arXiv subjects
Publications and source records attributed to Ankush Singh.
In this paper we discuss lung cancer detection using hybrid model of Convolutional-Neural-Networks (CNNs) and Support-Vector-Machines-(SVMs) in order to gain early detection of tumors, benign or malignant. The work uses this hybrid model by training upon the Computed Tomography scans (CT scans) as dataset. Using deep learning for detecting lung cancer early is a cutting-edge method.
Interference tests are used in shale reservoirs to evaluate the strength of connectivity between wells. The results inform engineering decisions about well spacing. In this paper, we propose a new procedure for interpreting interference tests. We fit the initial interference response with the solution to the 1D diffusivity equation at an offset observation point. It is advantageous to use the initial interference response, rather than the subsequent trend, because it is less affected by nonlinearities, time-varying boundary conditions, and uncertainties about flow geometry and regime. From the curve fit, we estimate the hydraulic diffusivity and conductivity of the fractures connecting the wells. For engineering purposes, it would be useful to quantify the impact of interference on well production. Thus, we seek a relationship between the degree of production interference (DPI) and an appropriate dimensionless quantity that can be derived from the estimate of fracture conductivity. Using simulations run under a wide range of conditions, we find that the classical definition for dimensionless fracture conductivity does not achieve a consistent prediction of DPI. This occurs because the dimensionless fracture conductivity is derived assuming radial flow geometry, but the dominant flow geometry during shale production is linear. We also find that the CPG (Chow Pressure Group) metric does not yield consistently accurate predictions of DPI. As an alternative, we derive a dimensionless quantity similar to the classical dimensionless fracture conductivity, but derived for linear, flow geometry. Using this approach, we calculate a dimensionless interference length that collapses all cases onto a single curve that predicts DPI as a function of fracture conductivity, well spacing, and formation properties. We conclude by applying the new method to field cases from the Anadarko and Delaware Basins.
ResFrac is a combined hydraulic fracturing, reservoir, and hydraulic fracturing simulator. It describes multiphase fluid flow (black oil or compositional), proppant transport, transport of non-Newtonian fluid additives, and thermal transport. It also includes stress shadowing from fracture propagation and porothermoelastic responses from pressure change in the matrix. It uses constitutive equations that smoothly transition between equations for flow through an open crack to flow through a closed crack (with or without proppant). This document provides a detailed technical description of the code, along with validation simulations to confirm numerical accuracy.