Related Experiment Videos
Visualization of a pharmaceutical unit operation: wet granulation.
Anna Cecilia Jørgensen1, Jukka Rantanen, Pirjo Luukkonen
1Pharmaceutical Technology Division, Faculty of Pharmacy, P.O. Box 56, FIN-00014 University of Helsinki, Helsinki, Finland.
Analytical Chemistry
|September 15, 2004
Summary
Combining physical and chemical data into a process vector enhances process understanding. Visualization tools like Self-Organizing Maps (SOM) offer a comprehensive overview, improving manufacturing sciences.
Area of Science:
- Process Engineering
- Manufacturing Sciences
- Data Visualization
Background:
- Advanced process engineering and manufacturing sciences offer deeper process insights.
- Extracting actionable understanding from vast data volumes presents a significant challenge.
Purpose of the Study:
- To develop a process vector integrating all pertinent information from a model process.
- To create a tool for effective combination and visualization of complex process data.
Main Methods:
- Collected physical (impeller torque, temperature) and chemical (near-infrared spectroscopy) data from high-shear granulation.
- Constructed process vectors using combined data.
- Visualized vectors using Principal Component Analysis (PCA) and Self-Organizing Maps (SOM).
Main Results:
- Individual measurement techniques alone were insufficient to fully describe the process state.
- Combined data visualization provided a comprehensive process overview.
- Self-Organizing Maps (SOM) offered advantages over PCA, including direct variable interpretation and nonlinear analysis.
Conclusions:
- Integrating diverse data streams into a process vector is crucial for enhanced process understanding.
- Both PCA and SOM are valuable for visualizing process progress and improving overall comprehension.
- The SOM approach demonstrates particular utility for analyzing complex, nonlinear process dynamics.