; Step 3: Based on the features extracted under different labels and the preprocessed label parameter data, and combined with the machine learning algorithm, construct the prediction function under the corresponding label; Step 4: Obtain a verification data set of the actual production environment and input it into the defect prediction model constructed by the prediction function under all labels, predict equipment defects and feed them back to the defect prediction model for model optimization and equipment defect evaluation.
[0006] On the other hand, in step 1, sample defect data of equipment in operation of chemical water is collected. include: According to the type and location of the equipment in the chemical water operation, a first number is assigned to the corresponding equipment and a second number is assigned to the target sensor of the corresponding equipment; The first number and the second number are matched according to the sensor-equipment type comparison table in the installation guide, the operation data of the equipment in operation of the chemical water is monitored, and the sample defect data is obtained.
[0007] On the other hand, in step 1, the sample defect data is classified based on the defect type. Class label processing to obtain label parameter data, including: Extract keywords from all sample defect data according to the defect names under the defect types of the equipment; The extracted keywords are synonymously clustered according to word similarity to obtain several defect clusters, and a classification label is configured for each defect cluster to obtain label parameter data, wherein one defect cluster corresponds to one classification label, and the number of defect clusters is less than the number of defect types.
[0008] On the other hand, in step 2, the label parameter data is preprocessed, including: Obtain standard parameter data corresponding to each label parameter data and construct a first matrix A; ; in, represents the standard parameter data corresponding to the i-th tag parameter data at the j-th monitoring time, where: , ; Perform matrix normalization on matrix A to obtain the normalized matrix .
[0009] On the other hand, in step 2, feature extraction is performed on parameter data under different labels according to the characteristics of equipment defects, including: According to the type-characteristic mapping table, obtaining a device defect characteristic corresponding to the defect type of the device; According to the standardized matrix , calculate the covariance matrix, and perform eigenvalue decomposition on the covariance matrix to obtain the standardized matrix Composition characteristics ; according to Determine the composition characteristics The contribution of each characteristic element in , where represents the fth characteristic element; m1 represents the component characteristic The number of characteristic elements in ; Indicates the contribution degree of the fth characteristic element; Select elements whose contribution is greater than a preset degree from all characteristic elements to form the main feature, and determine the defect characteristic function of the device: ; in, Indicates the device The defect characteristic function of l characteristic elements, Indicates the filtered l characteristic elements, Indicates the equipment defect characteristics corresponding to the defect type of the equipment, represents the logarithmic function, Indicates the conversion coefficient of the characteristic element; Indicates the conversion factor for equipment defect characteristics; Based on all defect characteristic functions, the defect characteristics of the device are determined.
[0010] On the other hand, in step 3, the prediction function under the corresponding label is constructed, including: Compare and analyze the extracted features with the preprocessed label parameter data to obtain a feature-data comparison list for each extracted feature, and calculate the comparison degree of the corresponding extracted feature according to the feature-data comparison list; ; in, represents the contrast of the dth extracted feature; represents the number of pre-processed label parameter data determined from the feature-data comparison list of the dth extracted feature; Indicates the amount of data corresponding to the i2th preprocessing label parameter data; represents the allowed usage weight of the dth extracted feature corresponding to the i2th preprocessed label parameter data; It indicates the usage rate of the d-th extracted feature using the corresponding i2-th preprocessed label parameter data; Perform machine learning on each feature-data comparison list based on a machine learning algorithm, and construct a prediction function according to the comparison degree and the learning rate of the machine learning algorithm.
