Coal mill fault diagnosis and prediction method and system based on big data analysis

Abstract

The invention provides a coal mill fault diagnosis and prediction method and system based on big data analysis, and relates to the technical field of big data analysis, and the method comprises the steps: collecting a vibration signal of a coal mill bearing through a vibration sensor, carrying out the multi-scale analysis through wavelet decomposition, extracting features, constructing a deep convolutional neural network model, and carrying out the fault diagnosis and prediction of a coal mill bearing. A maximum mean value difference adaptive layer is introduced to carry out feature alignment on data of a source domain and a target domain, identification of fault types and degrees is realized, a health state evaluation index including a fault severity coefficient and a development trend coefficient is established, and the remaining service life is calculated and early warning is carried out. According to the invention, the accuracy and generalization ability of fault diagnosis are improved, accurate prediction and timely early warning of the bearing fault of the coal mill are realized, and the service life of equipment is prolonged.

Description

Technical Field

[0001] The present invention relates to the technical field of big data analysis, and in particular to a coal mill fault diagnosis and prediction method and system based on big data analysis. Background Art

[0002] Coal mill is a key equipment in the coal powder preparation system of thermal power plants, and its operating status directly affects the safe and stable operation of power plants. Bearings are the core components of coal mills, and their failure rate is relatively high. If they are not discovered and handled in time, it will lead to equipment shutdown and even cause safety accidents. Therefore, real-time monitoring and fault diagnosis and prediction of coal mill bearings are of great significance; The existing coal mill fault diagnosis methods have the following main problems: 1. Traditional fault diagnosis methods mainly rely on manual experience and judgment, which makes it difficult to accurately identify faults and provide timely warnings; 2. Conventional machine learning models have poor generalization ability when processing fault data under different working conditions and are difficult to adapt to complex and changing working environments; 3. Existing life prediction methods often ignore the dynamic characteristics of fault development and cannot accurately evaluate the health status and remaining service life of the equipment; The technical problem to be solved by the present invention is to provide a coal mill fault diagnosis and prediction method based on big data analysis. Through deep learning and transfer learning technology, intelligent diagnosis and accurate prediction of coal mill bearing faults can be achieved, the accuracy and reliability of fault diagnosis can be improved, and a scientific basis can be provided for equipment maintenance and management. Summary of the invention

[0003] The embodiments of the present invention provide a coal mill fault diagnosis and prediction method and system based on big data analysis, which can solve the problems in the prior art.

[0004] According to a first aspect of the embodiments of the present invention, Provide coal mill fault diagnosis and prediction methods based on big data analysis, including: The vibration signal of the coal mill bearing is collected in real time by using a vibration sensor, and the vibration signal is subjected to multi-scale analysis by wavelet decomposition to extract the time domain features, frequency domain features and time-frequency features of the vibration signal, and the time domain features, frequency domain features and time-frequency features are normalized to obtain a feature vector matrix; a training data set of a deep convolutional neural network model is constructed based on the feature vector matrix, and the training data set is divided into a source domain data set and a target domain data set; Constructing a deep convolutional neural network model including a convolutional layer, a pooling layer, and a fully connected layer, wherein the input layer of the deep convolutional neural network model receives the feature vector matrix, and the output layer outputs the fault type and fault degree; introducing a maximum mean difference adaptive layer into the deep convolutional neural network model, and adaptively aligning the feature distributions of the source domain data set and the target domain data set through the maximum mean difference adaptive layer; pre-training the deep convolutional neural network model using the source domain data set, and fine-tuning the training using the target domain data set; The real-time collected vibration signal of the coal mill bearing is input into the trained deep convolutional neural network model to obtain the fault type and fault degree of the coal mill bearing; based on the fault type and the fault degree, a health status assessment index of the coal mill bearing is established, and the health status assessment index includes a fault severity coefficient and a fault development trend coefficient; the remaining service life of the coal mill bearing is calculated according to the health status assessment index, and a warning signal is issued when the remaining service life is lower than a preset health threshold.

[0005] Introducing a maximum mean difference adaptive layer into the deep convolutional neural network model, and adaptively aligning the feature distributions of the source domain dataset and the target domain dataset through the maximum mean difference adaptive layer includes: A Gaussian kernel function is used to construct a maximum mean difference adaptive layer; a multi-core mapping function is constructed based on the Gaussian kernel function, wherein the multi-core mapping function is obtained by linearly combining multiple Gaussian kernel functions with different bandwidths, each of the Gaussian kernel functions corresponds to a weight coefficient, and the sum of the weight coefficients is 1; Extracting deep feature representations of a source domain dataset and a target domain dataset, calculating channel attention weights of the source domain dataset and the target domain dataset, performing weighted processing on the deep feature representations according to the channel attention weights to obtain weighted source domain features and target domain features; inputting the weighted source domain features and the weighted target domain features into the multi-core mapping function, and calculating the maximum mean difference distance between the source domain features and the target domain features; The product of the maximum mean difference distance and the dynamic balance parameter is taken as the maximum mean difference loss, and the dynamic balance parameter gradually increases from an initial value to a preset value as the number of training rounds increases; the maximum mean difference loss is added to the total loss function of the deep neural network, and the network parameters of the deep neural network and the weight coefficient of the multi-core mapping function are updated through the back propagation algorithm until the maximum mean difference loss is minimized, thereby achieving adaptive alignment of source domain features and target domain features.

[0006] Inputting the real-time collected vibration signal of the coal mill bearing into the trained deep convolutional neural network model to obtain the fault type and fault degree of the coal mill bearing includes: The Softmax activation function is used to calculate the probability distribution of the output layer of the deep convolutional neural network model, and the temperature scaling parameter is introduced to calibrate the Softmax output to obtain the calibrated fault type probability distribution; the deep convolutional neural network model is sampled multiple times by the Monte Carlo dropout method to obtain the distribution of fault degree prediction values, the mean and standard deviation are calculated based on the distribution of fault degree prediction values, and the fault degree confidence interval is constructed; A priori probability distribution model of fault type is established based on Dirichlet distribution, and the influence of time decay factor on historical observation data is introduced to dynamically adjust, the calibrated fault type probability distribution is input into the priori probability distribution model of fault type, and the posterior probability is calculated using Bayesian formula to update the fault type probability; the fault degree confidence interval is combined with the updated fault type probability to obtain a comprehensive fault probability; The comprehensive fault probability is smoothed in time series by using an exponentially weighted moving average method to obtain a smoothed fault probability sequence; a working condition adjustment factor is constructed based on the equipment load, rotation speed and temperature, and the working condition adjustment factor changes dynamically with the equipment operation state; a local standard deviation of the smoothed fault probability sequence is weightedly combined with the working condition adjustment factor to generate a dynamic fault threshold; Perform continuity judgment and trend analysis on the smoothed fault probability sequence, and trigger an initial alarm signal when the fault probabilities at multiple consecutive time points exceed the dynamic fault threshold and show an upward trend; collect diagnostic results of multiple sensors, calculate the consistency of the diagnostic results of the multiple sensors, and convert the initial alarm signal into a final alarm signal when the consistency exceeds a preset consistency threshold; and perform fault diagnosis on the coal mill bearing based on the final alarm signal.

[0007] Based on the fault type and the fault degree, a health status evaluation index of the coal mill bearing is established, wherein the health status evaluation index includes a fault severity coefficient including: Based on the historical maintenance data of the coal mill, the production decline rate, energy consumption increase rate, and vibration amplitude change rate when the fault occurs are collected to construct a performance loss index, and the performance loss index is converted into a performance impact score through the principal component analysis method; the stress concentration area is determined based on the bearing stress-strain analysis, the crack propagation rate is calculated in combination with the material fatigue characteristics, and a Markov chain model is established to obtain the fault propagation probability value; the spare parts replacement cost, labor cost, equipment downtime loss, associated fault repair expenditure, and production quality loss are fuzzy comprehensively evaluated to obtain a relative cost coefficient; the hierarchical analysis method is used to combine the performance impact score, the fault propagation probability value, and the relative cost coefficient to form a fault type weight matrix; Constructing a piecewise nonlinear normalization function, extracting time domain features, frequency domain features, and time-frequency domain features from the vibration signal, combining the time domain features, the frequency domain features, the time-frequency domain features, temperature parameters, and acoustic emission parameters into a feature vector, using principal component analysis to reduce the dimension of the feature vector, and establishing a mapping relationship from the feature vector to the fault degree through support vector regression; The fault type weight matrix, fault diagnosis probability and fault degree processed by the piecewise nonlinear normalization function are weightedly combined to obtain an initial fault severity coefficient; the initial fault severity coefficient is exponentially smoothed, and the smoothing coefficient is determined based on the minimum prediction error criterion to obtain a final fault severity coefficient, and the final fault severity coefficient is used as a basis for fault assessment of a coal mill bearing.

