Approach for Wellbore Production Monitoring Using Distributed Acoustic Noise Measurements

The objective of this study was to develop a data processing flow for the oil and gas industry enabling the determination of inflow profile and fluid type identification for in well oil-water flows, and to quantify the phase and flow rates using distributed acoustic vibration data, in a near real-time wellbore monitoring scenario. The implementation of the approach will enable alarms to be raised in real-time in zones when changes in the flow rates and phase changes exceed predetermined levels, allowing quantifiable operational decisions. The acoustic data was acquired using a distributed fiber optical (FO) sensing technique and reference hydrophones in a laboratory flow loop equipped with accurate reference flowmeters. The acoustic noise created by the main pipe flow and inflow was studied. Total flow rate varied within the range 0 to 200 m3/d for water and a model of light oil. A set of numerical models was used to support development of the interpretation approaches through an enhanced understanding of the acoustic field in various real wellbore geometries. Two interpretation approaches based on laboratory correlations of acoustic noise energy in a selected frequency range and machine learning algorithms were developed to quantify phase rates from distributed acoustic vibration data induced by turbulent fluid flow in laboratory conditions. It is shown that correlation-based interpretation enables flow quantification and profiling within acceptable uncertainty levels, for a field dataset of distributed vibration measurements and a reference production logging tool (PLT) log in a producing wellbore. The goal of the software development is to provide a quantitative flow characterization from the interpretation of distributed acoustic vibration measurements. This method was tested using field fiber optic datasets combined with reference PLT log data.