Predicting of luminous flux for a LED array using artificial neural network

Because of its compact size, low power consumption, fast response, and high reliability, light-emitting diode (LED) becomes a new generation of light source to replace the traditional ones. To meet the needs of high illumination, the LED array is often used in light module. However, high temperature and high current will result in the luminous efficacy depreciation of the LED array module. This study proposes the artificial neural network (ANN) methods to predict the luminous flux of the LED array module and their accuracies are verified under different operation conditions. In detail, the Back Propagation (BP) and Genetic Algorithm-Back Propagation (GA-BP) NNs are firstly applied regarding the case temperature and current as inputs. Then, the feasibilities of the proposed methods are verified and compared. The results show that: 1) the luminous flux prediction accuracies of BP and GA-BP NNs are all higher than 98.5%; 2) compared to BP NN, GA-BP NN has higher prediction accuracy but costs more computation time.