Evren Ozbayoglu

TL;DR: Wang et al. as mentioned in this paper used RNN-LSTM models to continuously predict the incoming signals in real-time drilling data, and the prediction results were quantified and the probability of kicks calculated based on the different weights for each indicator.

TL;DR: In this article, a methodology is proposed in an attempt to obtain qualitative and quantitative predictions of displacement efficiency, which can provide an insight on the degree of cement contamination, and guidelines on how to minimize the amount of inter-mixing.

Abstract: Rock mechanics triaxial tests and well logs are the main source for determination of mechanical properties of reservoir. These methods obtain values for mechanical properties corresponding to parts of the formation close to the well area. The experimental approach is much more limited as the coring operation is costly and is performed in a handful of appraisal/exploratory wells. Here, a novel methodology, geomechanical well testing, is presented to estimate the elastic and plastic mechanical properties using pressure transient testing. In the first part of the paper, the theory of poroelasticity was utilized to develop a geomechanical well testing technique, which attains estimation of reservoir's poroelastic properties. The technique is on the basis of developing a diffusivity equation incorporating mechanical deformation into the fluid transport in porous media and via applying different mechanical boundary conditions, one can obtain several closed-form solutions for storage coefficient. The developed storage coefficients, however, include Poisson's ratio (PR) as the target variable to estimate from conventional well testing. As a verification, a case study was conducted using the pressure transient data from a sandstone aquifer. Adopting the conventional well testing principles, three different values for PR were obtained. The first impression to this point was to decide which value is the most representing value for the aquifer under investigation. We envisioned two methods for this critical step: 1) comparison of obtained PR values from the geomechanical well testing technique with the one estimated via another reliable source such as well log/seismic and/or 2) comparison of pressure-time measurements with available numerical simulations. We took this finding for granted and devise a way to obtain an estimation of dominant boundary conditions of a reservoir. Once the dominant deformational behavior was assessed, the technique may be re-implemented along with future well testing measuements to monitor the evolutions of elastic properties during the life of reservoir. The second part of the paper deals with another geomechanical well testing technique, which provides a tool for estimation of strength properies through developing a generalized diffusivity equation using poroplasticity theory with a plastic potential function based on Mohr-Coulomb yield/failure criterion. An analytical solution was obtained under uniaxial strain assumption, which can be used to estimate internal friction angle or dilation angle.