Better Understanding of Process Operation Using Performance Metrics and Visualization Plots

摘要

In an industrial facility, a large amount of data on the operation of the facility is collected. The data includes control variables used in monitoring and regulation, alarm variables displayed to notify operators of abnormalities, and operators' actions taken to correct abnormalities as well as to operate the facility. As the data is historized and is readily accessible, it can provide a wealth of knowledge on the operation of the facility. Thus, this paper tackles the following question: Can the historized data be used to extract information to better understand process operation? More specifically, the paper attempts to answer the question by its two contributions, namely: (i) the development of performance metrics that characterize the time and frequency relations of events, and (ii) the development of two types of visualization plots to present the performance metrics in a succinct fashion. The paper also discusses the incorporation of the performance metrics and the visualization plots into industrial automation platforms.