Stationary Subspace Analysis-Based Hierarchical Model for Batch Processes Monitoring

Abstract

Conventional batch process monitoring strategies implement phase partition using all the collected variables in high dimensions, which may result in high computation complexity and inaccurate division results. Besides, due to the time-varying characteristics, the acceptable operation regions of many monitoring models are generally too wide and thus their detection sensitivity may be compromised. In this brief, a stationary subspace analysis (SSA)-based hierarchical monitoring model is developed to solve the aforementioned issues. The proposed method extracts the global stationary features from the historical process data and establishes a global monitoring model for the time-invariant information throughout the whole batch process. Based on the remaining nonstationary global features, a phase partition method is developed to divide the process using dynamic information in low dimensions. According to the partition result, local monitoring models are constructed for each operation phase using equilibrium relationship and dynamic information. The operation status of the newly collected sample is codetermined by both the global and local models, and a physical interpretation is provided for better process understanding. The proposed method is illustrated using a simulated process and a real industrial process. The experimental results show that the proposed method can extract key features to accurately divide the batch process into different operation phases and effectively detect the incipient fault so that immediate and corrective actions can be taken.