Search

Full history match models in subsurface systems are challenging due to the large number of reservoir simulations required, and the need to preserve geological realism in matched models. This drawback increases significantly in big real fields due to the high heterogeneity of the geological models, the reservoir simulation computational time (which increases superlinearly). In this work, we propose a novel framework based on artificial intelligence to address these shortcomings. Our workflow is based on two main components: The first is the new combination of model order reduction techniques (e.g., principle component analysis (PCA), kernel-PCA (k-PCA)) and artificial intelligence for parameterizing complex three-dimensional (3D) geomodels, called "Geo-Net". Our new approach is able to create complex high dimensional heterogeneous reservoirs in seconds, providing better correspondence with the underlying geomodels, hard-data constraints and geological plausibility. The second component is a derivative-free optimization framework to complete the automatic history matching (AHM). This new approach allows us to perform local changes in the reservoir at the same time as we conserve geological plausibility. We have examined our methodology in a real field in Colombia. The Rubiales Oil Field is located in the Llanos Basin with original oil in place of around 6 billion barrels. The key finding here is that the Geo-Net is able to recreate the full geological workflow obtaining the same high order of statistics as traditional geo-statistical techniques. Nonetheless, our Geo-Net allows us to control the full process with a low-dimensional vector and reproduces the full geological workflow 10,000 times faster than commercial geo-statistical packages. Finally, the full optimization workflow has been applied to AHM. Results show an improvement with respect to best practice of traditional history match workflows.

A geologically plausible history matching workflow has been applied to a complex reservoir to improve reservoir characterization. Different structural interpretations have also been included in the formulation which allow with a single workflow to match petrophysical properties, structural interpretations, fluid properties and fault transmissibilities avoiding any regional multipliers or inconsistent discontinuities. A multiobjective optimization was formulated to assimilate production and pressure timeseries as well as well test data. An inhouse implementation of a Particle Swarm Optimization allows to efficiently solve the optimization problem and provide multiple matching solutions for an improved uncertainty quantification. The multiobjective formulation allows the decision maker to screen the matching realizations based on the degree of confidence on the difference data type as well have better control selected the most representative realizations. The workflow proposed shows good match with the observed quantities and allows a review of the initial model of the field based on the improved understanding of the dynamic response of the reservoir. It is the first step before a field development plan optimization with structural uncertainty.

Field development plan optimization under uncertainty requires a consistent analysis of well placement across the geological realizations to evaluate the selected cost function. Special care has to be taken in the well trajectory description in the target zone, to allow in the same formulation vertical, deviated and directional well assessment for a more effective decision making. In case of structural uncertainty well trajectories will cross different grid elements in each realization. The workflow proposes a methodology to screen well trajectories based on the expected productivity overall the realizations and the fulfillment of user defined constrains. Well constrains can include, inter-well distance, well length, distance from the closest fault. For a consistent uncertainty propagation and an efficient optimization a nested optimization loop has also been proposed to allow the well screening before the actual reservoir simulation evaluation and allow only the most promising strategies to be evaluated and, therefore, reducing the overall computational burden. The workflow has been tested on a real reservoir case showing the strength of the methodology in assessing the location for an infill and a sidetrack well and improving the understanding of the reservoir dynamic behavior.

The General Video Game AI (GVGAI) competition and its associated software framework provides a way of benchmarking AI algorithms on a large number of games written in a domain-specific description language. While the competition has seen plenty of interest, it has so far focused on online planning, providing a forward model that allows the use of algorithms such as Monte Carlo Tree Search. In this paper, we describe how we interface GVGAI to the OpenAI Gym environment, a widely used way of connecting agents to reinforcement learning problems. Using this interface, we characterize how widely used implementations of several deep reinforcement learning algorithms fare on a number of GVGAI games. We further analyze the results to provide a first indication of the relative difficulty of these games relative to each other, and relative to those in the Arcade Learning Environment under similar conditions., 8 pages, 4 figures, Accepted at the conference on Computational Intelligence and Games 2018 IEEE

