Modeling Approaches Applied to Pulse Jet Mixing Data - Part 2

• Activity Number: 313
• Type: Contributed
• Date/Time: Tuesday, August 4, 2009 : 10:30 AM to 12:20 PM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): Brett Amidan*+ and Greg Piepel and Alejandro Heredia-Langner and Perry Meyer and Beric Wells and James Fort and Judith Bamberger and William Kuhn
• Companies: Pacific Northwest National Laboratory and Pacific Northwest National Laboratory and Pacific Northwest National Laboratory and Pacific Northwest National Laboratory and Pacific Northwest National Laboratory and Pacific Northwest National Laboratory and Pacific Northwest National Laboratory and Pacific Northwest National Laboratory
• Keywords: Physical models ; nondimensional variables ; spurious correlation ; semi-empirical models ; multiple regression

Abstract

Pulse jet mixing tests were conducted to support the design of mixing systems for the Hanford Waste Treatment and Immobilization Plant. A physical approach (based on hydrodynamic behavior) and two semi-empirical (SE) approaches were applied to the data to develop models for predicting two response variables (critical-suspension velocity and cloud height). Part 1 discusses background information, the tests, the three modeling approaches, and spurious correlation. This Part 2 presents one physical and two SE models for each response. The SE models were developed by generalizing the physical model and using dimensional and/or nondimensional (ND) variables. The results of fitting the physical and SE models are presented and compared. Considering goodness-of-fit, prediction performance, spurious correlation, and the need to extrapolate, the SE models based on ND variables are recommended.