Petroleum Science >2015,??Issue 3:??492-500 DOI: https://doi.org/10.1007/s12182-015-0032-y
Model inference using the Akaike information criterion for turbulent flow of non-Newtonian crude oils in pipelines Open?Access
文章信息
作者：Ahmed H. Kamel,Ali S. Shaqlaih and Essam A. Ibrahim
作者單位：
College of Business & Engineering, University of Texas of the Permian Basin, 4901 E University, Odessa, TX 79762-0001, USA;Department of Mathematics & Information Sciences, University of North Texas at Dallas, 7400 University Hills Blvd, Dallas, TX 75241, USA;College of Business & Engineering, University of Texas of the Permian Basin, 4901 E University, Odessa, TX 79762-0001, USA
投稿時(shí)間：2015-07-09
引用方式：Kamel, A.H., Shaqlaih, A.S. & Ibrahim, E.A. Pet. Sci. (2015) 12: 492. https://doi.org/10.1007/s12182-015-0032-y
文章摘要
The friction factor is a crucial parameter in
calculating frictional pressure losses. However, it is a
decisive challenge to estimate, especially for turbulent flow
of non-Newtonian fluids in pipes. The objective of this
paper is to examine the validity of friction factor correlations
adopting a new informative-based approach, the
Akaike information criterion (AIC) along with the coefficient
of determination (R2). Over a wide range of measured
data, the results show that each model is accurate when it is
examined against a specific dataset while the El-Emam
et al. (Oil Gas J 101:74–83, 2003) model proves its superiority.
In addition to its simple and explicit form, it covers
a wide range of flow behavior indices and generalized
Reynolds numbers. It is also shown that the traditional
belief that a high R2 means a better model may be misleading.
AIC overcomes the shortcomings of R2 as a trade
between the complexity of the model and its accuracy not
only to find a best approximating model but also to develop
statistical inference based on the data. The authors present
AIC to initiate an innovative strategy to help alleviate
several challenges faced by the professionals in the oil and
gas industry. Finally, a detailed discussion and models’
ranking according to AIC and R2 is presented showing the
numerous advantages of AIC.
calculating frictional pressure losses. However, it is a
decisive challenge to estimate, especially for turbulent flow
of non-Newtonian fluids in pipes. The objective of this
paper is to examine the validity of friction factor correlations
adopting a new informative-based approach, the
Akaike information criterion (AIC) along with the coefficient
of determination (R2). Over a wide range of measured
data, the results show that each model is accurate when it is
examined against a specific dataset while the El-Emam
et al. (Oil Gas J 101:74–83, 2003) model proves its superiority.
In addition to its simple and explicit form, it covers
a wide range of flow behavior indices and generalized
Reynolds numbers. It is also shown that the traditional
belief that a high R2 means a better model may be misleading.
AIC overcomes the shortcomings of R2 as a trade
between the complexity of the model and its accuracy not
only to find a best approximating model but also to develop
statistical inference based on the data. The authors present
AIC to initiate an innovative strategy to help alleviate
several challenges faced by the professionals in the oil and
gas industry. Finally, a detailed discussion and models’
ranking according to AIC and R2 is presented showing the
numerous advantages of AIC.
關(guān)鍵詞
-
Friction factor Pipeline Informationtheory Non-Newtonian Turbulent