Analyses of pressure data from a Wolfcamp tight oil reservoir spacing optimization pilot area in the Midland Basin with a fractured producer at the center and four surrounding monitoring wells at distances of about 460 ft are presented. The center well had a single downhole memory gauge, while each of the four un-fractured monitoring wells had hydraulically isolated high-resolution permanent P/T gauges installed in eight layers. Data from the center well and from the thirty-two observation wells gauges have been studied to improve our understanding of reservoir characteristics and well performance. All wells are vertical, and production in the center well is commingled from eight hydraulically fractured stages that penetrate formations from Upper and Lower Spraberry to Upper, Lower, and Base Wolfcamp, as well as Strawn and Atoka (8 zones). Different pressure response from the center well and the four monitoring wells showed the different layers to be highly heterogeneous both vertically and laterally. In addition, pressure spikes were captured in monitoring wells when the outside offset wells were hydraulically fractured at different time periods. This paper presents interpretations of the multiple data sets to determine flow properties and identify and quantify pressure communication between wells in different layers. Multi-well numerical models with unstructured grids have been used in analyses of the pressure response from the center well and of interference data at the monitoring wells, with special focus on identifying possible discrete fracture networks. The results displayed dynamic evolution of well drainage and inter-well connectivity over time. The interference data revealed a highly complicated network of connections over large areas, far more complex than what we observed from high-resolution downhole four-array microseismic data. The analyses used both constant and pressure-dependent properties in an effort to get a better understanding of enhanced regions from stimulation activities, effective fracture half-lengths, drainage areas, well interferences, and discrete fracture networks. As we moved to horizontal well program, better characterizing natural-induced fracture networks has significant potentials in helping infill drilling, horizontal well planning, fracture hit mitigation and landing strategy in the stacked pay. In addition to the single-layer analyses of center-well and interference data we also used a general-purpose numerical simulator to match an extended production data set from the center well. To this end we used a single-layer, single-well approach with enhanced properties in a rectangular region around the hydraulic fracture. This analysis gave the best match with pressure-dependent properties, but an equivalent reasonable match is also shown with a time-dependent skin approach.
展开▼
