Perpustakaan Digital ITB

Conventional reservoir modeling has long relied on extensive petrophysical data acquisition and lengthy development timelines. While thorough, this approach is often too slow and resource-intensive for rapid field screening or early-stage opportunity identification. This study proposes a practical alternative: a dual- domain field screening framework built based on Top-Down Reservoir Modeling (TDRM) philosophy, which draws characterization entirely from historically accumulated field data rather than demanding new acquisition campaigns. The framework brings together two complementary analytical domains. The first employs a Gaussian Mixture Model (GMM) to characterize static petrophysical properties, while the second applies Fuzzy Pattern Recognition (FPR) through the Fuzzy C-Means (FCM) algorithm to capture dynamic reservoir behavior, using Decline Curve Analysis (DCA) parameters as its primary input features. The study was conducted in Field X, working with production and reservoir data from 56 wells across nearly 50 years of production history (1974–2024). The outputs of both domains were then integrated into a quadrant- based screening matrix, categorizing wells into four distinct classes: Sweet Spot, Underperformer, Fracture Anomaly, and Consistent Poor. The results show that this framework is capable of effectively capturing spatial heterogeneity and production performance variability across Field X, that offer an efficient basis for prioritizing field development decisions under conditions of limited subsurface information.