A flow-based pattern recognition algorithm for rapid quantification of geologic uncertainty

Abstract

Geologic uncertainties and limited well data often render recovery forecasting a difficult undertaking in typical appraisal and early development settings. Recent advances in geologic modeling algorithms permit automation of the model generation process via macros and geostatistical tools. This allows rapid construction of multiple alternative geologic realizations. Despite the advances in geologic modeling, computation of the reservoir dynamic response via full-physics reservoir simulation remains a computationally expensive task. Therefore, only a few of the many probable realizations are simulated in practice. Experimental design techniques typically focus on a few discrete geologic realizations as they are inherently more suitable for continuous engineering parameters and can only crudely approximate the impact of geology. A flowbased pattern recognition algorithm (FPRA) has been developed for quantifying the forecast uncertainty as an alternative. The proposed algorithm relies on the rapid characterization of the geologic uncertainty space represented by an ensemble of sufficiently diverse static model realizations. FPRA characterizes the geologic uncertainty space by calculating connectivity distances, which quantify how different each individual realization is from all others in terms of recovery response.Fast streamline simulations are employed in evaluating these distances. By applying pattern recognition techniques to connectivity distances, a few representative realizations are identified within the model ensemble for full-physics simulation. In turn, the recovery factor probability distribution is derived from these intelligently selected simulation runs. Here, FPRA is tested on an example case where the objective is to accurately compute the recovery factor statistics as a function of geologic uncertainty in a channelized turbidite reservoir. Recovery factor cumulative distribution functions computed by FPRA compare well to the one computed via exhaustive full-physics simulations.

Multivariate Analysis and Data Mining of Well-Stimulation Data by Use of Classification-and-Regression Tree With Enhanced Interpretation and Prediction Capabilities

“…However, other distance norms can be used as measures of dissimilarity, such as Minkowski, Canberra, chord, dominance, or city-block distance (Alpak et al 2010). Kernel-based clustering methods (Filippone et al 2008) overcome this limitation by embedding the data points into a high-dimensional nonlinear domain and defining their similarity by use of a nonlinear kernel distance function (Kim et al 2004).…”

“…This is again not an intuitive conclusion, suggesting that low flow rates can be seen as the cause of screenouts. However, in literature (Cleary et al 1993;Aud et al 1994), field case studies have been reported with such premature or near-well screenouts. This could be caused by poor connection between the wellbore and the created fracture because of well completions, near-wellbore tortuosity, and stress or presence of natural fractures (Urbancic and Maxwell 2002).…”

Robust Optimization of Well Placement in Geologically Complex Reservoirs

“…The objective function can be mathematically expressed as (1) where r o denotes the oil revenue per unit volume, r wp is the injection cost per unit volume, and r wi is the water production cost per unit volume. In Eq.…”

“…The representative realizations are determined using the flow-based pattern recognition technique documented in Alpak et al (2010). The latter takes advantage of rapid streamline simulation and the Kernel k-Means technologies.…”

A multi-resolution workflow to generate high-resolution models constrained to dynamic data

“…In recent years, several papers have been published showing the benefits of distance-based techniques applied to ensembles of models for uncertainty quantification, model selection and history matching [2,4,15,20,21]. Scheidt and Caers [20] show how a distance which is well correlated to the difference in flow response can be used to perform uncertainty quantification efficiently.…”