Data-Driven Modeling Approach for Recovery Performance Prediction in SAGD Operations

Dzurman, Peter; Leung, Juliana Y.; Zanon, Stefan; Amirian, Ehsan

doi:10.2118/165557-ms

Cited by 26 publications

(8 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Canada and Venezuela have the greatest amounts (1.7 trillion bbl and 1.8 trillion bbl, respectively) of bitumen and heavy-oil reserves (Burton et al 2005;Nasr and Ayodele 2006). More than 60% of the total natural bitumen resources are in the province of Alberta (Attanasi and Meyer 2007).…”

Section: Introductionmentioning

confidence: 99%

Characterizing the Effects of Lean Zones and Shale Distribution in Steam-Assisted-Gravity-Drainage Recovery Performance

Wang

Leung

2015

SPE Reservoir Evaluation &Amp; Engineering

Self Cite

View full text Add to dashboard Cite

Performance of steam-assisted gravity drainage (SAGD) is influenced significantly by the distributions of lean zones and shale barriers, which tend to impede the vertical growth and lateral spread of a steam chamber. Previous literature has partially addressed their effects on SAGD performance; however, a comprehensive and systematic investigation of the heterogeneous distribution (location, continuity, size, saturation, and proportions) of shale barriers and lean zones is still lacking.In this study, numerical simulations are used to model the SAGD process. Capillarity and relative permeability effects, which were ignored in many previous simulation studies, are incorporated to model bypassed oil. Numerous ranking schemes are formulated to analyze various aspects of SAGD performance. A detailed sensitivity analysis is performed by varying the location, continuity, size, proportions, and saturation of these heterogeneous features. Lean zones and shale lenses (imbedded in a region of degraded rock properties) with different sizes and degrees of continuity are placed in areas above the injector, below the producer, or in between the well pair. It is noted that among numerous parameters that influence the ultimate recovery, remaining bypassed oil, chamber advancement, and heat loss, continuity and position of these features in relation to the well pair play a particularly crucial role. Neural network modeling is subsequently used for constructing data-driven models to identify and propose a set of input variables for correlating relevant parameters or measures, which are descriptive of the heterogeneity and properties of the shale barriers and lean zones, to recovery and ranking results.This work provides a guideline for assessing the impacts of reservoir and saturation heterogeneities on SAGD performance. A set of input variables and parameters that have significant impacts on the ensuing recovery response is identified. One can define readily the proposed set of variables from well logs and apply immediately in data-driven models with field data and scaleup analysis of experimental models to assist field-operation design and evaluation. One can also extend the approach presented in this paper to analyze other solvent-assisted SAGD processes.

show abstract

Section: Introductionmentioning

confidence: 99%

Characterizing the Effects of Lean Zones and Shale Distribution in Steam-Assisted-Gravity-Drainage Recovery Performance

Wang

Leung

2015

SPE Reservoir Evaluation &Amp; Engineering

Self Cite

View full text Add to dashboard Cite

show abstract

“…In other words, ANNs are computational models developed on the principle of the biological nervous system [64]. According to [65], it is a virtual intelligence or machine learning technique which is useful for pattern recognition and prediction of a complicated non-linear relationship between input and output.…”

Section: Kriging Modelmentioning

confidence: 99%

Development of proxy models for petroleum reservoir simulation: a systematic literature review and state-of-the-art

Silva¹,

Avansi²,

Schiozer³

2020

IJAERS

View full text Add to dashboard Cite

Proxy models are derived mathematical functions developed as substitutes for reservoir flow simulators. Several types of proxy models are reported in the literature, for instance, response surface models, surrogate models, or metamodels. These models are fast methods, recommended for their efficient response time to approximate model responses and, therefore, useful in the decision-making process related to reservoir management. These studies focus on modelling a limited set of factors, applications, and case studies of any technique. A systematic literature review (SLR) is performed to gather the aspects prompting the modelling of proxy models in the literature and state-of-the-art. For this, a set of search keywords with appropriate string were utilised to extract the most important studies that satisfied all the criteria defined and classified under journal and conference paper categories. The papers were condensed after removing redundancy, repetition and similarity through a sequential and iterative process. From the analysis carried out, several gaps were identified, especially during the proxy model construction. Proxy models have already been discussed in petroleum engineering as a representation of the real system of reservoir flow simulator software. However, the proxy model response is faster but has yet to establish the issues of uncertainty in the outputs. There is a need for the integration of fast methods and reservoir simulators which can improve and accelerate results within acceptance criteria and accuracy in decisionmaking processes related to reservoir management.

