RMBatch: Intelligent real-time monitoring and diagnosis system for batch processes

Abstract

Batch Process-generally exhibits batch-to-batch variation that makes it difficult to produce a uniform high quality product. The variation arises from sources such as: composition disturbances, deviations from specified trajectories. equipment malfunction and heat transfer limitation. Extensive research has been done to develop a model for the efficient and reliable monitoring and diagnosis for these batch processes. Various algorithms such as multivariate statistical process control based on multhvay principal component analysis and partial least square have been proposed for the monitoring of batch processes. A prototype of intelligent Real-time Monitoring and diagnosis system for Batch process (“RMBatch”) has been developed as a tool that offers real-time monitoring and diagnosis for batch processes. Our developed RMBatch is composed of plant information module, data preprocessing module, data analysis and modeling module, monitoring and diagnosis module. The design and the various functions of RMBatch will be illustrated using its application to an industrial batch process. The application has shown that RMBatch can help operators monitor, diagnose and improve the operation of batch processes.