A Mechanistic Approach to Understanding Wellbore Phase Redistribution

Abstract

ABSTRACT Wellbore phase redistribution (WPR) frequently occurs during a shut-in test in wellbores having very compressible fluids, such as low-pressure, single-phase gas and high gas/liquid ratio multiphase fluid mixture, including steam. The consequence of WPR is a pressure buildup signature that deviates greatly from the commonly used constant-storage model, thereby precipitating test interpretation problems. A lack of physical understanding of the underlying process has led us to rely solely on empirical approaches — exponential and error function models — for analyzing field data. The existing methods have some drawbacks. For example, one may obtain nonunique results. Further, we cannot predict whether a particular well will exhibit WPR a priori. Thus, the available methods are restricted to the test analysis mode only. In this work, we present a mechanistic approach to understanding the principal causes for WPR. A simple physical model, consisting of a liquid column and a small pocket of segregated gas at the top, is assumed to mimic a wellbore. The model well cannot receive any fluid from the reservoir upon shut-in; however, backflow from the wellbore into the reservoir is permitted to relieve excess pressure, generated by a rising bubble. A mathematical model is developed for the idealized well by studying the rise velocity of a single bubble in the liquid column. Interestingly enough, a simplified analytical solution leads to an exponential form to describe the excess pressure behavior, caused by a single-bubble rise, as hypothesized by Fair. Thus, this work provides a basis for Fair’s intuitive exponential model. The results show that both the wellhead and bottomhole pressures increase with elapsed time as a single gas bubble ascends up the liquid column. The magnitude of pressure rise is a strong function of flow impediment or skin in the wellbore vicinity, well diameter and its orientation, and wellhead pressure. A comparison of the simplified (exponential) and the rigorous solutions shows good agreement between the two at early times. The proposed model provides a framework for understanding the physical mechanisms that underlie an anomalous pressure rise during a shut-in test. The model, although simplistic in nature, provides some justification for the use of existing empirical methods for interpreting field data.