Application of feature extraction and artificial intelligence techniques for increasing the accuracy of x-ray radiation based two phase flow meter

Abstract

The increasing consumption of fossil fuel resources in the world has placed emphasis on flow measurements in the oil industry. This has generated a growing niche in the flowmeter indus-try. In this regard, in this study, an artificial neural network (ANN) and various feature extractions have been utilized to enhance the precision of X-ray radiation-based two-phase flowmeters. The detection system proposed in this article comprises an X-ray tube, a NaI detector to record the pho-tons, and a Pyrex-glass pipe, which is placed between detector and source. To model the mentioned geometry, the Monte Carlo MCNP-X code was utilized. Five features in the time domain were de-rived from the collected data to be used as the neural network input. Multi-Layer Perceptron (MLP) was applied to approximate the function related to the input-output relationship. Finally, the intro-duced approach was able to correctly recognize the flow pattern and predict the volume fraction of two-phase flow’s components with root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE) of less than 0.51, 0.4 and 1.16%, respectively. The ob-tained precision of the proposed system in this study is better than those reported in previous works.