Share Email Print

Proceedings Paper • Open Access

Optimization, sensitivity analysis, and robust design using response surface modeling
Author(s): Egbert Lenderink; Peter Stehouwer

Paper Abstract

Response surface modeling (RSM) is described as a tool to perform optimizations and sensitivity analysis in optical modeling. With this method, the behavior of the system is first mapped out using a limited set of ray-tracing simulations, carefully spanning the full parameter space. This can already be done before the full merit function is known. The relation between design parameters and system performance is then approximated by fitting the simulation results to functional forms. All subsequent optimizations are then performed very time-efficiently on the functional descriptions of the dependencies. This contrasts with direct optimization, where the computationally intensive optical modeling is in the loop of the optimization algorithm, and where changes in merit function, mapping out trade-offs, and determining sensitivities, are very time consuming. The paper will discuss the advantages of RSM with respect to direct optimization and give recommendations for the type of problems that are preferentially addressed by RSM. The method will be illustrated by a case: how optical simulations were used in the design of LumiramicTM phosphor conversion components for LEDs.

Paper Details

Date Published: 20 September 2008
PDF: 8 pages
Proc. SPIE 7103, Illumination Optics, 710302 (20 September 2008); doi: 10.1117/12.797649
Show Author Affiliations
Egbert Lenderink, Philips Lighting (Netherlands)
Peter Stehouwer, Ctr. for Quantitative Methods (Netherlands)

Published in SPIE Proceedings Vol. 7103:
Illumination Optics
Tina E. Kidger; Stuart R. David, Editor(s)

© SPIE. Terms of Use
Back to Top
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?