Industrial process key parameter prediction method based on space-time diagram convolutional network

Abstract

The invention provides an industrial process key parameter prediction method based on a space-time diagram convolutional network, aiming at the problems of time delay characteristics, data redundancy, information missing and the like in industrial process data. The method comprises the following steps: firstly, carrying out data preprocessing and sequence division on collected industrial process data; modeling is carried out on internal features of the subsequences in variable dimensions through global and local graph convolution; then, time modeling between the subsequences is completed in the time dimension through continuous time modeling and multi-head attention, continuous time modeling is composed of a plurality of Shenchang differential equations and a gating mechanism, and continuous modeling of each subsequence in the whole time dimension is completed in sequence through circulation; and finally, predicting the key parameters of the industrial process by using a multi-layer perceptron. According to the method, the problems of industrial process data in modeling are solved, the prediction precision of the industrial process key parameter data is improved, and technical support is provided for optimization control of the industrial process.

Description

Technical Field

[0001] The present invention belongs to the field of industrial time series prediction, and specifically is a method for predicting key parameters of industrial processes based on a spatiotemporal graph convolutional network. Background Art

[0002] In modern industry, data-driven models offer broad application prospects for improving production efficiency and optimizing resource utilization. Deep learning and machine learning have become important tools for processing industrial data. In recent years, with the advancement of data acquisition technology, the amount of industrial process data has increased significantly, highlighting the limitations of shallow neural networks in capturing data dependencies. Therefore, deep learning methods have become the key to achieving accurate predictions. However, despite the significant progress in performance of existing deep learning models, there are still some urgent problems to be solved when facing the particularity and complexity of industrial data, which often affect the stability and generalization ability of the model in practical applications.

[0003] First, in industrial scenarios, there are usually complex nonlinear relationships between variables, such as the mutual influence of physical quantities such as temperature, pressure and flow. In recent years, graph neural networks have made progress in extracting potential dependencies and effectively solved this problem. For example, the adaptive graph convolutional recurrent network (AGCRN) captures process-related features through a learnable adjacency matrix, improving prediction accuracy and interpretability; the meta-graph learning network (MegaCRN) combines graph networks with meta-learning to enhance the adaptability and generalization ability of the model in complex scenarios. However, in actual industrial data, there is often a lag effect between variables. For example, in the process of tobacco rehumidification, there is a time delay between the outlet moisture and other variables, and existing methods mostly assume that the variable relationship is synchronous, ignoring this delayed transmission effect, resulting in a decrease in prediction accuracy.

[0004] Secondly, due to the complexity of sensor settings and working conditions, industrial process data often have significant redundancy. Some data need to reach a threshold to affect quality parameters, while others are only relevant to specific variables. The attention mechanism filters irrelevant information through weight distribution, effectively alleviating the problem of data redundancy. For example, the block temporal attention network (PatchTST) and the interactive attention network (iTransformer) capture key interactions and extract target-related features through dot product attention; the fully connected spatiotemporal graph neural network (FCSTGNN) introduces an attention mechanism in the graph network to adaptively highlight important neighbor information. However, industrial data has high noise due to sensor errors, environmental impacts, etc., and the attention mechanism is sensitive to noise, which may lead to unstable model predictions.

[0005] Finally, industrial data is often discrete, sparse, and lacks continuity due to the influence of complex environments. Sensors are easily damaged under high temperature, vibration, or corrosion conditions, resulting in interrupted data collection; layout restrictions and lack of monitoring of key locations further exacerbate sparsity, making time series data less coherent. Traditional interpolation or completion methods have difficulty capturing complex dynamic relationships. Neural ordinary differential equations (NODEs) solve the sparse sampling problem through continuous time modeling and capture more refined temporal dynamic features. For example, multi-block graph neural ODE networks (MTGODEs) integrate NODEs to achieve the continuity of graphs and multi-block structures. However, NODEs are limited by the dimensionality of the hidden layer space, which affects their ability to model long-term dependencies. To this end, the fast-slow dynamics ODE network (GODERN) introduces fast ODEs and slow ODEs based on the fast-slow dynamics theory to describe complex systems at different time scales. However, it still suffers from catastrophic forgetting problems when modeling continuity over long time spans.

[0006] In summary, researching and developing advanced prediction methods for key parameters of industrial processes has important theoretical significance and practical application value. Summary of the invention

[0007] In view of the defects or deficiencies of the prior art described above, the present invention provides a method for predicting key parameters of industrial processes based on spatiotemporal graph convolutional networks. First, the collected industrial process data is preprocessed and sequenced; then, the internal features of the subsequences are modeled in the variable dimension through global and local graph convolutions; then, the time modeling between subsequences is completed in the time dimension through continuous time modeling and multi-head attention, wherein the continuous time modeling consists of multiple neural ordinary differential equations and gating mechanisms, and the continuous modeling of each subsequence in the entire time dimension is completed in sequence through a cycle; finally, the prediction of key parameters of the industrial process is realized using a multilayer perceptron. The present invention solves the problems existing in the modeling of industrial process data, improves the prediction accuracy of key parameter data of industrial processes, and provides technical support for the optimization control of industrial processes.

