Control of Reservoir Heterogeneity in SAGD Bitumen Processes

Abstract

Uniform steam chamber growth (conformance) in a Steam Assisted Gravity Drainage (SAGD) process promotes enhanced bitumen recovery, project economics and environmental benefits. In the past, operators have implemented numerous alternative strategies to improve steam conformance in the SAGD process. Simultaneous injection in the inner tubing and annular space or use of dual-tubing completions are commonly used to provide some degree of injection control at the heel- and toe-regions of the horizontal well pair. However, this does not necessarily guarantee the uniformity and performance efficiencies sought. Distributed flow control devices (FCDs) can also contribute to more uniform production and injection. They can be incorporated in the horizontal production completion as restrictive elements to modify the pressure distribution along the length of the wellbore. A hybrid of these two technologies is proposed to provide superior (uniform and efficient) steam chamber development that benefits asset performance in highly heterogeneous bitumen reservoirs. It has been shown that using Proportional-Integral-Derivative (PID) feedback to control the steam injection can be beneficial. The feedback control is applied to each steam injection point in the horizontal well pair. Injection at these control points is regulated by a PID feedback controller which monitors temperature differences between injected and produced fluids in order to both enforce a specified subcool and to achieve uniform production along the entire length of the producer. This paper examines detailed wellbore simulations of a SAGD well pair with an FCD completion in the producer and PID controlled steam injection with dual tubing strings. A synthetic reservoir model, based on logs from the Athabasca region of Alberta, is employed and represents a highly heterogeneous formation with properties typical of a bitumen resource. Feedback controlled steam injection (FCSI) can be dynamically configured to target the worst-offending regions of the well pair in order to (i) start the PID control as early as possible after switchover, (ii) achieve a specified subcool target in those regions that would benefit most from the subcool and (iii) temporarily ignore regions which are bypassed and difficult to produce, but which may be dynamically included at a later time. Practical algorithms are presented to achieve these goals, based on existing technology.

Model-Based Adaptive-Predictive Control and Optimization of SAGD under Uncertainty

“…Given a particular operating point , it is desired to find a linear model which is locally equivalent to (17) around . This model has the form (18) where G ϭ [ 1 2 . .…”

“…injecting steam at a empirically specified constant rate, considering well-known features of SAGD processes, such as: (i) multiple injection (input variables) and production (output variables) segments, (ii) complex dynamics, i.e., nonlinear, slow, high order, time varying, potentially highly heterogeneous reservoirs, (iii) operational constraints, and (iv) model uncertainty and measurement noise. Subcool control for tackling conformance has been essentially addressed through the conventional Proportional-Integral-Derivative (PID) approach (Hin-Sum , Stone et al 2013, Stone and Bailey 2014. While PID will regularly drive and keep the subcool relatively close to the set-point, the control performance might not be completely satisfactory considering the SAGD process above-referenced complex features.…”

Mechanisms and Operational Strategies of Multi-Lateral Steam-Assisted Gravity Drainage (SAGD) for Heterogeneous Reservoirs

“…Researchers have made investigations including fast screening of 3D heterogeneous shale barrier configurations [4], data-driven models for characterizing shale barrier configuration [5], numerical studies of the effects of lean zones and evolution characteristics of the SAGD steam chamber [6], application of cluster analysis and artificial neural networks [7], and data analytics and machine learning to characterize shale barrier configuration [8]. They have carried out a 3D physical simulation on dual horizontal well SAGD in heterogeneous reservoirs [9], classified the impact of thermal and permeability heterogeneity on SAGD performance [10], and proposed methods to control reservoir heterogeneity [11].…”

Application of Temperature Fall-Off Interpretation Method in Superheavy Oil or Oil Sand SAGD Process

“…For superheavy oil or oil sand SAGD projects, the preheating stage is usually indispensable to form interwell hydraulic connection or mobile oil before conversion to the normal SAGD process [2]. The preheating ways usually include steam circulation, steam bullheading, and steam stimulation [3][4][5]. Affected by strong reservoir heterogeneity, shale interlayers or laminae, thief zones (e.g., bottom water, top water, and gas cap), and undulating well trajectory, there are great differences in both preheating time (usually 3-6 months) and steam consumption (usually 25,000-50,000 m 3 ) among different well pairs [6].…”