Patacchini, Leonardo (Abu Dhabi Marine Operating Company) | Mohmed, Farzeen (Abu Dhabi Marine Operating Company) | Lavenu, Arthur P. C. (Abu Dhabi Marine Operating Company) | Ouzzane, Djamel (Abu Dhabi Marine Operating Company) | Hinkley, Richard (Halliburton) | Crockett, Steven (Halliburton) | Bedewi, Mahmoud (Halliburton)
The classic method for initializing reservoir simulation models in the presence of a transition zone, based on primary drainage capillary-gravity equilibrium, is extended to account for partial reimbibition post oil migration. This tackles situations where structural events, such as trap tilting or caprock leakage, caused the current free-water level (FWL) to rise above deeper paleo-contacts. A preliminary primary drainage initialization is performed with zero capillary pressure at the paleo (or deepest historical) FWL, to obtain a minimum historical water saturation distribution. From a capillary pressure hysteresis model, it is then possible to determine the appropriate imbibition scanning curve for each gridblock, which are used to perform a second initialization with zero capillary pressure at the current FWL. With the proposed method, log-derived saturation profiles can be honored using a physically meaningful capillary pressure model. Furthermore, when relative permeability hysteresis is active, it is possible as a byproduct of the initialization to assign the correct scanning curves at time zero to each gridblock, which ensures that initial phase mobilities (hence reservoir productivity) and residual oil saturation (hence recoverable oil to waterflood) are modeled correctly. This is demonstrated with a synthetic vertical 1D model. The method was implemented in a commercial reservoir simulator to support modeling work for a giant undeveloped carbonate reservoir, where available data suggest that more than 3/4 of the initial oil in place could be located between the current FWL and a dome-shaped paleo-FWL. This work is used as a case study to illustrate the elegance of the proposed method in the presence of multiple (or tilted) paleo-FWLs, as only one set of capillary pressure curves per dynamic rock-type is required to honor the complex logderived saturation distribution.
Short-term production and injection optimization are best approached from an integrated surface/subsurface perspective, recognizing that well performance is driven by competition for an existing network hydraulic capacity.
This paper presents a tool for real-time optimization (RTO) of water-injection systems at the scheduling time scale (i.e., days to months). Its development stemmed from the observation that operations such as pigging or shutting manifolds for rig activity might disrupt the injection network balance; hence, injectors would benefit from quick control readjustments. Furthermore, an existing network is not necessarily able to distribute available water where desired, and control compromises best found by an optimizer should be sought.
It is assumed that reservoir conditions are stationary, and injection targets at any level of granularity (well, reservoir segment, or field level) have been established based on subsurface requirements. By use of performance curves for each injector and either a simplified or a full-fledged network model, the algorithm finds a set of optimal well controls with a steepest-descent method implemented in Microsoft (2016) Visual Basic for Applications (VBA). The interface is spreadsheet-based, facilitating updates in well-performance data or changes in reservoir requirements. When needed by the algorithm, a third-party hydraulic-flow simulator able to balance the system from the injection modules down to the manifolds is called through an application programming interface.
A case study is presented, illustrating how the tool has been used to estimate the benefits of installing wellhead chokes on the currently more than 200 active injection strings of a giant oil field offshore Abu Dhabi.
Su, Shi Jonathan (Schlumberger) | Patacchini, Leonardo (Abu Dhabi Marine Operating Company) | Mohmed, Farzeen (Abu Dhabi Marine Operating Company) | Farouk, Magdy (Abu Dhabi Marine Operating Company) | Ouzzane, Djamel (Abu Dhabi Marine Operating Company) | Draoui, Elyes (Abu Dhabi Marine Operating Company) | Torrens, Richard (Schlumberger) | Amoudruz, Pierre (Schlumberger)
Coupling is performed periodically at the wellhead, using a reservoir simulator in which the field manager controls the reservoir models by supplying well constraints and controls the network models by supplying well performance curves. Allocation strategies and pressure and flow constraints are imposed by the field manager, for which the different sub-models are black boxes; the models themselves are controlled hydraulically without embedded production or injection constraints. This explicit approach has been selected for its flexibility. In particular, by expressing rates at the surfacesubsurface interface at standard conditions, it is possible for the two reservoir models to have different equations of state and different treatments of injected water salinity, while the surface models use a blackoil fluid description. This project required ensuring rate continuity at the transition from history to forecast for over 600 active production and injection strings, even when the reservoir and network models are not perfectly historymatched. This was achieved by introducing pressure shifts in each vertical flow performance curve to ensure continuity of the choking margins (i.e., differences between wellhead pressures and backpressures) and by overriding the default guide rate flow allocation method of the field manager to prevent abrupt changes in the production split of wells currently producing below potential. The use cases described here are based on an eight-year (2015-2023) drilling schedule followed by no further activity. We focus on assessing the impact on production and injection arising from: replacing pipelines or changing network topologies; relaxing the constraint of producing at initial solution gas-oil ratio with and without reduction of separator pressures; and redistributing or increasing the water injection capacity. 2 SPE-183153-MS
Waterflood is the most commonly field development scheme used for hydrocarbon recovery in carbonate reservoirs. Successfully managing water injection scheme greatly depends on the reservoir characterization, in both structure and fluids distribution, that in turn requires a variety of subsurface measurements coupled with the ability to integrate them into a coherent subsurface models.
Electrical resistivity is an important tool in formation evaluation of oil and gas reservoirs, and frequently used in logging programs to evaluate near-well reservoir rocks, fluids saturations and to help design well completions. With the extended use and added complexity of reservoir modelling it has become imperative to extend this knowledge away from the wellbores and into the reservoir, thereby enhancing the understanding of fluids distribution inside the reservoirs, thus improving overall field management. This can be accomplished by applying tools that are sensitive to the interwell environment. One of these tools is inductive DeepLook Electromagnetics or Crosswell Electromagnetics (EM). Crosswell EM is an induction based tomography technology, which inductively measures the interwell resistivity between wells. This technology, particularly useful for tracking water and steam floods or mapping residual saturation, is also used to optimize sweep efficiency, identify bypassed pay, and predict fluid-related issues such as water breakthrough.
Integration of Crosswell EM interpreted results to reservoir characterization process may offer new wide range applications to deep reading measurements. The current Crosswell EM interpretations are the results of inverted resistivity and hence the water saturation in pseudo 3-D format between wells. Until now there are no methods, workflows or tools available to integrate Crosswell EM data into the reservoir characterization process through history matching.
This paper presents the methodology and investigation of integration of complex Crosswell EM results to history matching process in order to improve the reservoir characterization during dynamic modelling. The use of this data in reservoir characterization has not been yet considered by the industry. Our attempt is to lead the oil industry to utilize Crosswell EM data in the history matching process of the legacy data. The proposed workflow helps reduce and/or quantify uncertainties in the reservoir model. Therefore, the inclusion of time lapse Crosswell EM data into history matching will improve the understanding of the flow paths, barriers, as well as lateral and vertical transmissibility.