Compositional Methods for Viscosity 1

[carlos81]

I am carrying out a study to diminish the uncertainty presented by the process simulators in the estimation of the crude oil viscosity.
For the development of this work, I need some methods used by the process simulator. I appreciate any information you have about the estimation of crude oil viscosity, specifically the Black Oil Model, the Hydrocarbon Mixtures Model (both for PIPESIM)and the method for Liquid Hydrocarbon Mixtures (HYSYS).

I am focused on compositional methods.

Regards,

 

[sshep]

You probably already know this but, typical simulation methods for psuedo components (and assays) are based on the Watson Characterization Factor (K) which is a function of boiling point and specific gravity. The API technical data book chapters 2 and 11 explain how to determine K and use it to predict viscosity.

It was always my belief that the underlying method was empirical, however my copy of the API Tech Data Book references API Documentation Report No. 2-81 as documenting the basis for the method. I have never read that report but it could be of use to you.

 

[carlos81]

sshep,

Thank you for the reference, I found some of the methods of calculation for the prediction of crude oil viscosity. However, there is no mention of the compositional methods, which are the ones I need to compare results, although I've found some information in other articles.

Regards,

 

[sshep]

Carlos81,

Sorry, I had thought that you were interested in viscosity methods applied particularly to Assays where the only characterization is density and boiling point (trueBP, ASTM D86, etc) curves.

If you are using mixtures of conventional components and real data, viscosity can be easily regressed to the Andrade Equation: ln(mu)= a+b/T+cln(T). I suggest that the first 2 terms are sufficient in most cases. The normal mixing rule is: ln(mumx)=sum(xi*ln(mui)) + interaction term. The interaction term is usually unused (and hope for the best) because the parameters must be regressed with greater difficulty and interaction terms are poorly understood in advance (at least by me) because of the usual nested summations.

If you have a license to Hysys, all of this information can be obtained from the knowledge base and documentation on the Aspen website.

Good luck! sshep

 

[jmw]

The Andrade equation provides a general model for the relationship between the temperature and viscosity of a fluid.

e.g. at this link:

http://infosys.korea.ac.kr/kdb/kdb/hcprop/showcoef.php?cmpid=989&prop=VSL

However, for hydrocarbons, the version found in ASTM D341 is very well established and may be preferred. For example:

log10.log10(v+0.7)= A-Blog10(T+273)
where v is the kinematic viscosity at temperature T deg C (for 2 cSt and above; below 2cst additional terms are required).

Incidentaly, i am shortly going to put a spreadsheet for the ASTM calculations (2cst and above) onto my website (if i can figure out how). When it is there i will post again.

This does not, it seems, answer the question as to the prediction of the viscosity of a crude based on its composition.
Possibly the ASTM standard on molecular weight may assist. This relates the viscosity to molecular weight and the molecular weight is a function of the components?


JMW

http://www.viscoanalyser.com

Eng-Tips: Pro bono publico, by engineers, for engineers.

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[DickRussell]

For prediction of viscosity of a narrow-boiling pseudocomponent at two temperatures, there is the Twu method (Twu, C.H., IEC Proc.Des.Dev., 24, 1287 (1985). For mixture viscosity, the DIPPR Data Prediction manual recommends logarithmic mixing [ ln(vmix) = sum(ln of component v)] for general chemical systems. The API Technical Data Book recommends cubic averaging [cube root of vmix = molar weighted average of cube roots of component viscosities]. The old Maxwell Data Book on Hydrocarbons presented the blending index method for oil fractions. Finally, ASTM method D-341 (API Procedure 11A4.4) uses a double logarithmic method.

I once ran a crude column four times and had the simulator (PD-PLUS) calculate product viscosities using each of the methods described. The answers for blended viscosity were all fairly close.

RAR

 

[jmw]

There are also the Louis and Refutas methods. A search within Eng-Tips may find another thread on this.

JMW

http://www.viscoanalyser.com

Eng-Tips: Pro bono publico, by engineers, for engineers.

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.