Use of Advanced Multivariate Analysis to Solve Plant Problems

Cross-functional collaboration of statisticians and chemical engineers can be a winning strategy to solve chemical plant problems. In modern chemical engineering plants, we collect massive amounts of data. This plant data is highly correlated due to the underlying mass and energy balance. Multivariate analyses, often referred to as chemometrics, have been used to analyze highly correlated data to solve plant problems. This data analysis combined with the right subject matter expertise can be a powerful strategy for troubleshooting. In this presentation, we will give an overview of multivariate analysis using a commercial tool (SIMCA). We will provide examples using actual plant troubleshooting problems. The examples will highlight how using these advanced data analytics methods, we were able to compare the stable plant operating conditions with the problematic ones and quickly identify the main differences. This led to identifying the exact root cause and quick resolution of the issue.