Document Type

Thesis - Open Access

Award Date

2021

Degree Name

Master of Science (MS)

Department

Mechanical Engineering

First Advisor

Stephen Gent

Keywords

Computational Fluid Dynamics, Cryogenics, Inverse Problems, Natural Convection

Abstract

Computational fluid dynamics (CFD) is a branch of fluid mechanics which is employed to numerically solve complex fluid, heat transfer, and multiphysics problems. Traditionally, CFD techniques are used to solve “forward” problems—using some known information of a system as inputs to a representative model to predict experimental measurements or expected system behavior. The work presented here demonstrates how CFD may be used to solve an “inverse” problem—given limited experimental data and some model, predict (previously unidentified) “input” system (or system model) parameters. The case study for this research uses a validated CFD modeling approach of the liquid argon (LAr) region of the ProtoDUNE Single Phase neutrino detector. Incomplete experimental temperature data (which deviated from the expected, roughly-linear distribution with height for such a natural convection driven flow) are used to inform parametric changes to the base CFD model. Features such as the addition of previously neglected physical geometries and heat sources were parametrically added to the model in the commercial CFD program Star-CCM+, and the resulting temperature distributions were compared to the experimental data. Results of this study suggest that there are numerous possible causes for the abnormal experimental temperature distribution. Model inputs such as increased heat from the cold electronics and the field cage (Faraday cage) and lowering of the LAr height caused a more nonlinear temperature distribution in the sensor region, improving CFD agreement. The addition of previously neglected flow obstructions near the LAr surface do not directly improve the temperature agreement but are significant to the flow patterns and thus should be included in future modeling. Confidence in the inverse problem solution is limited by uncertainties in “known” system information.

Number of Pages

125

Publisher

South Dakota State University

Rights

Copyright © 2021 Cecilia Streff

Share

COinS