4.5 Article

Machine learning augmented dead oil viscosity model for all oil types

Journal

Publisher

ELSEVIER
DOI: 10.1016/j.petrol.2020.107603

Keywords

Viscosity; Aspahletene; Machine learning; Heavy oil; Light oil

Ask authors/readers for more resources

Dead oil viscosity is one of the most unreliable properties to predict with classical black oil correlations. This results mostly from the large effect that oil type has on viscosity. Two dead oil samples with identical APIs (https://www.eia.gov/today), (https://en.wikipedia.org/a) and can have even an order of magnitude difference in viscosity at the same temperature (Dindoruk and Christman, 2004). In this work, we tried to limit this spread to a certain degree by incorporating a parameter such as MW to capture additional information for the character of the oil. Limitations of the classical black oil correlations became even more prominent when a wide spectrum of viscosity values coupled with a wide range of temperatures are considered. Given the constraints of limited input variables, the problem becomes particularly challenging for heavy-extra heavy oils with high asphaltene content (Sinha et al., 2019), where prediction errors could easily be as high as a couple of log-cycles. Even though there are several viscosity correlations available in the industry (Ali, 2003)- (Bergman and Sutton, 2009), (De Ghetto and Villa, 1994), (Dindoruk and Christman, 2004), (Mehrotra, 1992), (Motahhari et al., 2013), (Naseri et al., 2005), (Lindeloff et al., 2003), (Lohrenz et al., 1964), (Pedersen and Fredenslund, 1987) (Petrosky and Farshad, 1995), (Peng and Robinson, 1976), (Standing, 1977), (Teja and Rice, 1981), (Yarranton et al., 2013) most of those correlations are only applicable to the oil samples belonging to specific geographical regions and/or for structurally similar oils, because of the inherent bias in the training datasets used in the development of correlations. Therefore, they are predictively valid for a relatively narrower range of oils and/or viscosity. In this work, we considered a very wide range of oils (6 degrees API to 50 degrees API). Therefore, producing two easy to use viscosity correlations for API (https://www.eia.gov/today), (https://en.wikipedia.org/a) gravities above and below 20 degrees API that can readily predict the viscosity at any desired temperature within an extended temperature range (15 degrees C to 160 degrees C). Also, the two sets of correlations were kept compatible in a region in the vicinity of switching points (20 API and 36 cp) so that they will have a proper transition from one branch to another. While the range of fluid properties is very wide, we were able to keep the input parameters to a minimum in terms of defining the character of the fluid (molecular weight and specific gravity). We demonstrated that the proposed correlation along with the given methodology performs much better than the leading correlations with the similar input proxies published in the literature for a wide range of viscosities (0.42 cp to 860, 000 cp). The use cases for the proposed correlation can be divided into three parts: 1) Prediction of the dead oil viscosity with limited input data, 2) use of limited in-hand viscosity or reference viscosity data to generate viscosities for the conditions that are hard to perform accurate experimentation or simply not having the physical sample in hand to do additional experiments(for example, for thermal recovery processes where viscosity is needed at elevated temperatures, or to construct the lift curves and/or tables for pipeline flow at lower temperatures)) and 3) it can be used to check the consistency and the quality of the existing data. In addition to classical correlation development efforts using known but limited physical control parameters, we have also attempted to model the viscosity with various machine learning methods K-Nearest Neighbor (KNN) (Cover and Hart, 1967) and Kernel-based Support Vector Machine (KSVM)) (https://docs.oracle.com/c, 2835), (Suykens and Vandewalle, 1999) and compared the outcome to each other and as well as against the proposed correlation. Based on calculated statistical parameters and cross-plots, the proposed correlation performed better than the other leading viscosity correlations (Bergman and Sutton, 2009), (Lindeloff et al., 2003), (Lohrenz et al., 1964), (Pedersen and Fredenslund, 1987), (Pedersen et al., 1984), (Petrosky and Farshad, 1995), (Standing, 1977), (Teja and Rice, 1981), (Yarranton et al., 2013) and as well as the selected supervised machine learning regression principles (https://en.wikipedia.org/b), (https://machinelearningma) such as KNN and KSVM. The subject correlation also helps in improving the accuracy as well as guiding the performance of these otherwise Blackbox machine learning principles as it can fill the gaps in the data especially in the context of extending the tuned viscosities based on a single point measurement (reference viscosity) in temperature domain. Furthermore, we also explain how it can be combined with Sinha et al. (Sinha et al., 2019) relative viscosity correlation to include the impact of asphaltene concentration to be able to estimate vertical or areal viscosity variations which can also ultimately help to improve the mobility cut-off predictions of the asphaltene/tar mat zones or heavier fluids.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.5
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available