[0008] The technical solution adopted by the present invention to achieve the above-mentioned purpose is: a method for predicting key parameters of industrial processes based on spatiotemporal graph convolutional networks, comprising the following steps:

[0009] S1. Collect key parameter data of industrial processes;

[0010] S2, preprocess the data, establish the data set and divide the sequence;

[0011] S3, variable dimension modeling within the subsequence: embed the features of a single sequence point within the subsequence into the spatial dimension to generate a new embedding vector, which is used for the subsequent network to capture the time lag characteristics in the data during modeling; then use global graph convolution to model all dependencies within the subsequence; at the same time, use local graph convolution to optimize the redundancy of the embedding vector so that the model only focuses on the features related to itself; finally, add the outputs of the two to complete the time series data modeling within the subsequence;

[0012] S4. Temporal dimension modeling between subsequences: For each subsequence, more fine-grained features are extracted through continuous time modeling to make up for the defect of insufficient feature extraction caused by missing data. Then, the temporal features of the entire sequence are extracted through multi-head attention to obtain the output features of temporal dimension modeling.

[0013] S5. Establish a multi-layer perceptron network to iteratively learn the mapping relationship between input features and output labels, so as to predict the future key parameter sequence of industrial processes;

[0014] S6. Using the data in the dataset as input, repeat steps 3 to 5 for iterative training, select the model parameters with the best accuracy as the final ideal model, and predict the future key parameter sequence based on the historical data of the industrial process.

[0015] The key parameter data of the industrial process are industrial status data and process control data collected by detection sensors and control equipment set in the industrial production environment.

[0016] The preprocessing includes removing outliers and normalizing the maximum and minimum values.

[0017] The data set is divided into training set, validation set and test set in proportion; the training set data is used for training and cutoff according to the set number of iterations; the validation set data is used to verify the model accuracy; and the test set data is used for testing. Furthermore, each sample data in each set contains L+T sampling values, of which the first L sampling values ​​constitute the historical input data for training, and the last T sampling values ​​serve as the label value of the prediction target during the model training and testing process.

[0018] The sequence division is to use a sliding window of a fixed size to divide the entire historical input data sequence of each sample into multiple subsequences according to a specific step size, and the length of each subsequence is equal to the size of the window.

[0019] Furthermore, the global graph convolution is to multiply the adjacency matrix A of the adaptive graph convolution by a time delay matrix π to obtain an enhanced matrix This enhances the model's perception of the time distance within the subsequence. Assuming there is an adjacency matrix A containing 3 time points, the delay matrix measures the time distance through the decay rate δ<1. The farther away from the current time step, the greater the degree of decay. Among them, the time delay matrix π is expressed in an 8-neighborhood with a step size of T as

[0020] Furthermore, the local graph convolution is to calculate the weights of neighbor nodes through local graph convolution of the adjacency matrix of the adaptive graph convolution and mask out the nodes with low weights to obtain a similarity matrix R, thereby masking the redundant information in the subsequence.

[0021] Specifically, a mask matrix M is first defined to remove smaller edge values ​​in the adjacency matrix A, as shown in Formula 1:

[0022]

[0023] Where τ is the mask threshold, i and j are the indices of the matrix elements. Then use M and A to define a new similarity matrix R, where each element R ij Calculate according to the following formula:

[0024]

[0025] The continuous-time modeling includes:

[0026] The data of a single subsequence at this moment and the output of the subsequence at the previous moment are modeled in continuous time using the Neural Ordinary Differential Equation to simulate the continuous evolution of the system.

[0027] After modeling for a period of time, the gating mechanism is used to capture the long-term dependencies in the data to alleviate the forgetting problem of continuous time modeling;

[0028] Repeat the above two steps until the time point of the next subsequence is modeled, and the output of the subsequence is obtained and used for the continuous time modeling of the next subsequence;

[0029] Repeat the above three steps until all subsequences are modeled.

[0030] The present invention has the following beneficial effects and advantages:

[0031] The present invention is a method for predicting key parameters of industrial processes based on a spatiotemporal graph convolutional network. When dealing with the time-lag characteristics in industrial process data, the time-lag features in the sequence are extracted through sequence partitioning and global graph convolution. The embedded vector is optimized for redundancy using local graph convolution, so that the model only focuses on the features related to itself, alleviating the interference of redundant information. By extracting finer-grained features from each subsequence through continuous time modeling, the problem of insufficient feature extraction caused by missing data in the model is compensated. In general, the present invention improves the prediction accuracy in predicting key parameters of industrial processes based on a spatiotemporal graph convolutional network. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 The figure is a flow chart of the method of the present invention.

