Volume fraction calculation in multiphase system such as oil-water-gas using neutron

Multi-phase flows are common in diverse industrial sectors and the attainment of the volume fraction of each element that composes the flow system presents difficulties for the engineering process, therefore, to determine them is very important. In this work is presented methodology for determination of volume fractions in annular three-phase flow systems, such as oil-water-gas, based on the use of nuclear techniques and artificial intelligence. Using the principle of the fast-neutron transmission/scattering, come from an isotopic 241Am-Be source, and two point detectors, is gotten measured that they are influenced by the variations of the volumec fractions of each phase present in the flow. An artificial neural network is trained to correlate such measures with the respective volume fractions. In order to get the data for training of the artificial neural network without necessity to carry through experiments, MCNP-X code is used, that simulates computational of the neutrons transport. The methodology is sufficiently advantageous, therefore, allows to develop a measurement system capable to determine the fractions of the phases (oil-water-gas), with proper requirements of each petroliferous installation and with national technology contributing, possibly, with reduction of costs and increase of productivity