Phase I: Modeling Steady State Conditions | Fault Detection and Diagnosis in Chemical and Petrochemical Processes, Part 2

The goal of process monitoring is to ensure that the planned operations are successful. You can achieve this goal by recognizing process upsets and faults using data-driven measures, such as PCA. These measures are derived directly from process data and aid in fault detection and diagnosis by transforming the high dimensional data into a lower dimension, and thereby capturing important information in the process.

Create a PCA model to describe the normal variability in the operation of a methanol-ethanol distillation column. Building an effective monitoring system requires a good data set that represents the steady-state, normal operating conditions. Use an application developed in MATLAB^® by GIEM as an aid for understanding the PCA-based MSPC strategy.