Frequency Domain Feature Extraction Investigation to Increase the Accuracy of an Intelligent Nondestructive System for Volume Fraction and Regime Determination of Gas-Water-Oil Three-Phase Flows

In this research, a methodology consisting of an X-ray tube, one Pyrex-glass pipe, and two NaI detectors was investigated to determine the type flow regimes volume fractions gas-oil-water three-phase flows. Three prevalent patterns—namely annular, stratified, homogenous—in various percentages—10% 80% with step 10%—were simulated by MCNP-X code. After simulating all states collecting signals, Fast Fourier Transform (FFT) used convert data frequency domain. The first second dominant amplitudes were extracted be as inputs neural networks. Radial Basis Function Neural Networks (RBFNN) trained for determining predicting gas water fractions. correct detection determination percentages Mean Relative Error (MRE) less than 2.02% shows that use characteristics in these important parameters can very effective. Although radiation-based two-phase flowmeters have lot advantages over radioisotope-based ones, they suffer from lower measurement accuracy. One reason might multi-energy spectrum recorded detector has been analyzed simple way. It is worth mentioning sources generate photons despite radioisotopes single energy photons, therefore analyzing radioisotope would easier ones. As mentioned, problems researchers encountered accuracy X-ray, flowmeters. aim present work resolve problem improving precision flowmeter using artificial network (ANN) feature extraction techniques.