PREDICTION OF PERFORMANCE AND EMISSION OF COMPRESSED NATURAL GAS (CNG) IN A SUPERCHARGED DIRECT INJECTION SPARK IGNITION ENGINE USING ARTIFICIAL NEURAL NETWORK

Abstract

Emissions of greenhouse gases such as carbon dioxide, nitrogen oxide and some hydrocarbons have been the main causes of global warming and have posed severe impacts on climate changes. Consequently, the focus on sustainable developments has swiftly grown in recent years. The utilisation of Compressed Natural Gas (CNG) in the spark ignition (SI) engine or using as dual fuel in compression ignition (CI) engine is getting remarkable attention nowadays. In this paper, performance and emission of CNG in supercharged direct injection spark ignition engine were predicted via Artificial Neural Network (ANN). The Levenberg Marquardt training algorithm was used due to its fast response, easy operation and high accuracy. The models’ results are compared with experimental results available from a previous study by the author. It was observed that R2 values for both performance and emission of CNG were higher than 98%, indicating good prediction. Following the training, high accuracy with value higher than 95% was observed for both analyses. Hence, it can be concluded that ANN is a promising technique for the prediction of performance and emission of CNG in direct injection spark ignition engine that consists of numerous inputs and output conditions.Keywords: Natural gas, CNG, supercharging, artificial neural network