Search

International audience; A neural network (NN) model is trained with a database widely used in the aerosol remote sensing community to rapidly predict the single-scattering optical properties of spheroidal dust particles. Analytical solutions for their Jacobians with respect to microphysical properties are derived based on the functional form of the NN. The Jacobian predictions are improved by adding Jacobians from a linearized T-matrix model into the training. Out-of-database testing implies that NN-based predictions perform better than the business-as-usual method that interpolates optical properties from the database. Independent validation further demonstrates the efficacy of the NN-based predictions by reducing computational costs while maintaining accuracy. This work represents the first use of machine learning-based function approximation to computationally expedite the application of the existing spheroidal dust properties database; the resultant NN model can be implemented in atmospheric models and satellite retrieval algorithms with high accuracy, computational efficiency, and the rigor of analytical solutions.

Waterflooding in deepwater reservoirs typically involves injecting seawater or produced water from the surface via pumps into injection wells. This technique is often cost-prohibitive for many reservoirs and poses significant mechanical/operational risks. This paper discusses how one Gulf of Mexico (GOM) operator overcame all these challenges using smart well technology to implement the first controlled dumpflood in deepwater GOM and boosted the injection rate, reservoir pressure, and recovery from a reservoir at a depth of 20,000 ft.In a typical dumpflood project, uncontrolled water production from the aquifer and subsequent injection into the target zone occurs downhole within the same wellbore. Therefore, typical surface and downhole complexities associated with conventional waterflood projects can be avoided. In this first deepwater GOM controlled dumpflood well, the controlled water flow (≥20,000 bbl/d) is directed from the source aquifer to the target oil zone via inflow control valves (ICV). The ICV, downhole permanent pressure gauges, and the downhole flowmeter provide complete surveillance and control of the injection operation to achieve reservoir management and optimize the waterflood objectives.A world-class Pliocene oil reservoir in the deepwater GOM underwent significant pressure depletion due to a weak water-drive mechanism. Extensive subsurface studies and modeling suggested great rock quality and reservoir connectivity, favorable oil-water mobility ratios, and significant upside potential making this reservoir a perfect candidate for waterflooding. Given topsides facility space constraints, a topsides injection was ruled out. Seawater injection via subsea pumping was deemed risky and marginally economical given the high cost and low commodity prices. The asset team then brainstormed ways to minimize the cost and overcome the associated risks and challenges. The asset team envisioned a dumpflood scenario would overcome all the challenges, but a dumpflood had not previously been implemented in the deepwater GOM. From a technical standpoint, all the known risks were identified and addressed, and a low risk factor was determined for this project.After a complex well completion job, the injection rate was ramped-up to ≥20,000 bwpd water via the ICV. An immediate uptick in reservoir pressure and production rate was observed in the producer well 3,000 ft away. Continuous injection has resulted in reservoir pressure and flowrate increases by at least 1,000 psi and 4,000 bopd, respectively, consistent with reservoir modeling estimates.The operator was successful in implementing an existing technology in a unique way in the deepwater environment. A naturally occurring water source at a depth of 19,000 ft was efficiently harvested to increase recovery from a reservoir at a fraction of the cost of a conventional deepwater waterflood project. Great interdisciplinary collaboration and forward thinking enabled the success of this unique project, opening up tremendous possibilities to increase recovery from other fields where a conventional waterflood may not be justifiable.