Use of Wavelet Transform and Signal Processing Techniques for Inferring Interwell Connectivity in Waterflooding Operations

Abstract

Abstract Inter-well connectivity (IWC) is one of the most significant properties when evaluating the success of a waterflood. This connectivity has been obtained from various physics-based methods such as simulations, tracers and using heuristics and semi-analytical tools like capacitance-resistance model (CRM). Production and injection data are a key piece of information required to compute the IWC. In this study, we present a new method for estimating IWC using signal processing techniques on the wavelet transform of the injection and production rate data. First, the injection and production rates are subjected to multiresolution analysis using the wavelet transform to determine the detail coefficients. The variance of the detail coefficients is then computed and is ready to be processed using various signal processing techniques. Signal processing techniques such as cross-correlation, time lag, Spearman correlation, and Kendal correlation are used to identify the level of relationship between the processed injection and production data in wavelet scale space. Based on the correlation coefficients, a new IWC link parameter is proposed for characterizing the IWC between well pairs. The IWC link parameters between well pairs are then plotted for visual representation. We created several simulation models for multi-well systems, established water-flood patterns, and for randomly placed wells to establish the new IWC link parameter. The resulting injection and production rates were analyzed using the methodology above and the new IWC link parameter is established in terms of cross-correlation coefficient. We also performed several simulations for a heterogenous reservoir to compute and compare the accuracy of the new IWC link parameter. Finally, the methodology is subjected to real field waterflooding, and compared against the CRM results, which shows a good agreement. The visual representation gives new insight into whether the connectivity is being affected by the reservoir or from near wellbore events (such as changes in skin). This study integrates signal processing techniques and waterflood IWCs. Novel use of wavelet transforms coupled with variance for processing the injection and production rate data is proposed. It must be emphasized that wavelet is used in this context for processing and not for smoothing or data compression. Ultimately, this method can be implemented as a real-time automated monitoring system. Moreover, the new IWC link parameter provides insights by identifying problematic IWC, well-completion issues, and high perm channels for taking timely operational decisions.