A coal mill fault diagnosis method based on multivariate time series classification

Abstract

The application discloses a coal mill fault diagnosis method based on multivariate time series classification, relates to the technical field of mechanical fault diagnosis, and comprises the following steps: collecting time series data of a coal mill, pre-processing the time series data and storing the time series data into a data set; constructing a self-supervised contrast learning model, and performing contrast learning training on different views in the data set; inputting features output by the self-supervised contrast learning model into a boosted linear classifier, and performing fault classification and diagnosis. The method disclosed by the application improves the fault diagnosis effect under different working conditions by constructing a self-supervised contrast learning model, and enhances the adaptability and stability of the method. The method mines associated features in multivariate time series data through a boosted linear classifier, performs comprehensive analysis, improves the comprehensive understanding of the running state of the coal mill, and thus realizes more accurate fault diagnosis. The application achieves better effects in terms of precision, reliability and generalization ability.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mechanical fault diagnosis, in particular to a coal mill fault diagnosis method based on multivariate time series classification. BACKGROUND

[0002] The coal mill is a key equipment in the power plant, and its running state directly affects the safety and economy of the power plant. Due to the complex running environment and frequent load changes of the coal mill, faults are prone to occur. If the faults cannot be detected and diagnosed in a timely and accurate manner, not only the normal operation of the power plant will be affected, but also serious economic losses may be caused. Therefore, the coal mill fault diagnosis method has important practical significance.

[0003] Traditional fault diagnosis methods mainly rely on experience and univariate analysis, and have the problems of low diagnosis accuracy and high false positive rate. With the development of sensing technology and information processing technology, multivariate time series analysis methods have been gradually applied in the field of fault diagnosis. Multivariate time series can comprehensively reflect the running state of equipment, and through joint analysis of multiple monitoring parameters, the accuracy and reliability of fault diagnosis can be improved. SUMMARY

[0004] In view of the above problems, the present application is proposed.

[0005] Therefore, the technical problem solved by the present application is that the existing coal mill fault diagnosis method has the problems of univariate analysis, low diagnosis accuracy, high false positive rate, and how to effectively improve the accuracy and reliability of coal mill fault diagnosis through multivariate time series classification.

[0006] To solve the above technical problems, the present application provides the following technical scheme: a coal mill fault diagnosis method based on multivariate time series classification, comprising collecting coal mill time series data, preprocessing the time series data and storing it in a data set; constructing a self-supervised contrast learning model to train different views in the data set; based on the output features of the self-supervised contrast learning model, input the features into a reinforced linear classifier for fault classification and diagnosis.

[0007] As a preferred scheme of the coal mill fault diagnosis method based on multivariate time series classification, the collection of coal mill time series includes the collection of time series of different fault running and normal running of the coal mill, and the collected time series includes the current of the coal mill, the outlet temperature of the coal mill, the inlet primary air pressure of the coal mill, the differential pressure of the coal mill, the primary air flow of the coal mill, the inlet temperature of the coal mill and the primary air temperature of the coal mill.

[0008] As a preferred scheme of the coal mill fault diagnosis method based on multivariate time series classification of the application, wherein: the storage to the data set includes length segmentation of the collected time series data to form continuous time series samples, Fourier transform and time series decomposition technology, and conversion of the time series samples into original time series views, frequency views and seasonal views; and the preprocessed data is classified and stored in the data set.

[0009] As a preferred scheme of the coal mill fault diagnosis method based on multivariate time series classification of the application, wherein: the construction of the self-supervised contrast learning model includes designing three neural networks with the same structure and not sharing parameters as feature extractors based on the original time series views, frequency views and seasonal views in the data set, and representing as:

[0010] g i =G(x i ), f i =F(x i ), s i =S(x i )

[0011] Wherein, G is the original time series view, F is the frequency view, S is the seasonal view, x i is the original data of the i-th time series sample, g i is the original time series view feature sample of the i-th time series sample, f i is the frequency view feature sample of the i-th time series sample, s i is the seasonal view feature sample of the i-th time series sample; based on the data set samples, the instance contrast loss function of the samples g i , f i and s i is constructed and represented as:

[0012]

[0013] Wherein, L Inst is the instance contrast loss function, L is the indicator function, exp is the exponential function, sim is the similarity function, N is the total number of data set samples, g i ′ is the enhanced sample of g i , f i ′ is the enhanced sample of f i , s i ′ is the enhanced sample of s i , k is the sample other than i, and t is the temperature parameter.

