Insights from Transforming Factorial Variables into Mixture Variables (2025-US-30MP-2524)

This presentation examines an initially routine DOE aimed at elucidating the effects of various resin synthesis variables on the properties of a coating. Preliminary data analysis uncovered several promising correlations between the synthesis parameters and the resulting coating characteristics. To enhance our understanding, we developed a theoretical framework to predict the individual components of the resin based on the underlying reaction mechanisms and straightforward probability calculations. This innovative approach enabled us to reinterpret the original DOE variables as mixture variables, facilitating a more nuanced analysis.

By reanalyzing the data within the context of these mixture variables, we uncovered new insights into how resin composition affects coating performance, leading to novel and counterintuitive ideas for resin enhancements.

We demonstrate the application of JMP software throughout this study, highlighting several analytical tools, including the DOE platform, prediction profiler, ternary plot, scatter plot matrix, and mixture profiler.