Date of Thesis

5-4-2017

Thesis Type

Masters Thesis (Bucknell Access Only)

Degree Type

Master of Science in Mechanical Engineering

Department

Mechanical Engineering

First Advisor

Nathan Siegel

Second Advisor

M. Laura Beninati

Abstract

The purpose of this thesis is to develop a general modeling approach to guide the design of high-flux solar simulators based on metal-halide arc lamps. Currently, high flux solar simulators are each custom built, which results in high costs and long lead times, both of which would be reduced using the results presented in this thesis. In general, the current design process involves parametric simulations of a lamp/reflector pair, the result of which is a flux map (i.e. the spatial variation of irradiance) on a target plane some distance from the reflector. The current methods for modeling light sources (i.e. arc lamps) in solar simulator design calculations include treating these non-uniform volumetric emitters as either spherical or cylindrical surface sources or uniform volumetric emitters. This simplification of the complex arc shape and power distribution leads to a disagreement between theoretical and experimental results in both the predicted power delivered and peak irradiance, without agreement, it is not possible to design solar simulators based strictly on theoretical results in the absence of experimental data. Developing a general approach for designing solar simulators therefore first requires a theoretical model for the light source that replicates the performance of actual hardware. To create such a model four arc lamps have been experimentally characterized using an imaging technique to elucidate the variation in the intensity of emitted light within the arc. From these characterizations, a semi-empirical, general model for arc lamp light sources was developed and combined with an optical model for the elliptical reflector common to high flux solar simulators. The combined lamp/reflector model was validated using experimental flux maps. The model has been shown to predict the performance of high-flux solar simulators within 8.6% of power delivered and 19.7% of peak irradiance. With the general model in place a solar simulator design methodology was developed. The design methodology requires only the lamp power, lamp geometry, and reflector properties as the basis of a complete optical model of a solar simulator module. Using the methodology any lamp/reflector pair can be used to design a simulator and the performance can be predicted within 10% of total power delivered and 20% of peak flux. This methodology will improve the accuracy of design calculations and help to reduce the cost of custom-built solar simulators.

Share

COinS