[0033] Figure 2 The figure is a flow chart of the modeling method within a subsequence and between subsequences of the present invention.

[0034] Figure 3 Schematic diagram of the delay matrix in the global graph convolution of the present invention.

[0035] Figure 4 This is a flow chart of the continuous-time modeling method of the present invention. DETAILED DESCRIPTION

[0036] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation method of the present invention is described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the invention, so the present invention is not limited by the specific implementation disclosed below.

[0037] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The present invention is further described in detail below in conjunction with the accompanying drawings and examples.

[0038] like Figure 1 The flowchart of the method of the present invention is shown in FIG. A method for predicting key parameters of industrial processes based on a spatiotemporal graph convolutional network. The programming language used in the program execution steps of the present invention is not limited to MATLAB, Python, etc.

[0039] The specific steps of the present invention are as follows:

[0040] Step 1, collect industrial key parameter data (taking the loosening and rehumidification process of the food processing industry as an example, the variable data includes 24 variables such as the moisture content of the material at the drum inlet, the moisture content of the material at the drum outlet, and the actual value of water addition, the water valve opening, the actual value of the outlet temperature, the water addition setting value, the actual temperature of the barrel 1, the actual temperature of the barrel 2, and the water addition coefficient. The key parameter is the moisture content of the material at the drum outlet). The present invention can predict the moisture value of the material at the drum outlet based on the 24 variables to provide a reference for industrial control.

[0041] Step 2: For each data, remove outliers according to the set error range, and then perform maximum and minimum normalization processing to normalize the data to between (0, 1).

[0042] Step 3: Create a sample data set for model training and testing. Each sample contains L+T sampling values, where the first L sampling values ​​constitute the historical input data for training, and the last T sampling values ​​serve as prediction targets, that is, label values ​​during model training and testing.

[0043] Step 4: Use a sliding window of fixed size to divide the sample data generated in step 3 into multiple subsequences according to a specific step size. The length of each subsequence is equal to the size of the window.

[0044] Step 5: Model the variable dimensions within the subsequence as follows Figure 2 As shown in the figure. The data after data partitioning will use global graph convolution to extract all dependencies within the entire subsequence. It can more effectively extract and utilize global features in the data through flexible modeling of graph structure. At the same time, the data after data partitioning also uses local graph convolution to calculate the weights of neighbor nodes and mask out nodes with low weights, thereby alleviating the problem of information redundancy in industrial process data.

[0045] Furthermore, the global graph convolution is based on the adaptive graph convolution, and its adjacency matrix A is multiplied by a time delay matrix π to obtain an enhanced matrix This enhances the model's perception of the temporal distance within a subsequence. Assuming there is an adjacency matrix A containing three time points, the delay matrix is ​​as follows: Figure 3 As shown, the matrix measures the time distance through the decay rate δ<1. Specifically, the farther away from the current time step, the greater the degree of decay. Among them, the time delay matrix π is in an 8-neighborhood with a step size of T.

[0046]

[0047] Furthermore, the local graph convolution is based on the adaptive graph convolution, and its adjacency matrix is ​​replaced with a similarity matrix to mask the redundant information in the subsequence. The weights of neighbor nodes are calculated through local graph convolution and nodes with low weights are masked. Specifically, we first define a mask matrix M to remove smaller edge values ​​in the adjacency matrix A, as shown in Formula 1:

[0048]

[0049] Where τ is the mask threshold, i and j are the indices of the matrix elements. Then use M and A to define a new similarity matrix R, where each element R ij Calculate according to the following formula:

[0050]

[0051] Step 6: Model the time dimension between subsequences as follows Figure 2 As shown in Figure 1. For each subsequence, continuous time modeling is used to extract more fine-grained features, which makes up for the problem of insufficient feature extraction caused by missing data in the model; then, the time features of the entire sequence are extracted through multi-head attention to obtain the output of time dimension modeling. The specific process of continuous time modeling is as follows: Figure 4 shown.

[0052] Step 6-1, using the Neural Ordinary Differential Equation to perform continuous time modeling on the data of a single subsequence at this moment and the output of the subsequence at the previous moment, so as to simulate the continuous evolution of the system;

[0053] Step 6-2: After modeling for a period of time, capture the long-term dependencies in the data through the gating mechanism to alleviate the forgetting problem of continuous time modeling;

[0054] Step 6-3, repeat the above steps 6-1 and 6-2 until the time point of the next subsequence is modeled, and the output of the subsequence is obtained and used for the continuous time modeling of the next subsequence;

[0055] Step 6-4: Repeat step 6-3 above until all subsequences are modeled.

