Flow Measurement Through Orifice‐Metre With Decision Tree Regression Method

The effective design of orifice is a major concern for the flow field measurement. The error of the flow occurs downstream of the orifice due to instability and back flow. Accurate orifice design boosts the precise flow field measurements. Predicting the pressure drop (Δp) across an orifice and estimating the flow rate (ṁ) in multiphase fluid flows is the propriety of the study. The key output parameters influencing Δp or ṁ are identified as Reynolds number (Re), area ratio (AR), and volume fraction ( R ) of the air‐water mixture flow. Numerical analysis is conducted by varying the parameters within specific ranges (Re < 100,000, 0.2 < AR < 0.6, and 0.2 < R < 0.8). Machine learning is a new regression method to process a large number of data to produce a correlation between input and output data. Decision tree regression (DTR) is employed to build a model that predicts Δp based on the input parameters (Re, AR, R). This model is trained using the numerical data generated from simulations based studies carried out with ANSYS software. The trained DTR model is evaluated for its accuracy and reliability in predicting Δp. The performance of the ML‐based predictions is compared against existing numerical data for concentric orifice flow to assess the effectiveness of the ML approach. The findings of the present study will be a foundation for the further advancement in CFD applications and integration with machine learning in fluid flow analysis.