Hydraulic fracturing is a stimulation treatment routinely performed to create fracture network on low permeability reservoirs to enhance the productivity. Such induced fracture network has much higher conductivity and generally is treated through either local grid refinement (LGR) to capture the transient phenomenon or embedded discrete fracture model (EDFM). Both approaches require complex gridding meshes, leading to heavy computational time. LGR also requires the orthogonal orientation of hydraulic fracture with horizontal wellbore trajectory. Another challenge for LGR and EDFM comes from the dynamic meshing over the time, embedded from the request of infill drilling and re-fracturing due to the nature of the fast production decline in those hydraulically fractured wells. In the case of infill drilling or re-fracturing, the grids for the well or completion stages have to be generated from the beginning of simulation, causing computational inefficiency. In addition, sensitivity evaluation of well landing point, spacing and completion optimization needs easy preprocessing of model input and quick simulation time.
In this paper, we handled the above challenges through representing hydraulic fracture network with multisegment well (MSW) concept, offering an improved description of the wellbore physics over the conventional well modeling. Since MSW node system is independent upon reservoir grid system, fracture orientation can be at any angle freely with wellbore trajectory, which avoids the complex LGR or EDFM and reduces the number of grids. Meanwhile, MSW provides a flexibility of fracture geometry representation, enabling easy addition and alteration of fractures at any simulation time. We developed a user-friendly interface to process all the grid generation and seamlessly feed into reservoir simulator. We also linked the interface to our in-house uncertainty and optimization package to perform experimental design and automatically submit all the necessary simulations and statistically analyze the simulation results based on EUR and NPV objective functions.
Our MSW approach has been validated by comparing with the LGR through several benchmarking studies of a tight reservoir. A field case was demonstrated for infill drilling among existing vertical wells and re-fracturing operations. While the infill drilling of a well with hydraulic fractures is difficult by the other modeling approach, the MSW option makes it easy by just opening the wellbore with fractures at certain simulation time. The re-fracturing operation is modeled as the opening of the completion. These settings can be done for each well. Our approach also has potential to dynamically couple with stimulation design tools.