Wet Gas Metering Using a Venturi-meter and Neural Networks

Lijun Xu,Hui Li,Shaliang Tang,Cheng Tan,Bo Hu
DOI: https://doi.org/10.1109/imtc.2008.4547139
2008-01-01
Abstract:In this paper, a novel approach is presented to the measurement of wet gas flows using a Venturi meter and neural network technique. Results obtained on a laboratory test rig suggest that the flowrate of wet gas flowing in a throat-extended Venturi meter is related to the characteristic features of the differential pressures across the converging section and the extended throat section, the static pressure and temperature signals within the Venturi-meter. The relation between the signal features and gas/liquid flowrates of wet gas is established through the use of back-propagation (BP) neural networks. The experimental test carried out within static pressure range of 0.1-0.8 MPa, gas flowrate range of 50~160 m3/d and oil flowrate range of 1.1-5.3 m3/h suggested that it is a simple and viable method to solve the problem of wet gas metering by combining a revised Venturi meter and neural networks techniques.
What problem does this paper attempt to address?