[0008] Based on the fault type and the fault degree, a health status evaluation index of the coal mill bearing is established. The health status evaluation index includes a fault development trend coefficient including: Determine the length of the sliding time window through autocorrelation analysis, and calculate the first-order difference sequence and the second-order difference sequence of the fault severity coefficient within the sliding time window; determine the polynomial fitting order based on the Akaike information criterion, perform local weighted polynomial fitting on the first-order difference sequence and the second-order difference sequence, and obtain a fault development trend curve; Collect the speed parameters, load rate parameters and feed rate parameters of the coal mill, perform dimensionality reduction processing on the speed parameters, load rate parameters and feed rate parameters through principal component analysis to obtain the working condition characteristics, establish a multivariate regression model of the working condition characteristics and the fault development rate, and calculate the ratio of the actual working condition development rate to the standard working condition development rate to obtain the working condition compensation factor; Constructing a fuzzy neural network, inputting temperature parameters and humidity parameters into the input layer of the fuzzy neural network, using Gaussian fuzzy membership function to perform fuzzy reasoning in the hidden layer of the fuzzy neural network, obtaining an environmental impact coefficient in the output layer of the fuzzy neural network, and using a loss function with a regularization term for training the fuzzy neural network; Performing exponential weighted averaging on the first-order difference sequence to obtain short-term variation characteristics, calculating the derivative of the fault development trend curve to obtain long-term trend characteristics, and multiplying the working condition compensation factor by the environmental impact coefficient to obtain a comprehensive compensation term; The short-term change characteristics, the long-term trend characteristics, and the comprehensive compensation items are weightedly combined to obtain an initial trend coefficient, and the Kalman filter algorithm is used to eliminate noise from the initial trend coefficient to obtain a final fault development trend coefficient.

[0009] Calculating the remaining service life of the coal mill bearing according to the health status assessment index includes: The fault severity coefficient and the fault development trend coefficient are nonlinearly weighted and fused, and the cross term of the fault severity coefficient and the fault development trend coefficient is introduced to establish a comprehensive health index model, wherein the comprehensive health index model includes a linear weight coefficient of the fault severity coefficient, a linear weight coefficient of the fault development trend coefficient, and a cross term weight coefficient; A Weibull process model is established based on the comprehensive health index, and a bearing force analysis model is established to calculate the stress distribution under different working conditions. The life loss caused by changes in working conditions is calculated based on Miner's linear cumulative damage theory. The life loss is input into the Weibull process model as a correction term, and a remaining life prediction equation considering the influence of working conditions is established to calculate the remaining service life of the coal mill bearing.

[0010] The method further comprises: Taking the predicted value of the remaining life prediction equation as the state variable and the health status assessment index as the observation variable, respectively establishing a state equation containing process noise and an observation equation containing measurement noise, and constructing an adaptive particle filter correction model; Constructing an adaptive noise covariance matrix, wherein the adaptive noise covariance matrix adopts a recursive form with a forgetting factor, controls the decay rate of historical information through the forgetting factor, constructs an innovation covariance term based on the prediction error at the current moment, and dynamically estimates the process noise in the state equation and the measurement noise in the observation equation; Inputting the adaptive noise covariance matrix into the adaptive particle filter correction model, correcting the state prediction, calculating the likelihood value of each particle based on the observation data of the observation equation, and updating the weight of the particle; The updated particle weights are used to perform weighted averaging on the state estimation values ​​to obtain a revised prediction value of the remaining life of the coal mill bearing, wherein the revised prediction value of the remaining life of the coal mill bearing includes the influence of operating condition changes and random fluctuation characteristics during equipment operation.

[0011] According to a second aspect of the embodiments of the present invention, Provide coal mill fault diagnosis and prediction system based on big data analysis, including: The first unit is used to collect the vibration signal of the coal mill bearing in real time by using a vibration sensor, and perform multi-scale analysis on the vibration signal by wavelet decomposition, extract the time domain characteristics, frequency domain characteristics and time-frequency characteristics of the vibration signal, and normalize the time domain characteristics, frequency domain characteristics and time-frequency characteristics to obtain a feature vector matrix; construct a training data set of a deep convolutional neural network model based on the feature vector matrix, and divide the training data set into a source domain data set and a target domain data set; The second unit is used to construct a deep convolutional neural network model including a convolutional layer, a pooling layer and a fully connected layer, wherein the input layer of the deep convolutional neural network model receives the feature vector matrix, and the output layer outputs the fault type and fault degree; a maximum mean difference adaptive layer is introduced into the deep convolutional neural network model, and the feature distributions of the source domain data set and the target domain data set are adaptively aligned through the maximum mean difference adaptive layer; the deep convolutional neural network model is pre-trained using the source domain data set, and fine-tuned using the target domain data set; The third unit is used to input the real-time collected vibration signal of the coal mill bearing into the trained deep convolutional neural network model to obtain the fault type and fault degree of the coal mill bearing; establish the health status assessment index of the coal mill bearing based on the fault type and the fault degree, and the health status assessment index includes a fault severity coefficient and a fault development trend coefficient; calculate the remaining service life of the coal mill bearing according to the health status assessment index, and issue a warning signal when the remaining service life is lower than a preset health threshold.

[0012] According to a third aspect of the embodiments of the present invention, An electronic device is provided, comprising: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0013] A fourth aspect of the embodiments of the present invention is: A computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the aforementioned method is implemented.

[0014] The beneficial effects of this application are as follows: 1. Improve the accuracy of diagnosis and prediction: By combining multi-scale feature extraction of wavelet decomposition and deep convolutional neural network, the fault information contained in the vibration signal of coal mill bearing can be mined more comprehensively and deeply. In particular, the adaptive layer of maximum mean difference is used to adaptively align the distribution of source domain and target domain data features, which effectively solves the model generalization problem caused by data distribution differences under different working conditions, thereby improving the accuracy of fault diagnosis and prediction.

[0015] 2. Realize the health status assessment and life prediction of coal mill bearings: construct health status assessment indicators (fault severity coefficient and fault development trend coefficient) based on fault type, fault degree, etc., and calculate the remaining service life based on this, so as to realize a comprehensive assessment of the health status of coal mill bearings and predict their remaining service life, providing a scientific basis for formulating reasonable maintenance plans.

[0016] 3. Improve the reliability and safety of equipment operation and reduce maintenance costs: Through real-time monitoring and fault warning functions, potential faults of coal mill bearings can be discovered in advance, avoiding unexpected downtime and major economic losses caused by faults, and improving the reliability and safety of equipment operation. At the same time, by predicting the remaining service life, maintenance strategies can be optimized to avoid excessive maintenance, thereby reducing maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic diagram of a flow chart of a coal mill fault diagnosis and prediction method based on big data analysis according to an embodiment of the present invention; Figure 2 It is a structural schematic diagram of a coal mill fault diagnosis and prediction system based on big data analysis according to an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0019] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0020] Figure 1 FIG. 1 is a flow chart of a method for diagnosing and predicting coal mill faults based on big data analysis according to an embodiment of the present invention. Figure 1 As shown, the method includes: S101. Use a vibration sensor to collect the vibration signal of the coal mill bearing in real time, and perform multi-scale analysis on the vibration signal through wavelet decomposition to extract the time domain characteristics, frequency domain characteristics and time-frequency characteristics of the vibration signal, and normalize the time domain characteristics, frequency domain characteristics and time-frequency characteristics to obtain a feature vector matrix; construct a training data set of a deep convolutional neural network model based on the feature vector matrix, and divide the training data set into a source domain data set and a target domain data set; S102. Construct a deep convolutional neural network model including a convolutional layer, a pooling layer and a fully connected layer, wherein the input layer of the deep convolutional neural network model receives the feature vector matrix, and the output layer outputs the fault type and fault degree; introduce a maximum mean difference adaptive layer into the deep convolutional neural network model, and adaptively align the feature distributions of the source domain data set and the target domain data set through the maximum mean difference adaptive layer; pre-train the deep convolutional neural network model using the source domain data set, and perform fine-tuning training using the target domain data set; S103. Input the coal mill bearing vibration signal collected in real time into the trained deep convolutional neural network model to obtain the fault type and fault degree of the coal mill bearing; establish the health status assessment index of the coal mill bearing based on the fault type and the fault degree, and the health status assessment index includes a fault severity coefficient and a fault development trend coefficient; calculate the remaining service life of the coal mill bearing according to the health status assessment index, and issue a warning signal when the remaining service life is lower than a preset health threshold.

[0021] The specific implementation of the coal mill fault diagnosis and prediction method is as follows: First, vibration sensors, such as acceleration sensors, are placed at key locations of the coal mill bearings to collect the vibration signals of the coal mill in real time. The collection frequency of the vibration signals is set to 20kHz and the sampling time is 10 seconds.

