Application of Artificial Neural Network for Internal Combustion Engines: A State of the Art Review

Area of Science:

• Automotive Engineering
• Computational Fluid Dynamics
• Artificial Intelligence

Background:

• The automotive industry is transitioning to electric vehicles, but internal combustion engines (IC engines) will remain crucial for decades.
• Optimizing IC engine performance with alternative fuels is vital for sustainability.
• Traditional engine experiments are costly, time-consuming, and challenging due to global economic and pandemic-related constraints.

Purpose of the Study:

• To summarize the application of artificial neural networks (ANNs) in predicting and optimizing IC engine characteristics.
• To investigate various ANN network topologies used for engine modeling.
• To statistically evaluate the performance of developed ANN models and compare them with other prediction methods.

Main Methods:

• Review of existing literature on ANN applications in IC engine research.
• Analysis of different ANN network architectures and their suitability for engine performance prediction.
• Statistical evaluation and comparison of ANN models against traditional prediction techniques.

Main Results:

• ANNs effectively predict and optimize complex IC engine characteristics like performance, combustion, and emissions.
• Different network topologies and deep learning approaches enhance ANN model performance.
• ANN models demonstrate competitive or superior accuracy compared to other prediction methods.

Conclusions:

• ANNs provide a powerful computational tool to reduce experimental costs and time in IC engine research.
• Further research into ANN design and topology can lead to improved engine efficiency and reduced emissions.
• ANNs are a viable solution for the automotive industry's challenges during the transition to new technologies.