Application of Statistical Methodology to Chemical Pilot Plant Processes

For a chemical reaction process, to achieve a targeted yield, multiple catalysts are investigated and compared over a certain reaction period. For some scenarios, such as a batch reactor, different catalysts can be employed and compared over the same reaction period. For some other scenarios, such as a continuous pilot plant, different catalysts may have to be employed at different reaction periods to achieve the final targeted yield. During the process, several questions are of interests, such as: Does the selectivity of product/by-product change across the period of time? Is there an effect from the different catalysts? Is there a time effect for those catalysts? If there are some differences among the catalysts over the reaction time, are those differences are statistically significant? To address those questions, there are several challenges to develop a meaningful approach to data analysis in this system due to the change of the product/by-product selectivity over time. The work here is to present a developed statistical method to address those questions. Two examples are presented to demonstrate how to apply the statistical methodology to the chemical reaction processes.