A field scale polymer flood has been in operation since early 2010 in a major oil field of the Sultanate of Oman. The project comprises 27 patterns where water flood was on-going prior to initiation of Polymer flood in 2010. A polymer flood project has high chemical operating expenditure (Opex). Thus, optimization of a polymer flood requires continuous tracking of mass of polymer injected per unit volume of incremental oil produced for individual polymer flood patterns and then polymer throughput in individual patterns needs to be dynamically altered. To meet this objective, a full-field streamline simulation model has been built, history matched and is being used for optimizing the polymer-flood. Full-field simulation allows the proper modeling of each pattern and their interactions with off-set patterns, and these simulations can be performed in a reasonable computation time because of the efficiency of streamline modeling. Computational efficiency of streamline simulation has facilitated use of the model for routine well and reservoir management decisions. This would not have been possible with a finite difference model because of excessive run time and inability to clearly establish injection-production relationship as in a streamline model. The model has facilitated optimization of polymer flood patterns, specifically when to stop polymer injection, slug size, and slug concentration. Individual pattern performance can be visualized effectively and their efficiency can be compared. The model is also being used for ranking the existing water-flood patterns for the next phase of polymer-flood implementation and carrying out short term production forecast.

A field-scale polymer flood has been in operation since early 2010 in a major oil field of the Sultanate of Oman. The project is composed of 27 mature waterflood patterns that were converted to polymer flood in 2010. Because a polymer-flood project has high chemical operating expenditure, optimization of a polymer flood requires continuous tracking of the mass of polymer injected per unit volume of incremental oil produced (relative to waterflood) for each polymer-flood pattern. To meet these objectives, a fullfield streamline simulation model was built, was history matched, and is being used for optimizing the polymer flood. Full-field simulation allows the proper modeling of each pattern and their interactions with offset patterns. However, full-field simulations can be expensive, so we use a streamline-based simulator to run forecast scenarios in a reasonable computation time on reasonable hardware. Streamlines have the added benefit of determining the time-varying well-rate allocation factors per pattern, meaning that pattern-level diagnostics are relatively easy to compute and are based on the dynamic flow characteristics of the model. Computational efficiency and quantification of patterns have facilitated use of the model for routine well and reservoir-management decisions. We show that one can determine the effectiveness of the polymer flood on a pattern-by-pattern basis over the historical polymer-injection period with a standard oil-produced vs. polymer-injected ranking. In forecasting, we show how to quantify the incremental recovery caused by polymer, above base waterflood, on a pattern-by-pattern basis to facilitate optimization of polymer-flood patterns and more specifically to determine when to stop polymer injection and which new patterns to move polymer injection to.

Water flooding has been used successfully in heavy and viscous oil fields since last fifty years. Though characterized by low recoveries and high cost due to adverse mobility ratio, waterflood is still an attractive option before embarking on expensive EOR solutions. One of the key elements to success of waterfloods in heavy oil is well and reservoir management. This paper illustrates the use of daily production data to set up systems that indicate the health of waterflood through use of classical techniques so that they are readily available on a demand to know basis for any hierarchy of production systems; field pattern, groups, individual patterns etc. The observations and results can then be applied to take proactive measures for preventive management. To manage water floods in a dynamic scenario, reservoir engineers need to watch them closely, analyze them for anomalous behavioral trends in a continuous fashion and be able to apply remedial measures as they manifest themselves. This can not be done through numerical simulation models, first, because of the high cost and time constraints and secondly because the models usually are highly simplified and are constructed using an average property data. Numerous techniques have been developed by various people working on waterfloods around the world in last 4 decades. Most of these techniques are in the form of diagnostic plots based on easily available data. This paper is an effort to put many such methodologies in a structured format which will enable the engineer to monitor the floods in a systematic and step wise manner. It is intended to move personnel involved in waterfloods up a learning curve and provide them with a toolset to not only measure but improve the flood efficiency. It will assist reservoir engineer's production technologists and development geologists who are involved in or are contemplating waterflood operations. Introduction Reservoirs communicate with engineers through production data. This language, if interpreted correctly can lead the engineer to understand the sub-surface flow process and thereby improve the efficiency of the system. This paper uses only production data with general use software applications like spreadsheet/ data base systems etc to develop tools that help the engineer to understand the message of reservoirs. The paper reviews and explains in detail along with application on live data, techniques to estimate volumetric sweep efficiency, prediction of pressure response (using net voidage), interpretation of injection behavior (Hall's Plot), predicting water source in a oil producer (WOR techniques), using heterogeneity index to priorities options and prediction of ultimate recovery through X-plot calculations and a host of minor techniques that help in understanding & visualization of waterflood behavior.