**Source**

**Conference**

**Theme**

**Author**

- Avansi, Guilherme Daniel (1)
- Caiado, Camila (1)
- Davolio, Alessandra (1)
- Ferreira, Carla J. (1)
**Goldstein, Michael (3)**- Heincke, Björn (1)
- Hobbs, Richard W. (1)
- Jegen, Marion (1)
- Maschio, Célio (1)
- Moorkamp, Max (1)
- Nandi Formentin, Helena (1)
- Roberts, Alan W. (1)
- Schiozer, Denis J. (1)
- Schiozer, Denis José (1)
- Vernon, Ian (2)

**Concept Tag**

- application (1)
- Artificial Intelligence (3)
- Bayesian (1)
- calibration (1)
- canonical correlation (1)
- canonical correlation technique (1)
- canonical function (1)
- coefficient (1)
- correlation (1)
- cycle (1)
- dataset (1)
- discrepancy (1)
- distribution (1)
- emulation (1)
- emulator (3)
- equation (1)
- evaluation (1)
- frequency (1)
- function (1)
- gravity (1)
- grid cell (1)
- historical data (1)
- implausibility (2)
- implausibility measure (1)
- information (1)
- Information Index (1)
- input parameter (1)
- input space (1)
- joint inversion (1)
- layer (1)
- Lyr (1)
- machine learning (3)
- management and information (1)
- model (1)
- plausible model (1)
- procedure (1)
- quantity (1)
- reduction (1)
- Reservoir Characterization (2)
- reservoir description and dynamics (1)
- reservoir simulation (2)
- resistivity (1)
- Scenario (2)
- search space (1)
- seismic data (1)
- seismic processing and interpretation (1)
- simulator (2)
- space (1)
- Thickness (1)
- uncertainty reduction (2)
- Upstream Oil & Gas (3)
- water saturation (1)
- water saturation map (1)

**File Type**

Nandi Formentin, Helena (Durham University and University of Campinas) | Vernon, Ian (Durham University) | Avansi, Guilherme Daniel (University of Campinas) | Caiado, Camila (Durham University) | Maschio, Célio (University of Campinas) | Goldstein, Michael (Durham University) | Schiozer, Denis José (University of Campinas)

Reservoir simulation models incorporate physical laws and reservoir characteristics. They represent our understanding of sub-surface structures based on the available information. Emulators are statistical representations of simulation models, offering fast evaluations of a sufficiently large number of reservoir scenarios, to enable a full uncertainty analysis. Bayesian History Matching (BHM) aims to find the range of reservoir scenarios that are consistent with the historical data, in order to provide comprehensive evaluation of reservoir performance and consistent, unbiased predictions incorporating realistic levels of uncertainty, required for full asset management. We describe a systematic approach for uncertainty quantification that combines reservoir simulation and emulation techniques within a coherent Bayesian framework for uncertainty quantification.

Our systematic procedure is an alternative and more rigorous tool for reservoir studies dealing with probabilistic uncertainty reduction. It comprises the design of sets of simulation scenarios to facilitate the construction of emulators, capable of accurately mimicking the simulator with known levels of uncertainty. Emulators can be used to accelerate the steps requiring large numbers of evaluations of the input space in order to be valid from a statistical perspective. Via implausibility measures, we compare emulated outputs with historical data incorporating major process uncertainties. Then, we iteratively identify regions of input parameter space unlikely to provide acceptable matches, performing more runs and reconstructing more accurate emulators at each wave, an approach that benefits from several efficiency improvements. We provide a workflow covering each stage of this procedure.

The procedure was applied to reduce uncertainty in a complex reservoir case study with 25 injection and production wells. The case study contains 26 uncertain attributes representing petrophysical, rock-fluid and fluid properties. We selected phases of evaluation considering specific events during the reservoir management, improving the efficiency of simulation resources use. We identified and addressed data patterns untracked in previous studies: simulator targets, ^{−11}%). The systematic procedure showed that uncertainty reduction using iterative Bayesian History Matching has the potential to be used in a large class of reservoir studies with a high number of uncertain parameters.

