Title: Predicting the different engine parameters of a rubber seed oil-ethanol dual fuel engine using artificial neural networks

Authors: J.S. Femilda Josephin; V. Edwin Geo; Ankit Sonthalia; C. Bharatiraja; Fethi Aloui

Addresses: Department of Software Engineering, SRM University, Kattankulathur, 603203, Tamil Nadu, India ' Department of Automobile Engineering, SRM University, Kattankulathur, 603203, Tamil Nadu, India ' Department of Mechanical and Automobile Engineering, SRM University, NCR Campus, Modi Nagar 201204, India ' Department of Electrical and Electronics Engineering, SRM University, Kattankulathur, 603203, Tamil Nadu, India ' Department of Mechanical Engineering, University of Valenciennes (UVHC), Campus Mont Houy, F-5931, LAMIH UMR CNRS 8201, Valenciennes Cedex 9, France

Abstract: The present study investigates the potential of artificial neural network for predicting the performance and emission characteristics of a compression ignition (CI) engine. A number of experiments are performed using diesel, rubber seed oil (RSO) and its methyl ester (RSOME) as the primary fuel, and ethanol as the secondary fuel in a dual fuel engine. The experimental data obtained is used for training and testing the neural network. The predictions are performed using feed forward-back propagation training algorithm. Engine load and ethanol energy share is used as network input parameters, and brake thermal efficiency (BTE), brake specific energy consumption (BSEC), NOx, HC, CO and smoke are the predicted parameters. The prediction performance of the network is measured by comparing it with experimental data. The measurement of statistical error shows that ANN can predict BTE, BSEC, NOx, HC, CO and smoke for a dual fuel engine with high accuracy.

Keywords: artificial neural network; ANN; ethanol; dual fuel diesel engine; rubber seed oil; RSO; rubber seed oil methyl ester; RSOME.

DOI: 10.1504/IJGW.2018.095995

International Journal of Global Warming, 2018 Vol.16 No.4, pp.485 - 506

Received: 27 Nov 2017
Accepted: 20 Apr 2018

Published online: 06 Nov 2018 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article