[0011] On the other hand, in step 4, predicting equipment defects and feeding back to the defect prediction model for model optimization and equipment defect assessment includes: Acquire real-time parameter data of the equipment from the actual production environment, and obtain a verification data set after preprocessing and label classification of the real-time parameter data, wherein the verification data set contains sub-data sets under different label classifications; Input the verification data set into the defect prediction model constructed by all prediction functions to obtain a first prediction array under each classification label, wherein the first prediction array is obtained by performing defect prediction analysis on the verification data set based on the defect prediction model; At the same time, each sub-data set is input into the prediction function under the corresponding classification label to obtain a corresponding second prediction array, wherein the second prediction array is obtained by performing defect prediction analysis on the corresponding sub-data set based on the prediction function under the corresponding classification label; The first prediction array, the second prediction array and the actual defect array under the same classification label are presented in the same coordinate system, and three curves are obtained by drawing; Obtain the defect arrays under the same defect feature in the three curves respectively, and calculate the first defect difference, the second defect difference, the third defect difference and the first defect variance in the defect array respectively, and assign defect coefficients to the corresponding defect arrays ; ; in, is the first defect difference in the corresponding defect array, is the second defect difference in the corresponding defect array; is the third defect difference in the corresponding defect array; Based on Calculate the first defect variance; According to the standard that the defect coefficient is greater than the preset coefficient, the three curves are intercepted, and the second defect variance of the intercepted curve is determined. ; ; in, is the i3th first defect variance involved in the intercept curve; is the variance of all first defect variances involved in the intercept curve; Get the appearance time difference array of adjacent defect features in the three curves , and determine the corresponding third defect variance, and then calculate the fourth defect variance of the intercept curve, wherein t01, t11 and t21 are the occurrence times of the first predicted defect, the second predicted defect and the actual defect under the first defect feature in the adjacent defect features respectively; t02, t12 and t22 are the occurrence times of the first predicted defect, the second predicted defect and the actual defect under the second defect feature in the adjacent defect features respectively; Obtaining a first area of the first curve corresponding to the first prediction array above the third curve corresponding to the actual defect array and a second area below the third curve corresponding to the actual defect array in the three curves; At the same time, obtaining a third area of the second curve corresponding to the second prediction array above the third curve corresponding to the actual defect array and a fourth area below the third curve corresponding to the actual defect array in the three curves; Obtaining a fifth defect variance according to a first ratio of the first area to the third area, a second ratio of the second area to the fourth area, a first overlap ratio of the first area to the third area, and a second overlap ratio of the second area to the fourth area; According to the second defect variance, the fourth defect variance, and the fifth defect variance, obtaining a function optimization standard of a prediction function under a corresponding classification label from a variance-label-optimization mapping table; All function optimization criteria are sequentially input into the standard analysis model to obtain the optimization vector; The defect prediction model is optimized according to the optimization vector, and the device defect assessment is performed on the verification data set according to the optimized model.
[0012] On the other hand, it includes: Data collection module: collects sample defect data of equipment in operation of chemical water, and classifies and labels the sample defect data based on the defect type to obtain label parameter data; Feature parameter module: pre-processes the tag parameter data and extracts features of the parameter data under different tags according to the characteristics of equipment defects; Prediction function module: Based on the features extracted under different labels and the preprocessed label parameter data, combined with machine learning algorithms, the prediction function under the corresponding label is constructed; Evaluation module: obtains the verification data set of the actual production environment and inputs it into the defect prediction model constructed by the prediction function under all labels, predicts equipment defects and feeds them back to the defect prediction model for model optimization and equipment defect evaluation.
[0013] The present invention provides a method and system for predicting equipment defects based on machine learning, which collects equipment defect data, trains an iterative prediction model using machine learning, and finally evaluates the prediction model and issues defect warnings based on real-time data. The defect location is accurately located, improving the accuracy of equipment defect prediction. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0015] Figure 1 It is a flow chart of a method for predicting equipment defects based on machine learning provided in an embodiment of the present invention.
[0016] Figure 2 It is a structural diagram of a machine learning-based equipment defect prediction system provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0017] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0018] Embodiment 1: like Figure 1 As shown, an embodiment of the present invention provides a method for predicting equipment defects based on machine learning, comprising: Step 1: Collect sample defect data of equipment in operation of chemical water treatment, and classify and label the sample defect data based on the defect type of the equipment to obtain label parameter data; Step 2: preprocess the tag parameter data, and extract features of the parameter data under different tags according to the characteristics of equipment defects; Step 3: Based on the features extracted under different labels and the preprocessed label parameter data, and combined with the machine learning algorithm, construct the prediction function under the corresponding label; Step 4: Obtain a verification data set of the actual production environment and input it into the defect prediction model constructed by the prediction function under all labels, predict equipment defects and feed them back to the defect prediction model for model optimization and equipment defect evaluation.
[0019] In this embodiment, chemical water operation refers to the process of using water as a carrier medium to carry out reactions, mass transfer, separation and other operations in the chemical industry production process.
[0020] In this embodiment, the sample defect data is a defect information record collected from chemical water operation equipment, including a detailed description of equipment failure, damage, abnormality and other problems and relevant information of the equipment.
[0021] In this embodiment, the defect type refers to various problems or defect types that may occur in chemical water operation equipment, such as equipment failure, leakage, wear, blockage, corrosion, etc.
