Advanced interpretation of Venturi meter measurements in multiphase flow

The Venturi meter is commonly used in combination with other sensors for the measurement of gas fraction and water cut as part of a multiphase flow meter system for the measurement of oil, water and gas flowrates. The Venturi differential pressure signals are generally interpreted using empirical corrections to a momentum equation to give a solution for volumetric flowrate.
In this project the use of advanced processing techniques and neural networks have been used to examine the potential improvement in measurement performance which may be obtained from a system incorporating a standard Venturi tube in combination with a gamma densitometer. The data is pre-processed before being applied to a neural network using simple statistical methods combined with some signal processing techniques.
It has been found that this neural network solution performs extremely well when only predicting the gas and liquid phases. Poorer performance was registered when trying to predict all three phases, indicating the need for a water cut sensor.