SummaryDecision making under uncertainty can be quite challenging, especially when complex numerical simulations are considered in the work flow and the decision has to be made relatively fast (e.g., in hours). This is the case when one needs to rank a given field portfolio within a limited budget and with acquisition constraints. If the ranking measure associated with each field is properly and rapidly evaluated, new prospect opportunities, which may lead to a favorable strategic position, can be readily identified.In this paper, we propose an efficient methodology for computing a “production-potential” measure that can be used to rank greenfield portfolios in the presence of geological uncertainty, quantifying both uncertainty and risk propagation. Next, we briefly describe the basics of the method proposed. First, uncertainty in sedimentary variability and flow behavior has to be characterized by a number of representative geological realizations. Sampling techniques are used to significantly reduce the number of realizations while preserving accuracy in the description and uncertainty propagation. Thereafter, multiple and varied field-development plans, based on primary/secondary-recovery mechanisms, are automatically generated while accounting for key parameters related to the number, drilling locations, and drilling sequence of wells. In these plans the reservoir is clustered by areas with similar production/injection potential, and the well locations and drilling schedules are obtained accordingly. The well controls are determined through estimations of the field-recovery factor. By means of experimental-design techniques a relatively small number of field-development plans are selected to capture the most significant production profiles. Each of these development plans is simulated for the realizations sampled previously, and the production-potential measure [e.g., average net present value (NPV) over all sampled realizations] is computed for all the plans. The highest of these measures (i.e., the best development plan) can be used for ranking the greenfield in the portfolio. Response-surface procedures are considered to perform additional analysis computations within iterative optimization procedures. It is important to note that other statistics related to the exploitation potential (e.g., standard deviation of the NPV) can also be used to complement the ranking, thereby mitigating the decision makers’ risk tolerance. The methodology has been tested on the Brugge Field benchmark, which presents 104 realizations of multiple geological parameters. The benchmark has been modified to simulate a greenfield scenario. The ranking measure is the (discounted) NPV averaged over the 104 realizations. The proposed work flow yields a ranking measure of USD 5.43 billion, and the computational cost is approximately 1,900 simulations (performed in a parallel-computing environment). This NPV is somewhat higher than those found for the Brugge benchmark (with similar modified settings) by other researchers. To validate the results, we performed more-exhaustive checking by use of approximately 17,000 simulations, and the ranking measure found was USD 5.51 billion.The new work flow presented allows one to efficiently and in a sufficiently accurate manner support decision making in greenfield-portfolio evaluation. Fast reservoir-performance-evaluation engines open new prospect opportunities that, with traditional decision-making techniques, may be frequently lost.

Hanabi is a cooperative card game with hidden information that has won important awards in the industry and received some recent academic attention. A two-track competition of agents for the game will take place in the 2018 CIG conference. In this paper, we develop a genetic algorithm that builds rule-based agents by determining the best sequence of rules from a fixed rule set to use as strategy. In three separate experiments, we remove human assumptions regarding the ordering of rules, add new, more expressive rules to the rule set and independently evolve agents specialized at specific game sizes. As result, we achieve scores superior to previously published research for the mirror and mixed evaluation of agents., Comment: IEEE Computational Intelligence and Games (CIG) conference, 2018, Maastricht. 8 pages, 1 figure, 8 tables

Economic systems are rife with heterogeneous risk events that have the potential to cause disruption. The diversity of risk types makes it challenging for companies to conduct comprehensive risk analysis for any chosen business opportunity. The current practice is laborious and expensive, involving internal risk analysts and external risk advisory services. In this paper, we present a cognitive system that augments human abilities, with the objective of drastic improvements in the productivity of risk analysis efforts. Our system is provided with a comprehensive risk taxonomy and its textual description along with an extensive corpus of textual data such as news articles. Using a series of textual analysis, knowledge extraction and machine learning techniques, the data corpus is annotated with risk-related information and indexed in a risk store for flexible query and retrieval. Our system interfaces with the risk analyst using a query orchestrator which translates analyst queries that are posed at a high level into lower level queries that are expanded to exploit the system's risk-related knowledge. It also enables formulating a graphical model and assessing the required probabilities; we introduce a particular family of models that can succinctly represent risk events modeled as stochastic processes over a long time horizon. We illustrate how a risk analyst can query the system to build a risk model with the help of a case study.

