This paper presents methodologies to create simple reservoir models that reproduce the hydraulic behavior of complex networks of fractures in unconventional reservoirs. The methods use the results of the rate transient analysis (RTA) methodology recently developed by Acuña (2017). This approach shows that the Characteristic Flow Volume (CFV) is the fundamental property that defines fluid flow behavior for complex networks of fractures and that different reservoirs with the same CFV give the same hydraulic behavior. We use this premise to construct alternative models that we call butterfly models that share the same CFV of complex fracture systems but look different and are simpler to construct and simulate numerically while preserving the same hydraulic behavior. We also show how to construct and calibrate models based on the fractured models for each flow type proposed by Acuña et al. (2018). These fracture models offer a second alternative for construction of simple models. Comparison between the two types of simplified models is performed. Simulation work with the simplified fracture models is used to explain flow behavior commonly seen in unconventional wells without the need for enhanced permeability regions, compaction or desorption. This new approach has been validated with actual data and reservoir modeling and it can match not only the flow behavior of complex fracture networks but also has the capability to forecast the long-term performance. We also explore their use in multi-phase analysis and present a field case.
Acuña et al. (2018) proposed fracture network configurations that explain the three types of behavior for unconventional wells: sub-linear, linear and sub-radial. Figure 1 shows these fracture models. Figure 1 (left) shows a schematic of a sub-linear flow fracture model where a complex network of highly conductivity fractures (purple and blue) creates matrix fragments of many different sizes. We have observed sub-linear flow in oil and in gas wells. Sublinear flow behavior is characterized by large initial production but large decline rate. For oil wells, the gas-oil ratio (GOR) sometimes increases to values several times the value for gas in solution and remains constant for a long time (Zhang and Ayala, 2015). When modeling sub-linear behavior with a linear flow model (Figure 1 center) we find that the calculated fracture length is not enough to deliver the initial production with the prevailing matrix permeability. This leads to the inclusion on enhanced permeability regions (Wang and Karaoulanis, 2016) next to the fractures to improve initial production. Matching the high decline rate, however, requires mechanisms such as stress-dependent permeability implemented in the form of compaction tables that may reduce fracture permeability up to 90% (Ali and Sheng, 2015) as the reservoir depletes. We demonstrate with simulation examples how sub-linear flow also results from the combined effect of matrix fragments of different size flowing together without compaction mechanisms or enhanced permeability regions. This fracture model is consistent with the observations in cores through the SRV by Raterman et al. (2017) and their statement that flow behavior is produced by fracture complexity alone and that matrix changes near the hydraulic fractures that may enhance permeability are extremely limited or absent. Acuña et al. (2018) add that reduction of fracture and matrix permeability with pressure depletion (compaction) should most of the time be a second order effect. The rock fracturing processes that may lead to this fracture network geometry are discussed in Acuña et al. (2018).
