Application of Artificial Intelligence for Determining the Volume Percentages of a Stratified Regime’s Three-Phase Flow, Independent of the Oil Pipeline’s Scale Thickness

As time passes, scale builds up inside the pipelines that deliver oil or gas product from source to processing plants storage tanks, reducing diameter and ultimately wasting energy efficiency. A non-invasive system based on gamma-ray attenuation is one of most accurate diagnostic methods detect volumetric percentages in different conditions. including two NaI detectors dual-energy gamma sources (241Am 133Ba radioisotopes) recommended requirement for modeling a volume-percentage detection using Monte Carlo N particle (MCNP) simulations. Oil, water, form three-phase flow stratified-flow regime volume percentages, which flows scaled pipe with thicknesses. Gamma rays are emitted side, photons absorbed other side by scintillator detectors, finally, three features names count under Photopeaks 241Am first detector total second were obtained. By designing MLP neural networks said inputs, can be predicted an RMSE less than 1.48 independent thickness. This low error value guarantees effectiveness intended method usefulness this approach petroleum petrochemical industries.