Poroelastic Effects on the Interpretation of Diagnostic Fracture Injection Tests

A fully coupled poroelastic model is developed to simulate hydraulic fracture initiation, propagation, and closure using the finite element method. The pressure decline data after shut-in is analyzed to estimate the leak-off coefficient, in-situ stress, permeability, and pore pressure. Results are compared to previous DFIT techniques developed based on linear elastic fracture mechanics and Carter's leak-off model. During the before-closure analysis, the pressure decline data is analyzed to estimate the leak-off coefficient and in-situ stress through a special dimensionless time function called G-function. With poroelastic effects taken into account, the leak-off coefficient can be twice as large as the one obtained from previous DFIT techniques. Also, higher in-situ stress estimates are obtained in the fully coupled model. The after-closure analysis involves the identification of linear flow and radial flow regimes. The linear flow analysis can significantly overestimate reservoir permeability, while the radial flow analysis can obtain a reliable permeability estimate. Both the linear and radial flow analyses can provide accurate estimates of pore pressure. This study demonstrates that poroelastic effects cannot be ignored in some cases. The previous DFIT techniques may result in inaccurate estimation of properties such as leak-off coefficient and permeability. 1. INTRODUCTION Diagnostic Fracture Injection Tests (DFIT) are techniques that have been widely used in hydraulic fracturing to estimate reservoir properties and parameters, such as leak-off coefficient, in-situ stress, permeability, and pore pressure. The DFIT procedure involves pumping a small volume of fluid into reservoir at a constant rate for a short of time to create a fracture, then the well is shut-in and the pressure falloff is recorded and analyzed. However, existing DFIT techniques are developed based on Carter's oversimplified leak-off model, which can result in inaccurate estimation of reservoir parameters. Pressure interpretation prior to fracture closure is known as before-closure analysis (BCA), which was pioneered by Nolte's work (1979 and 1986) with the development of G-function. Assuming that the fracture surface growth follows the power law, Nolte noted that pressure decreases linearly with the dimensionless G-function before the fracture closes. Analytical techniques were also provided to estimate the fluid loss coefficient, fluid efficiency, and time for the fracture to close. Barree and Mukherjee (1996) and Barree et al. (2009) proposed a method to identify closure pressure qualitatively using G-function and its derivatives. The closure is identified by the deviation of the semi-log derivative of pressure (Gdp/dG) versus G-function from the straight line. This method is referred to as the tangent line method and has become the standard practice for DFIT analysis. Mayerhofer (1993 and 1995) introduced an alternative model to interpret the leak-off process instead of using the Carter's leak-off model. In this method, pressure response is considered from the fracture into the formation. By representing the pressure falloff data in a specialized plot which is referred to as the Mayerhofer plot, the filter cake resistance and the reservoir permeability can be estimated from the intercept and slope of the straight line in the Mayerhofer plot. McClure et al. (2016) presented the compliance method to estimate fracture closure pressure, which is picked up at the point where the system stiffness starts to increase. The compliance method results in an earlier and higher closure pressure estimate than the tangent line method. Wang and Sharma (2017) proposed the variable compliance method to estimate minimum in-situ stress by averaging the G-time of the tangent line method and compliance method. The variable compliance method leads to a more accurate estimation of minimum in-situ stress without considering poroelastic effects. Zanganeh (2017) reinterpreted the facture closure dynamics using the cohesive zone model for fracture initiation, propagation, and closure. A leak-off model that is pressure-dependent and related to formation properties is used for simulation. The closure pressure is picked up at the start of the tip closure, which is consistent with the compliance method proposed by McClure.