Improved Oil Viscosity Characterization by Low-Field NMR Using Feature Engineering and Supervised Learning Algorithms

Abstract

Conventional methods for determining and monitoring the viscosity of oils are time-consuming, expensive, and in some instances, technically unfeasible. These limitations can be avoided using low-field nuclear magnetic resonance (LF-NMR) relaxometry. However, due to the chemical dissimilarity of oils and various temperatures these oils are exposed to, as well as LF-NMR equipment limitations, the commonly used models fail to perform at a satisfactory level, making them impractical for use in heavy oil and bitumen reservoirs and in environments with large temperature oscillations (e.g., mechanical systems). We present a framework that combines supervised learning algorithms with domain knowledge for synthesizing new features to improve model forecasts using only one NMR parameterT 2 geometric mean. Two principal methods were considered, support vector regression (SVR) and gradient boosted trees (GBRT). Models were trained using the experimental data from our previous studies and literature data combining conventional oils, heavy oils, and bitumens from various reservoirs in Canada and United States. The models’ performance was compared against four other intelligent algorithms and four well-known empirical NMR models against which the SVR- and GBRT-based models achieved the highest statistical scores. These two models can be used for oil viscosity prediction in conventional and heavy oil reservoirs with a wide range of oil viscosities and in situations where high precision is needed, such as in the determination of viscosity of petroleum distillates or for monitoring of oil viscosity in mechanical systems. The proposed framework can also be applied to determine other physicochemical properties of oils by LF-NMR, where the application of supervised learning is usually impractical due to the limited volume of experimental data.

Application of time domain nuclear magnetic resonance (TD‐NMR) for study of the distillation curve of petroleum

“…The transverse relaxation time (T 2 ) provides information about both chemical composition and physicochemical M ac (%) properties of crude oils. [4,[28][29][30] Oils with small T 2 values have a pronounced spin-spin interaction, suggesting a higher viscosity, which in turn implies the presence of heavier and long chain chemical compounds. [30,31] In this context, the T 2 values can provide molecular information from crude oils and their distillates.…”

Comprehensive Review of Hydrogen Leakage in Relation to Fuel Cell Vehicles and Hydrogen Refueling Stations: Status, Challenges, and Future Prospects

“…In the data processing phase, for known leak events, the level of the leak is annotated in the corresponding data segments to serve as a reference for model training. 213 The process of establishing and training the model for leak prediction is similar to that for leak localization. Finally, this trained model is used to predict the level of leaks.…”

“…Alternatively, the leak level can be predicted based on the concentration of diffused hydrogen, in conjunction with data such as pressure and temperature from relevant equipment. In the data processing phase, for known leak events, the level of the leak is annotated in the corresponding data segments to serve as a reference for model training . The process of establishing and training the model for leak prediction is similar to that for leak localization.…”

Debris Flow Scale Prediction Based on Correlation Analysis and Improved Support Vector Machine

“…To analyze the overall performance of the prediction model, this paper selects root mean square error (RMSE), average absolute error (MAE), and coefficient of determination (R 2 ) to evaluate the above four prediction models [28].…”