Investigation of gas-liquid flow using electrical resistance tomography and wavelet analysis techniques for early kick detection

Gas kick is a well control problem and is defined as the sudden influx of formation gas into the wellbore. This sudden influx, if not controlled, may lead to a blowout problem. An accidental spark during a blowout can lead to a catastrophic oil or gas fire. This makes early gas kick detection crucial to minimize the possibility of a blowout. The conventional kick detection methods such as the pit gain and flow rate method have very low sensitivity and are time-consuming. Therefore, it is required to identify an alternative kick detection method that could provide real-time readings with higher sensitivity. In this study, Electrical Resistance Tomography (ERT) and dynamic pressure techniques have been used to investigate the impact of various operating parameters on gas volume fraction and pressure fluctuation for early kick detection. The experiments were conducted on a horizontal flow loop of 6.16 mERT with an annular diameter ratio of 1.8 for Newtonian fluid (Water) with varying pipe inclination angle (0 - 10^o) and annulus eccentricity (0 – 30%), liquid flow rate (165 – 265 kg/min), and air input pressure (1 – 2 bar). The results showed that ERT is a promising tool for the measurement of in-situ gas volume fraction. It was observed that the liquid flow rate, air input pressure and inclination has a much bigger impact on gas volume fraction whereas eccentricity does not have a significant influence. An increase in the liquid flow rate and eccentricity by 60% and 30% decreased the gas volume fraction by an average of 32.8% and 5.9% respectively, whereas an increase in the inclination by 8⁰ increased the gas volume fraction by an average 42%. Moreover, it was observed that the wavelet analysis of the pressure fluctuations has good efficacy for real-time kick detection. Therefore, this study will help provide a better understanding of the gas-liquid flow and potentially provide an alternative method for early kick detection.