The prospect of dynamic reservoir characterization using flow and pressure data gathered during underbalanced drilling (UBD) is a powerful driver for implementation of UBD. The mathematical aspects of this complex, ill-posed, inverse problem have been the subject of research in the past decade. This paper focuses on practical, field implementation of UBD reservoir characterization, and the problems that consequently arise. Interpretation of data from UBD is made difficult by transducer errors, operational transients, and noise in data. It is therefore often very difficult to interpret the reservoir characteristics from the instantaneous productivity index (PI). In this paper, we introduce a parameter known as the Rate Integral Productivity Index (RIPI), which borrows from the theory of rate-transient analyses. The mathematical and physical basis of RIPI and its relationship to the instantaneous PI are presented. The behavior of RIPI and its implications for reservoir characterization are discussed. RIPI de-noises the data, and scales the problem such that the trends in data are more obvious, enabling robust interpretation of UBD data, and increasing the confidence in calls made regarding reservoir characteristics. Application of RIPI to field data is illustrated through several examples. Data acquisition, processing, and preparation for UBD reservoir characterization are discussed. In particular, the importance of filtering, de-noising, and identifying and excluding operationally induced transients is described. Limitations imposed by the data gathering methods are highlighted. It is shown that the ability of RIPI to reduce noise in raw PI data allows trends to be read more easily. The use of RIPI for static and dynamic characterization of super-matrix features (such as fractures, thief zones, etc.) is illustrated. The limitations of the approach and future trends are discussed.
Introduction
When drilling underbalanced in permeable reservoir rock, the return fluid carries reservoir fluids and hence, it is presumed, a signature of the reservoir being drilled. Moreover, the signature is available while drilling, which allows discrimination between progressively exposed reservoir sections. If unraveled, this signature gives us information about the reservoir that is otherwise not easily available with the immediacy with which UBD operations provide it. The prospect of unraveling this signature and thus improving reservoir knowledge is increasingly seen as an important driver for application of UBD.
During UBD, pressures and rates at the inlet (injection into drill string and/or concentric gas injection annulus between casing and tie-back) and outlet (choke) are usually measured. Downhole conditions (temperature, flowing wellbore pressure) are also often captured while drilling. In general, all of these measurements are time-varying. If the inverse problem is formulated properly, these known parameters can deliver the pore pressures, permeability and other production related characteristics of the reservoir. This can have significant benefits in reservoir characterization, production optimization, and in justification of UBD. In particular, the ability to characterize the reservoir during drilling enables the Petroleum Engineer to make immediate use of the knowledge in designing completions that optimize the performance of the well being drilled. Additionally, the evaluation and exploitation of opportunities can occur almost simultaneously, with powerful implications for decision making and maximizing asset value.