Design of a freeform reflector for LED sources based on a feedback method

Abstract

We propose a method for the design of a freeform reflector by means of feedback methods to create light-emitting diode (LED) sources with uniform illuminance. A reflector is first designed to obtain a uniform circular light distribution for a point source. Then the optical performance of the reflector is investigated for both the point sources and extended sources based on a Monte Carlo ray tracing method. Comparing the illumination differences, a new reflector is generated by adding feedback. Repeated iterations of feedback can improve the reflector performance. To validate the method, a reflector is designed for an extended LED source. The results demonstrate that the illuminance uniformity can reach 88% and system efficiency can reach 97.4% simultaneously, after six feedback cycles. These are certainly able to meet the needs of general lighting applications.

Investigation on characteristic of LED power density via tapered fiber

“…In this paper, the simulation software based on the Monte Carlo Ray Tracing Method is used to simulate the LED white light [21][22][23][24][25], the transmission characteristics of light in tapered fiber are simulated and a method for improving LED light power density in the microenvironment is presented. It is worth mentioning that the tapered fiber can not increase the total luminous flux of LED, but it increases the luminous flux per unit area on the small face and improves the optical power density.…”

Designing double freeform surfaces for point source and extended sources using a construction, iteration and optimization process

“…The simultaneous multiple surface (SMS) method 20,21 allows the simultaneous construction of two FOSs that can exactly transform two input bundles of rays into two prescribed output bundles. The feedback compensation method, [22][23][24][25] which developed from point source algorithms, modifies the target distribution according to the simulated result iteratively till matching the desired distribution. But the design process is fussy and time-consuming, because in every loop of the process, Monte-Carlo simulations have to be completed and the simulation data need to be output for calculation in the next loop.…”