Abstract
By the invention of common rail system, thermal efficiency and environmental performance of a diesel engine have been improved. However, the present diesel engines are controlled by look-up-tables, called control maps. It requires a lot of experimental data, so it takes a lot of money and effort to make maps. Also, controller based on maps have weak robustness with respect to changes of environment. In this research, we propose a combustion controller by using feedback error learning. By using this control system, the two problems are solved. In addition, we use cerebellar model articulation controller (CMAC) as learning algorithm to reduce calculation load and learn quickly. Simulation and experimental results show the effectiveness of the proposed system.
