Qualitative trend analysis for the monitoring of process plant

Time series analysis is of great importance in many areas of science and technology. Monitoring of process plant would be a typical problem of industrial time series analysis. Very often, a qualitative analysis is enough to solve the problem. This paper introduces a methodology to analyze qualitative trend of time series data. Based on the methodology, a method to diagnose a control valve in a process control loop is developed. The diagnosis method is successfully applied to several data sets in industrial chemical processes.