Support vector machine-based exergetic modelling of a DI diesel engine running on biodiesel–diesel blends containing expanded polystyrene. (5th February 2016)

• Record Type: Journal Article
• Authors: Shamshirband, Shahaboddin; Tabatabaei, Meisam; Aghbashlo, Mortaza; Yee, Por Lip; Petković, Dalibor
• Is Part Of: Applied thermal engineering. Volume 94(2016:Feb.)
• Page Start: 727
• Page End: 747
• Publication Date: 2016-02-05
• Subjects: Diesel/biodiesel blends -- Exergetic performance modelling -- Expanded polystyrene (EPS) -- Support vector machine -- Wavelet transform; Heat engineering -- Periodicals; Heating -- Equipment and supplies -- Periodicals; Periodicals; 621.40205
• DOI: 10.1016/j.applthermaleng.2015.10.140
• Languages: English
• ISSNs: 1359-4311

Abstract

Highlights: SVM-based thermodynamic modelling of a DI diesel engine working with diesel/biodiesel blends containing EPS. Comparison of SVM-WT, SVM-FFA, SVM-RBF, SVM-QPSO, and ANN approaches for exergetic modelling of the engine. Satisfactory performance of the SVM-WT for performance modelling of the engine over the other approaches. Abstract: In the present study, four Support Vector Machine-based (SVM-based) approaches and the standard artificial neural network (ANN) model were designed and compared in modelling the exergetic parameters of a DI diesel engine running on diesel/biodiesel blends containing expanded polystyrene (EPS) wastes. For this aim, the SVM was coupled with discrete wavelet transform (SVM-WT), firefly algorithm (SVM-FFA), radial basis function (SVM-RBF) and quantum particle swarm optimization (SVM-QPSO). The exergetic data were computed using mass, energy, and exergy balance equations for the engine at different speeds and loads as well as various biodiesel and EPS wastes quantities. Three statistical indicators namely root means square error, coefficient of determination and Pearson coefficient were used to access the capability of the developed approaches for exergetic performance modelling of the DI diesel engine. The modelling results indicated that the SVM-WT approach was more efficient in exergetic modelling of the engine than the other three approaches. Moreover, the results obtained confirmed the effectiveness of the SVM-WT model in identifying the most exergy-efficient combustion conditions and the best fuel composition for achieving the most cost-effective and eco-friendly combustion process.