Radial Basis Function Neural Network (RBFNN)Based Modeling in Liquified Petroleum Gas (LPG)-Diesel Dual Fuel Engine with Exhaust Gas Recirculation (EGR)

Abstract

In this paper the application of Radial Basis Function Neural Network (RBFNN) for modeling the performance and emission parameters of a LPG-Diesel dual fuel engine with EGR is presented. Three different centre selection strategies namely fixed centre selected at random (FCSR), Fuzzy c-means Algorithm (FCM) and Conditional Fuzzy c-means Algorithm (CFCM) are used in the design of Radial Basis Function (RBF) network. The performance parameters included are Brake Power (BP), Brake Specific Energy Consumption (BSEC) and Brake Thermal Efficiency (BTE) and emission parameters included are Exhaust Gas Temperature (EGT), smoke, Hydro-Carbons (HC) and Nitrogen Oxides (NOx).The results showed that there is a good correlation between the RBFNN predicted values and the experimental values for various engine performance and emission parameters with R2 value ranging from 0.90 to 0.99 and the Mean Relative Error (MRE) values are within 5% which is acceptable. A comparison with MLP modeling demonstrated that RBFNN is more efficient than Multi-Layer Perceptron Neural Network (MLPNN) for modeling in this application.