Spark-assisted HCCI engine using hydrous methanol as a fuel: An ANN approach

Abstract

The main purpose of the experiment is to investigate the performance and emission characteristics of a hydrous methanol (85% methanol and 15% water) fueled homogeneous charge compression ignition (HCCI) engine through various spark timings (SPT) (32°, 34°, 36°, 38° and 40° before Top Dead Centre). In this study a spark plug is used for assisting auto-ignition. Experimental investigation reveals that the brake thermal efficiency (BTE) of an HCCI engine increases when the SPT is increased and a maximum BTE of 28.5% was obtained for 40°SPT. Emission analysis reveals a significant decrease in nitrogen oxides (NOx) as there are slightly higher emissions of hydrocarbon (HC) and carbon monoxide (CO). An artificial neural network (ANN) was developed with brake mean effective pressure (BMEP), SPT, and energy share as the input data and BTE, NOx, HC, CO, and rate of pressure rise as output data. In this prediction technique, about 80% of experimental data obtained were used in training and 20% of data were used to test the model developed. The performance in the developed ANN models was compared with experimental data, and statistically evaluated; they are seen to have good agreement.