[0022] Next, the collected vibration signal is preprocessed, including removing noise and outliers. The wavelet decomposition method is used to perform multi-scale analysis on the vibration signal. Select a suitable mother wavelet, such as the db4 wavelet, and perform a 5-layer decomposition of the vibration signal. Extract time domain features, frequency domain features, and time-frequency features from the decomposed wavelet coefficients of different frequency bands. Time domain features include mean, standard deviation, peak, kurtosis, etc. Frequency domain features include energy and frequency center of different frequency bands. Time-frequency features include wavelet energy entropy, wavelet packet energy entropy, etc. For example, the features extracted from the original vibration signal include a mean of 0.5g, a standard deviation of 0.1g, a kurtosis of 3.2, a 1kHz band energy of 0.8J, and a wavelet energy entropy of 0.6. All extracted features are normalized and scaled to between 0 and 1 to form a feature vector matrix.

[0023] Then, the normalized eigenvector matrix is ​​used to construct the training data set of the deep convolutional neural network model. The data set is divided into a source domain data set and a target domain data set. The source domain data set contains a large amount of vibration data of the coal mill under normal operation, as well as vibration data under some common fault conditions, such as bearing inner ring fault, outer ring fault, rolling element fault, etc. The target domain data set contains a small amount of vibration data of the coal mill to be diagnosed, which may contain unknown fault types. Assume that the source domain data set contains 10,000 samples and the target domain data set contains 100 samples.

[0024] A deep convolutional neural network model is constructed. The model contains multiple convolutional layers, pooling layers, and fully connected layers. The convolutional layer is used to extract features, the pooling layer is used to reduce the feature dimension, and the fully connected layer is used for classification. A maximum mean difference adaptive layer is introduced in the model to align the feature distribution of the source domain dataset and the target domain dataset to reduce the impact of domain shift. For example, the model contains 3 convolutional layers, each of which contains 32 convolution kernels with a convolution kernel size of 3x3. The model also contains 2 pooling layers, using maximum pooling, and the pooling window size is 2x2. Finally, the model contains 2 fully connected layers, the first fully connected layer contains 128 neurons, and the second fully connected layer contains the number of neurons in the output layer. The number of neurons in the output layer is the same as the number of fault types, for example, 4, representing normal, inner race fault, outer race fault, and rolling element fault, respectively.

[0025] The source domain dataset is used to pre-train the deep convolutional neural network model. The Adam optimizer is used, the learning rate is set to 0.001, the batch size is set to 32, and the number of training rounds is set to 100. After the pre-training is completed, the model is fine-tuned using the target domain dataset. The learning rate of the fine-tuning training is set to 0.0001, and the number of training rounds is set to 50.

[0026] The real-time collected vibration signal of the coal mill bearing is input into the trained deep convolutional neural network model to obtain the fault type and fault degree of the coal mill bearing. For example, the model output result is [0.1, 0.8, 0.05, 0.05], which means that there is an 80% probability that the bearing has an inner ring fault.

[0027] According to the fault type and fault severity, the health status assessment index of the coal mill bearing is established, including the fault severity coefficient and the fault development trend coefficient. The fault severity coefficient indicates the severity of the current fault, and the value range is 0 to 1. The larger the value, the more serious the fault. The fault development trend coefficient indicates the speed of fault development, and the value range is -1 to 1. A positive value indicates that the fault is getting worse, and a negative value indicates that the fault is getting better. For example, according to the model output results and historical data, the fault severity coefficient is calculated to be 0.7 and the fault development trend coefficient is 0.2.

[0028] The remaining service life of the coal mill bearing is calculated based on the health status assessment indicators. For example, based on the fault severity coefficient and fault development trend coefficient, as well as the coal mill operation history data, the remaining service life of the bearing is predicted to be 1,000 hours.

[0029] When the remaining service life is lower than the preset health threshold, such as 500 hours, an early warning signal is issued to remind maintenance personnel to perform maintenance in a timely manner.

[0030] Beneficial effects: 1. Improve the accuracy of fault diagnosis: Through the deep convolutional neural network model and the maximum mean difference adaptive layer, the fault features can be effectively extracted, the impact of domain offset can be reduced, and the accuracy of fault diagnosis can be improved.

[0031] 2. Realize fault prediction and health management: Through health status assessment indicators and remaining service life prediction, fault prediction and health management of coal mill bearings can be realized to avoid unexpected downtime.

[0032] 3. Reduce maintenance costs: Through early warning, maintenance personnel can be guided to carry out targeted maintenance, avoid excessive maintenance, and reduce maintenance costs.

[0033] In an optional implementation, introducing a maximum mean difference adaptive layer into the deep convolutional neural network model, and adaptively aligning the feature distributions of the source domain dataset and the target domain dataset through the maximum mean difference adaptive layer includes: A Gaussian kernel function is used to construct a maximum mean difference adaptive layer; a multi-core mapping function is constructed based on the Gaussian kernel function, wherein the multi-core mapping function is obtained by linearly combining multiple Gaussian kernel functions with different bandwidths, each of the Gaussian kernel functions corresponds to a weight coefficient, and the sum of the weight coefficients is 1; Extracting deep feature representations of a source domain dataset and a target domain dataset, calculating channel attention weights of the source domain dataset and the target domain dataset, performing weighted processing on the deep feature representations according to the channel attention weights to obtain weighted source domain features and target domain features; inputting the weighted source domain features and the weighted target domain features into the multi-core mapping function, and calculating the maximum mean difference distance between the source domain features and the target domain features; The product of the maximum mean difference distance and the dynamic balance parameter is taken as the maximum mean difference loss, and the dynamic balance parameter gradually increases from an initial value to a preset value as the number of training rounds increases; the maximum mean difference loss is added to the total loss function of the deep neural network, and the network parameters of the deep neural network and the weight coefficient of the multi-core mapping function are updated through the back propagation algorithm until the maximum mean difference loss is minimized, thereby achieving adaptive alignment of source domain features and target domain features.

[0034] In order to better achieve adaptive alignment of feature distributions of source domain datasets and target domain datasets, a deep convolutional neural network model based on the maximum mean difference (MMD) adaptive layer is proposed. By introducing multi-core mapping functions and channel attention mechanisms, the model effectively reduces the distribution difference between the source domain and the target domain, and improves the generalization performance of the model in the target domain.

[0035] First, a multi-core maximum mean difference adaptive layer is constructed. The adaptive layer uses a linear combination of multiple Gaussian kernel functions with different bandwidths to obtain a multi-core mapping function. Each Gaussian kernel function corresponds to a weight coefficient, and the sum of all weight coefficients is 1. For example, three Gaussian kernel functions with bandwidths of 1, 5, and 10 can be selected, and the initial weight coefficients are set to 0.2, 0.3, and 0.5 respectively. These parameters can be adjusted according to actual conditions.

[0036] Then, deep feature representations are extracted from the source domain dataset and the target domain dataset respectively. Assume that the source domain dataset has 1000 samples, the target domain dataset has 500 samples, and the extracted feature dimension is 256. Then, the channel attention weights of the source domain dataset and the target domain dataset are calculated respectively. The channel attention weights can be calculated by global average pooling or other methods. For example, the features of each channel are averagely pooled to obtain 256 values, and then passed through a fully connected layer and a Sigmoid activation function to obtain the attention weight of each channel, which ranges from 0 to 1.

[0037] According to the calculated channel attention weights, the deep feature representations of the source domain and the target domain are weighted. The features of each channel are multiplied by the corresponding channel attention weight to obtain the weighted source domain features and target domain features.

[0038] The weighted source domain features and target domain features are input into the constructed multi-core mapping function to calculate the maximum mean difference distance between the source domain features and the target domain features. This distance represents the degree of difference between the distribution of the two domain features.

[0039] In order to balance the impact of the maximum mean difference loss and other losses (such as classification loss), a dynamic balance parameter is introduced. This parameter gradually increases from the initial value to the preset value as the number of training rounds increases. For example, the initial value is set to 0.1, the preset value is set to 1, and as the training progresses, the dynamic balance parameter gradually increases linearly from 0.1 to 1. The product of the maximum mean difference distance and the dynamic balance parameter is taken as the maximum mean difference loss.

[0040] The maximum mean difference loss is added to the total loss function of the deep neural network. The network parameters of the deep neural network and the weight coefficients of the multi-core mapping function are updated through the back-propagation algorithm until the maximum mean difference loss is minimized. Assuming that the total loss function consists of the classification loss and the maximum mean difference loss, by minimizing the total loss function, the classification performance and domain adaptation ability of the model can be optimized simultaneously.

[0041] Through the above steps, adaptive alignment of source domain features and target domain features can be achieved, thereby improving the generalization performance of the model in the target domain.

[0042] The beneficial effects can be summarized into the following three aspects: 1. Improved the cross-domain generalization ability of the model: By minimizing the difference between the feature distributions of the source domain and the target domain, the performance degradation problem caused by domain shift is effectively solved, and the generalization performance of the model in the target domain is improved.

[0043] 2. Enhanced model robustness: The multi-core maximum mean difference adaptation layer can better capture feature differences at different scales, making the model more robust to different types of domain shifts.