[0014] As a preferred scheme of the coal mill fault diagnosis method based on multivariate time series classification of the application, wherein: the contrast learning training includes constructing a prototype clustering method using joint learning, denoted as:

[0015]

[0016]

[0017] wherein, is the prototype clustering center vector of the category c in the original time series view feature, is the prototype clustering center vector of the category c in the frequency view feature, is the prototype clustering center vector of the category c in the seasonal view feature, is the i-th cluster center of the frequency view feature, is the i-th cluster center of the seasonal view feature, is the i-th cluster center of the original time series view feature; based on the prototype, the joint contrast learning is constructed sample g i , f i and s i The joint learning contrast loss function is denoted as:

[0018]

[0019] wherein, L Prot is the joint learning contrast loss function, is the joint clustering center of the original time series view and the seasonal view, is the joint clustering center of the seasonal view and the original time series view, is the joint clustering center of the frequency view and the seasonal view, is the joint clustering center of the original time series view and the frequency view, is the joint clustering center of the frequency view and the original time series view, is the joint clustering center of the seasonal view and the frequency view, M is the number of contrast samples of the i-th coal mill sample, M contains 2 positive samples and M-2 negative samples, and j is the contrast sample serial number; based on the instance contrast loss function and the joint learning contrast loss function, the overall objective function is constructed, denoted as:

[0020]

[0021] wherein, L' is the overall objective function, and λ is a parameter for balancing the two types of loss functions.

[0022] As a preferred scheme of the coal mill fault diagnosis method based on multivariate time series classification, the reinforced linear classifier comprises output features based on a contrast learning algorithm, the features are input into the reinforced linear classifier, the classification and diagnosis of the coal mill fault are completed, a linear classifier decision function is constructed based on logistic regression, and a predicted label is calculated, which is represented as:

[0023]

[0024] wherein, is the predicted label, w is a weight vector, b is a bias term, sigma is a sigmoid function, z is an input feature vector, and z comprises an original time series view feature vector, a frequency view vector and a seasonal view vector; a cross-entropy loss function is constructed to train the linear classifier, which is represented as:

[0025]

[0026] wherein, L" is a cross-entropy loss function, y is a real fault type label belonging to a normal data label.

[0027] As a preferred scheme of the coal mill fault diagnosis method based on multivariate time series classification, the fault classification and diagnosis comprises output results based on the reinforced linear classifier, and the coal mill operation state is divided into four modes of normal mode, coal mill coal blocking, coal mill coal breaking and coal mill internal fault.

[0028] Another object of the present application is to provide a coal mill fault diagnosis system based on multivariate time series classification, which can input features into a reinforced linear classifier based on output features of a self-supervised contrast learning model, perform fault classification and diagnosis, and solve the problems of low diagnosis accuracy and high false positive rate in the current coal mill fault diagnosis method.

[0029] As a preferred scheme of the coal mill fault diagnosis system based on multivariate time series classification, it comprises a data processing module, a self-supervised contrast learning module and a fault analysis module; the data processing module is used for collecting coal mill time series data, pre-processing the time series data and storing the time series data into a data set; the self-supervised contrast learning module is used for constructing a self-supervised contrast learning model and performing contrast learning training on different views in the data set; and the fault analysis module is used for inputting features into a reinforced linear classifier based on output features of the self-supervised contrast learning model, performing fault classification and diagnosis.

[0030] A computer device comprises a memory and a processor, the memory stores a computer program, and the processor executes the computer program to realize the steps of the coal mill fault diagnosis method based on multivariate time series classification.

[0031] A computer-readable storage medium having stored thereon a computer program, the computer program implementing the steps of the coal mill fault diagnosis method based on multivariate time series classification when executed by a processor.