[0022] In this embodiment, the classification label processing is to classify and label the sample data.
[0023] In this embodiment, the label parameter data refers to data obtained after classification and label processing of the collected sample defect data.
[0024] In this embodiment, preprocessing refers to a series of cleaning, conversion and processing operations performed on raw data before performing machine learning tasks.
[0025] In this embodiment, the equipment defect characteristics refer to the characteristics of different equipment defects in the sample data, including: change trends of parameters such as temperature, pressure, flow, and other parameters, range and other data anomalies, etc.
[0026] In this embodiment, feature extraction refers to selecting and extracting information that best describes data characteristics from raw data as new features for use in machine learning tasks.
[0027] In this embodiment, the machine learning algorithm is an algorithm that enables the system to learn from data and continuously improve performance by using data and statistical techniques, including supervised learning algorithms, unsupervised learning algorithms, reinforcement learning algorithms, semi-supervised learning algorithms, and other types.
[0028] In this embodiment, the prediction function is a function obtained through a machine learning training model, which is used to map input data to a predicted output result.
[0029] In this embodiment, the actual production environment refers to a real industrial production environment in which equipment is running and generating data.
[0030] In this embodiment, the verification data set is a data sample collected from an actual production environment, including data under normal operating conditions and data with known defects.
[0031] In this embodiment, the defect prediction model is a model for predicting future defects of a device.
[0032] In this embodiment, model optimization is a process of continuously adjusting and improving the parameters of the machine learning model to improve the performance and accuracy of the model on actual data.
[0033] In this embodiment, the equipment defect assessment refers to the process of comprehensively evaluating and analyzing the equipment defect conditions predicted by the prediction model.
[0034] The working principle and beneficial effects of the above technical solution are: by collecting and processing equipment defect data, building a prediction model and optimizing the evaluation, accurate prediction and monitoring of chemical water operation equipment defects can be achieved, problems can be discovered in advance, production risks can be reduced, and the location of defects can be accurately located, thereby improving the accuracy of equipment defect prediction.
[0035] Embodiment 2: Based on the above embodiment 1, in step 1, collecting sample defect data of equipment in operation of chemical water includes: According to the type and location of the equipment in the chemical water operation, a first number is assigned to the corresponding equipment and a second number is assigned to the target sensor of the corresponding equipment; The first number and the second number are matched according to the sensor-equipment type comparison table in the installation guide, the operation data of the equipment in operation of the chemical water is monitored, and the sample defect data is obtained.
[0036] In this embodiment, the equipment types include various water treatment equipment, such as water pumps, filters, heaters, reactors, etc.
[0037] In this embodiment, the device location refers to the specific installation location of the device in the chemical water treatment system, such as a water supply inlet, a water outlet, a circulation pipeline, etc.
[0038] In this embodiment, the first number refers to a unique identification number of a corresponding device, which is used to distinguish and identify different devices in the monitoring system.
[0039] In this embodiment, the second number is a unique identification number pointing to the target sensor configuration of the corresponding device, and is used to distinguish and identify different sensors in the monitoring system.
[0040] In this embodiment, the sensor-device type comparison table is a comparison table indicating the mapping relationship between sensor types and device types.
[0041] The working principle and beneficial effects of the above technical solution are: by configuring unique numbers and sensors, the monitoring of chemical water operation equipment and the acquisition of defect data are realized, which provides a data basis for accurately locating the location of defects and improves the accuracy of equipment defect prediction.
[0042] Embodiment 3: On the basis of the above-mentioned embodiment 1, in step 1, the sample defect data is classified and labeled based on the defect type to obtain label parameter data, including: Extract keywords from all sample defect data according to the defect names under the defect types of the equipment; The extracted keywords are synonymously clustered according to word similarity to obtain several defect clusters, and a classification label is configured for each defect cluster to obtain label parameter data, wherein one defect cluster corresponds to one classification label, and the number of defect clusters is less than the number of defect types.
[0043] In this embodiment, the defect name is to classify and name the defects that occur during the operation of the equipment, such as leakage, blockage, wear, oxidation, etc.
[0044] In this embodiment, keyword extraction refers to extracting important words or phrases describing defects from sample defect data.
[0045] In this embodiment, synonym clustering refers to the process of grouping semantically similar words or phrases into the same cluster according to certain rules.
[0046] In this embodiment, defect clustering refers to the process of classifying similar defects into the same category.