[0044] 3. Optimized the training efficiency of the model: The channel attention mechanism can highlight important feature channels and reduce the impact of noise, thereby accelerating the training convergence of the model.

[0045] In an optional implementation, inputting the real-time collected coal mill bearing vibration signal into a trained deep convolutional neural network model to obtain the fault type and fault degree of the coal mill bearing includes: The Softmax activation function is used to calculate the probability distribution of the output layer of the deep convolutional neural network model, and the temperature scaling parameter is introduced to calibrate the Softmax output to obtain the calibrated fault type probability distribution; the deep convolutional neural network model is sampled multiple times by the Monte Carlo dropout method to obtain the distribution of fault degree prediction values, the mean and standard deviation are calculated based on the distribution of fault degree prediction values, and the fault degree confidence interval is constructed; A priori probability distribution model of fault type is established based on Dirichlet distribution, and the influence of time decay factor on historical observation data is introduced to dynamically adjust, the calibrated fault type probability distribution is input into the priori probability distribution model of fault type, and the posterior probability is calculated using Bayesian formula to update the fault type probability; the fault degree confidence interval is combined with the updated fault type probability to obtain a comprehensive fault probability; The comprehensive fault probability is smoothed in time series by using an exponentially weighted moving average method to obtain a smoothed fault probability sequence; a working condition adjustment factor is constructed based on the equipment load, rotation speed and temperature, and the working condition adjustment factor changes dynamically with the equipment operation state; a local standard deviation of the smoothed fault probability sequence is weightedly combined with the working condition adjustment factor to generate a dynamic fault threshold; Perform continuity judgment and trend analysis on the smoothed fault probability sequence, and trigger an initial alarm signal when the fault probabilities at multiple consecutive time points exceed the dynamic fault threshold and show an upward trend; collect diagnostic results of multiple sensors, calculate the consistency of the diagnostic results of the multiple sensors, and convert the initial alarm signal into a final alarm signal when the consistency exceeds a preset consistency threshold; and perform fault diagnosis on the coal mill bearing based on the final alarm signal.

[0046] The training process of the deep convolutional neural network model is as follows: First, the vibration signals of the coal mill bearings under different working conditions and fault states are collected, and data preprocessing is performed, such as denoising and normalization. Then, a deep convolutional neural network model is constructed, which contains multiple convolutional layers, pooling layers, and fully connected layers. The model is trained using the preprocessed vibration signal data, and the model parameters are optimized using appropriate loss functions and optimization algorithms, so that the model can accurately identify the fault type and fault degree of the coal mill bearings.

[0047] Next, the real-time collected coal mill bearing vibration signal is input into the trained deep convolutional neural network model. The output layer of the model uses the Softmax activation function to calculate the probability distribution of the fault type. In order to improve the reliability of the model output, the temperature scaling parameter is introduced to calibrate the Softmax output. The function of the temperature scaling parameter is to control the smoothness of the probability distribution. A smaller temperature scaling parameter will make the probability distribution sharper, and vice versa. For example, assuming that the original probability distribution of the model output is [0.2, 0.3, 0.5], after calibration with a temperature scaling parameter of 0.5, the probability distribution becomes [0.15, 0.25, 0.6].

[0048] In order to evaluate the uncertainty of the fault degree, the Monte Carlo dropout method is used to sample the deep convolutional neural network model multiple times. Specifically, when the model makes a prediction, a part of the neurons are randomly disabled, and then multiple predictions are made to obtain multiple fault degree prediction values. For example, 100 samples are taken to obtain 100 fault degree prediction values. Based on these prediction values, their mean and standard deviation are calculated, and a confidence interval for the fault degree is constructed. For example, if the mean of the 100 prediction values ​​is 0.8 and the standard deviation is 0.05, a 95% confidence interval of [0.7, 0.9] can be constructed.

[0049] In order to combine historical observation data, a fault type prior probability distribution model based on Dirichlet distribution is established. The model can estimate the probability of different fault types based on historical data. In order to dynamically adjust the impact of historical data, a time decay factor is introduced. The time decay factor will gradually decrease over time, making the recent data have a greater impact on the prior probability. For example, assuming that the current time is t, the time decay factor can be set to exp(-λ(t-t0)), where λ is the decay rate and t0 is the initial time. The calibrated fault type probability distribution is input into the fault type prior probability distribution model, and the posterior probability is calculated using the Bayesian formula to update the fault type probability. For example, assuming that the prior probability is [0.1, 0.2, 0.7] and the calibrated probability distribution is [0.15, 0.25, 0.6], the posterior probability can be calculated as [0.12, 0.22, 0.66].

[0050] Combine the fault level confidence interval with the updated fault type probability to get the comprehensive fault probability. For example, assuming the fault types are A, B, and C, the corresponding updated probabilities are 0.12, 0.22, and 0.66, respectively, and the fault level confidence interval is [0.7, 0.9]. The comprehensive fault probability can be expressed as: A: [0.084, 0.108], B: [0.154, 0.198], C: [0.462, 0.594].

[0051] The exponentially weighted moving average method is used to perform time series smoothing on the comprehensive failure probability to obtain a smoothed failure probability sequence. For example, assuming that the comprehensive failure probability at the current time point is 0.8, the smoothed failure probability at the previous time point is 0.7, and the attenuation factor is 0.9, then the smoothed failure probability at the current time point is 0.79.

[0052] The operating condition adjustment factor is constructed based on the operating condition parameters such as equipment load, speed and temperature. The operating condition adjustment factor changes dynamically with the equipment operation status. For example, the higher the equipment load, the larger the operating condition adjustment factor. The local standard deviation of the smoothed fault probability sequence is weighted and combined with the operating condition adjustment factor to generate a dynamic fault threshold. For example, if the local standard deviation is 0.05, the operating condition adjustment factor is 1.2, and the weights are 0.8 and 0.2 respectively, the dynamic fault threshold is 0.046.

[0053] The smoothed fault probability sequence is subjected to continuity judgment and trend analysis. When the fault probability at multiple consecutive time points exceeds the dynamic fault threshold and shows an upward trend, the initial alarm signal is triggered. For example, if the fault probability at 5 consecutive time points exceeds the dynamic fault threshold and shows an upward trend, the initial alarm signal is triggered.

[0054] The diagnostic results of multiple sensors are collected and the consistency of these diagnostic results is calculated. When the consistency exceeds the preset consistency threshold, the initial alarm signal is converted to a final alarm signal. For example, if the diagnostic results of the three sensors all indicate a fault, the consistency is high and the initial alarm signal is converted to a final alarm signal.

[0055] The coal mill bearing fault diagnosis is carried out based on the final alarm signal.

[0056] The beneficial effects of this method can be summarized into the following three parts: 1. Improve the accuracy of fault diagnosis: Through technologies such as deep learning models, temperature scaling parameters, Monte Carlo dropout methods, and Bayesian formulas, fault types can be identified and fault severity can be assessed more accurately.

[0057] 2. Enhance the reliability of fault diagnosis: Through technologies such as time attenuation factor, time series smoothing, working condition adjustment factor and multi-sensor diagnosis, the false alarm rate and missed alarm rate can be effectively reduced, and the reliability of fault diagnosis can be improved.

[0058] 3. Improve the efficiency of fault diagnosis: Through technologies such as dynamic fault thresholds and continuity judgment, faults can be discovered in a timely manner and alarms can be issued, thus avoiding equipment downtime and economic losses caused by faults.

[0059] In an optional implementation, a health status evaluation index of a coal mill bearing is established based on the fault type and the fault degree, and the health status evaluation index includes a fault severity coefficient including: Based on the historical maintenance data of the coal mill, the production decline rate, energy consumption increase rate, and vibration amplitude change rate when the fault occurs are collected to construct a performance loss index, and the performance loss index is converted into a performance impact score through the principal component analysis method; the stress concentration area is determined based on the bearing stress-strain analysis, the crack propagation rate is calculated in combination with the material fatigue characteristics, and a Markov chain model is established to obtain the fault propagation probability value; the spare parts replacement cost, labor cost, equipment downtime loss, associated fault repair expenditure, and production quality loss are fuzzy comprehensively evaluated to obtain a relative cost coefficient; the hierarchical analysis method is used to combine the performance impact score, the fault propagation probability value, and the relative cost coefficient to form a fault type weight matrix; Constructing a piecewise nonlinear normalization function, extracting time domain features, frequency domain features, and time-frequency domain features from the vibration signal, combining the time domain features, the frequency domain features, the time-frequency domain features, temperature parameters, and acoustic emission parameters into a feature vector, using principal component analysis to reduce the dimension of the feature vector, and establishing a mapping relationship from the feature vector to the fault degree through support vector regression; The fault type weight matrix, fault diagnosis probability and fault degree processed by the piecewise nonlinear normalization function are weightedly combined to obtain an initial fault severity coefficient; the initial fault severity coefficient is exponentially smoothed, and the smoothing coefficient is determined based on the minimum prediction error criterion to obtain a final fault severity coefficient, and the final fault severity coefficient is used as a basis for fault assessment of a coal mill bearing.

