Prediction in a biodiesel Engine using MLP and RBFFNN Networks-A Comparision

Shivakumar, . and Pai, Srinivasa P and Rao, Sriniviasa B R (2013) Prediction in a biodiesel Engine using MLP and RBFFNN Networks-A Comparision. International Journal of Research in Engineering and Technology (IJRET), 2 (5). pp. 287-291. ISSN 2277 – 4378

IJRET026063(PSRC Journal Paper).pdf - Published Version

Download (763kB) | Preview


In the present study engine performance and emission characteristics from biodiesel engine fuelled with WCO (Waste Cooking Oil) was modelled by using ANN (Artificial neural networks). MLP (Multilayer Perceptron) and RBFNN (Radial Basis Function neural networks) were used for modeling. MLP modeling was done by training the network using different algorithms. RBF center initialization was done by using random initialization and by Cluster Dependent Weighted Fuzzy C Means (CDWFCM) method. In MLP, the use of trainlm algorithm gave minimum MRE (Mean relative error) and an average prediction accuracy of 98% for performance parameters and 82% for emission parameters. A comparison of two ANN models showed that RBFNN based model performed better than MLP. In the RBFNN, CDWFCM algorithm gave the best results with slightly lower MRE and higher prediction accuracy compared with MLP, thus establishing its usefulness for modeling the biodiesel engine performance and emission parameters

Item Type: Article
Uncontrolled Keywords: Artificial neural networks, Multilayer perceptron, Emissions, Radial basis function, Mean relative error
Subjects: Engineering > MIT Manipal > Mechanical and Manufacturing
Depositing User: MIT Library
Date Deposited: 30 May 2015 11:29
Last Modified: 30 May 2015 11:29

Actions (login required)

View Item View Item