Experimental and modeling analysis of the transient spray characteristics of cyclopentane at sub- and transcritical conditions using a machine learning approach

Although fuel spray parameters, such as cone angle and penetration length, are crucial for developing high-efficiency high-performance combustion engines, general models predicting transient characteristics of these parameters have not been suggested. In this study, the cyclopentane at sub- transcritical conditions relevant IC engine rocket injections were experimentally evaluated. A single simplified model angles lengths over time was developed by adopting artificial neural networks (ANN). Spray measurements conducted shadowgraphy Mie scattering techniques to recognize phase change behavior spray, changing injection chamber conditions. The ANN using a multi-layer network with six normalized inputs four outputs. It trained five datasets two subcritical three validated one dataset. prediction assessed, proposed approach represents According results, predicted in good agreement experimental results useful pressure temperature range 40–55 bar 465–564 K, mean absolute percentage errors 2.25% (shadowgraphy) 4.92% (Mie) angles, 1.11% 3.44% lengths. Moreover, can predict ratio, providing information on changes sprays. study is expected become universal near critical point. By adding database various types conditions, be used develop engines or other combustors.