[0060] The specific implementation steps of the coal mill bearing health status assessment method are as follows: First, collect the historical maintenance records of the coal mill, including the type of each failure, the time of occurrence, the percentage of production decline, the percentage of energy consumption increase, and the percentage of change in vibration amplitude. For example, a bearing wear failure occurred on January 10, 2023, and the production decreased by 5%, the energy consumption increased by 8%, and the vibration amplitude increased by 15%. Collect multiple samples of similar data.

[0061] Then, based on the collected historical data, statistical methods were used to calculate the average values ​​of the production decline rate, energy consumption increase rate, and vibration amplitude change rate for each fault type. These average values ​​reflect the degree of impact of different fault types on coal mill performance. For example, the average production decline rate of bearing wear faults is 6%, the average energy consumption increase rate is 9%, and the average vibration amplitude increase rate is 16%.

[0062] Next, the average values ​​of the production decline rate, energy consumption increase rate, and vibration amplitude change rate for each fault type are combined into a performance loss index. For example, the performance loss index for a bearing wear fault can be expressed as (6%, 9%, 16%).

[0063] Then, the principal component analysis method was used to convert the multidimensional performance loss indicators into a single performance impact score. For example, the performance impact score of a bearing wear fault is 0.8, and the performance impact score of a bearing fracture fault is 0.95. The higher the score, the greater the impact of the fault on performance.

[0064] At the same time, stress-strain analysis is performed on the bearing to determine the stress concentration area. For example, through finite element analysis, it is found that the inner ring raceway of the bearing is a stress concentration area. Then, combined with the fatigue properties of the bearing material, such as the SN curve, the crack growth rate is calculated. For example, it is calculated that under a certain stress level, the crack grows 1 mm per year.

[0065] A Markov chain model is established to predict the probability of a fault developing from the current state to a more serious state. For example, the probability of a bearing with slight wear developing into severe wear is 0.1, and the probability of severe wear developing into fracture is 0.2.

[0066] Next, evaluate the relative cost coefficient of each failure type. This requires considering spare parts replacement costs, labor costs, equipment downtime losses, repair expenses for other failures that may be caused, and production quality losses. For example, the relative cost coefficient of bearing wear is 0.5, and the relative cost coefficient of bearing fracture is 1.

[0067] Using the hierarchical analysis method, the performance impact score, fault extension probability value and relative cost coefficient are combined into a fault type weight matrix. For example, the weight of bearing wear is 0.3, and the weight of bearing fracture is 0.7.

[0068] Extract time domain features, frequency domain features, and time-frequency domain features from the vibration signal of the coal mill bearing. Time domain features include root mean square value, peak value, etc.; frequency domain features include characteristic frequency amplitude, etc.; time-frequency domain features include wavelet energy, etc. At the same time, collect the temperature and acoustic emission parameters of the bearing. Combine all features with temperature and acoustic emission parameters into a feature vector.

[0069] The principal component analysis method is used to reduce the dimension of the feature vector, for example, reducing the dimension of a 10-dimensional feature vector to 3 dimensions. Then, support vector regression is used to establish the mapping relationship between the feature vector and the degree of fault. For example, the feature vector (0.1, 0.2, 0.3) corresponds to a fault degree of 0.5. The piecewise nonlinear normalization function is used to process the fault degree so that it falls between 0 and 1.

[0070] The fault type weight matrix, fault diagnosis probability and fault degree processed by piecewise nonlinear normalization function are weighted and combined to obtain the initial fault severity coefficient. The initial fault severity coefficient is subjected to exponential smoothing, for example, a smoothing coefficient of 0.8 is selected to obtain the final fault severity coefficient. This coefficient is the basis for fault assessment of coal mill bearings.

[0071] The beneficial effects of this method are reflected in the following three aspects: 1. Improve assessment accuracy: Comprehensive consideration of performance loss, failure development probability and cost factors can more comprehensively and accurately assess the health status of bearings.

[0072] 2. Achieve early warning: Through the crack growth rate and Markov model, the fault development trend can be predicted, early warning can be achieved, and major failures can be avoided.

[0073] 3. Optimize maintenance strategy: Based on the fault severity coefficient, a more reasonable maintenance strategy can be formulated to avoid over-maintenance or under-maintenance and reduce maintenance costs.

[0074] In an optional implementation, a health status evaluation index of a coal mill bearing is established based on the fault type and the fault degree, and the health status evaluation index includes a fault development trend coefficient including: Determine the length of the sliding time window through autocorrelation analysis, and calculate the first-order difference sequence and the second-order difference sequence of the fault severity coefficient within the sliding time window; determine the polynomial fitting order based on the Akaike information criterion, perform local weighted polynomial fitting on the first-order difference sequence and the second-order difference sequence, and obtain a fault development trend curve; Collect the speed parameters, load rate parameters and feed rate parameters of the coal mill, perform dimensionality reduction processing on the speed parameters, load rate parameters and feed rate parameters through principal component analysis to obtain the working condition characteristics, establish a multivariate regression model of the working condition characteristics and the fault development rate, and calculate the ratio of the actual working condition development rate to the standard working condition development rate to obtain the working condition compensation factor; Constructing a fuzzy neural network, inputting temperature parameters and humidity parameters into the input layer of the fuzzy neural network, using Gaussian fuzzy membership function to perform fuzzy reasoning in the hidden layer of the fuzzy neural network, obtaining an environmental impact coefficient in the output layer of the fuzzy neural network, and using a loss function with a regularization term for training the fuzzy neural network; Performing exponential weighted averaging on the first-order difference sequence to obtain short-term variation characteristics, calculating the derivative of the fault development trend curve to obtain long-term trend characteristics, and multiplying the working condition compensation factor by the environmental impact coefficient to obtain a comprehensive compensation term; The short-term change characteristics, the long-term trend characteristics, and the comprehensive compensation items are weightedly combined to obtain an initial trend coefficient, and the Kalman filter algorithm is used to eliminate noise from the initial trend coefficient to obtain a final fault development trend coefficient.

[0075] The core of the coal mill bearing health status assessment method is to construct an indicator that can reflect the development trend of faults. This indicator comprehensively considers multiple factors such as fault type, fault severity, working environment, etc., aiming to more accurately predict the remaining life of the bearing and avoid unexpected downtime.

[0076] First, collect the vibration signal of the coal mill bearing and analyze it according to the known fault type. For example, the characteristic frequency associated with the specific fault type can be extracted by using methods such as envelope analysis. Then, calculate the fault severity coefficient based on the characteristic frequency of the fault type. For example, the root mean square value, peak value and other indicators of the vibration signal can be used to quantify the severity of the fault.

[0077] Next, determine a suitable sliding time window. The length of this window needs to be determined based on the autocorrelation characteristics of the vibration signal to ensure that the signal within the window can fully reflect the development trend of the fault. For example, the autocorrelation function of the vibration signal can be calculated, and the appropriate window length can be selected based on the attenuation characteristics of the autocorrelation function. In one example, if the autocorrelation function decays to near zero after 100 data points, the length of the sliding time window can be set to 100.

[0078] In the sliding time window, the first-order difference sequence and second-order difference sequence of the fault severity coefficient are calculated. The first-order difference reflects the rate of change of the fault severity, and the second-order difference reflects the change of the rate of change, that is, the acceleration of the fault development. For example, assuming that the fault severity coefficients in a window are 1, 2, 4, 7, and 11, respectively, then the first-order difference sequence is 1, 2, 3, 4, and the second-order difference sequence is 1, 1, 1.

[0079] The first-order difference sequence and the second-order difference sequence are locally weighted polynomial fitted to obtain the fault development trend curve. The order of fitting can be determined according to the Akaike Information Criterion (AIC), and the order with the smallest AIC value is selected. The idea of ​​local weighting is that in the fitting process, more attention is paid to samples close to the current point, while smaller weights are given to samples at farther points. Assuming that the fitting order is determined to be 3 according to the AIC criterion, the fault development trend curve can be obtained by using a locally weighted cubic polynomial fitting.

[0080] At the same time, the speed, load rate, feed rate and other operating parameters of the coal mill are collected. These parameters are subjected to principal component analysis to reduce their dimensions to one or several main operating characteristics. For example, suppose two principal components are extracted through principal component analysis, representing the load condition and feed condition of the coal mill respectively.

[0081] A multivariate regression model of operating condition characteristics and fault development rate is established. The fault development rate can be obtained by taking the derivative of the fault development trend curve. This model can be used to predict the speed of fault development under different operating conditions. For example, a simple linear regression model can be expressed as: Fault development rate = a*load condition + b*feed condition + c, where a, b, c are regression coefficients.

[0082] The ratio of the actual operating condition development rate to the standard operating condition development rate is calculated to obtain the operating condition compensation factor. The standard operating condition can be the ideal operating condition for the coal mill design operation. For example, if the actual operating condition development rate is 1.2 times the standard operating condition development rate, the operating condition compensation factor is 1.2.

