Industrial Process Monitoring Method For Fused Magnesia Based On Discriminant Analysis And Bilinear Regression

The fused magnesia smelting process is complicated and prone to fault.Because it is difficult to establish an accurate mathematical model,the data-based process monitoring method is used to monitor the fused magnesia smelting process.In order to solve the problem of obtaining the marking data of fused magnesia smelting process and how to use soft marking data,a process monitoring method based on linear discriminant analysis of trusted soft label is proposed.In order to solve the problem that the image data dimension is too high during the fused magnesia melting process,and the calculation speed is slow in the modeling or online monitoring process,a process monitoring method based on two-dimensional two-dimensional sparse orthogonal discriminant analysis is proposed.Because using one type of data for modeling information is less or incorrect,and the connection between multiple data is ignored.A method of monitoring based on heterogeneous data manifold maintenance and bilinear regression is proposed.In response to the above questions,the research done in this paper is as follows:(1)In order to solve the problem that the marked data is less,the label propagation method is used to obtain the soft label of the unmarked data,and the problem of the traditional positive label propagation method is solved,and the soft label with higher accuracy is obtained.In order to solve the problem of how to use soft marker data,the linear discriminant analysis method of trusted soft label is used to reduce the dimension of the training data to obtain the projection matrix from high dimensional space to low dimensional space.The process monitoring model of fused magnesia smelting process was established.The experimental results show that the algorithm can effectively identify the operating state of fused magnesia smelting process.(2)In order to solve the problem that the image data dimension is too high during the fused magnesia smelting process,the feature of the image data is selected based on the two-dimensional two-dimensional sparse orthogonal discriminant analysis method,and the selected features of the image data are retained,and the unselected feature is set to zero.The features selected by this method are the features that best represent the differences between multiple types of data.The feature matrix is obtained by dimension reduction of the sparse matrix data by the lateral projection matrix and the vertical projection matrix.The feature matrix is reduced by the method of bilinear regression to obtain the left projection vector and the right projection vector of each type of data.The process monitoring model of fused magnesia smelting process was established.The experimental results show that the method can quickly and effectively identify the operating state of fused magnesia smelting process.(3)In order to solve the problem that using one type of data for modeling information is less or incorrect,and the connection between multiple data is ignored,the feature matrix of sparse matrix data is reduced by using heterogeneous data manifold maintenance and bilinear regression method.Modeling based on heterogeneous data makes full use of information about image data and physical data.The lateral projection matrix and the vertical projection matrix of the original matrix data are obtained by the two-dimensional two-dimensional sparse orthogonal discriminant analysis method.The left projection vector and right projection vectors of each type of data of the feature matrix is obtained by the bilinear regression method based on heterogeneous data manifold maintenance.The monitoring model of fused magnesia smelting process was established,and the effectiveness of the method was proved by experimental simulation.