[0032] The coal mill fault diagnosis method based on multivariate time series classification provided by the present application can more accurately identify the fault type and fault location of the coal mill by using the contrast learning technology and multivariate time series data analysis, thereby improving the accuracy and reliability of fault diagnosis. BRIEF DESCRIPTION OF DRAWINGS

[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0034] Figure 1 A multivariate time series classification-based coal mill fault diagnosis method provided for the first embodiment of the present application;

[0035] Figure 2 A multiview learning method structure diagram of the multivariate time series classification-based coal mill fault diagnosis method provided for the first embodiment of the present application;

[0036] Figure 3 A module schematic diagram of the multivariate time series classification-based coal mill fault diagnosis system provided for the third embodiment of the present application. DETAILED DESCRIPTION

[0037] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should fall within the scope of protection of the present application.

[0038] Embodiment 1, reference Figures 1-2 For an embodiment of the present application, a coal mill fault diagnosis method based on multivariate time series classification is provided, comprising:

[0039] S1: Collecting coal mill time series data, preprocessing time series data and storing it in a data set.

[0040] Further, collecting coal mill time series includes collecting time series of different fault operation and normal operation of the coal mill, and the collected time series includes coal mill current, coal mill outlet temperature, coal mill inlet primary air pressure, coal mill differential pressure, coal mill primary air flow, coal mill inlet temperature and coal mill primary air temperature.

[0041] It should be noted that storing in the data set includes length segmentation of the collected time series data to form continuous time series samples, Fourier transform and time series decomposition technology, converting the time series samples into original time series view, frequency view and seasonal view, and classifying and storing the preprocessed data in the data set.

[0042] It should also be noted that the frequency signal sample of the time series signal sample is generated by Fourier transform using frequency, and in addition to the frequency signal, the time signal is decomposed into trend, seasonal and residual signals by using time series decomposition technology method. In the model of the present application, the seasonal signal decomposed by the time series decomposition technology is selected as a new view together with the time signal, which allows comparative learning of the seasonal signal at the same time as the frequency signal and the time signal, thereby providing more sensitive insights into the time series. By preprocessing the collected data through Fourier transform and time series decomposition technology, the time series data is converted into original time series view, frequency view and seasonal view, providing multi-view input for the model, thereby improving the representation and learning ability of the model. The features related to the running state of the coal mill can be extracted, thereby improving the accuracy of fault diagnosis, reducing the influence of noise and outliers on fault diagnosis, and improving the precision and reliability of fault diagnosis.

[0043] S2: Constructing a self-supervised contrastive learning model to train different views in the data set.

[0044] Furthermore, the self-supervised contrastive learning model is constructed by designing three neural networks with identical structures but without sharing parameters, based on the original time-series view, frequency view, and seasonal view in the dataset, as feature extractors, represented as follows:

[0045] g i =G(x) i ),f i =F(x) i ),s i =S(x i )

[0046] Where G is the original time series view, F is the frequency view, S is the seasonal view, and x i For the original data of the i-th time series sample, g i f is the original time series view feature sample of the i-th time series sample. i For the frequency view feature sample of the i-th time series sample, s i Let g be the seasonal view feature sample of the i-th time series sample; based on the dataset samples, construct sample g. i f i and s i The instance contrast loss function is expressed as:

[0047]

[0048] Among them, L Inst Let L be the instance contrast loss function, exp be the indicator function, sim be the similarity function, N be the total number of samples in the dataset, and g be the similarity loss function. i ′ is the augmented sample of gi, f i ′ for f i Augmented samples, s i ′ for s i The enhanced samples are k, which are samples other than i, and t is the temperature parameter.

[0049] It should be noted that comparative learning training includes using joint learning to construct a prototype clustering method, expressed as:

[0050]

[0051] in, Let c be the prototype cluster center vector of category c under the original time series view features. Let c be the prototype cluster center vector of category c under the frequency view features. Let c be the prototype cluster center vector of category c under seasonal view features. Let i be the i-th cluster center of the frequency view features. the i-th cluster center of the seasonal view feature, the i-th cluster center of the original time series view feature; the joint contrast learning based on the prototype, constructing the sample g i , f i and s i The joint learning contrast loss function is represented as:

[0052]

[0053] Wherein, L Prot is the joint learning contrast loss function, is the joint cluster center of the original time series view and the seasonal view, is the joint cluster center of the seasonal view and the original time series view, is the joint cluster center of the frequency view and the seasonal view, is the joint cluster center of the original time series view and the frequency view, is the joint cluster center of the frequency view and the original time series view, is the joint cluster center of the seasonal view and the frequency view, M is the number of contrast samples of the i-th coal mill sample, M contains 2 positive samples and M-2 negative samples, and j is the sample serial number; based on the instance contrast loss function and the joint learning contrast loss function, the overall objective function is constructed, represented as:

[0054]

[0055] Wherein, L' is the overall objective function, and λ is the parameter for balancing the two types of loss functions.

