ANN Application for Prediction of Diesel Engine Heat with Nano-Additives on Diesel-Biodiesel Blends

Authors

  • Seyyed Hassan Hosseini Department of Mechanical Biosystems, Gorgan University of Agricultural Science and Natural Resources , P. O. Box 386, Phone: +98 173 4426942, Fax: +98 173 4426942, Gorgan, Iran. Email: ho3eini1991@gmail.com
  • Ahmad Taghizadeh Alisaraei Department of Mechanical Biosystems, Gorgan University of Agricultural Science and Natural Resources , P. O. Box 386, Phone: +98 173 4426942, Fax: +98 173 4426942, Gorgan, Iran. Email: Ahmadtza@yahoo.com
  • Barat Ghobadian Department of Mechanical Biosystems, Faculty of Agriculture, Tarbiat Modares University, P. O. Box 14115-336, Phone: +98 21 44196522, Fax: +98 21 44196524, Tehran, Iran. Email: Ghobadib@modares.ac.ir
  • Ahmad Abbaszadeh-Mayvan Department of Mechanical Biosystems, Gorgan University of Agricultural Science and Natural Resources , P. O. Box 386, Phone: +98 173 4426942, Fax: +98 173 4426942, Gorgan, Iran. Email: abbaszadeh62@gmail.com

Abstract

     Biodiesel is renewable clean bioenergy as it can be produced from vegetable oils, animal fats and micro-algal oil and also it can be applied instead of diesel fuel without any special modifications to the engines. In recent years, Nano-catalysts or Nano-additives in fuels improve the thermo-physical properties of fuels. In this study, the Carbon Nanotubes (CNT) as additive were mixed with the B5 and B10 fuel blends to evaluate the cylinder head and cylinder block temperature of a CI single-cylinder engine with an artificial neural network. carbon Nanotubes with concentrations of 30, 60, and 90 ppm were used for each fuel blends. Assessed characteristics were cylinder head and cylinder block temperature for full load engine at three speeds of 1800, 2300, and 2800 rpm. The results for optimum ANN model showed that the training algorithm of Back-Propagation with 24 neurons in a hidden layer was sufficient enough in predicting engine cylinder head and cylinder block temperature for different engine speeds and different fuel blends ratios. The MSE error and R-value for training, validation and testing of optimum ANN model were 0.00095, 10.40, 9.71 and 0.9999, 0.9487 and 0.9726 respectively. It can be concluded neural network is a powerful tool to predict diesel engine cylinder head and cylinder block temperature parameter with reasonable accuracy.

Downloads

Published

2017-08-18

Issue

Vol. 19 No. 2 (2017): CIGR Journal

Section

IV-Energy in Agriculture