Improved Correlations for Heavy-Oil Viscosity Prediction with NMR

Sandor, Magdalena; Cheng, Yuesheng; Chen, Songhua

doi:10.1016/j.petrol.2016.09.004

Cited by 19 publications

(15 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Empirical NMR Models. The performances of GBRT-FE and SVR-FE models are compared with four well-known empirical NMR viscosity models based on T 2gm , TE, and T. These models were developed by Straley et al, 37 Nicot et al, 47 Cheng et al, 6 and Sandor et al 5 Previous research showed that tuning by nonlinear least squares (NLS) improves the performance of empirical models. 25 However, the dataset in the present study has long-tailed distribution with many outliers at higher viscosities, which dominate the sum of squares minimization, thus ultimately leading to erroneous model fit and misleading statistical scores.…”

Section: Resultsmentioning

confidence: 99%

“…19 Conversely, in petrophysical studies of unconventional oil reservoirs, the viscosity of the oil may vary up to 3 orders of magnitude, meaning that flow dynamics through the pore space will vary drastically. 5,20 In this sense, the spatial variation of oil viscosity influences not only the calculation of recoverable reserves but also the technical aspects of the recovery scheme, selection of enhanced oil recovery (EOR) methods, and the way the numerical simulation of the reservoirs are performed. Studies have shown that LF-NMR tools can help resolve this problem using the empirical NMR viscosity models, which can account for chemical complexity and a wide span of temperatures with the help of NMR-derived parameters such as the relative hydrogen index (RHI).…”

Section: Introductionmentioning

confidence: 99%

“…In recent years, the wave of innovation has led to the application of low-field nuclear magnetic resonance (LF-NMR) tools for the characterization of hydrogen-bearing liquids due to their ability to convey series of contactless, noninvasive experiments rapidly. Although this technology has been proved to be a viable tool for observing differences in variable viscosity oils, numerous constraints arise as a consequence of not only the embedded chemical complexity of oils and limitations of LF-NMR devices but also from analytical tools and models used for the interpretation of experimental results. − Since the former two are technologically challenging to change, one can attempt to improve the analytical tools and frameworks using new mathematical approaches. In such circumstances, the supervised learning (SL) methods have been proven to be useful in developing more reliable mathematical models in many relevant fields such as fuel processing, petrophysical studies of porous mediums, and oil viscosity monitoring equipment in mechanical systems. − …”

Section: Introductionmentioning

confidence: 99%

“…Conversely, in petrophysical studies of unconventional oil reservoirs, the viscosity of the oil may vary up to 3 orders of magnitude, meaning that flow dynamics through the pore space will vary drastically. , In this sense, the spatial variation of oil viscosity influences not only the calculation of recoverable reserves but also the technical aspects of the recovery scheme, selection of enhanced oil recovery (EOR) methods, and the way the numerical simulation of the reservoirs are performed. Studies have shown that LF-NMR tools can help resolve this problem using the empirical NMR viscosity models, which can account for chemical complexity and a wide span of temperatures with the help of NMR-derived parameters such as the relative hydrogen index (RHI). , However, the determination of RHI requires a recovery of the representative oil sample from the given formation, preferably with the preserved gas content, for subsequent laboratory measurements, which is often a technically challenging and expensive task. , Moreover, oil saturation volumes must be determined independently, which is necessary for normalization of measured oil NMR response by the amplitude of an equal quantity of water .…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

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

et al. 2020

View full text Add to dashboard Cite

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.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

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

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Their main advantage over the other ones lies in the ability to carry out continuous measurements in the ease of their automation. However, the existing types of rotational viscometers have a limited lower value of the measured viscosity and a high measurement error due to the inability to take into account the flow of the fluid at the end of the movable cylinder [10][11][12][13][14].…”

Section: Introductionmentioning

confidence: 99%

Method and instruments to measure dynamic viscosity of oil products in pipeline

Belogusev

Kozlov

Егоров

et al. 2018

Engineering for Rural Development

View full text Add to dashboard Cite

Abstract. The purpose of this paper is to develop a new method to determine the dynamic viscosity of oil products in a pipeline. Its main difference in comparison with the existing methods is that the viscosity can be determined directly in the pipeline, which allows excluding the human factor that occurs, when using the existing methods and measuring instruments, and improving the accuracy of measuring the dynamic viscosity of oil products due to taking into account the cavitation processes inside a pipeline. The gist of a new method lies in the fact that the rotating element -an ellipse that has the form of a spindle -is placed inside a pipeline in a horizontal position and set in motion by an asynchronous electric motor. Determination of the dynamic viscosity of oil products is carried out by measuring the moment of fluid resistance against the motion of an asynchronous electric motor: the higher the dynamic viscosity, the greater the moment of resistance is developed when the ellipse rotates. The main criteria for estimating the dynamic viscosity of oil products are the moment of inertia of the rotating masses of the drive mechanism and their angular accelerations. The developed hardware-software complex with a rotation angle sensor allows registering and processing the measured parameters. The experimental data obtained in the course of the study demonstrated sufficient convergence of the dynamic viscosity values obtained by the developed method and by the existing methods. In future, it is planned to obtain the dependences of the moments of resistance of various types of oil products on the angular velocity of an asynchronous electric motor and, based on the results obtained, determine the range of speeds, at which the greatest measurement accuracy is achieved.

show abstract

NMR Self‐Diffusion and Transverse Relaxation Time in Bitumen: The Effect of Aging

Porto,

Angelico,

Calandra

et al. 2023

ChemPhysChem

View full text Add to dashboard Cite

In this investigation the dynamics of two types of bitumens with different penetration grade were tested by using dynamic shear rheometry (DSR) and Nuclear Magnetic Resonance (NMR) at unaged conditions, and upon both short‐ and long‐term artificial aging. The gel‐sol transition temperature was found to increase with increasing the time of aging treatment. Arrhenius parameters of the viscosity were found, unexpectedly, to be correlated with those of simple liquids, suggesting that the two kinds of systems, although chemically and physically quite different, share the same basic process at the molecular level. The molecular dynamics has been then investigated by NMR Pulsed Field Gradient Stimulated‐Echo (PFGSE) and relaxometry (Carr‐Purcell‐Meiboom‐Gill, CPMG, spin‐echo pulse sequence) to capture the effect of aging upon dynamics variables such as self‐diffusion coefficients D and transverse relaxation times T2. The translational diffusion at T > of the light molecular components of both types of bitumens was characterized by broad distributions of D which were found independent of the experimental time scale up to 0.2 s. Similarly, T2 data could be described as a continuous unimodal distributions of relaxation times determined both at T < and T > .

show abstract

Improved Correlations for Heavy-Oil Viscosity Prediction with NMR

Cited by 19 publications

References 12 publications

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

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

Method and instruments to measure dynamic viscosity of oil products in pipeline

NMR Self‐Diffusion and Transverse Relaxation Time in Bitumen: The Effect of Aging

Contact Info

Product

Resources

About