Forecasting electrical parameters of mass-produced LED luminaire using the Monte Carlo method

Abstract

This paper describes the various input variables of an LED system (electronic driver and LED), such as component behavior, tolerance, thermal effects, and manufacturing variability - which all contribute to important end-use system parameters such as watts and total lumen output. A mathematical model with probability functions is constructed in Microsoft Excel to compute electrical parameters. Sensitivity analysis and parameter forecasting is then performed via the Monte Carlo method, using the Crystal Ball add-on software package by Oracle. Forecasted results for power are then compared and verified against test results of mass production parts.