[0056] It should also be noted that, as Figure 2 shown in the present application, three views of time signal, frequency signal and seasonal signal are used, instance-level contrast learning is used within a single view to obtain the hidden layer feature representation within the view, after obtaining the hidden layer feature representation of the three views, the contrast learning based on the prototype is used to jointly different view feature representations, which includes calculating the cluster index within the view, represented as v g , v f , v s , and the index from the other two views is calculated, represented as Three neural networks with the same structure and no shared parameters are designed as feature extractors, and the encoder of each individual view is composed of a 3-layer CNN pyramid network with a max pooling module added after each convolution block. The enhanced samples corresponding to each individual view are obtained by performing dropout after each convolution block. A self-supervised contrastive learning model is constructed, and through comprehensive analysis and contrastive learning of multivariate time series data, the diagnostic model is adaptively adjusted to improve the fault diagnosis effect under different working conditions and enhance the adaptability and stability of the method. The model learns the feature representation of time series data to capture the pattern differences of the coal mill under different operating conditions. Using normal data and different fault mode data, the model automatically learns the feature differences between data through a contrastive learning strategy without explicit labels. By maximizing the feature consistency of the same class data and minimizing the feature difference of different class data, the model can effectively learn the data patterns without labels.

[0057] S3: Based on the output features of the self-supervised contrastive learning model, the features are input into the enhanced linear classifier for fault classification and diagnosis.

[0058] Furthermore, the enhanced linear classifier includes outputting features based on the contrastive learning algorithm, inputting the features into the enhanced linear classifier, and completing the classification and diagnosis of the coal mill fault. Based on logistic regression, a linear classifier decision function is constructed to calculate the predicted label, which is represented as:

[0059]

[0060] wherein, is the predicted label, w is the weight vector, b is the bias term, σ is the sigmoid function, z is the input feature vector, and z includes the original time series view feature vector, the frequency view vector, and the seasonal view vector; a cross-entropy loss function is constructed to train the linear classifier, which is represented as:

[0061]

[0062] wherein, L" is the cross-entropy loss function, and y is the true fault type label belonging to the normal data label.

[0063] It should be noted that the fault classification and diagnosis includes dividing the coal mill operating state into four modes: normal mode, coal mill coal blockage, coal mill coal breakage, and coal mill internal fault based on the output results of the enhanced linear classifier.

[0064] It should be noted that the linear classifier is supervised learning trained by using the coal mill operation data labeled with labels, further optimizing the model, fully mining the correlation features in the multivariate time series data, conducting comprehensive analysis, improving the comprehensive understanding of the coal mill operation state, so as to realize more accurate fault diagnosis, so that it can accurately classify the operation state of the coal mill, and the trained linear classifier can classify the input coal mill data into four modes of normal mode, coal mill coal blocking, coal mill coal breaking and coal mill internal fault.

[0065] In example 2, an embodiment of the present application provides a coal mill fault diagnosis method based on multivariate time series classification. In order to verify the beneficial effects of the present application, economic benefit calculation and simulation experiment are used for scientific demonstration.

[0066] The MATLAB / Simulink software is used for simulation verification of the system. The parameters in the simulation are shown in Table 1.

[0067] Table 1 Simulation parameters

[0068] Parameter settings Parameter values Training data set (number) 1196 Test data set (number) 697 Feature dimension 256 Time series length 128 Number of variables (channels) 7 G, F, S backbone model Pyramid three-layer convolution Pyramid convolution first layer input / output (7，32) Pyramid convolution second layer input / output (32，64) Pyramid convolution third layer input / output (64，128) Pyramid convolution kernel size 7 Negative sample clipping rate 0.6

[0069] Referring to Table 2, the experimental data is recorded and analyzed.

