Optimization of Diesel Engine Parameters using Genetic Algorithm
This work involves in selection of parameters for a compression ignition engine which influence fuel economy and harmful emissions such as carbon monoxide and oxides of nitrogen. Experiments are conducted for performance and emission by varying the operating parameters within the operating range. Full factorial experimentation is done and this large data is analyzed by non-traditional soft computing techniques namely Genetic Algorithm (GA). Mathematical models are formed using MINITAB software and the same are used for optimization of settings using GA. A single layer Levenberg-Marquardt back propagation network has been trained using the experimental data. By using the trained network the output for the optimum set of parameters obtained from GA is predicted. The outputs from experiments, GA are compared and the outcome is discussed. This optimized set of parameters, when applied in the engine, reduces the harmful emissions of the engine and increases its performance, thus conserving fuel and promoting a cleaner environment.