[0083] The environmental temperature and humidity parameters are collected and input into a fuzzy neural network. The hidden layer of the network uses Gaussian fuzzy membership function for fuzzy reasoning, and the output layer outputs the environmental impact coefficient. When training the fuzzy neural network, a loss function with a regularization term is used to prevent overfitting. For example, assuming the temperature is 25°C and the humidity is 60%, after the fuzzy neural network calculation, the environmental impact coefficient is 0.9.

[0084] Perform exponential weighted averaging on the first-order difference sequence to obtain the short-term change characteristics. The exponential weighted average pays more attention to recent data and can better reflect the short-term change trend of the fault. For example, assuming the attenuation factor is 0.8 and the first-order difference sequence is 1, 2, 3, 4, the short-term change characteristic is 3.328.

[0085] Calculate the derivative of the fault development trend curve to obtain the long-term trend characteristics. This reflects the long-term development trend of the fault. Multiply the working condition compensation factor by the environmental impact coefficient to obtain the comprehensive compensation term. For example, if the working condition compensation factor is 1.2 and the environmental impact coefficient is 0.9, the comprehensive compensation term is 1.08.

[0086] The short-term change characteristics, long-term trend characteristics, and comprehensive compensation items are weighted and combined to obtain the initial trend coefficient. The weight can be determined based on experience or expert knowledge. For example, assuming that the weight of the short-term change characteristics is 0.4, the weight of the long-term trend characteristics is 0.5, and the weight of the comprehensive compensation item is 0.1, then the initial trend coefficient = 0.4*short-term change characteristics + 0.5*long-term trend characteristics + 0.1*comprehensive compensation item.

[0087] Finally, the Kalman filter algorithm is used to eliminate the noise of the initial trend coefficient and obtain the final fault development trend coefficient. Kalman filtering can effectively remove the noise in the signal and improve the accuracy of the evaluation results.

[0088] The beneficial effects of this method can be summarized in the following three aspects: 1. Improved the accuracy of the assessment: This method takes into account multiple factors such as fault type, fault severity, working environment, etc., and can more accurately assess the health status of the coal mill bearing.

[0089] 2. Enhanced reliability of evaluation: The use of advanced technologies such as local weighted polynomial fitting, fuzzy neural network, Kalman filtering, etc. effectively reduces the impact of noise and interference and improves the reliability of evaluation results.

[0090] 3. Improved practicality of the assessment: This method can provide a quantitative fault development trend coefficient, which can be used to predict the remaining life of the bearing, guide maintenance work, avoid unexpected downtime, and thus improve production efficiency.

[0091] In an optional implementation, calculating the remaining service life of the coal mill bearing according to the health status assessment index includes: The fault severity coefficient and the fault development trend coefficient are nonlinearly weighted and fused, and the cross term of the fault severity coefficient and the fault development trend coefficient is introduced to establish a comprehensive health index model, wherein the comprehensive health index model includes a linear weight coefficient of the fault severity coefficient, a linear weight coefficient of the fault development trend coefficient, and a cross term weight coefficient; A Weibull process model is established based on the comprehensive health index, and a bearing force analysis model is established to calculate the stress distribution under different working conditions. The life loss caused by changes in working conditions is calculated based on Miner's linear cumulative damage theory. The life loss is input into the Weibull process model as a correction term, and a remaining life prediction equation considering the influence of working conditions is established to calculate the remaining service life of the coal mill bearing.

[0092] The remaining service life assessment method of coal mill bearings is specifically implemented as follows: First, collect the vibration signal of the coal mill bearing. For example, use an acceleration sensor to collect the vibration signal of the bearing during operation, set the sampling frequency to 20kHz, and set the sampling time to 10 seconds.

[0093] Then, the collected vibration signal is preprocessed. The preprocessing process includes operations such as removing trend items, filtering and noise reduction. For example, a high-pass filter is used to remove low-frequency noise, and the cutoff frequency is set to 500Hz; a wavelet threshold denoising method is used to remove high-frequency noise, and the wavelet basis function is db4, and the number of decomposition layers is set to 5.

[0094] Next, extract the bearing fault features. Based on the preprocessed vibration signal, extract the characteristic indicators that reflect the severity and development trend of the bearing fault. For example, the extracted fault severity indicators include the root mean square value, peak value, and kurtosis; the extracted fault development trend indicators include energy entropy and singular value. Assume that the extracted root mean square value is 0.8g, the peak value is 2.5g, the kurtosis is 4.2, the energy entropy is 0.7, and the singular value is 0.3.

[0095] Subsequently, the fault severity coefficient and the fault development trend coefficient are calculated. According to the extracted fault characteristic indicators, the fault severity coefficient and the fault development trend coefficient are calculated respectively. For example, the fuzzy membership function is used to map the root mean square value, peak value and kurtosis to the [0,1] interval, and then the fault severity coefficient is obtained by weighted average; the normalization method is used to map the energy entropy and singular value to the [0,1] interval, and then the fault development trend coefficient is obtained by weighted average. Assume that the calculated fault severity coefficient is 0.6 and the fault development trend coefficient is 0.4.

[0096] Next, a comprehensive health index model is established. The fault severity coefficient and the fault development trend coefficient are nonlinearly weighted and fused, and the cross term of the two is introduced to establish a comprehensive health index model. The model contains the first-order weight coefficient of the fault severity coefficient, the first-order weight coefficient of the fault development trend coefficient, and the cross-term weight coefficient. For example, assuming that the first-order weight coefficient of the fault severity coefficient is 0.7, the first-order weight coefficient of the fault development trend coefficient is 0.3, and the cross-term weight coefficient is 0.2. Then the comprehensive health index = 0.7*0.6+0.3*0.4+0.2*0.6*0.4=0.588.

[0097] Then, a Weibull process model is established. Based on the calculated comprehensive health index, a Weibull process model is established. The shape parameter and scale parameter of the Weibull process model can be determined by the maximum likelihood estimation method. Assume that the determined shape parameter is 2.5 and the scale parameter is 10,000 hours.

[0098] Next, establish a bearing force analysis model. According to the actual working conditions of the coal mill, establish a bearing force analysis model to calculate the stress distribution under different working conditions. For example, the finite element analysis method is used to establish a bearing force analysis model, and the maximum stress under rated working conditions is calculated to be 500MPa, the maximum stress under light load conditions is 300MPa, and the maximum stress under heavy load conditions is 700MPa.

[0099] Then, calculate the life loss caused by the change in operating conditions. Based on Miner's linear cumulative damage theory, calculate the life loss caused by the change in operating conditions. For example, assume that the coal mill runs 80% of the time under rated conditions, 10% of the time under light load conditions, and 10% of the time under heavy load conditions. Based on the stress distribution under different operating conditions and the corresponding fatigue life curve, the life loss under each operating condition can be calculated, and then the total life loss can be accumulated. Assume that the calculated life loss is 1000 hours.

[0100] Next, establish a remaining life prediction equation that takes into account the influence of operating conditions. Input the calculated life loss as a correction term into the Weibull process model to establish a remaining life prediction equation that takes into account the influence of operating conditions. For example, subtract the life loss from the scale parameter of the Weibull process model, and the corrected scale parameter is 9000 hours.

[0101] Finally, the remaining service life of the coal mill bearing is calculated. According to the modified Weibull process model, the remaining service life of the coal mill bearing is calculated. For example, assuming that the current operating time is 5000 hours, the remaining service life can be calculated according to the Weibull distribution function.

[0102] The beneficial effects of this method can be summarized in the following three aspects: 1. Improved remaining life prediction accuracy: By introducing the cross term of the fault severity coefficient and the fault development trend coefficient, and considering the impact of operating condition changes, the degradation process of the bearing can be more accurately reflected, thereby improving the remaining life prediction accuracy.

[0103] 2. Realize online evaluation of remaining life: This method can perform real-time calculation based on online monitoring data, thereby realizing online evaluation of remaining life and providing a decision-making basis for equipment maintenance.

[0104] 3. Strong practicability: This method does not require complicated mathematical formulas, is easy to understand and apply, and can be widely used in various types of coal mill bearings.

[0105] In an optional implementation, the method further includes: Taking the predicted value of the remaining life prediction equation as the state variable and the health status assessment index as the observation variable, respectively establishing a state equation containing process noise and an observation equation containing measurement noise, and constructing an adaptive particle filter correction model; Constructing an adaptive noise covariance matrix, wherein the adaptive noise covariance matrix adopts a recursive form with a forgetting factor, controls the decay rate of historical information through the forgetting factor, constructs an innovation covariance term based on the prediction error at the current moment, and dynamically estimates the process noise in the state equation and the measurement noise in the observation equation; Inputting the adaptive noise covariance matrix into the adaptive particle filter correction model, correcting the state prediction, calculating the likelihood value of each particle based on the observation data of the observation equation, and updating the weight of the particle; The updated particle weights are used to perform weighted averaging on the state estimation values ​​to obtain a revised prediction value of the remaining life of the coal mill bearing, wherein the revised prediction value of the remaining life of the coal mill bearing includes the influence of operating condition changes and random fluctuation characteristics during equipment operation.