We advance the applicability of Bayesian History Matching for reservoir studies with four deliveries: (a) a general workflow for systematic BHM, (b) the use of phases to progressively evaluate the historical data; and (c) the integration of two-class emulators in the BHM formulation. Finally, we demonstrate the internal discrepancy as a source of error in the reservoir model.

application, Artificial Intelligence, Bayesian, calibration, discrepancy, emulator, evaluation, historical data, implausibility, implausibility measure, information, Information Index, machine learning, procedure, quantity, reservoir simulation, Scenario, search space, simulator, uncertainty reduction, Upstream Oil & Gas

SPE Disciplines:

Technology: Information Technology > Artificial Intelligence > Machine Learning > Statistical Learning (1.00)

In petroleum engineering, simulation models are used in reservoir performance prediction and in the decision-making process. These models are complex systems, typically characterized by a vast number of input parameters. Typically, the physical state of the reservoir is highly uncertain and, thus, the appropriate parameters of the input choices are also highly uncertain. 4D seismic data can reduce significantly the uncertainty of the reservoir because it has a high area resolution, as opposed to the observed well rates and pressure. However, two main challenges are faced to calibrate the simulation model using 4D seismic data. The process can be time consuming because most models go through a series of iterations before being considered sufficiently accurate to give an adequate representation of the physical system. The consideration of 4D seismic data as an observed parameter in the form of maps would lead to an unfeasibly large number of variables to be matched. To overcome such issues, the construction of an emulator that represents the simulation model and the use of the canonical correlation technique to incorporate 4D seismic data can be used. The present study constructed a stochastic representation of the computer model called an emulator to quantify the reduction in the parameter input space. 4D seismic data was incorporated in the procedure through the canonical correlation technique. The water saturation map derived from seismic data was converted into seven canonical functions. Such functions represent the observable characteristics to be matched in the uncertainty reduction process. A high number of evaluations was necessary to identify the range of input parameters whose outputs matched the historical data (4D seismic data). The large number of evaluations justifies the use of an emulator and the reduction of uncertainties with areal characteristics shows that 4D seismic data was successfully incorporated. The emulator methodology represents a powerful tool in the analysis of complex physical problems such as history matching. The incorporation of 4D seismic data as an observable output to be matched leads to a difficult problem to be solved. However, the canonical correlation permitted a successful incorporation of such data into the problem.

Artificial Intelligence, canonical correlation, canonical correlation technique, canonical function, correlation, emulator, grid cell, implausibility, input parameter, input space, machine learning, reduction, Reservoir Characterization, reservoir simulation, Scenario, seismic data, uncertainty reduction, Upstream Oil & Gas, water saturation, water saturation map

Oilfield Places: Africa > Angola > Angola Offshore > Lower Congo Basin > Block 17 > Girassol Field (0.99)

SPE Disciplines:

Technology: Information Technology > Artificial Intelligence > Machine Learning > Statistical Learning (0.70)

The main obstacle to carrying out a full Bayesian search over all plausible model space is computational expense, because the time taken to run one forward modelling step for a 3D model can be of order an hour or more. The purpose of an emulator is to use a computationally cheap function to approximate the output of the model simulator for a given set of input parameters. The strategy is to train the emulator using a number of the full forward simulator runs. This is similar to training a neural network. However, in training the emulator we also calibrate how uncertain the emulator is. This is important, particularly because we aim to use it in order to reject implausible model space, and a lack of knowledge about the uncertainty of the emulator would make this task impossible. Our approach is to build an emulator for each of the datasets; travel-time vs offset (

Artificial Intelligence, coefficient, cycle, dataset, distribution, emulation, emulator, equation, frequency, function, gravity, joint inversion, layer, Lyr, machine learning, management and information, model, plausible model, Reservoir Characterization, reservoir description and dynamics, resistivity, seismic processing and interpretation, simulator, space, Thickness, Upstream Oil & Gas

SPE Disciplines:

Technology: Information Technology > Artificial Intelligence > Machine Learning > Statistical Learning (1.00)

Thank you!