Investigation into Driving Techniques and Color Control of LED Systems with Multiple Color Channels

Abstract

Artificial lighting consumes a large portion of the worldwide electricity. High power LED lamps are replacing traditional incandescent and fluorescent lamps due to higher energy efficiency and longer lifespan. Although LED systems with multiple color channels may provide a large gamut of colors and quality lightings with high color rendering indexes, the driving techniques and color control of multi-color LED systems are critical concerns for their wide applications. This thesis presents the findings of research on driving techniques and color control of LED systems with multiple color channels, to provide researchers and engineers better understandings and control methods of multi-color LED systems.

To control the light of multi-color LED lamps, the LED drivers are required to provide controlled power to each LED channel. Existing driving techniques with parallel-connected channel structure are challenged by current spikes, limited utilization of LED channels and current sensing problems. To solve such problems and reduce the form factor of LED driver, using a single inductor to drive multiple series-connected LED channels is investigated. Besides driving techniques, color consistency is another critical issue for LED systems with multiple color channels. The color control methods are required to handle the change of luminous flux and color with temperature and aging so as to maintain the synthesized light at the desired color point. Novel color control methods for LED systems with multiple color channels are proposed, including spectral peak control and neural-network-based color control. Spectral peak control achieves color control of multi-color LED systems with high color rendering indexes by controlling the peaks in the spectral power distributions of LED channels to approximate the spectral power distribution of black bodies. Experimental results have verified the effectiveness and robustness with ambient temperature of spectral peak control. Artificial neural network is applied to assist the color control of multi-color LED systems in order to deal with ambient light. Compared with conventional control methods, the proposed methods have the merits of 1) not requiring an accurate electrical-optical system model, 2) requiring only RGB color sensors for online operations of tetra-color and penta-color LED systems, 3) providing consistent color with high color rendering index with or without ambient light, 4) being applicable to all multi-color LED systems.

Date of Award	12 Mar 2019
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Shu Hung Henry CHUNG (Supervisor)