[0047] The working principle and beneficial effects of the above technical solution are: through steps such as keyword extraction, synonymous clustering and configuration of classification labels, defect data can be better organized and managed, providing a data basis for accurately locating the location of defects and improving the accuracy of equipment defect prediction.
[0048] Embodiment 4: On the basis of the above embodiment 1, in step 2, the tag parameter data is preprocessed, including: Obtain standard parameter data corresponding to each label parameter data and construct a first matrix A; ; in, represents the standard parameter data corresponding to the i-th tag parameter data at the j-th monitoring time, where: , ; Perform matrix normalization on matrix A to obtain the normalized matrix .
[0049] In this embodiment, the standard parameter data is the parameter value when the device operates normally under certain specific conditions.
[0050] In this embodiment, the first matrix is a matrix composed of label parameter data and corresponding standard parameter data.
[0051] In this embodiment, matrix standardization processing refers to converting the data in the matrix according to certain rules so that the data meets the standard.
[0052] The working principle and beneficial effects of the above technical solution are: by acquiring standard parameter data, constructing matrix A, and performing standardized processing, data comparison and analysis are achieved, providing a data basis for accurately locating the location of defects and improving the accuracy of equipment defect prediction.
[0053] Embodiment 5: On the basis of the above-mentioned embodiment 4, in step 2, feature extraction is performed on parameter data under different tags according to the characteristics of equipment defects, including: According to the type-characteristic mapping table, obtaining a device defect characteristic corresponding to the defect type of the device; According to the standardized matrix , calculate the covariance matrix, and perform eigenvalue decomposition on the covariance matrix to obtain the standardized matrix Composition characteristics ; according to Determine the composition characteristics The contribution of each characteristic element in , where represents the fth characteristic element; m1 represents the component characteristic The number of characteristic elements in ; Indicates the contribution degree of the fth characteristic element; Select elements whose contribution is greater than a preset degree from all characteristic elements to form the main feature, and determine the defect characteristic function of the device: ; in, Indicates the device The defect characteristic function of l characteristic elements, Indicates the filtered l characteristic elements, Indicates the equipment defect characteristics corresponding to the defect type of the equipment, represents the logarithmic function, Indicates the conversion coefficient of the characteristic element; Indicates the conversion factor for equipment defect characteristics; Based on all defect characteristic functions, the defect characteristics of the device are determined.
[0054] In this embodiment, the type-characteristic mapping table is a mapping table that represents the relationship between the defect types of the device and the defect characteristics corresponding to each defect type.
[0055] In this embodiment, the device defect characteristics are defect phenomena that occur during the operation or use of the device, including various physical characteristics, performance parameters, etc.
[0056] In this embodiment, the covariance matrix is a matrix that describes the strength and direction of the relationship between defect types and defect characteristics.
[0057] In this embodiment, eigenvalue decomposition is used to decompose the covariance matrix into a set of eigenvectors and corresponding eigenvalues.
[0058] In this embodiment, the component features refer to new features obtained by performing eigenvalue decomposition on the covariance matrix.
[0059] In this embodiment, the contribution degree is an indicator used to measure the contribution degree of each characteristic element in the component feature to the overall data change.
[0060] In this embodiment, the main feature refers to a component feature whose sharing degree is greater than a preset degree and is selected according to the contribution degree in eigenvalue decomposition.
[0061] In this embodiment, the characteristic element refers to each element in the component feature obtained by performing eigenvalue decomposition of the standardized matrix.
[0062] In this embodiment, the conversion coefficient is a coefficient used to convert the characteristic element and the device defect characteristic.
[0063] The working principle and beneficial effects of the above technical solution are: through covariance matrix decomposition and contribution degree analysis, combined with equipment defect characteristics and conversion coefficients, a function model that can better reflect the characteristics of equipment defects is constructed. It provides a data basis for accurately locating the location of defects and improves the accuracy of equipment defect prediction.
[0064] Embodiment 6: On the basis of the above embodiment 1, in step 3, a prediction function under a corresponding label is constructed, including: Compare and analyze the extracted features with the preprocessed label parameter data to obtain a feature-data comparison list for each extracted feature, and calculate the comparison degree of the corresponding extracted feature according to the feature-data comparison list; ; in, represents the contrast of the dth extracted feature; represents the number of pre-processed label parameter data determined from the feature-data comparison list of the dth extracted feature; Indicates the amount of data corresponding to the i2th preprocessing label parameter data; represents the allowed usage weight of the dth extracted feature corresponding to the i2th preprocessed label parameter data; It indicates the usage rate of the d-th extracted feature using the corresponding i2-th preprocessed label parameter data; Perform machine learning on each feature-data comparison list based on a machine learning algorithm, and construct a prediction function according to the comparison degree and the learning rate of the machine learning algorithm.
