Intelligent identification system of flow regime of oil–gas–water multiphase flow

Abstract

Instantaneous differential pressure signals of oil–gas–water multiphase flow in a horizontal pipe are measured with a piezo-resistance differential pressure transducer with fast response. The signals are denoised by using wavelet theory and then the characteristic vectors of various flow regimes are obtained from the denoised differential pressure signals with fractal theory. The characteristic vectors of known flow regimes are fed into a neural network for training and later on weight coefficients of neural network are obtained through training. Then, the characteristic vector of some kind of unknown flow regime of oil–gas–water multiphase flow is fed into the neural network and the neural network can automatically send out the information in respect to the classification of flow regime, thus the intelligent identification of flow regime of oil–gas–water multiphase flow is realized. Practice shows that this new method for identifying flow regimes of multiphase flow and the system constructed with the method has the merits of high accuracy, fast response and automatic identification without artificial intervention etc. It will have promising application prospect.