Ensemble SVM for characterisation of crude oil viscosity

Oloso, Munirudeen A.; Hassan, Mohamed; Bader-El-Den, Mohamed; Buick, James

doi:10.1007/s13202-017-0355-x

Cited by 14 publications

(5 citation statements)

References 55 publications

(35 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Regarding what was said, this has graveled the way for modification and adoption of already existing empirical correlations over a period of time. Furthermore, to overcome these challenges, some machine learning (ML) and artificial intelligence techniques (AI) have also been used to improve the prediction of oil viscosity, including radial basis function neural network (RBFNN) [30], artificial neural network (ANN) [68][69][70][71][72][73], functional networks (FN) [46,71], genetic algorithm (GA) [74], support vector machine (SVM) [75], and group method of data handling (GMDH) [76] and Ensemble models [77]. The literature argues that the lowest average absolute relative error can be achieved when viscosity is predicted by AI models and the highest correlation coefficient as compared to existing empirical correlations [70].…”

Section: Introductionmentioning

confidence: 99%

Prediction of Dead Oil Viscosity: Machine Learning vs. Classical Correlations

et al. 2021

View full text Add to dashboard Cite

Dead oil viscosity is a critical parameter to solve numerous reservoir engineering problems and one of the most unreliable properties to predict with classical black oil correlations. Determination of dead oil viscosity by experiments is expensive and time-consuming, which means developing an accurate and quick prediction model is required. This paper implements six machine learning models: random forest (RF), lightgbm, XGBoost, multilayer perceptron (MLP) neural network, stochastic real-valued (SRV) and SuperLearner to predict dead oil viscosity. More than 2000 pressure–volume–temperature (PVT) data were used for developing and testing these models. A huge range of viscosity data were used, from light intermediate to heavy oil. In this study, we give insight into the performance of different functional forms that have been used in the literature to formulate dead oil viscosity. The results show that the functional form f(γAPI,T), has the best performance, and additional correlating parameters might be unnecessary. Furthermore, SuperLearner outperformed other machine learning (ML) algorithms as well as common correlations that are based on the metric analysis. The SuperLearner model can potentially replace the empirical models for viscosity predictions on a wide range of viscosities (any oil type). Ultimately, the proposed model is capable of simulating the true physical trend of the dead oil viscosity with variations of oil API gravity, temperature and shear rate.

show abstract

Section: Introductionmentioning

confidence: 99%

Prediction of Dead Oil Viscosity: Machine Learning vs. Classical Correlations

et al. 2021

View full text Add to dashboard Cite

show abstract

“…In [16] of viscosity the developed ensemble support vector regression models could potentially replace the empirical correlation for viscosity prediction. The developed ensemble support vector regression models for viscosity prediction could potentially replace the empirical correlation, as the improved predictions of viscosity.…”

Section: Related Workmentioning

confidence: 99%

“…Since these laboratory experiments is costly and much time is required, alternative approaches are using. These approaches used in the prediction of oil viscosity in the available literature [15][16][17] are classified into three main categories:…”

Section: Introductionmentioning

confidence: 99%

Evaluation of Oil Viscosity Based Various Empirical Correlations for Azerbaijan Crude Oils

İmamverdiyev¹,

Hajirahimova²

2019

IJITCS

View full text Add to dashboard Cite

In the oil industry, the evaluation of oil viscosity is one of the important issues. Generally, the viscosity of crude oil depends on pressure and temperature. In this study, the prediction issue of oil viscosity has been viewed applying empirical correlations as Beggs-Robinson, Labedi, modified Kartoatmodjo, Elsharkawy and Alikhan, Al-Khafaji. Original field data reports have been obtained from Guneshli oil field of Azerbaijan sector of Caspian Basin. The correlation models used in the evaluation of viscosity of Azerbaijan oil have been implemented in the Python software environment. The obtained values on empirical correlations have been compared to experimental data obtained from Guneshli oil field. Statistical analysis in terms of percent absolute deviation (% AD) and the percent absolute average deviation (% AAD), mean absolute error (% MAE), correlation coefficient (% 2 R ), root mean square error (% RMSE) are used to subject the evaluation of the viscosity correlations. According to statistical analysis, it has been known that the Beggs-Robinson model has shown the lowest value on AAD (10.5614%), MAE (12.4427 %), RMSE (20.0853 %). The Labedi model has presented the worst result on every four criterions. Even though the Elsharkawy-Alikhan model has presented the highest result (99.9272%) on correlation coefficient, in the evaluation of viscosity of Azerbaijan crude oil, the Beggs-Robinson model can be considered more acceptable.

show abstract

“…To predict crude oil pressure volume temperature (PVT) properties, Al-Marhoun et al [8] and Nagi et al [9] proposed support vector machines, and Makrufa et al [10] used ANN. Oloso et al [11] developed an ensemble ML model for the prediction of dead oil, saturated and undersaturated viscosities. Recently, Gonzalez et al [12] compared popular ML algorithms and evaluated their potential to develop state-ofthe-art models to predict spark-ignition fuel properties.…”

Section: Introductionmentioning

confidence: 99%

Machine Learning Applied to Predict Key Petroleum Crude Oil Constituents

Dhankar,

Sharma,

Mohanta

et al. 2023

Chem Eng & Technol

View full text Add to dashboard Cite

Sulfur compounds are the most important inorganic constituents of petroleum and require to be estimated beforehand because of their corrosive nature and other processing anomalies during crude oil processing. Paraffins, naphthene, and aromatics form the bulk of crude oil. Machine learning (ML) predictions of these constituents were made by training the ML model with a diverse industrial data set of 515 oils. The XGBoost model gave an excellent R2 in the range 0.88–0.99 for the bulk compounds. R2 for sulfur was in the modest range of 0.45–0.6, which improved significantly to 0.8 for additional inputs. ML applicability was thereby found to depend on the nature of the constituent. This work furthers ML‐based predictions, with the incentive of reducing expensive spectroscopic analytical methods.

show abstract

Ensemble SVM for characterisation of crude oil viscosity

Cited by 14 publications

References 55 publications

Prediction of Dead Oil Viscosity: Machine Learning vs. Classical Correlations

Prediction of Dead Oil Viscosity: Machine Learning vs. Classical Correlations

Evaluation of Oil Viscosity Based Various Empirical Correlations for Azerbaijan Crude Oils

Machine Learning Applied to Predict Key Petroleum Crude Oil Constituents

Contact Info

Product

Resources

About