Almost all the multi-phase flow meters (MPFMs) using gamma-ray attenuation, are calibrated for liquid
and gas phases with constant density and pressure. When operational conditions such as temperature
and pressure change in pipelines, the radiation-based multi-phase flow meters would measure the flow
rate with error. Therefore, performance of MPFMs would be improved by eliminating any dependency on
the fluid properties such as density. In this work, a method based on dual modality densitometry
combined with Artificial Neural Network (ANN) is proposed in order to estimate the void fraction in
homogenous regime of gas-liquid two-phase flows in unstable operational conditions (changeable
temperature and pressure) in oil industry. An experimental setup was implemented to generate the
optimum required input data for training the network. ANNs were trained on the registered counts of the
transmission and scattering detectors in various liquid phase densities and void fractions. Void fractions
were predicted by ANNs with mean relative error of less than 0.78% in density variations range of 0.735
up to 0.98 g/cm3.
کلید واژگان :
Gamma ray; Artificial neural network; Multi layer perceptron; Void fraction.