[0065] In this embodiment, the control analysis is a process of analyzing the interactive relationships formed by the variables to express the connection between the variables.
[0066] In this embodiment, the feature-data comparison list is a list obtained by comparing and arranging the extracted features with the pre-processed label parameter data, and records the association relationship between each feature and the corresponding data.
[0067] In this embodiment, the comparison degree refers to a measure of the degree of association between a feature and data in the feature-data comparison list, indicating the degree of association between each extracted feature and the corresponding data.
[0068] In this embodiment, the weight allowed to be used is a coefficient of the degree of influence of the feature on the label data.
[0069] In this embodiment, the learning rate refers to a hyperparameter used to control the update step size of model parameters in a machine learning algorithm.
[0070] The working principle and beneficial effects of the above technical solution are: analyzing the feature-data correspondence relationship through the feature-data comparison list, calculating the comparison degree, and applying the machine learning algorithm to construct the prediction function, combined with the learning rate optimization model training, to provide a data basis for accurately locating the defect location and improve the accuracy of equipment defect prediction.
[0071] Embodiment 7: Based on the above-mentioned embodiment 4, in step 4, predicting equipment defects and feeding back to the defect prediction model for model optimization and equipment defect assessment includes: Acquire real-time parameter data of the equipment from the actual production environment, and obtain a verification data set after preprocessing and label classification of the real-time parameter data, wherein the verification data set contains sub-data sets under different label classifications; Input the verification data set into the defect prediction model constructed by all prediction functions to obtain a first prediction array under each classification label, wherein the first prediction array is obtained by performing defect prediction analysis on the verification data set based on the defect prediction model; At the same time, each sub-data set is input into the prediction function under the corresponding classification label to obtain a corresponding second prediction array, wherein the second prediction array is obtained by performing defect prediction analysis on the corresponding sub-data set based on the prediction function under the corresponding classification label; The first prediction array, the second prediction array and the actual defect array under the same classification label are presented in the same coordinate system, and three curves are obtained by drawing; Obtain the defect arrays under the same defect feature in the three curves respectively, and calculate the first defect difference, the second defect difference, the third defect difference and the first defect variance in the defect array respectively, and assign defect coefficients to the corresponding defect arrays ; ; in, is the first defect difference in the corresponding defect array, is the second defect difference in the corresponding defect array; is the third defect difference in the corresponding defect array; Based on Calculate the first defect variance; According to the standard that the defect coefficient is greater than the preset coefficient, the three curves are intercepted, and the second defect variance of the intercepted curve is determined. ; ; in, is the i3th first defect variance involved in the intercept curve; is the variance of all first defect variances involved in the intercept curve; Get the appearance time difference array of adjacent defect features in the three curves , and determine the corresponding third defect variance, and then calculate the fourth defect variance of the intercept curve, wherein t01, t11 and t21 are the occurrence times of the first predicted defect, the second predicted defect and the actual defect under the first defect feature in the adjacent defect features respectively; t02, t12 and t22 are the occurrence times of the first predicted defect, the second predicted defect and the actual defect under the second defect feature in the adjacent defect features respectively; Obtaining a first area of the first curve corresponding to the first prediction array above the third curve corresponding to the actual defect array and a second area below the third curve corresponding to the actual defect array in the three curves; At the same time, obtaining a third area of the second curve corresponding to the second prediction array above the third curve corresponding to the actual defect array and a fourth area below the third curve corresponding to the actual defect array in the three curves; Obtaining a fifth defect variance according to a first ratio of the first area to the third area, a second ratio of the second area to the fourth area, a first overlap ratio of the first area to the third area, and a second overlap ratio of the second area to the fourth area; According to the second defect variance, the fourth defect variance, and the fifth defect variance, obtaining a function optimization standard of a prediction function under a corresponding classification label from a variance-label-optimization mapping table; All function optimization criteria are sequentially input into the standard analysis model to obtain the optimization vector; The defect prediction model is optimized according to the optimization vector, and the device defect assessment is performed on the verification data set according to the optimized model.
