PFI SI engine fueled with ethanol with exhaust gas rebreathing: Thermal efficiency optimization with thermodynamic modeling and ANN-PSO application

An ethanol-fueled Atkinson cycle PFI SI engine using exhaust rebreathing (ER) was evaluated under 110 different operating conditions. While ethanol is an important biofuel to reduce greenhouse gas emissions, Atkinson cycle engines have been studied as low-consumption alternatives for hybrid architec...

Full description

Saved in:
Bibliographic Details
Published in:International journal of engine research 2024-01, Vol.25 (1), p.214-230
Main Authors: Fagundez, Jean Lucca Souza, Lanzanova, Thompson Diórdinis Metzka, Martins, Mario Eduardo Santos, Salau, Nina Paula Gonçalves
Format: Article
Language:English
Subjects:
Air flow
Artificial neural networks
Biofuels
Combustion
Emissions
Ethanol
Exhaust gases
Greenhouse gases
Nodes
Optimization
Otto cycle
Rebreathing
Spark ignition
Thermodynamic efficiency
Thermodynamic models
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:An ethanol-fueled Atkinson cycle PFI SI engine using exhaust rebreathing (ER) was evaluated under 110 different operating conditions. While ethanol is an important biofuel to reduce greenhouse gas emissions, Atkinson cycle engines have been studied as low-consumption alternatives for hybrid architecture vehicles. The operating conditions were validated with GT-Power® and the resulting combustion and performance parameters were trained by a series of artificial neural network (ANN) structures in order to learn and reproduce the non-linear correlations. The ANN structures were analyzed by number of nodes, hidden layers, and optimization methods. TPA results showed that ER tests had higher indicated thermal efficiency and lower combustion temperatures than both the standard Atkinson cycle and Otto cycle conventional throttled valve strategy with a similar engine setup. The best ANN to accurately predict all six outputs with easy-to-measure operating conditions was a combined ANN-PSO model, with two hidden layers with [30 10] nodes to predict net IMEP, airflow, total mass trapped and burned mass percentage, and a complementary 10-nodes PSO hidden layer, to predict PMEP, and the average exhaust runner temperature. The final model can predict and calibrate an Atkinson cycle PFI SI engine operating with ER with significant accuracy and no need for further TPA simulations.
ISSN:1468-0874
2041-3149
DOI:10.1177/14680874231194383