show abstract

“…Several areas of application included reservoir characterization (Artun and Mohaghegh 2011;Raeesi et al 2012;Alizadeh et al 2012), candidate well selection for hydraulic fracturing treatments (Mohaghegh et al 1996), well-placement/trajectory optimization (Centilmen et al 1999;Doraisamy et al 2000;Johnson and Rogers 2001;Guyaguler and Horne 2000;Yeten et al 2003;Gokcesu et al 2005;Mohaghegh et al 2006), screening and optimization of secondary/enhanced oil recovery processes (Ayala and Ertekin 2005;Patel et al 2005;Demiryurek et al 2008;Artun et al 2010Artun et al , 2012Parada and Ertekin 2012;Amirian et al 2013), history matching (Cullick et al 2006Silva et al 2007;Zhao et al 2015), reservoir modeling, monitoring and management (Zangl et al 2006;Mohaghegh 2011;Mohaghegh et al 2014;Zhao et al 2015;Kalantari-Dhaghi et al 2015;Esmaili and Mohaghegh 2016). Most of these problems presented in the literature are based on development of artificial neural network (ANN) based proxy models that can accurately mimic reservoir models within a reasonable amount of accuracy and computational efficiency.…”

Section: Data-driven Modeling Approach Using Artificial Neural Networkmentioning

confidence: 99%

Characterizing Reservoir Connectivity and Forecasting Waterflood Performance Using Data-Driven and Reduced-Physics Models

Artun¹

2016

All Days

View full text Add to dashboard Cite

Closed-loop reservoir management requires representative models that can be updated quickly. Such kind of models that can be used for characterization of reservoir connectivity would greatly contribute to decisions in waterflooding operations. In heterogeneous reservoirs that are characterized with thief-zones/ fractures or flow-barriers, that may enhance or prevent connectivity, it becomes challenging to construct geological models with both reasonable accuracy and computational efficiency. In this study, practical and computationally-efficient methods are presented to first accurately characterize the connectivity between producer and injector wells, and then to forecast the waterflood performance. Two classes of modeling approaches are evaluated in terms of their applicability to characterize the connectivity: 1) Data-driven modeling -artificial neural networks (ANN), in which there is not any presumed functional relationship and connectivity is derived from the weights on connection links of neural networks; 2) Reduced-physics modeling -capacitance-resistance models (CRM), in which physics are incorporated with certain assumptions and connectivity is derived from the fractional flow parameter in the formulation. In both approaches; a producer-based control volume is considered to focus on producers and their connectivity to different injectors, and both models are trained with injection/production histories. Connectivity values derived from these methods are fed into a short-term, pattern-based forecasting methodology, which quantified ultimate recoveries at the desired oil-cut values. Two synthetic reservoir models are used to test the applicability of aforementioned methods; which are characterized with high-permeability streaks and flow-barriers. These key features that affect connectivity are identified by both modeling approaches with reasonable accuracies, but with varying interpretations of fluid movement in the reservoir. Pattern-based forecasting methodology resulted in ultimate recovery predictions very close to the numerical-model based results, with errors within 3%. The workflow presented is quick and reliable to be implemented in closed-loop reservoir management protocols, which requires models that can be updated easily with recent data, and can be used in decision making processes with confidence.

show abstract

Data-Driven Modeling Approach for Recovery Performance Prediction in SAGD Operations

Cited by 26 publications

References 15 publications

Characterizing the Effects of Lean Zones and Shale Distribution in Steam-Assisted-Gravity-Drainage Recovery Performance

Characterizing the Effects of Lean Zones and Shale Distribution in Steam-Assisted-Gravity-Drainage Recovery Performance

Development of proxy models for petroleum reservoir simulation: a systematic literature review and state-of-the-art

Characterizing Reservoir Connectivity and Forecasting Waterflood Performance Using Data-Driven and Reduced-Physics Models

Contact Info

Product

Resources

About