Gas-liquid two-phase flow pattern recognition method based on acoustic emission-BP neural network

Abstract

A gas-liquid two-phase flow pattern recognition method based on acoustic emission-BP neural network measurement comprises the steps that an acoustic emission sensor is arranged on the outer wall of apipeline, the maximum value in blank pipe acoustic signals is set as threshold voltage, the two-phase flow acoustic signals are collected, and acoustic signals under all flow patterns are obtained; acoustic signal characteristic parameters under each flow pattern are calculated, wherein the acoustic signal characteristic parameters comprise an acoustic signal average voltage level, a root-mean-square value, an absolute energy value and ringing count; wavelet packet decomposition analysis of the acoustic signal under each flow pattern is calculated; four features of the normalized acoustic signal and norm squares of reconstructed waveforms at each frequency band are taken as neural network input quantities; optimization training is conducted on the neural network; and then flow pattern identification can be carried out. Real-time, rapid and online identification can be realized, typical flow patterns of inclined and vertical pipe gas-liquid two-phase flows, i.e., a bubble flow, a slug flow, a stirring flow and an annular flow, can be identified, the identification accuracy reaches up to 95% or above, and the method is of great significance to flow safety guarantee and monitoring ofhigh-pressure thick-wall oil and gas pipelines.

Description

technical field [0001] The invention relates to the field of flow pattern detection of gas-liquid two-phase flow, in particular to a method for real-time on-line recognition of the flow pattern of gas-liquid two-phase flow in an inclined or vertical (20-90°) standpipe of an offshore oil-gas mixed transportation system. Background technique [0002] The phenomenon of gas-liquid two-phase flow widely exists in long-distance onshore mixed transportation pipeline systems and offshore oilfield mixed transportation production systems. The distribution of two-phase flow medium flowing in the pipe, that is, the flow pattern, is the basic problem of gas-liquid flow. Accurate flow pattern identification has an important impact on the prediction and monitoring of the total pressure drop in the pipeline, the stagnant liquid in the pipeline, and the generation of sediment, which are very concerned about the operation of the mixed transportation system. It is the basic information for the...

AI Technical Summary

Problems solved by technology

US005353627A discloses a device and method for detecting multiphase flow pattern using non-invasive acoustic emission detection technology. The method adopts acoustic emission technology to collect acoustic signals, and finally judges flow pattern by manual experience according to the signal spectrogram. However, this method There is no detailed and feasible method for the recognition process, which is greatly affected by the subjective factors of the recognizer

Images