project 4

Pioneer research involving generation of ML based Chemical Kinetic Model for Computational Fluid Dynamics (CFD)

• Created an efficient pre-processing and sampling pipeline to obtain data from 0-D statistical simulation using PaSR Simulator
• Successfully tuned an ANN based model to obtain accurate predictions for 34/39 chemical species with an R2 score of >~ 0.996
• Constructed and fine-tuned a bottleneck network designed to reduce feature space of Detailed Chemical Kinetic Data