Injection Profiling for Intelligently Completed Wells Equipped with Fiber Optics

Abstract

Abstract Injectors with intelligent well completions are installed in a geologically complex reservoir setting to inject fluid into multiple zones. For effective reservoir management, it is imperative to know the injection profile along the wellbore and to detect injection defects. Distributed Temperature Sensing (DTS) fiber installed along the length of the well provides high spatial resolution and high frequency temperature data. This paper presents a methodology to estimate zonal-level fluid injection volumes using these DTS measurements. When cold fluid is injected into a hotter reservoir, the region near the wellbore cools down. When injection ceases, the wellbore region slowly warms back to the far field geothermal temperature. The proposed methodology models heat transfer in the radial direction during this warmback period using 1-D energy equations. The model is tuned to DTS data from the warm-back using regression techniques applied independently at each depth. Only the middle time region (MTR) of the warmback is considered for the model since, at early times, we have thermal effects from adiabatic expansion and, at late times, the 1-D modeling assumption is no longer valid. This approach is analogous to the MTR concept often used in pressure transient analysis. Model parameters at each depth are related directly to the volume of fluid injected in the corresponding zone and their relative values can be used to allocate injected fluid volumes. The conventional use of the warm-back model looks at the relative rate that individual zones warm up during a shut-in following cooler water injection to estimate the relative volume of fluid injected into that particular zone. This conventional methodology considers all the data points from warm-back to fit the model. The improved method presented in this paper acknowledges that the early time region (ETR) and late time region (LTR) are not really addressed in the model. Rather it carves out the information by modeling the middle time region (MTR). An alternate approach presented looks at the asymptotic behavior of the middle-time warm-back to infer the cumulative injection volume in each zone. A few test cases are presented comparing the results of the various methodologies and the insights we gain from this analysis. Injection logging is an expensive, conventional method for injection profiling. With the help of fiber technologies and methodologies presented in the paper, we can achieve an accurate understanding of injection profiling without the need for logging. Additionally, we show a methodology where subsequent analysis of warmback at two or more shut-in times gives insight into the injection profile when the original geotherm is disturbed by prior injection or not well known.