[0072] In this embodiment, the real-time parameter data refers to various parameter data collected in real time during the operation of the device, such as temperature, pressure, humidity, current, etc.
[0073] In this embodiment, the sub-dataset is a small-scale data set extracted from the validation data set according to different label classifications.
[0074] In this embodiment, the first prediction array is the result of defect prediction analysis on the verification data set based on the defect prediction model, and includes the prediction of the device defects under each classification label.
[0075] In this embodiment, defect prediction analysis refers to a process of predicting and evaluating defects of a device by analyzing and processing real-time parameter data of the device using a prediction model.
[0076] In this embodiment, the second prediction array is a set of prediction results obtained by performing defect prediction analysis on the corresponding sub-data set based on the prediction function under the corresponding classification label.
[0077] In this embodiment, the defect array refers to the real-time parameter data of the equipment that is obtained by analyzing the verification data set according to the defect prediction model after being preprocessed and labeled.
[0078] In this embodiment, the first defect difference refers to a defect array under the same defect feature after the first prediction array, the second prediction array and the actual defect array under the same classification label are presented in the same coordinate system.
[0079] In this embodiment, the second defect difference refers to the difference between the second prediction array and the actual defect array obtained by the prediction function under each classification label after the defect prediction model analyzes the verification data set.
[0080] In this embodiment, the third defect difference first prediction array, the second prediction array and the actual defect array are presented in the same coordinate system, and three curves are plotted to obtain the difference between the third curve and the other arrays.
[0081] In this embodiment, the defect coefficient is a quantitative evaluation index of equipment defects, which measures the severity of defects of the equipment under different defect characteristics.
[0082] In this embodiment, the first defect variance is a difference metric calculated by analyzing the first prediction array, the second prediction array, and the actual defect array under the same classification label.
[0083] In this embodiment, the second defect variance is obtained by comparing the difference between the second prediction array and the actual defect array corresponding to different prediction functions under the same classification label in the same coordinate system.
[0084] In this embodiment, the third defect variance is obtained by comparing the differences between the third prediction array and the actual defect array corresponding to different prediction functions under the same classification label in the same coordinate system.
[0085] In this embodiment, the fourth defect variance is obtained by analyzing and comparing three different curves and combining the variance of the intercepted curve obtained by the third defect variance.
[0086] In this embodiment, the first area refers to the area above the third curve corresponding to the first prediction array.
[0087] In this embodiment, the second area refers to the area under the third curve corresponding to the actual defect array.
[0088] In this embodiment, the third area is the area of the second curve corresponding to the second prediction array among the three curves that is located above the third curve corresponding to the actual defect array.
[0089] In this embodiment, the fourth area is the area under the third curve corresponding to the second prediction data.
[0090] In this embodiment, the first ratio refers to the ratio of the first area to the third area.
[0091] In this embodiment, the second ratio refers to the ratio of the second area to the fourth area.
[0092] In this embodiment, the first overlap ratio refers to the ratio of the area of the first curve corresponding to the first prediction array above the third curve corresponding to the actual defect array to the total area above the third curve corresponding to the actual defect array, that is, the first area / (first area+third area).
[0093] In this embodiment, the second overlapping ratio refers to the second area / (the second area+the fourth area).
[0094] In this embodiment, the fifth defect variance is calculated based on a first ratio of the first area to the third area, a second ratio of the second area to the fourth area, a first overlap ratio of the first area to the third area, and a second overlap ratio of the second area to the fourth area.
[0095] In this embodiment, the variance-label-optimization mapping table is a table that records the relationship between variance and label and the corresponding optimization criteria.
[0096] In this embodiment, the function optimization criterion refers to the optimization criterion of the corresponding variance and label data found in the variance-label-optimization mapping table according to the label classification.
[0097] In this embodiment, the standard analysis model is a specific model used to determine the function optimization standard.
[0098] In this embodiment, the optimization vector refers to a set of vectors calculated according to all optimization criteria combined with the standard analysis model, and is used to optimize the defect prediction model.
[0099] The working principle and beneficial effects of the above technical solution are: utilizing real-time acquisition of equipment parameter data, and performing defect prediction and optimization through defect prediction models and prediction functions, which provides a data basis for accurately locating the location of defects and improves the accuracy of equipment defect prediction.