[0070] Table 2 Experimental data record table

[0071] Comparison model Accuracy Precision Recall F1 score AUROC AUPRC KNN 0.7479 0.8437 0.8171 0.8334 0.8307 0.8428 TS2vec 0.9121 0.8228 0.8518 0.8215 0.9389 0.8278 TS-TCC 0.9548 0.9344 0.9094 0.9344 0.9767 0.9363 Our invention 0.9656 0.9698 0.9703 0.9536 0.9789 0.9367

[0072] By comparing the diagnostic accuracy, precision, recall, F1 score, AUROC and AUPRC of different fault types, the beneficial effects of the present invention can be seen, for example, the diagnostic accuracy of the present invention model for coal mill blockage, coal mill coal breakage, coal mill internal fault, etc. Different fault types are all more than 95%, compared with other comparative models (KNN: about 70%, TS2vec: about 90%, TS-TCC: about 95%), which has a significant improvement, which shows that the present invention model can more accurately identify the coal mill fault, and has high precision, recall and F1 score, which can effectively reduce the risk of misdiagnosis and missed diagnosis, improve the reliability of coal mill fault diagnosis, through this comprehensive evaluation, the potential fault type of the coal mill can be accurately identified, and targeted maintenance and maintenance can be carried out, thereby improving the operation efficiency and safety of the coal mill. The present invention is superior to the traditional coal mill fault diagnosis method, because the present invention not only considers the running state and fault characteristics of the coal mill, but also introduces multivariate time series analysis and self-supervised contrast learning technology, which can more comprehensively and accurately identify the potential fault of the coal mill, and through the enhanced linear classifier, accurate fault classification and diagnosis can be carried out. Compared with the prior art, the present invention provides a more advanced and reliable coal mill fault diagnosis method, which can more accurately identify the potential fault of the coal mill by integrating multivariate time series analysis and self-supervised contrast learning technology, and through the enhanced linear classifier, accurate fault classification and diagnosis can be carried out. This not only improves the efficiency of coal mill fault diagnosis, but also helps to improve the operation efficiency and safety of the coal mill, so the present invention has creativity.

[0073] Example 3, refer to Figure 3 For an embodiment of the present invention, a coal mill fault diagnosis system based on multivariate time series classification is provided, including a data processing module, a self-supervised contrast learning module, and a fault analysis module.

[0074] The data processing module is used to collect coal mill time series data, pre-process the time series data and store it in the data set; the self-supervised contrast learning module is used to construct a self-supervised contrast learning model, and the different views in the data set are trained by contrast learning; the fault analysis module is used to output features based on the self-supervised contrast learning model, input the features into the enhanced linear classifier, and perform fault classification and diagnosis

[0075] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0076] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0077] More specific examples (a non-exhaustive list) of computer-readable media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0078] It should be understood that portions of the present application can be implemented with hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented with software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, implementation can be with any or a combination of the following technologies, which are all well known in the art: a discrete logic circuit having logic gates for implementing logic functions upon an application of data signals, an application specific integrated circuit having appropriate combinational logic gates, a programmable gate array (PGA), a field programmable gate array (FPGA), etc. It should be understood that the foregoing embodiments are merely illustrative of the present application and are not to be used to limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, it will be apparent to those skilled in the art that various changes and modifications can be contributed to the present application without departing from the spirit and scope of the present application, and such changes and modifications should be encompassed within the scope of the appended claims.

[0079] It should be understood that the foregoing embodiments are merely illustrative of the present application and are not to be used to limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, it will be apparent to those skilled in the art that various changes and modifications can be contributed to the present application without departing from the spirit and scope of the present application, and such changes and modifications should be encompassed within the scope of the appended claims.

