"Smart" or "intelligent" wells are advanced wells with sensors and valves installed downhole to allow for easy monitoring and regulation. The valves are adjusted remotely based on conditions detected by the sensors.  The first smart well was WellDynamics' Surface Controlled Reservoir Analysis and Management System, implemented in 1997 at Saga Petroleum's Snorre oil field in the North Sea. Since then, hundreds of smart well systems have been put into operation around the world. 
Smart wells enhance reservoir management in several ways:
The ability to alter reservoirs remotely, without intervention, is both economical and expedient. Smart well technology was specifically pioneered for this purpose, although its applications have been expanded since then.  While traditional wirelines are cheap and effective when dealing with easily accessible wells, they are unfit for subsea and extended reach completions, which have become increasingly prevalent. Maintaining deepwater work platforms from which to perform the interventions is costly. Furthermore, mechanically manipulating valves thousands of feet below the surface of the ocean can be risky.  Smart well technology utilizes hydraulic and electrical power, and provides a way to adjust valves without direct intervention. [3,4]
While originally conceived for the above reasons, smart wells have become an important tool for accelerating recovery. The valves can be adjusted to minimize the production of water and other undesirable effluents and to maximize oil recovery. The downhole sensors provide real time collection of flow rate, pressure, temperature, and even seismic data, allowing analysis of each production zone.  This data can be used in one of two ways: "reactive mode," with valves reset whenever problems are detected, or, less commonly, "defensive mode," where the data is used in conjunction with production models and simulations to optimize performance and lift costs. [1,5]
Smart well technology is particularly useful in its ability to regulate different branches or segments of the well independently, an ability which is valuable for multilateral wells (wells with multiple branches) and wells with complex or fragmented structure.  It provides a solution to the problem of small, closely spaced reservoirs. Before smart wells, such reservoirs usually had to be produced sequentially, which was wasteful and sometimes economically unviable. However, with smart well technology, such reservoirs can be commingled, with each branch controlled separately using the variable valves. 
The main challenge of smart wells is optimizing their operation. Many competing optimization algorithms, both deterministic and stochastic, have been put to this task. Deterministic optimization algorithms are generally classified as either derivative-free or gradient-based; the former is simpler but can be quite slow, while the latter is potentially very quick to converge but sometimes requires too much calculation to be practical.  However, deterministic algorithms, by definition, always converge to the same optimum given the same initial conditions, which may or may not be the global optimum. [5,6] The advantage of stochastic algorithms is that, sooner or later, the global optimum will be reached, assuming the algorithm is not terminated prematurely. Stochastic derivative-free methods are a popular choice for smart well optimization, particularly genetic algorithms. 
As reserves disappear, innovation becomes the key to avoiding a decline in oil production. Smart wells are an exciting new technology with a high rate of success.
© Ellie Kitanidis. The author grants permission to copy, distribute and display this work in unaltered form, with attribution to the author, for noncommercial purposes only. All other rights, including commercial rights, are reserved to the author.
 I. Aitokhuehi and L.J. Durlofsky , "Optimizing the Performance of Smart Wells in Complex Reservoirs Using Continuously Updated Geological Models," J. Petroleum Sci. Eng. 48, 254 (2005).
 Y. Shuai et al., "Using Multiscale Regularization to Obtain Realistic Optimal Control Strategies," One Petro 142043-MS, Proc. SPE Reservoir Simulation Symposium, 21 Feb 11.
 M. Konopczynski, "Intelligent Wells Can Improve Reservoir Performance," Drilling Contractor, March/April 2004, p. 37.
 S. Dyer et al., "Intelligent Completions - A Hands-Off Management Style," Oilfield Review, Winter 2007/2008, p. 4.
 B. Yeten et al., "Decision Analysis under Uncertainty for Smart Well Deployment," J. Petroleum Sci. Eng. 43, 183 (2004).
 B. Yeten, L.J. Durlofsky and K. Aziz, "Optimization of Nonconventional Well Type, Location, and Trajectory," Soc. Petrol. Eng. J. 8, 200 (2003).