[0100] Embodiment 8: like Figure 2 As shown, an embodiment of the present invention provides a machine learning-based equipment defect prediction system, characterized in that it includes: Data collection module: collects sample defect data of equipment in operation of chemical water, and classifies and labels the sample defect data based on the defect type to obtain label parameter data; Feature parameter module: pre-processes the tag parameter data and extracts features of the parameter data under different tags according to the characteristics of equipment defects; Prediction function module: Based on the features extracted under different labels and the preprocessed label parameter data, combined with machine learning algorithms, the prediction function under the corresponding label is constructed; Evaluation module: obtains the verification data set of the actual production environment and inputs it into the defect prediction model constructed by the prediction function under all labels, predicts equipment defects and feeds them back to the defect prediction model for model optimization and equipment defect evaluation.
[0101] The working principle and beneficial effects of the above technical solution are: by collecting and processing equipment defect data, building a prediction model and optimizing the evaluation, accurate prediction and monitoring of chemical water operation equipment defects can be achieved, problems can be discovered in advance, production risks can be reduced, and the location of defects can be accurately located, thereby improving the accuracy of equipment defect prediction.
[0102] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for predicting equipment defects based on machine learning, characterized in that: include: Step 1: Collect sample defect data of equipment in operation of chemical water treatment, and classify and label the sample defect data based on the defect type of the equipment to obtain label parameter data; Step 2: preprocess the tag parameter data, and extract features of the parameter data under different tags according to the characteristics of equipment defects; Step 3: Based on the features extracted under different labels and the preprocessed label parameter data, and combined with the machine learning algorithm, construct the prediction function under the corresponding label; Step 4: Obtain a verification data set of the actual production environment and input it into the defect prediction model constructed by the prediction function under all labels, predict equipment defects and feed them back to the defect prediction model for model optimization and equipment defect evaluation.
2. The device defect prediction method based on machine learning according to claim 1 is characterized in that: In step 1, sample defect data of equipment in operation of chemical water is collected, including: According to the type and location of the equipment in the chemical water operation, a first number is assigned to the corresponding equipment and a second number is assigned to the target sensor of the corresponding equipment; The first number and the second number are matched according to the sensor-equipment type comparison table in the installation guide, the operation data of the equipment in operation of the chemical water is monitored, and the sample defect data is obtained.
3. The device defect prediction method based on machine learning according to claim 1 is characterized in that: In step 1, the sample defect data is classified and labeled based on the defect type to obtain label parameter data, including: Extract keywords from all sample defect data according to the defect names under the defect types of the equipment; The extracted keywords are synonymously clustered according to word similarity to obtain several defect clusters, and a classification label is configured for each defect cluster to obtain label parameter data, wherein one defect cluster corresponds to one classification label, and the number of defect clusters is less than the number of defect types.
4. The device defect prediction method based on machine learning according to claim 1, characterized in that: In step 2, the label parameter data is preprocessed, including: Obtain standard parameter data corresponding to each label parameter data and construct a first matrix A; ； in, represents the standard parameter data corresponding to the i-th tag parameter data at the j-th monitoring time, where: , ; Perform matrix normalization on matrix A to obtain the normalized matrix .
5. The method for predicting equipment defects based on machine learning according to claim 4, characterized in that: In step 2, feature extraction is performed on parameter data under different labels according to the characteristics of equipment defects, including: According to the type-characteristic mapping table, obtaining a device defect characteristic corresponding to the defect type of the device; According to the standardized matrix , calculate the covariance matrix, and perform eigenvalue decomposition on the covariance matrix to obtain the standardized matrix Composition characteristics ; according to Determine the composition characteristics The contribution of each characteristic element in , where represents the fth characteristic element; m1 represents the component characteristic The number of characteristic elements in ; Indicates the contribution of the fth characteristic element; Select elements whose contribution is greater than a preset degree from all characteristic elements to form the main feature, and determine the defect characteristic function of the device: ； in, Indicates the device The defect characteristic function of l characteristic elements, Indicates the filtered l characteristic elements, Indicates the equipment defect characteristics corresponding to the defect type of the equipment, represents the logarithmic function, Indicates the conversion coefficient of the characteristic element; Indicates the conversion factor for equipment defect characteristics; Based on all defect characteristic functions, the defect characteristics of the device are determined.
