Batch Multivariate Statistical Process Control (MSPC) for PAT

Course Description

Today's highly instrumented chemical and manufacturing processes produce a tremendous amount of data, much of which is archived and only reviewed after a major process upset or fault. Many of these processes, particularly in pharmaceutical applications, are run in batch mode, which adds additional complexity to the process modeling problem. Using chemometric models in the Process Analytical Technolgy (PAT) framework involves some unique considerations, including regulatory compliance, on-line model deployment logistics, and model performance monitoring.

Batch MSPC for PAT covers methods and strategies for dealing with this data overload and extracting critical information about process health. The course covers monitoring and fault detection in batch chemical and manufacturing processes. Using diagnostic plot to track down root causes is covered. Implementation and deployment issues in the PAT environment are considered at length. The course includes hands-on computer time for participants to work example problems using PLS_Toolbox.


Linear Algebra for Chemometricians, MATLAB for Chemometricians or equivalent experience. Chemometrics I--PCA or equivalent experience highly recommended.

Course Outline

  1. Introduction
    General Principles of SPC and Fault Detection
    Use of Models
    The Multivariate Advantage
  2. A favorite tool: Principal Components Analysis
    Some examples of PCA for MSPC
    Diagnostic Plots for Interpreting and Sourcing Faults
    PCA Scores and Loadings
    Q and T2 Statistics
    Contribution plots
  3. Theoretical basis for MSPC
    Time Series Models and Lagged Variables
    More examples
  4. Modeling Batch Processes
    Unfold PCA (aka Multi-way PCA)
    Parallel Factor Analysis (PARAFAC)
    Tucker Models
  5. Strategies for Dealing with Unequal Batch Record Length
    Summary Variables
    Correlation Optimized Warping (COW)
    Dynamic Time Warping (DTW)
    Extent of Reaction and other Indicator Variables
    Batch Maturity Index
    Comparison of methods on some example data
  6. Implementing Models for PAT
    Sampling issues, calibration protocols
    Cleaning “messy” data
    Developing and Optimizing Models
    Validating and Testing Models (QA)
    Model updating: “augment, or replace?”
  7. Model Deployment Logistics
    Review of Different Deployment Scenarios/”Landscapes”
    Enabling IT Technologies
    DCS Integration Issues
    Organizational and Compliance Issues in Deployment
    Deployment Solutions
    Implementation Checklists
    Documentation, and Database management
  8. Examples
    On-line deployment demos
    Case Studies