Prediction of Heat Energy Release Rate for Ammonia Combustion in a Constant Volume Combustion Chamber: A Machine Learning Approach.

Addressing global warming requires immediate and concerted action. Transitioning to non-conventional energy sources, such as solar and Aeolian power, reducing emissions from transportation and industry, and implementing sustainable land-use practices are crucial steps. Therefore, Ammonia attracts researchers when used as a fuel, and offers several potential benefits such as higher energy density. However, due to poor combustion speed of Ammonia offers slower energy release. To enhance the average flame velocity of Ammonia, the current investigation employed the exhaust gas expelled from the secondary chamber to compress the mixture within the primary chamber and facilitate its HCCI combustion. The main goal of research is the optimization on air-fuel ER to contribute to the development of more efficient and sustainable constant volume combustion. However, the time required for constructing physical prototypes and conducting experiments can be considerable, thus, to mitigate these challenges, the authors turned to machine learning tools as an alternative approach. In the proposed research, results were evaluated through Support Vector Regression. However, due to poor prediction accuracy, authors tried KNeighborsRegressor model for predicting and validating the results. Finally, it is revealed from the KNN model that data sets are highly accurate.