Predictive modeling and optimization of SI engine performance and emissions with GEM blends using ANN and RSM

Area of Science:

• Internal Combustion Engines
• Alternative Fuels
• Computational Modeling

Background:

• Optimizing engine performance and emissions is crucial for fuel efficiency and environmental impact.
• Gasoline, Ethanol, and Methanol (GEM) blends offer potential as alternative fuels for internal combustion engines.
• Predictive modeling and optimization techniques can accelerate the development of efficient fuel blends.

Purpose of the Study:

• To predict the performance and emissions of a single-cylinder SI engine using GEM blends.
• To optimize engine operating conditions for maximum performance and minimal emissions.
• To evaluate the effectiveness of Artificial Neural Network (ANN) and Response Surface Methodology (RSM) in engine optimization.

Main Methods:

• An Artificial Neural Network (ANN) model was developed to predict engine performance and emissions.
• Response Surface Methodology (RSM) was employed to optimize engine parameters and fuel blends.
• Experimental validation was performed to assess the accuracy of the ANN predictions.

Main Results:

• The ANN model achieved less than 5% error in predicting engine performance and emissions.
• Optimal operating conditions were identified as 2992.9 rpm engine speed and an E20-equivalent GEM blend.
• The optimized conditions yielded a brake thermal efficiency (B_The) of 34.63%, BSFC of 243.7 g/kW-hr, and low emissions (1.5% CO, 108.13 ppm HC, 1211.8 ppm NOₓ).

Conclusions:

• ANN and RSM are effective tools for optimizing engine performance with alternative fuel blends.
• The E20-equivalent GEM blend offers a favorable balance between engine efficiency and emissions.
• The study demonstrates a highly desirable combination of performance and emissions characteristics for the optimized engine setup.