Mirroring algorithm for gas–water two-phase gas holdup based on an array of optical fiber probes

Abstract Dynamic monitoring of gas–water two-phase production profiles of vertical wells is critical for effective oilfield development. Conventional methods, which rely on production logging tools sampling at a single position along the center of the wellbore, fail to capture the complete fluid distribution across the wellbore cross-section. To address this limitation, this paper presents a new method for gas holdup imaging based on data from an array of optical fiber probes that leverages the inherent symmetry of vertical well gas–water flows. In a simulated wellbore environment, a flow imager tool was used to measure vertical well gas–water two-phase flows under various conditions, including gas–water flow rates (100–600 m 3 d −1 ) and water cuts (10%–90%). Results revealed four primary flow patterns—bubble flow, plug flow, slug flow, and froth flow, with the gas and water phases distributed symmetrically along the well’s central axis. Measurement data from six optical fiber probes were then used in an interpolation algorithm to predict fluid information in unmeasured regions, thereby reconstructing a cross-sectional image of the gas–water distribution. Based on the symmetrical attributes of the flow patterns, a mirroring technique was then proposed whereby the six optical fiber probes were rotated counterclockwise along the radial direction by specific angles (e.g. 180°, 60° and 6°) to generate mirrored probe coordinates, thus forming a gas holdup cross-sectional image derived from both the original and mirrored probe data. Comparative analysis between the reconstructed cross-sectional gas-holdup images and experimental flow-pattern images across the entire experimental range demonstrated a flow pattern identification consistency rate of 83.3%. The method was further applied to an actual vertical gas–water well (Well A) in Northeast China, which produces 36 523 m 3 d −1 of gas and 75.04 m 3 d −1 of water. The reconstructed images accurately captured the slug-flow structure within the full-flow layer (3785–3798 m). These results verify the quantitative accuracy and field applicability of the method, demonstrating its potential for real-time gas–water distribution imaging and production profile interpretation in vertical wells.