Harmonic Characteristics Data-Driven THD Prediction Method for LEDs Using MEA-GRNN and Improved-AdaBoost Algorithm

Light-emitting Diode (LED) lamps have been widely used due to versatility and energy efficiency. However, LEDs are nonlinear loads, the massive usage will inject harmonics into lighting system, which has influenced power quality. Total Harmonic Distortion (THD) is an important parameter evaluate quality, but prediction of THD for a challenging task. This paper addresses this issue by designing harmonic characteristics detection experiment using artificial intelligence algorithm. Firstly, LED with different driving circuits were tested, relevant data each sampled analyzed. Then, method based on improved AdaBoost algorithm proposed. In method, Generalized Regression Neural Network (GRNN) model established, its parameters optimized Mind Evolution Algorithm (MEA) improve search ability GRNN. On basis, utilized integrate multiple MEA-GRNN individuals form strong predictor, improves generalization model. To avoid integration failure caused improper selection threshold value, sigmoid adaptive factor added accuracy Finally, Ada-MEA-GRNN trained simulated collected experiment. The simulation results show that proposed better than BP GRNN, can reach 95.48%. Meanwhile, even if input dimension reduced, error still small.