Usage of radial basis function neural network for dual-energy radiative detection system for measuring the oil pipelines scale layer

Abstract

Scaling oil pipelines over time leads to issues including diminished flow rates, wasted energy, and decreased efficiency. In order to take preventative measures in a timely manner and avoid the aforementioned issues, it is crucial to get a precise gauge of the scale within the pipe. Non-invasive gamma attenuation systems are one of the most accurate detection methods. In this investigation, he Monte Carlo N Particle (MCNP) technique was used to model a scale thickness measuring system including a test pipe, two sodium iodide detectors, and a dual-energy gamma source. The three-phase flow was created by simulating water, gas and oil in a homogenous flow regime in the test pipe. Additionally, a scale of varying thicknesses was installed inside the conduit. The photon intensity on the opposite side of the pipe was measured by detectors when gamma rays shone through it. Photopeaks of 241 Am and 133 Ba of the transmission detector and the total count of the scattering detector were derived from the signal received by the detectors. Since there are different variables such as scale layer and volume fractions in the present study, it is not possible to use regular known attenuation formulas for calculating the thickness. So, in this paper, RBF was implemented to solve this issue. By training and optimizing a RBF network, we were able to predict the thickness of a scale with an RMSE of 0.057. Having such a small margin for error confirms the validity of the suggested technology and its practical use in the petroleum and petrochemical industries.