Morning Tutorial

"Practical Aspects of Batch Process Monitoring"

 
Sunday, October 5 • 8AM-NOON
Instructor: Neal B. Gallagher, Ph.D.

Batch processes (such as metal etch) produce data that is a function of three modes: time step, measurement (variable) and batch number. This  data can be analyzed with a number of multivariate techniques, such as Principal Components Analysis (PCA), with different organizations of the data, i.e. various unfolding schemes. Batch data can also be  converted to descriptor or summary variables that can be used with PCA or other conventional multivariate models. True multi-way models, such as PARAFAC and Tucker, can also be used. These methods each have different  degrees of flexibility and can accommodate different non-idealities. Thus, for optimal model performance, the best model form should be matched to the process. Data preprocessing, such as warping, centering and scaling can also greatly affect model performance. In this half-day tutorial we review model forms and preprocessing methods and show how they can be used most effectively for batch process monitoring.

About the Instructor

Neal B. Gallagher, Ph.D. is vice president and co-founder of Eigenvector Research, Inc., a Washington-based company specializing in chemometrics consulting, algorithm development, training, and software for the semiconductor and chemical process industries, consumer product manufacturers and analytical instrument developers. Gallagher is actively involved in the development of new software products that enable the use of the newest research and methodology in a practical environment, including PLS_Toolbox advanced chemometrics software for use with MATLAB.

Gallagher received B.S. degrees in Chemical Engineering and Engineering Physics from the University of Colorado in 1985, an M.S. in Chemical Engineering from the University of Washington in 1987, and a Ph.D. in Chemical Engineering from the University of Arizona in 1992.

He has authored and contributed to several scientific publications and is currently a member of Sigma Xi, AAAS and IEEE.

More information about Eigenvector Research can be found at www.eigenvector.com.

Registration Information

You must be a registered symposium attendee to register for the tutorials. The combined fee for both tutorials $75 if registered on or before Sept. 1 . After that date, a registration fee of $125 will apply. To register for this event, check the corresponding "Tutorial" box on the Symposium Registration form. Please see the AEC/APC Symposium refund/cancellation policy for refund and cancellation deadlines.