In this work we present an approach to determine geologically plausible reservoir models that are consistent with well-log data and multimodal observations (e.g., production and seismic data). The methodology is based on optimization techniques that are not invasive to the simulators used to reproduce observations. The geologically plausible reservoir models obtained facilitate reliable probabilistic production forecasts and lead to more robust optimization and decision-making support tools. The methodology approaches model inversion as an optimization problem where the cost function is a measure of the discrepancy between simulated responses and measurement data. Reservoir model uncertainty quantification is addressed by means of multi-start optimization, i.e., the use of local optimization methods with multiple initial guesses. Parameter reduction techniques allow a more effective and efficient global exploration of the optimization space. Geological realism is incorporated in the approach through multiple-point statistics. The set of reservoir models, determined as the output of the multi-start optimization, is thereafter included in a framework for robust optimization under uncertainty (e.g., to determine the drilling location of a new well). The optimization algorithms used are noninvasive with respect to the simulators used in the workflow and are implemented on a distributed- computing cluster. Different aspects of this approach are illustrated by means of a comprehensive case example that uses well-log and production data. It is important to note that the use of simulator-noninvasive (black box, derivative-free) optimization algorithms reduces significantly the dependence of the methodology on the simulators. The optimization methods tested include Generalized Pattern Search, Mesh Adaptive Direct Search, Sequential Quadratic Programming (with numerical derivatives) and Particle Swarm Optimization, a population-based procedure. These algorithms may yield efficient performance in terms of wall-clock time when they are combined with parameter reduction techniques and distributed computing (as is the case in this work). When parallel computing resources are scarce or nonexistent one may resort to methods suited to serial implementations (examples of these simulator-noninvasive algorithms are Hooke-Jeeves Direct Search and model-based approaches). The parameter reduction techniques studied here are based on Principal Component Analysis. In particular, the recently introduced Optimization-based Principal Component Analysis has been observed to deal with complex geological models satisfactorily. Uncertainty in future production is then incorporated in the robust optimization of well placement and of reactive well controls. The Ensemble Kalman Filter is widely used in industry for reservoir model uncertainty quantification. This method fails in some occasions (e.g., the ensemble collapses) for reasons that are not yet well understood. The multistart methodology proposed in this work can be an alternative to the Ensemble Kalman Filter in these problematic situations. Seismic data could be included in this methodology to mitigate uncertainty in probabilistic forecasts.

In this paper, we show how to use hexahedral and prismatic elements for the efficient development of uniaxial anisotropic dispersive media (UPML) around an initial unstructured tetrahedral primal mesh using the generalized finite-difference (GFD) method. The initial mesh is assumed to have plane external boundaries. Taking into account the properties of the UPML medium, we build a special discrete Hodge in order to have a simple scheme for the time-domain solution of Maxwell's equations. The method is presented for the general 3-D case and numerical results show its efficiency.

The low recovery factor associated with carbonate reservoir and the availability of affordable gas has drawn new attention to a Water Alternative Gas (WAG) strategy to enhance hydrocarbon recovery for reservoir at early stage of production. The use of WAG allows to improve sweep efficiency limiting fingering and hydrocarbon trapping at macroscopic (Water injection - WI) and microscopic (pore) (Gas Injection - GI) level.The purpose of this work is twofold: present a new field development plan optimization formulation generalized for continuous and/or WAG injection and its suggested application to a real reservoir test case. Optimum field development plans for different strategies are presented and the advantages of a WAG strategy as opposite to a continuous water injection are further discussed. The proposed optimization problem covers well placement, control, schedule and gas lift under uncertainty. This problem is formulated as Mixed Integer Nonlinear Programing (MINLP) with non-linear constrains in order to take into account operational restrictions.A carbonate deep water reservoir at the early stages of its exploitation has been used to evaluate this new workflow. An in-house Particle Swarm Optimization (PSO) algorithm has been used to solve this nonlinear optimization problem. The use of a real field as test bench poses additional strength on the robustness of the formulation in presence of a large number of decision variables and constrains. Special care has been taken in the field development plan selection, taking into account the optimal solution as well as additional objectives and sentiments added by the decision makers which increase the realism of the final solution.The optima selected were compared to a WAG base case adjusted using standard commercial tools according to best practice. The results obtained for the current reservoir highlight key performances of the optimization algorithm applied to the WAG strategy. WAG increases the recovery factor and enhance an early production (paramount to improve the NPV according to the economic model) with respect to a continuous WI. The optimization of WAG cycles allows to find the best hydrocarbon production under the constrain imposed by the facilities and the gas availability. The optimum inter-well-distance and schedule results from the compromise between a strong sweep yet avoiding an early water breakthrough and reducing interference among producers. The well layout is therefore an output of the NPV optimization, the carbonate geological characterization and the operation risk. The optimal field development plan has shown a 35% of increment in NPV compared to the base case.This workflow presented and discussed presents an innovative and general formulation for the optimization of a WAG and/or continuous injection for real reservoir under uncertainty. The optimization algorithm allows fast convergence and accurate results. The extension to multiple objective functions and the inclusion of the human decision in the optimization process is also considered.

