Introducing a New Capacitance-Resistance Model and Solutions to Current Modeling Limitations

Abstract

Abstract Numerical reservoir simulation offers the best representation for reservoir fluid flow. However, uncertainty in assigning and distributing static parameters between wells adds a dimension of weakness to its implementation and cultivates modeling doubts. Capacitance-resistance modeling is one of the recent evolving technologies that provides opportunities to study fluid dynamic signatures that can help unveil uncertainties for some aspects of static reservoir parameters such as fractures and faults. Capacitance-resistance modeling (CRM) is a tool that relies on signal processing in associating actions (i.e. injection and drawdown) to reactions (i.e. production). This is accomplished through means of multivariate nonlinear regression. There are many documented limitations to this tool that reduce the model reliability, or prevent it from attaining an acceptable match. Capacitance-resistance modeling limitations result from varying model parameters, or missing input data. Examples of these limitations include changes to the number of active producers, significant change in well productivity index (e.g. well workover), high variation of fluid compressibility, and presence of aquifer support. Most reservoirs will encounter many of these limitations which makes it important to find a rectifying solution. In this paper, CRM limitations are addressed and a new modeling equation is introduced to produce a flexible capacitance-resistance model. Shutting-in producers, or introducing them, changes the number of active production wells. This limitation is solved by modifying the injection rates in a way that honors the mass balance and keeps the modeling weights constant. Changes in a well's productivity index are addressed by considering the well as two separate wells (before/after change), while the presence of aquifer support is engaged by adding a pseudo injection well, to mimic influx support, and adjusting the injection rate to reach the best possible match quality. As for the new modeling equation, it performs very well when compared to current CRM equations, has fewer variables, and has more applications due to the incorporation of a drainage control-volume that acts as a filter to all supporting streams. This new formulation also gives clearer indications to noncontributing injectors; the information from which can be used to detect the presence of faults and fractures.