(499f) Integration of Principal Component Modeling Methods in a Model Predictive Control Architecture

Manufacturing processes and measurement systems are becoming increasingly more complex. Recently it has become a common practice to use scores and residual statistics from principal component analysis (PCA) and partial least squares (PLS) models to summarize and monitor the state of a process. This session describes the integration of these PCS and PLS statistical metrics in a model predictive control (MPC) structure to provide the ability to maintain a process within a desired operating region. This is particularly relevant for the regulated pharmaceutical and biotech manufacturing environments where quality by design and design space quality systems require processes to be maintained within operating windows that assure quality.