Viscosity Conversion 2

[Falcon03]

Hello,

Can some one help me, I would like to have some sort of correlation to convert the heavy petroleum cuts (such as Vacuum bottom) Viscosity @ 100 °C to 50 °C. Any reference that I can refer to? Thank you.

 

[Falcon03]

Me make it simple;

The typical viscosity of our Vacuum is 2800 cts @ 100 °C, I would like to know the viscosity @ 50 °C. Thank you.

 

[TomSD]

I believe you give the -kinematic- viscosity in centistokes.

I haven't got this paper which sounds relevant for your problem
S.A. Beg, M.B. Amin and I. Hussain, "Generalized Kinematic Viscosity - Temperature Correlation for Undefined Petroleum Fractions", The Chemical Engineering Journal, 38, 123-136, 1987.

Perry's Chem.Eng. Handbook includes a section on correlation/prediction of absolute viscosity that might be of use. A generalized viscosity vs temperature chart is presented which, if your viscosity was 2800 centipoise at 100C, would indicate the viscosity at 50C to be approaching 1,000,000 centipoise, so I would check out alternative references.
Regards

 

[25362]

As far as I know there is no correlation because viscosities differ from crude to crude (paraffinicity, aromaticity, etc.). Thus your VB could even reach a KV of 20,000,000 cS @ 50 celsius, becoming almost a solid bitumen. A crude assay may help to predict viscosities of plant fractions. However, blends do not follow a linear arithmetical relationship. Maxwell's Data Book on Hydrocarbons may be of help in providing blending indices.

 

[kbander]

I remember that Twu had a viscosity correlation that was (and probably still is) very popular. Also look for jounal papers by Perdersen or Lorentz-Bray-Clark.

While some methods may be a bit complex, the correlations are available. Viscosity is pretty difficult to predict with great accuracy though...so I'm not sure if anything will give you what you want for very viscous materials (i.e., atmospheric resid). Someone else may be able to comment on this.

Good luck!

Bob

 

[TD2K]

A lot of fractions have a reasonably linear relationship plotted on log viscosity versus temperature curves. I'd try to get a viscosity of your sample at another temperature, say 210F/99C, and then extrapolate off that. Maxwell, as mentioned, has those graphs to use as well as a bunch of other references.

The other option is to find plotted vis data for similar fraction. Most similar fractions have similar slopes. Thus, you can use your data point and their slope to extrapolate, I've done this in the past.

As mentioned, if you are getting up into the 1 MM cst range, the accuracy of the results is going to be only so so at the best.

 

[25362]

The formula given (ChE, Perry) for hydrocarbon fractions is for absolute viscosities in cP, visc. = bT^m

By plotting a gas oil of the same crude oil origin on log-log paper you'd get a straight line. By drawing a parallel for VB, you probably would estimate the viscosity @ 50 Celsius. To transform back to cS must divide by densities.

Better, as TD2K suggests, use two actually measured kinematic viscosities at two different temperatures and by linear extrapolation on a Refutas type of chart, find the viscosities at other temperatures.

 

[krsmani77]

Hope your Vac resid's pour point is well below 50Deg C. Otherwise no correlation/extrapolation will give anything useful. Normally Vac resid will have some amount of lighters in it, depending on degree of fractionation, cracking in the furnace,etc, So it is best to do a lab check on each specific sample at two temperatures and use the data as mentioned above to get the viscosity at 50Deg C by drawing a st line thru two points on a semilog graph sheet. If you are doing a study, use the data from the crude book for the particular Vac bottom cut.

 

[TD2K]

I went back to some of my reference material. 2800 cst at 100C is in the range of asphalts (between a 25 and 100 pen asphalt). Assuming I can use the same slope as the asphalts shown on my reference, I get a viscosity at 50C of close to 1,000,000 cst for this material.

I don't know what you need this value for or what accuracy you need. Are you doing a fuel oil blend calculation?

 

[jmw]

We use the relationship log.log(v+0.7) = A-Blog T where V is kinematic viscosity in cst and T is degrees Kelvin. We use a variety of methods based on this equation (From ASTM D341) to determine the viscosity at a reference temperature from the measurement of viscosity in the process at the process temperature. We have used it with PIB, with bitumens, asphalts, vacuum tar rundown, stripper bottoms etc to good effect.
If you visit

http://www.solartronmobrey.com

and follow the bunker fuel link you can download a spreadsheet which allows you to run this equation on any product for which you have the viscosity at two temperatures. You can also download a powerpoint which explains how we use a number of calibration curves to use a single viscosity measurement to find the viscosity at base temperature. Of course, we use this in our viscosity computers for continuous process measurement but you could adapt the spreadsheet to allow you to do this with lab sample data.
However, i suspect you could use a cone and plate viscometer in the lab to measure the viscosity at 50degC (or if it is solid, at two other temps and use the spreadsheet to assess the theoretical viscosity at 50degC. Cone and plate viscometers are supplied by london research equipment (

http://www.london-research.com

I think)