Claims

1. A coal mill fault diagnosis method based on multivariate time series classification, characterized by, The method comprises the following steps: Collecting time series data of the coal mill, preprocessing the time series data, and storing the time series data into a data set; The storing into the data set comprises length segmentation of the collected time series data to form continuous time series samples, Fourier transform and time series decomposition technology, conversion of the time series samples into original time series views, frequency views and seasonal views, and classification and storage of the preprocessed data in the data set; Building a self-supervised contrast learning model and performing contrast learning training on different views in the data set; The building of the self-supervised contrast learning model comprises designing three neural networks with the same structure and not sharing parameters as feature extractors based on the original time series views, frequency views and seasonal views in the data set, and representing the three neural networks as: g i = G(x i ), f i = F(x i ), s i = S(x i ) where G is the original time series view, F is the frequency view, S is the season view, x i is the original data for the i-th time series sample, g i is the original time series view feature sample for the i-th time series sample, f i is the frequency view feature sample for the i-th time series sample, s i is the season view feature sample for the i-th time series sample. Based on the dataset samples, construct a sample g i , f i and s i The instance contrast loss function is represented as: wherein L Inst is an example contrastive loss function, L is an indicator function, exp is an exponential function, sim is a similarity function, N is a total number of samples in a dataset, g i is an augmented sample of g i , f i ′ is an augmented sample of f i , s i ′ is an augmented sample of s i , k is a sample other than i, and t is a temperature parameter. Based on the output features of the self-supervised contrast learning model, the features are input into a boosted linear classifier to perform fault classification and diagnosis; The contrast learning training comprises building a prototype clustering method by joint learning, and representing the prototype clustering method as: wherein, is a prototype cluster center vector for class c in the original time series view feature, is a prototype cluster center vector for class c in the frequency view feature, is a prototype cluster center vector for class c in the seasonal view feature, is the i-th cluster center for the frequency view feature, is the i-th cluster center for the seasonal view feature, is the i-th cluster center for the original time series view feature; joint contrastive learning based on a prototype clustering method, constructing a sample g i , f i and s i joint learning contrastive loss function, denoted as: wherein L Prot is a joint learning contrastive loss function, is a joint clustering center of the original time series view and the seasonal view, is a joint clustering center of the seasonal view and the original time series view, is a joint clustering center of the frequency view and the seasonal view, is a joint clustering center of the original time series view and the frequency view, is a joint clustering center of the frequency view and the original time series view, is a joint clustering center of the seasonal view and the frequency view, M is the number of contrastive samples of the i-th coal mill sample, M contains 2 positive samples and M-2 negative samples, j is the serial number of the contrastive sample, denotes a sample taken in class c; Based on the instance contrast loss function and the joint learning contrast loss function, an overall objective function is built, and represented as: Wherein, L' is the overall objective function, and λ is a parameter for balancing the two types of loss functions.

2. The multivariate time series classification based coal mill fault diagnosis method of claim 1, wherein: The collecting of the time series of the coal mill comprises collecting time series of different fault operations and normal operations of the coal mill, and the collected time series comprises coal mill current, coal mill outlet temperature, coal mill inlet primary air pressure, coal mill differential pressure, coal mill primary air flow, coal mill inlet temperature and coal mill primary air temperature.

3. The multivariate time series classification based coal mill fault diagnosis method of claim 2, wherein: The boosted linear classifier comprises inputting the features output by the contrast learning algorithm into the boosted linear classifier to complete the classification and diagnosis of the coal mill faults, building a decision function of the boosted linear classifier based on logistic regression, and calculating a predicted label, and the decision function of the boosted linear classifier is represented as: wherein, is the predicted label, w is the weight vector, b is the bias term, σ is the sigmoid function, z is the input feature vector, z comprises the original time series view feature vector, the frequency view vector, and the season view vector; A cross-entropy loss function is built to train the boosted linear classifier, and represented as: where L" is a cross-entropy loss function, y i is the true fault type label belongs to normal data label.

4. The multivariate time series classification based coal mill fault diagnostic method as claimed in claim 3, wherein: The fault classification and diagnosis comprises dividing the running state of the coal mill into four modes of normal mode, coal mill coal blocking, coal mill coal breaking and coal mill internal fault based on the output results of the boosted linear classifier.

5. A system employing the coal mill fault diagnostic method based on multivariate time series classification according to any one of claims 1 to 4, characterized in that: The method comprises a data processing module, a self-supervised contrast learning module and a fault analysis module; The data processing module is used for collecting time series data of the coal mill, preprocessing the time series data, and storing the time series data into a data set; The self-supervised contrast learning module is used for building a self-supervised contrast learning model and performing contrast learning training on different views in the data set; The fault analysis module is used for inputting features output by the self-supervised contrast learning model into a boosted linear classifier to perform fault classification and diagnosis. 6.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-5 when the computer program is executed by the processor. The processor executes the computer program to realize the steps of the coal mill fault diagnosis method based on multivariate time series classification in any one of claims 1 to 4.

7. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the coal mill fault diagnosis method based on multivariate time series classification in any one of claims 1 to 4.

Images