Machine_Learning_Technique_For_Flow_Regime_Classification

Two Phase Flow Regime Classification in a Straight Horizontal Pipe

About the project

• In the present study the 7 different two-phase flow regimes were classified for refrigerant flow occurring in a horizontal straight pipe are made with the application of Random Forest Classifier (RFC), and Xtreme Gradient Boosting Classifier (XGBC)

Experimental Flow Visualization (Padilla 2011 et al )

Different Two Phase Flow Regime Occurrence in a Straight Pipe

Image Credit: Mudawar et. al.

Methods

• More than 390+ experimental two-phase flow regimes data is taken from the literature with features- mass flux, vapor quality, pipe diameter, saturation pressure and saturation temperature

• RFC and XGBC models trained with 22 input features to classify 7 different flow regimes

Conclusion

• It is concluded that the application of RFC is shown best results in classifying flow regimes