The New Jersey Institute of Technology's
Electronic Theses & Dissertations Project

Title:	Numerical modeling and optimization of waterjet based surface decontamination
Author:	Babets, Konstantin
View Online:	njit-etd2001-044 (xviii, 204 pages ~ 28.7 MB pdf)
Department:	Department of Mechanical Engineering
Degree:	Doctor of Philosophy
Program:	Mechanical Engineering
Document Type:	Dissertation
Advisory Committee:	Geskin, E. S. (Committee chair) Harnoy, Avraham (Committee member) Ji, Zhiming (Committee member) Sodhi, R. S. (Committee member) Grebel, Haim (Committee member)
Date:	2001-01
Keywords:	Waterjet Cleaning Waterjet Nozzle Artificial Intelligence Neural Networks Genetic Algorithm
Availability:	Unrestricted
Abstract:	The mission of this study is to investigate the high-pressure waterjet based surface decontamination. Our specific objective is to develop a practical procedure for selection of process conditions at given constraints and available knowledge. This investigation is expected to improve information processing in the course of material decontamination and assist in the implementation of the waterjet decontamination technology into practice. The development of a realistic procedure for processing of a chaotic and non-accurate information constitutes the main accomplishment of this study. The research involved acquisition of representative information about removal of brittle, elastic and viscous deposits. As a result an extended database representing jet based decoating has been compiled and feasibility of the damage free decontamination of various surfaces including highly sensitive ones is demonstrated. Artificial Intelligence techniques (Fuzzy Logic, Artificial Neural Networks, Genetic Computing) have been applied for processing of the acquired information and a realistic procedure of such an application has been developed and demonstrated. This procedure enables us to integrate available information about surface in question and existing numerical models. The developed procedure allows a user to incorporate both qualitative (linguistic) and quantitative (crisp) information into a process model and to predict operational conditions for treatment of an unknown surface using a readily detectable single experimental parameter that characterizes a deposit/substrata combination. The suggested technique is shown to perform reliably in the case of incomplete and chaotic information, where the traditional regression based methods fail. Numerical simulations of the two-phase flow inside a waterjet nozzle are conducted. Numerical solutions of the partial differential equations of the two-phase turbulent jet flow are obtained using FLUENT package. The numerical prediction of jet velocity profiles and the interface between the two phases (water - air) inside a nozzle are in good agreement with experimental data available in the literature. Thus the current problem setup and the results of simulations can be applied to improvement in the nozzle design. A realistic procedure for the design of the jet based surfaces decontamination developed, as a result of this study, is applied for optimization of the removal of the paint, rust, tar and rubber from the steel surface.