Dynamic-scale graph neural network for fault detection

Traditional graph-based dynamic fault detection methods describe the dynamic characteristic through con-structing a single neighborhood graph at the current sample with some history samples. However, they ignore the diversity of dynamic properties of the variables in complex chemical processes. To overcome this problem, a novel neural network structure combining multiscale subgraphs is proposed, named dynamic-scale graph neural network (DSGNN), which divides variables into multiple groups according to their dynamic properties. DSGNN constructs a subgraph in each group. In traditional graph-based methods, the scale of the graph is usually manually designed. In DSGNN, the scale of each subgraph is decided by the dynamic properties of the variables in this subgraph. To aggregate the dynamic information, DSGNN utilizes convolution operations. The weights assigned to the neighbors in each subgraph are determined according to the similarity between the current data and its neighbors. Low-dimensional features are extracted through the back-propagation technique from the updated high-dimensional features produced by convolution operations. Two case studies on a multivariate dynamic process and the Tennessee Eastman process are conducted to show the superiority of DSGNN.

Automated summary

Traditional graph-based dynamic fault detection methods overlook the diversity of dynamic properties of variables in complex chemical processes. To address this issue, a novel neural network structure named DSGNN is proposed, which divides variables into groups based on their dynamic properties and constructs subgraphs for each group. DSGNN utilizes convolution operations to aggregate dynamic information and extracts low-dimensional features using back-propagation technique. Two case studies demonstrate the superiority of DSGNN in multivariate dynamic processes and the Tennessee Eastman process.