[0106] First, the operating status data of the coal mill bearings are collected, including vibration signals, temperature, current, etc., as well as health status evaluation indicators that reflect the degree of bearing wear, such as energy index, kurtosis index, etc. Taking the collected vibration signal as an example, the vibration signal is collected every 10 minutes, and 1024 data points are collected each time. At the same time, the temperature and current values ​​of the bearings are recorded every 1 hour, and the energy index and kurtosis index are calculated based on the vibration signal as health status evaluation indicators.

[0107] Next, the existing coal mill bearing life data, such as historical failure data or accelerated life test data, is used to establish a coal mill bearing remaining life prediction equation. This equation can be a model based on regression analysis, neural network and other methods. For example, the support vector machine method can be used to establish a coal mill bearing remaining life prediction equation based on historical failure data and corresponding health status assessment indicators. Assume that the remaining life prediction equation established based on historical data is: remaining life = 1000-5*energy index-2*kurtosis index; Then, the predicted value of the remaining life prediction equation is used as the state variable, and the health status assessment index is used as the observation variable to construct an adaptive particle filter correction model. The state equation describes the evolution of the remaining life over time, which includes process noise, which is used to represent the impact of unconsidered factors on the remaining life. The observation equation describes the relationship between the observation variable and the state variable, which includes measurement noise, which is used to represent the impact of measurement errors. For example, the state equation can be set as: the remaining life at the current moment = the remaining life at the previous moment - wear rate + process noise; the observation equation can be set as: the energy index at the current moment = the function related to the remaining life at the current moment + measurement noise.

[0108] Construct an adaptive noise covariance matrix. This matrix uses a recursive form with a forgetting factor, which controls the decay rate of historical information. The forgetting factor is usually between 0.9 and 0.99. Construct the new information covariance term based on the current prediction error, and dynamically estimate the process noise in the state equation and the measurement noise in the observation equation. For example, assuming that the current prediction error is 20 hours, the forgetting factor is 0.95, and the process noise covariance of the previous moment is 10, the process noise covariance of the current moment can be updated to 0.95*10+(1-0.95)*20*20=29.5.

[0109] The adaptive noise covariance matrix is ​​input into the adaptive particle filter correction model to correct the state prediction. First, the state prediction is performed according to the state equation to obtain the predicted value of each particle. Then, based on the observation data of the observation equation, such as the energy index and kurtosis index at the current moment, the likelihood value of each particle is calculated. The likelihood value reflects the degree of consistency between the particle and the actual observation data. After that, the weight of the particle is updated according to the likelihood value, and the particle with a large weight indicates that it is closer to the true state.

[0110] The updated particle weights are used to perform weighted averaging on the state estimation values ​​to obtain the revised prediction value of the remaining life of the coal mill bearing. For example, assuming there are 100 particles, each particle represents a possible remaining life value. The 100 remaining life values ​​are weighted averaged according to the updated weights to obtain the revised prediction value of the remaining life of the coal mill bearing. This revised prediction value includes the impact of operating condition changes and the random fluctuation characteristics during equipment operation.

[0111] The beneficial effects of this method can be summarized into the following three parts: 1. Improve prediction accuracy: By using the adaptive particle filter to correct the model, the influence of process noise and measurement noise can be effectively reduced, and the accuracy of the prediction of the remaining life of the coal mill bearing can be improved.

[0112] 2. Adapt to changes in operating conditions: The adaptive noise covariance matrix can dynamically adjust the noise estimation according to the actual operating conditions, so that the prediction model can better adapt to changes in operating conditions and improve the reliability of the prediction.

[0113] 3. Extend bearing life: Accurate remaining life prediction can help formulate a more reasonable maintenance plan and avoid replacing bearings too early or too late, thereby extending the service life of bearings and reducing maintenance costs.

[0114] Figure 2 FIG. 1 is a schematic diagram of the structure of a coal mill fault diagnosis and prediction system based on big data analysis according to an embodiment of the present invention. Figure 2 As shown, the system comprises: The first unit is used to collect the vibration signal of the coal mill bearing in real time by using a vibration sensor, and perform multi-scale analysis on the vibration signal by wavelet decomposition, extract the time domain characteristics, frequency domain characteristics and time-frequency characteristics of the vibration signal, and normalize the time domain characteristics, frequency domain characteristics and time-frequency characteristics to obtain a feature vector matrix; construct a training data set of a deep convolutional neural network model based on the feature vector matrix, and divide the training data set into a source domain data set and a target domain data set; The second unit is used to construct a deep convolutional neural network model including a convolutional layer, a pooling layer and a fully connected layer, wherein the input layer of the deep convolutional neural network model receives the feature vector matrix, and the output layer outputs the fault type and fault degree; a maximum mean difference adaptive layer is introduced into the deep convolutional neural network model, and the feature distribution of the source domain data set and the target domain data set are adaptively aligned through the maximum mean difference adaptive layer; the deep convolutional neural network model is pre-trained using the source domain data set, and fine-tuned using the target domain data set; The third unit is used to input the real-time collected vibration signal of the coal mill bearing into the trained deep convolutional neural network model to obtain the fault type and fault degree of the coal mill bearing; establish the health status assessment index of the coal mill bearing based on the fault type and the fault degree, and the health status assessment index includes a fault severity coefficient and a fault development trend coefficient; calculate the remaining service life of the coal mill bearing according to the health status assessment index, and issue a warning signal when the remaining service life is lower than a preset health threshold.

[0115] According to a third aspect of the embodiments of the present invention, An electronic device is provided, comprising: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the aforementioned method.

[0116] A fourth aspect of the embodiments of the present invention is: A computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the aforementioned method is implemented.

[0117] The present invention may be a method, an apparatus, a system and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present invention.

[0118] 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 replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A coal mill fault diagnosis and prediction method based on big data analysis, characterized in that: include: The vibration signal of the coal mill bearing is collected in real time by using a vibration sensor, and the vibration signal is subjected to multi-scale analysis by wavelet decomposition to extract the time domain feature, frequency domain feature and time-frequency feature of the vibration signal, and the time domain feature, frequency domain feature and time-frequency feature are normalized to obtain a feature vector matrix; Constructing a training data set of a deep convolutional neural network model based on the feature vector matrix, and dividing the training data set into a source domain data set and a target domain data set; Constructing a deep convolutional neural network model including a convolutional layer, a pooling layer, and a fully connected layer, wherein the input layer of the deep convolutional neural network model receives the feature vector matrix, and the output layer outputs the fault type and fault degree; introducing a maximum mean difference adaptive layer into the deep convolutional neural network model, and adaptively aligning the feature distributions of the source domain data set and the target domain data set through the maximum mean difference adaptive layer; Pre-training the deep convolutional neural network model using the source domain dataset, and performing fine-tuning training using the target domain dataset; The real-time collected vibration signal of the coal mill bearing is input into the trained deep convolutional neural network model to obtain the fault type and fault degree of the coal mill bearing; based on the fault type and the fault degree, a health status assessment index of the coal mill bearing is established, and the health status assessment index includes a fault severity coefficient and a fault development trend coefficient; the remaining service life of the coal mill bearing is calculated according to the health status assessment index, and a warning signal is issued when the remaining service life is lower than a preset health threshold.

2. The method according to claim 1, characterized in that Introducing a maximum mean difference adaptive layer into the deep convolutional neural network model, and adaptively aligning the feature distributions of the source domain dataset and the target domain dataset through the maximum mean difference adaptive layer includes: A maximum mean difference adaptive layer is constructed using a Gaussian kernel function; the maximum mean difference adaptive layer constructs a multi-core mapping function based on the Gaussian kernel function, the multi-core mapping function is obtained by linearly combining multiple Gaussian kernel functions with different bandwidths, each of the Gaussian kernel functions corresponds to a weight coefficient, and the sum of the weight coefficients is 1; Extracting deep feature representations of a source domain dataset and a target domain dataset, calculating channel attention weights of the source domain dataset and the target domain dataset, performing weighted processing on the deep feature representations according to the channel attention weights to obtain weighted source domain features and target domain features; inputting the weighted source domain features and the weighted target domain features into the multi-core mapping function, and calculating the maximum mean difference distance between the source domain features and the target domain features; The product of the maximum mean difference distance and the dynamic balance parameter is taken as the maximum mean difference loss, and the dynamic balance parameter gradually increases from an initial value to a preset value as the number of training rounds increases; the maximum mean difference loss is added to the total loss function of the deep neural network, and the network parameters of the deep neural network and the weight coefficient of the multi-core mapping function are updated through the back propagation algorithm until the maximum mean difference loss is minimized, thereby achieving adaptive alignment of source domain features and target domain features.

