Prediction of Optimum Ignition Timing in a Natural Gas-Fueled Spark Ignition Engine Using Neural Network 2006-01-1347
Ignition timing control is a key problem in the conventional spark ignition engine due to the nonlinear nature of its variation with the engine conditions. The problem is much more pronounced in natural gas fueled engines due to the frequent variation of the operating parameters. The variation in the air-to-fuel ratio, gas composition, engine load and speed during engine operation require different spark timings for better engine performance. The ability of neural networks to reasonably handle such a complicated control problem is introduced. The training and testing of the predictor of the optimum spark timing for maximum brake torque (MBT) is conducted. The training was carried out using a large set of experimental data collected from natural gas fueled engine. The obtained predictor is tested and possible causes of robustness problem are discussed. The predicted MBT spark timing is compared with a different set of experimental data for the same gas fueled engine. Reasonable agreement between the predicted spark timing and the experimental data has been observed with correlation coefficient of 0.947. The comparison showed that the predictor is reasonably robust to operating parameter variation.