Generate a plausible geological model for a green-field as well as rank reservoirs with different level of associated uncertainty can be quite challenging. In this paper, we propose a new approach to generate predictive static models in the first stages of reservoir appraisal. At this stage, structural, sedimentary and petro physical information is very limited. The workflow proposed aim to use any available information whether coming from analogues reservoirs and/or field-data to reduce uncertainty. Initially the structural uncertainty is characterized and propagated generating an ensemble of plausible scenarios based on the perturbation of structural types. Subsequently, a novel formulation is presented to define a conceptual sedimentary (optimum number and percentage of the facies) and petro-physical model for the generation of the 3-D geological plausible scenarios. The methodology has been tested on the Brugge field benchmark modified as green-field. The proposed workflow yields a good approximation of the reservoir static and dynamic behavior (e.g. OOIP and NPV) compared with the original model. In addition, the sedimentary definition of the facies is well characterized with respect to the core interpretation from laboratory. The new workflow presented allows to generate plausible geological scenarios with limited amount of information, quantifying and propagating the uncertainty.

The economical performance of an oilfield operation is uncertain and highly influenced by strategic and operational decisions variables such as well placement, scheduling and control. Based on numerically intensive reservoir simula- tors, the evaluation of an extensive list of possible decisions across all possible realizations becomes computationally intractable and additional mathematical techniques are required. A common approach to dealing with this problem is the Response Surface Methodology (RSM) coupled with Design of Experiments (DoE) and sampling techniques. Existing approaches to construct surrogates depend on specific statistical/risk measures such as expected value and standard deviation. For example, in order to construct a surrogate for the standard deviation of NPV, one would compute the standard deviation associated with the simulation results over the selected geological realizations for each candidate production strategy and then fit a mathematical model to it. In this case, the idiosyncratic response of each geological realization with respect to the production strategy is lost, which may lead to a bad risk assessment and, consequently an inappropriate decision making process. In this paper, we propose Stochastic Response Surface Methodology (SRSM) to enhance the decision-making process over the determination of oil & gas production strate- gies while properly taking into consideration geological uncertainty. The SRSM does not depend on any pre-defined risk measure providing the necessary flexibility to evaluate the intrinsic risk-return trade-off associated with the economical performance of the reservoir. Our approach is based on the construction of surrogates for each geolog- ical realization selected by sampling procedure. We argue that constructing a different surrogate for each selected realization captures the idiosyncratic behavior of each representative geological setting and provides the flexibility of choosing any set of risk measures after the surrogate construction has been done. Based on the Brugge field, an SPE benchmark case study, we provide a numerical example to illustrate our methodology.