[0056] Step 6-5: Input all the above subsequences into the multi-head attention to extract the temporal dependency of the entire sequence from a global perspective.

[0057] Step 7: Use the extracted features combined with the multi-layer perceptron to predict the future key parameter sequence of the industrial process.

[0058] Step 8: Repeat steps 4, 5, 6, and 7, select the model parameters with the best accuracy as the final optimization model, and predict the future key parameter sequence based on the historical data of the industrial process.

[0059] The data set is set as the loose rehydration process of a food processing industry, the historical data step size L = 16, the prediction data step size is 6, and the evaluation indicators are the mean absolute error (MAE), root mean square error (RMSE) and weighted absolute percentage error (WAPE). The performance results of the method and the other 7 methods are shown in Table 1, indicating that the method of the present invention improves the prediction accuracy of key parameter data of the industrial process compared with other existing methods, and provides technical support for the optimization control of the industrial process.

[0060] Table 1

[0061]

[0062] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any way. Any simple modification, change and equivalent structural change made to the above embodiment based on the technical essence of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A method for predicting key parameters of industrial processes based on spatiotemporal graph convolutional networks, characterized in that: The steps include: S1. Collect key parameter data of industrial processes; S2, preprocess the data, establish the data set and divide the sequence; S3, variable dimension modeling within the subsequence: embed the features of a single sequence point within the subsequence into the spatial dimension to generate a new embedding vector, which is used for the subsequent network to capture the time lag characteristics in the data during modeling; then use global graph convolution to model all dependencies within the subsequence; at the same time, use local graph convolution to optimize the redundancy of the embedding vector so that the model only focuses on the features related to itself; finally, add the outputs of the two to complete the time series data modeling within the subsequence; S4. Temporal dimension modeling between subsequences: For each subsequence, more fine-grained features are extracted through continuous time modeling to make up for the defect of insufficient feature extraction caused by missing data. Then, the temporal features of the entire sequence are extracted through multi-head attention to obtain the output features of temporal dimension modeling. S5. Establish a multi-layer perceptron network to iteratively learn the mapping relationship between input features and output labels, so as to predict the future key parameter sequence of industrial processes; S6. Using the data in the dataset as input, repeat steps 3 to 5 for iterative training, select the model parameters with the best accuracy as the final ideal model, and predict the future key parameter sequence based on the historical data of the industrial process.

2. The method for predicting key parameters of industrial processes based on spatiotemporal graph convolutional networks according to claim 1 is characterized in that: The key parameter data of the industrial process are industrial status data and process control data collected by detection sensors and control equipment set in the industrial production environment.

3. The method for predicting key parameters of industrial processes based on spatiotemporal graph convolutional networks according to claim 1 is characterized in that: The preprocessing includes removing outliers and normalizing the maximum and minimum values.

4. The method for predicting key parameters of industrial processes based on spatiotemporal graph convolutional networks according to claim 1 is characterized in that: The data set is divided into training set, validation set and test set in proportion; the training set data is used for training and cutoff according to the set number of iterations; the validation set data is used to verify the model accuracy; and the test set data is used for testing. Furthermore, each sample data in each set contains L+T sampling values, of which the first L sampling values ​​constitute the historical input data for training, and the last T sampling values ​​serve as the label value of the prediction target during the model training and testing process.

5. The method for predicting key parameters of industrial processes based on spatiotemporal graph convolutional networks according to claim 1 is characterized in that: The sequence division is to use a sliding window of a fixed size to divide the entire historical input data sequence of each sample into multiple subsequences according to a specific step size, and the length of each subsequence is equal to the size of the window.

6. The method for predicting key parameters of industrial processes based on spatiotemporal graph convolutional networks according to claim 1 is characterized in that: The global graph convolution is to multiply the adjacency matrix A of the adaptive graph convolution by a time delay matrix π to obtain an enhanced matrix 7. The method for predicting key parameters of industrial processes based on spatiotemporal graph convolutional networks according to claim 1 is characterized in that: The local graph convolution calculates the weights of neighbor nodes through the local graph convolution of the adjacency matrix of the adaptive graph convolution and masks the nodes with low weights to obtain a similarity matrix R, thereby masking the redundant information in the subsequence.

8. The method for predicting key parameters of industrial processes based on spatiotemporal graph convolutional networks according to claim 1 is characterized in that: The continuous-time modeling includes: The data of a single subsequence at this moment and the output of the subsequence at the previous moment are modeled in continuous time using the Neural Ordinary Differential Equation to simulate the continuous evolution of the system. After modeling for a period of time, the gating mechanism is used to capture the long-term dependencies in the data to alleviate the forgetting problem of continuous time modeling; Repeat the above two steps until the time point of the next subsequence is modeled, and the output of the subsequence is obtained and used for the continuous time modeling of the next subsequence; Repeat the above three steps until all subsequences are modeled.

Images