3. The method according to claim 1, characterized in that Inputting the real-time collected vibration signal of the coal mill bearing into the trained deep convolutional neural network model to obtain the fault type and fault degree of the coal mill bearing includes: The real-time collected coal mill bearing vibration signal is input into the trained deep convolutional neural network model, the Softmax activation function is used to calculate the probability distribution of the output layer of the deep convolutional neural network model, and the temperature scaling parameter is introduced to calibrate the Softmax output to obtain the calibrated fault type probability distribution; the deep convolutional neural network model is sampled multiple times by the Monte Carlo dropout method to obtain the distribution of fault degree prediction values, the mean and standard deviation are calculated based on the fault degree prediction value distribution, and the fault degree confidence interval is constructed; A priori probability distribution model of fault type is established based on Dirichlet distribution, and the influence of time decay factor on historical observation data is introduced to dynamically adjust, the calibrated fault type probability distribution is input into the priori probability distribution model of fault type, and the posterior probability is calculated using Bayesian formula to update the fault type probability; the fault degree confidence interval is combined with the updated fault type probability to obtain a comprehensive fault probability; The comprehensive fault probability is smoothed in time series by using an exponentially weighted moving average method to obtain a smoothed fault probability sequence; a working condition adjustment factor is constructed based on the equipment load, rotation speed and temperature, and the working condition adjustment factor changes dynamically with the equipment operation state; a local standard deviation of the smoothed fault probability sequence is weightedly combined with the working condition adjustment factor to generate a dynamic fault threshold; Perform continuity judgment and trend analysis on the smoothed fault probability sequence, and trigger an initial alarm signal when the fault probabilities at multiple consecutive time points exceed the dynamic fault threshold and show an upward trend; collect diagnostic results of multiple sensors, calculate the consistency of the diagnostic results of the multiple sensors, and convert the initial alarm signal into a final alarm signal when the consistency exceeds a preset consistency threshold; and perform fault diagnosis on the coal mill bearing based on the final alarm signal.

4. The method according to claim 1, characterized in that: Based on the fault type and the fault degree, a health status evaluation index of the coal mill bearing is established, wherein the health status evaluation index includes a fault severity coefficient including: Based on the historical maintenance data of the coal mill, the production decline rate, energy consumption increase rate, and vibration amplitude change rate when the fault occurs are collected to construct a performance loss index, and the performance loss index is converted into a performance impact score through the principal component analysis method; the stress concentration area is determined based on the bearing stress-strain analysis, the crack propagation rate is calculated in combination with the material fatigue characteristics, and a Markov chain model is established to obtain the fault propagation probability value; the spare parts replacement cost, labor cost, equipment downtime loss, associated fault repair expenditure, and production quality loss are fuzzy comprehensively evaluated to obtain a relative cost coefficient; the hierarchical analysis method is used to combine the performance impact score, the fault propagation probability value, and the relative cost coefficient to form a fault type weight matrix; Constructing a piecewise nonlinear normalization function, extracting time domain features, frequency domain features, and time-frequency domain features from the vibration signal, combining the time domain features, the frequency domain features, the time-frequency domain features, temperature parameters, and acoustic emission parameters into a feature vector, using principal component analysis to reduce the dimension of the feature vector, and establishing a mapping relationship from the feature vector to the fault degree through support vector regression; The fault type weight matrix, fault diagnosis probability and fault degree processed by the piecewise nonlinear normalization function are weightedly combined to obtain an initial fault severity coefficient; the initial fault severity coefficient is exponentially smoothed, and the smoothing coefficient is determined based on the minimum prediction error criterion to obtain a final fault severity coefficient, and the final fault severity coefficient is used as a basis for fault assessment of a coal mill bearing.

5. The method according to claim 1, characterized in that Based on the fault type and the fault degree, a health status evaluation index of the coal mill bearing is established. The health status evaluation index includes a fault development trend coefficient including: Determine the length of the sliding time window through autocorrelation analysis, and calculate the first-order difference sequence and the second-order difference sequence of the fault severity coefficient within the sliding time window; determine the polynomial fitting order based on the Akaike information criterion, perform local weighted polynomial fitting on the first-order difference sequence and the second-order difference sequence, and obtain a fault development trend curve; Collect the speed parameters, load rate parameters and feed rate parameters of the coal mill, perform dimensionality reduction processing on the speed parameters, load rate parameters and feed rate parameters through principal component analysis to obtain the working condition characteristics, establish a multivariate regression model of the working condition characteristics and the fault development rate, and calculate the ratio of the actual working condition development rate to the standard working condition development rate to obtain the working condition compensation factor; Constructing a fuzzy neural network, inputting temperature parameters and humidity parameters into the input layer of the fuzzy neural network, using Gaussian fuzzy membership function to perform fuzzy reasoning in the hidden layer of the fuzzy neural network, obtaining an environmental impact coefficient in the output layer of the fuzzy neural network, and using a loss function with a regularization term for training the fuzzy neural network; Performing exponential weighted averaging on the first-order difference sequence to obtain short-term variation characteristics, calculating the derivative of the fault development trend curve to obtain long-term trend characteristics, and multiplying the working condition compensation factor by the environmental impact coefficient to obtain a comprehensive compensation term; The short-term change characteristics, the long-term trend characteristics, and the comprehensive compensation items are weightedly combined to obtain an initial trend coefficient, and the Kalman filter algorithm is used to eliminate noise from the initial trend coefficient to obtain a final fault development trend coefficient.

6. The method according to claim 1, characterized in that Calculating the remaining service life of the coal mill bearing according to the health status assessment index includes: The fault severity coefficient and the fault development trend coefficient are nonlinearly weighted and fused, and the cross term of the fault severity coefficient and the fault development trend coefficient is introduced to establish a comprehensive health index model, wherein the comprehensive health index model includes a linear weight coefficient of the fault severity coefficient, a linear weight coefficient of the fault development trend coefficient, and a cross term weight coefficient; A Weibull process model is established based on the comprehensive health index, and a bearing force analysis model is established to calculate the stress distribution under different working conditions. The life loss caused by changes in working conditions is calculated based on Miner's linear cumulative damage theory. The life loss is input into the Weibull process model as a correction term, and a remaining life prediction equation considering the influence of working conditions is established to calculate the remaining service life of the coal mill bearing.

7. The method according to claim 6, characterized in that The method further comprises: Taking the predicted value of the remaining life prediction equation as the state variable and the health status assessment index as the observation variable, respectively establishing a state equation containing process noise and an observation equation containing measurement noise, and constructing an adaptive particle filter correction model; Constructing an adaptive noise covariance matrix, wherein the adaptive noise covariance matrix adopts a recursive form with a forgetting factor, controls the decay rate of historical information through the forgetting factor, constructs an innovation covariance term based on the prediction error at the current moment, and dynamically estimates the process noise in the state equation and the measurement noise in the observation equation; Inputting the adaptive noise covariance matrix into the adaptive particle filter correction model, correcting the state prediction, calculating the likelihood value of each particle based on the observation data of the observation equation, and updating the weight of the particle; The updated particle weights are used to perform weighted averaging on the state estimation values ​​to obtain a revised prediction value of the remaining life of the coal mill bearing, wherein the revised prediction value of the remaining life of the coal mill bearing includes the influence of operating condition changes and random fluctuation characteristics during equipment operation.

8. A coal mill fault diagnosis and prediction system based on big data analysis, used to implement the method described in any one of claims 1 to 7, characterized in that: include: The first unit is used to collect the vibration signal of the coal mill bearing in real time by using a vibration sensor, and perform multi-scale analysis on the vibration signal by wavelet decomposition, extract the time domain feature, frequency domain feature and time-frequency feature of the vibration signal, and normalize the time domain feature, frequency domain feature and time-frequency feature to obtain a feature vector matrix; Constructing a training data set of a deep convolutional neural network model based on the feature vector matrix, and dividing the training data set into a source domain data set and a target domain data set; The second unit is used to construct a deep convolutional neural network model including a convolutional layer, a pooling layer and a fully connected layer, wherein the input layer of the deep convolutional neural network model receives the feature vector matrix, and the output layer outputs the fault type and fault degree; a maximum mean difference adaptive layer is introduced into the deep convolutional neural network model, and the feature distribution of the source domain data set and the target domain data set are adaptively aligned through the maximum mean difference adaptive layer; Pre-training the deep convolutional neural network model using the source domain dataset, and performing fine-tuning training using the target domain dataset; The third unit is used to input the real-time collected vibration signal of the coal mill bearing into the trained deep convolutional neural network model to obtain the fault type and fault degree of the coal mill bearing; establish the health status assessment index of the coal mill bearing based on the fault type and the fault degree, and the health status assessment index includes a fault severity coefficient and a fault development trend coefficient; calculate the remaining service life of the coal mill bearing according to the health status assessment index, and issue a warning signal when the remaining service life is lower than a preset health threshold.

9. An electronic device, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to call the instructions stored in the memory to execute the method described in any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images