Practical Implementation of Knowledge-Based Approaches for SAGD Production Analysis

Abstract

Quantitative appraisal of different operating areas and assessment of uncertainty due to reservoir heterogeneities are crucial elements in optimization of production and development strategies in oil sands operations. Although detailed compositional simulators are available for recovery performance evaluation for SAGD, the simulation process is usually deterministic and computationally demanding, and it not quite practical for real-time decision-making and forecasting. Data mining and machine learning algorithms provide efficient modeling alternatives, particularly when the underlying physical relationships between system variables are highly complex, non-linear, and possibly uncertain.In this study, a comprehensive training set encompassing SAGD field data compiled from numerous publicly-available sources is studied. Exploratory data analysis is carried out to interpret and extract relevant attributes describing characteristics associated with reservoir heterogeneities and operating constraints. Because of their ease of implementation and computational efficiency, knowledge-based techniques including artificial neural networks (ANN) are employed to facilitate SAGD production performance prediction. Predicting (input) variables including porosity, net-to-gross ratio, saturation, gross pay, normalized shale barrier thickness and distance to well pair, and initial production rate are formulated. Measures such as cumulative production over discrete time intervals are considered as prediction (output) variables. Data records that are comprised of both input and output variables are assembled; the network is trained using the data set to identify all significant patterns and relationships that exist between the input and the output variables. The model is subsequently validated using a cross-verification procedure, during which records that have been excluded at the training stage are presented to the model. This paper demonstrates that knowledge-based techniques can be implemented in a practical manner to analyze large amount of competitor data efficiently. The approach can be integrated directly into most existing reservoir management routines. It can also be readily updated when new information has become available. Given that robust reservoir management and real-time decision-making are major challenges faced by the industry, the data-driven models presented in this paper has great potential to be applied in other recovery projects such as solvent-aided steam injection.

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

“…One can summarize variability among these 120 cases as follows: (1) top-water thickness is between 0 and 100 m (h Dwt ¼ 0 to 3.333) (a zero value refers to the absence of lean zone); bottomwater thickness is between 0 and 300 m (h Dwb ¼ 0 to 10); (2) oil saturation in the lean zones is between 0 and 0.95; (3) length of the shale barriers is between 3 and 30 m (l Ds ¼ 0.06 to 0.59); and (4) total number of shale barriers is between 2 and 15 (equivalent to 0 to 15% probability of shale occurrence). The total data set (m) is partitioned into two parts: (1) k samples are designated for training and validation of the BPNN model; an n-fold cross validation is implemented to identify the optimal network architecture (Ma et al 2014); and (2) the remaining (m-k) samples are assigned for final testing in a prediction mode with the previously trained network parameters. In this study, the entire data set is subdivided into 70 cases for training, 20 cases for validating, and the remaining 30 cases for testing.…”

“…The implementation of SI was tested recently with an SAGD field data set compiled from numerous publicly available sources (Ma et al 2014). In addition to the typical petrophysical variables including porosity, net-to-gross ratio, saturation, and gross pay, it was concluded that the dimensionless SI, which is defined from logs as the shale-barrier thickness divided by distance to well pair, is pertinent in describing the characteristics associated with reservoir heterogeneities and facilitating SAGD-production-performance prediction.…”

“…This procedure is to mimic a practical field data set consisting of petrophysical logs available only at vertical delineation wells (Ma et al 2014). The intentions are to (1) assess the performance of ANN models when only limited petrophysical information is available; and to (2) see the potential implications of ignoring interwell heterogeneities.…”

“…They also become the basis for formulating numerous input attributes in subsequent data-driven modeling: artificial neural networks (ANNs) or fuzzy-based models can be used to construct data-driven models that relate input (predicting) attributes describing reservoir/fluid/operational variables to output (target) attributes representing recovery performance (Amirian et al 2013;Ma et al 2014).…”

Correlating Stochastically Distributed Reservoir Heterogeneities with Steam-Assisted Gravity Drainage Production

“…ANN has also been used to analyze heavy oil recovery in a number of previous works (Queipo et al, 2002;Ahmadloo et al, 2010;Karambeigi et al, 2011;Popa et al, 2011;Zerafat et al, 2011;Popa and Patel, 2012;Amirian et al, 2015;Ma et al, 2016). In particular, ANN was implemented in Ma et al (2015) recently to construct a series of data-driven models from an actual SAGD field dataset.…”