A proper quantification and propagation of geological and geophysical uncertainty during phase of technical and economic appraisal of a reservoir is of primary importance to evaluate the potential and the risks associated with a new prospect. In this paper we propose an innovative methodology to define a conceptual model for different levels of available information particularly appealing for cases with limited data – such as that of a green field. The proposed methodology aims to construct multiple reservoir realizations which are geologically consistent with the the available data and constrain, and capable to propagate the uncertainty inherent to a given target reservoir specifications. The uncertainty is characterized using analogous reservoirs which are, in the present case, evaluated from an in-house database. The methodology can be resumed in a sequence of well defined steps. First, based on the quality and availability of the information, an optimization problem is formulated and solved to ensure that the generated realizations obey the petro-physical property statistics of the target reservoir. Outputs of the optimization process are the facies proportions yielding to the conditional probability distributions of properties. The physical distribution of the petrophysical properties are then generated based on these inputs using multipoint geostatistical techniques. To correctly propagate the uncertainty in the reservoir model the sensitive controlling variables input to the geostatistical algorithm are defined within a variability range. The selection of the relevant input parameters can be selected based on sensitivity analysis techniques. The stochasticity proper of the geostatistical algorithms used for the property distribution requires special care for the correct application and interpretation of the problem reduction techniques. Static and dynamic objective functions have been used to evaluate the statistical distribution of the generated realizations. From a static point of view the reservoir volumetric has been ranked using the field original-oil-in-place. The dynamic response of the generated realizations upon a simple pattern has been implemented as measure of connectivity. As proof of concept the field-oil-production-rate for an inverted-five-spot pattern under voidage replacement was chosen as dynamic ranking. The methodology has been applied to the Brugge field benchmark which presents 104 geological realizations (Peters et al (2010), Peters et al (2009)). The benchmark was modified in order to simulate a green field, therefore only mean values for the petro-physical properties were extracted and used as input to the methodology whils the complete geological realizations results of the full-field-model (FFM) have been used for validation. A list of equiprobable analogous reservoirs has been identified from the database. Thereof, a number of conceptual models have been created respecting the key static parameters statistics defined. To validate the workflow first the predicted original-oil-in-place from the analogues and from the generated realizations respectively has been positively compared with the benchmark realizations. Due to the absence of structural uncertainty in the benchmark realizations the brute volume and the mean original-oil-in-place compute from the analogues have been set and a large number of geological realizations have been simulated. Results indicate that the trends obtained by the conceptual model for a given production strategy is in agreement with the FFM and the FFM results are located within the produced realizations. The computational time associated with the conceptual model, due to its simplicity, was also very attractive compared to the full model. Noteworthy, the methodology provides important insight on the sensitivity of the geological model to several uncertain static and dynamic parameter and configurations which can be the base for a risk analysis or to quantify the value of additional information.

Decision-making under uncertainty can be quite challenging, especially when complex numerical simulations are considered in the workflow and the decision has to be made relatively fast (e.g. in hours). This is the case when one needs to rank a given field portfolio within a limited budget and with acquisition constraints. If the ranking measure associated with each field is properly and rapidly evaluated, new prospect opportunities, that may lead to a favorable strategic position, can be readily identified.In this paper, we propose an efficient methodology for computing a ‘production potential’ measure that can be used to rank field portfolios in the presence of geological uncertainty, quantifying both uncertainty and risk propagation. Next, we briefly describe the basics of the method proposed. In the first place, uncertainty in sedimentary variability and flow behavior has to be characterized by a number of representative geological realizations. Sampling techniques are used to significantly reduce the number of realizations while preserving accuracy in the description and uncertainty propagation. Thereafter, multiple and varied field development plans, based on primary/secondary recovery mechanisms, are automatically generated whilst accounting for key parameters related to the number, drilling locations and drilling sequence of wells. In these plans the reservoir is clustered by areas with similar production/injection potential, and the well locations and drilling schedules are obtained accordingly. The well controls are determined through estimations of the field recovery factor. By means of experimental-design techniques a relative small number of field development plans are selected to capture the most significant production profiles. Each of these development plans is simulated for the realizations sampled previously, and the production potential measure [e.g. average net present value (NPV) over all sampled realizations] is computed for all the plans. The highest of these measures (i.e. the best development plan) can be used for ranking the field in the portfolio. Response surface procedures are considered to perform additional analysis computations within iterative optimization procedures. It is important to note that other statistics related to the exploitation potential (e.g. standard deviation of the NPV) can also be used to complement the ranking, thereby mitigating the decision-makers risk tolerance. The methodology has been tested on the Brugge field benchmark which presents 104 realizations of multiple geological parameters. The benchmark has been modified in order to simulate a green-field scenario. The ranking measure is the (discounted) NPV averaged over the 104 realizations. The proposed workflow yields a ranking measure of $5.43 billion, and the computational cost is around 1,900 simulations (performed in a parallel computing environment). This NPV is somewhat higher than those found for the Brugge benchmark (with similar modified settings) by other researchers. In order to validate the results, we performed more exhaustive checking using around 17,000 simulations, and the ranking measure found was $5.51 billion.The new workflow presented allows one to efficiently, and in a sufficiently accurate manner, support decision making in field portfolio evaluation. Fast reservoir performance evaluation engines open new prospect opportunities that, with traditional decision-making techniques, may be frequently lost.