6. The device defect prediction method based on machine learning according to claim 1, characterized in that: In step 3, construct the prediction function under the corresponding label, including: Compare and analyze the extracted features with the preprocessed label parameter data to obtain a feature-data comparison list for each extracted feature, and calculate the comparison degree of the corresponding extracted feature according to the feature-data comparison list; ； in, represents the contrast of the dth extracted feature; represents the number of pre-processed label parameter data determined from the feature-data comparison list of the dth extracted feature; Indicates the amount of data corresponding to the i2th preprocessing label parameter data; represents the allowed usage weight of the dth extracted feature corresponding to the i2th preprocessed label parameter data; It indicates the usage rate of the d-th extracted feature using the corresponding i2-th preprocessed label parameter data; Perform machine learning on each feature-data comparison list based on a machine learning algorithm, and construct a prediction function according to the comparison degree and the learning rate of the machine learning algorithm.
7. The device defect prediction method based on machine learning according to claim 1, characterized in that: In step 4, equipment defects are predicted and fed back into the defect prediction model for model optimization and equipment defect assessment, including: Acquire real-time parameter data of the equipment from the actual production environment, and obtain a verification data set after preprocessing and label classification of the real-time parameter data, wherein the verification data set contains sub-data sets under different label classifications; Input the verification data set into the defect prediction model constructed by all prediction functions to obtain a first prediction array under each classification label, wherein the first prediction array is obtained by performing defect prediction analysis on the verification data set based on the defect prediction model; At the same time, each sub-data set is input into the prediction function under the corresponding classification label to obtain a corresponding second prediction array, wherein the second prediction array is obtained by performing defect prediction analysis on the corresponding sub-data set based on the prediction function under the corresponding classification label; The first prediction array, the second prediction array and the actual defect array under the same classification label are presented in the same coordinate system, and three curves are obtained by drawing; Obtain the defect arrays under the same defect feature in the three curves respectively, and calculate the first defect difference, the second defect difference, the third defect difference and the first defect variance in the defect array respectively, and assign defect coefficients to the corresponding defect arrays ; ； in, is the first defect difference in the corresponding defect array, is the second defect difference in the corresponding defect array; is the third defect difference in the corresponding defect array; Based on Calculate the first defect variance; According to the standard that the defect coefficient is greater than the preset coefficient, the three curves are intercepted, and the second defect variance of the intercepted curve is determined. ; ； in, is the i3th first defect variance involved in the intercept curve; is the variance of all first defect variances involved in the intercept curve; Get the appearance time difference array of adjacent defect features in the three curves , and determine the corresponding third defect variance, and then calculate the fourth defect variance of the intercept curve, wherein t01, t11 and t21 are the occurrence times of the first predicted defect, the second predicted defect and the actual defect under the first defect feature in the adjacent defect features respectively; t02, t12 and t22 are the occurrence times of the first predicted defect, the second predicted defect and the actual defect under the second defect feature in the adjacent defect features respectively; Obtaining a first area of the first curve corresponding to the first prediction array above the third curve corresponding to the actual defect array and a second area below the third curve corresponding to the actual defect array in the three curves; At the same time, obtaining a third area of the second curve corresponding to the second prediction array above the third curve corresponding to the actual defect array and a fourth area below the third curve corresponding to the actual defect array in the three curves; Obtaining a fifth defect variance according to a first ratio of the first area to the third area, a second ratio of the second area to the fourth area, a first overlap ratio of the first area to the third area, and a second overlap ratio of the second area to the fourth area; According to the second defect variance, the fourth defect variance, and the fifth defect variance, obtaining a function optimization standard of a prediction function under a corresponding classification label from a variance-label-optimization mapping table; All function optimization criteria are sequentially input into the standard analysis model to obtain the optimization vector; The defect prediction model is optimized according to the optimization vector, and the device defect assessment is performed on the verification data set according to the optimized model.
8. A machine learning-based equipment defect prediction system, characterized in that: include: Data collection module: collects sample defect data of equipment in operation of chemical water, and classifies and labels the sample defect data based on the defect type to obtain label parameter data; Feature parameter module: pre-processes the tag parameter data and extracts features of the parameter data under different tags according to the characteristics of equipment defects; Prediction function module: Based on the features extracted under different labels and the preprocessed label parameter data, combined with machine learning algorithms, the prediction function under the corresponding label is constructed; Evaluation module: obtains the verification data set of the actual production environment and inputs it into the defect prediction model constructed by the prediction function under all labels, predicts equipment defects and feeds them back to the defect prediction model for model optimization and equipment defect evaluation.
Citation Information
Cited By
Machine learning-based device defect prediction method and system
WO2026166148A1