FORMULATING THE LEAST-SQUARE REGRESSION FOR CONTINUOUS ANALYSIS.
LOUISIANA STATE UNIV BATON ROUGE COLL OF ENGINEERING
Pagination or Media Count:
Numerous techniques have been developed to adapt to changes in process parameters, most of which require the system to be taken off control and disturbed so that data may be collected for either on-line or off-line analysis. These techniques in many cases cannot be justified even though adaptive modeling is needed. The requirement that the process be intentionally disturbed in most cases is undesirable and frequently completely impractical. Also, the actual computer programs required to perform the analysis are generally long both in execution time and computer storage. This report demonstrates that in most cases these requirements are not necessary for either on-line or off-line analysis and that the regression problem can be formulated so that changes in model parameters can be continually monitored under normal controlled operation without extensive use of valuable computer time or storage. Author
- Industrial Chemistry and Chemical Processing
- Statistics and Probability
- Test Facilities, Equipment and Methods