A methodology to construct deep neural network- (DNN) and recurrent neural network- (RNN) based proxy flow models is presented; these can reduce computational time of the flow simulation runs in the routine reservoir engineering workflows, such as history matching or optimization. A comparison of these two techniques shows that the DNN model generates predictions more quickly, but the RNN model provides better quality. In addition, RNN-based proxy flow models can make predictions for times after those included in the training data set. Both approaches can reduce computational time by a factor of up to 100 in comparison to the full-physics flow simulator. An example of the proxy flow model application is successfully demonstrated in an exhaustive search history matching exercise. All developments are shown on a synthesized Brugge petroleum reservoir.

The digital transformation journey provides new opportunities for running simulations through cloud computing, with flexibility in hardware resources and availability of a wide array of software tools that can enhance decision-making. This work reviews the benefits of running reservoir simulations in a cloud environment and demonstrates the efficiency and cost savings. Additionally, a workflow for uncertainty analysis and history matching that integrates data analysis and machine-learning tools is presented. First, the hardware architecture must be designed to meet parallel reservoir simulation needs: significant message passing occurs between computer nodes, and for satisfactory performance, these nodes must be connected by a low-latency network, rather than be randomly located. Second, to ensure portability and easy replication across multiple cloud sites and platforms, the software performing the simulations must be containerized. Third, to reduce the time required to start a new simulation run, the Kubernetes platform is used to optimize resource allocation. Finally, reservoir simulation in the cloud is no longer merely the running of the simulation model, but it is integrated with data management and data analysis tools for decision-making. The cloud-based simulation services discussed herein exhibit good results during scale up, when a simulation operation requires a larger number of central processing units and/or greater memory, and also during scale out, when thousands of operation scenarios are necessary for history matching. The "pay as you go" pricing model reduces the time and capital costs of acquiring the new computing infrastructure to nearly zero, and the effectively unlimited scale-out capability can reduce the elapsed time for history matching by 80%. The availability of data centers in different regions is good for team collaborations. It serves the data management tool well to track history data, perform data mining, extract more information, and make decisions. Compared to traditional reservoir simulation, the cloud-based reservoir simulation software as a service model simplifies the process and reduces hardware acquisition and maintenance costs. Integrating intelligent data analysis with simulation helps quantify the uncertainty in the model and enables improved decisions.

Unconventional reservoirs have characteristics that differ from traditional conventional reservoirs. The productivity profile of an unconventional well can be significantly different from a conventional well, in that the production rate declines faster in an unconventional well. Therefore, properly planned well operations are crucial to optimize costs and production from these shale assets. This includes understanding the reservoir physics, planning optimal well spacing, improving well performance from completions, and simulating refracturing effects on well production. This paper presents a new multidisciplinary method to help improve field performance and productivity in unconventional reservoirs. A multidisciplinary approach is necessary for economical and successful operations in unconventional reservoirs. In unconventional fields, wells are drilled quickly; therefore, rapid decision-making is necessary. Currently, fracture modeling is performed using either fine local grid refinements or dual-porosity dual-permeability models, which can be cumbersome and time-consuming. This paper presents a new approach that uses multiple shale-specific features and unstructured models, which allows users to specify discrete natural fracture networks (NFNs) and hydraulic fractures with arbitrary orientations connected to practical well trajectories. The automated gridding technique significantly simplifies the workflow, thus allowing users to focus on addressing issues in the engineering space by streamlining the setting up of complex reservoir simulation models. The approach is applicable to black oil and compositional models of all fluid types. Using parallel capability, performance can be enhanced severalfold. The new approach helps enable modeling of multiple scenarios by modifying parameters easily; thereby, results are readily available to help operators plan optimal well and fracture spacing and length. This paper highlights how well productivity can be improved by optimizing well placement and incorporating the effect of NFNs and hydraulic fractures.

Reservoir simulation is becoming increasingly complex because of more advanced wells in the fields, including intelligent and multilateral wells. Advanced completions are also evolving to increase recovery efficiencies. This increasing complexity presents two difficulties, which include the design of advanced completions within reservoir simulators and increased simulation runtime. To describe a well in a reservoir model, a reservoir engineer typically defines a network of hundreds of nodes using keywords and specifies properties for each node. This is a cumbersome and error-prone process. Additionally, detailed well models can slow down reservoir simulation and often cause poor convergence. A new iterative round-trip approach has been implemented, in which an engineer imports an initial reservoir model into a nodal analysis simulator that models flow from the reservoir through complex completions to the wellhead. The simulator accurately models well production in steady state and designs completion strings in detail. After the design is complete, the nodal analysis simulator converts the well model into reservoir simulator keywords that are imported into full-scale simulations for transient analysis. Using this method, reservoir simulators can also model multiple annuli, which was not feasible previously. Highly detailed well models of several thousand nodes that accurately describe completion strings can be generated automatically. Reservoir engineers typically do not possess complex knowledge of well design because this is usually performed by completion/production engineers, who seldom have access to a reservoir simulator. Consequently, they have a limited ability to experiment with different well designs. This paper presents an approach that helps facilitate reservoir and production engineer collaboration, thus helping enable fine-tuning of final completion designs to maximize production, prevent early water/gas breakthrough, and increase overall recovery. This paper describes the application of the new process in an openhole well and presents various completion designs of the same well.