Application of Several JMP Tools for Process Improvement: A Purification Case Study (2026-EU-30MP-2782)

This presentation demonstrates how JMP tools support process optimization and online NIR implementation for reliable product quality monitoring. It includes a case study in which SPC and capability analysis are applied to assess product purity, while process control ranges are defined using DOE, Genreg/SVEM modeling, and the Design Space Profiler.

For NIR-based purity modeling, PCA and FDA scores serve as inputs for PLS, Genreg, SVEM, and ANN models. Results show that Genreg and machine learning approaches clearly perform better. Process capability simulations with non-normal purity distributions are best when SVEM is used for model calibration. However, excessive variation in purity and insufficient capacity are predicted, highlighting the limitations of current models.

The study concludes that collecting additional NIR spectra is essential to fill gaps in PCA and FPCA score spaces, thereby enabling more robust and reliable ML model development.