ANN and Multilayer-ELM based prediction of combustion, performance and emission characteristics of a diesel engine fuelled with Diesel-DTBP blends
Most of the earlier studies have used cetane improver (CI) with diesel engine fuels in order to compensate for the reduction in cetane number (CN) when biodiesel or alcohol fuels are added into neat diesel (ND) fuel. The studies which blended CI with ND are much limited in available literature. The...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-08-01
|
| Series: | Case Studies in Thermal Engineering |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2214157X25005830 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Most of the earlier studies have used cetane improver (CI) with diesel engine fuels in order to compensate for the reduction in cetane number (CN) when biodiesel or alcohol fuels are added into neat diesel (ND) fuel. The studies which blended CI with ND are much limited in available literature. The presented work investigates diesel engine combustion, performance and emission characteristics using ND blended with CI. Di-tert-butyl peroxide (DTBP) was used as CI and blended with ND in proportions of 0.7 % and 1.5 % by volume and blends were designated as A1 and A2 respectively. The experiments were carried out at a fixed engine load of 80 % and three different engine speeds of 1600 rpm, 2000 rpm and 2400 rpm respectively. Two machine learning (ML) models: Artificial Neural Network (ANN) and Multilayer Extreme Learning Machine (MELM) were developed to predict the key engine characteristics such as brake specific fuel consumption (BSFC), maximum cylinder pressure (Pmax), smoke emissions and nitrogen oxides (NOx) emissions. The experimental results showed that increasing engine speed deteriorates the overall combustion process and impacts fuel consumption as well as smoke emissions negatively. BSFC and smoke emissions were increased by 14.83 % and 28.95 % when speed increased from 1600 rpm to 2400 rpm. Moreover, ML models achieved least error metrics for all three key engine performance metrics. The prediction results validate the robustness of ANN and MELM in terms of capturing dynamics in predicting engine characteristics and emphasize its potential for estimation of engine performance and emissions. |
|---|---|
| ISSN: | 2214-157X |