Viscosity Calculator predictions...

Status
Not open for further replies.
Joined
Dec 12, 2002
Messages
43,887
Location
'Stralia
Was messing with this paper in my spare time today
https://www.sciencedirect.com/science/article/pii/S1110062117300673#f0010

Take a base oil, add viscosity modifiers to it, and measure the properties.

Some recent discussion has been as to how the viscosity calculators work with non pure, non Newtonian oils....fortunately the paper gives for the various combinations the KV40, and KV100 of a handful of the mixes...then fills it in with the KV60 and KV80...

I've tabled them below, and then used the Widman Calculator to predict what these intermediate viscosities are...

viscosity calculator versus measured.jpg
 
I think you're wasting your time by using this paper as a reference.

Polymers additive for improving the flow properties of lubricating oil
S. I. Shara, E. A. Eissa, and J. S. Basta, Egyptian Petroleum Research Institute, Analysis and Evaluation Department, Egypt

They start from a 4.28 cSt base oil but KV100 hardly moves with 2% VII content and even 4% VII content. This is not what we see in real oils. Typically you get 15% boost in viscosity (KV100) for every 1% of an OCP VII. I'm not sure what's going on but I'm not confident about the results in this paper.
 
Originally Posted by Gokhan
I think you're wasting your time by using this paper as a reference.

Polymers additive for improving the flow properties of lubricating oil
S. I. Shara, E. A. Eissa, and J. S. Basta, Egyptian Petroleum Research Institute, Analysis and Evaluation Department, Egypt

They start from a 4.28 cSt base oil but KV100 hardly moves with 2% VII content and even 4% VII content. This is not what we see in real oils. Typically you get 15% boost in viscosity (KV100) for every 1% of an OCP VII. I'm not sure what's going on but I'm not confident about the results in this paper.


What we have in finished oils are single, binary, ternary or n-ary mixtures of base oil viscosities, but I see nothing wrong with the paper.

They may have wanted to experiment with a LV oil to show the VII polymer did what was predicted.

As starting points, you have to have data on the base oil properties and then characterize the resulting mix, which is what they did.

EPDM's have been used as a Multi-functional VII-PPD's.
 
Last edited:
Shannow,

I had a very quick squint at this paper...

I think I must have tested every commercial liquid VII, ever produced at 10 wt% in Esso 150SN (Group I, KV100 about 5.5 cst). It was one of the quick & dirty 'look-see' tests I did to anything & everything that came through the lab door.

Although there was some variation depending on SSI, dilution ratio, chemical type, straight or functionalised, etc, but typically 10% liquid VII will give you a KV100 bump of 5.5 cst & a KV40 lift of 40 cst. The variation is 'dampened' by the fact that most commercial liquid VII are cut-back to a KV100 of around 1,000 cst (think thick but handleable).

With this in mind, the numbers in the study look strange. However this might be explainable by the fact that...

a) they cut the EPDM back in xylene (which is ultra low viscosity!), not base oil.

b) they don't actually state the 'active' (ie solid) concentration of the VII (I suspect it must be VERY dilute!)

c) they blend the VII into base oil cold which is terrible practice! The rule with VIIs is always store hot, pump hot & blend hot. Only once the VII is 'in' can you let the oil cool to ambient.

As I recall, VIIs, along with all other additive components behave predictability with temperature from about 0°C upwards. HTHS is something of a special case because you're applying lots of shear. Take away the shear at 150°C & it's KV is predictable using the classic double log relationship. Below -10°C however, that's when wax typically starts to precipitate (regardless of whether you add PPD or not) & things become seriously non-Newtonian. You will always massively under-predict CCS viscosity however hard you try.
 
Regarding the error in KV, several years ago, Doug Bogart, STLE CLS, Laboratory Director, Wear Check USA, had replied to me as:

"Normally +/- 2 cSt for given SAE. With new oils, it's less than +/- 0.5 cSt, not much significance."

If we take the absolute error as ±0.5 cSt, this corresponds to a ±5% relative error for KV100 = 10 cSt.

Therefore, it looks like the the KV values reported in the paper satisfy ASTM D341 well within the error bars.
 
Originally Posted by Garak
For the extrapolated values, that won't hold.

OK, rather than opposing everything without doing any calculation yourself, why don't you extrapolate the KV100 from the KV40 and KV80 values in the above table and then see if it holds or not?
 
Originally Posted by Gokhan
Originally Posted by Garak
For the extrapolated values, that won't hold.
OK, rather than opposing everything without doing any calculation yourself, why don't you extrapolate the KV100 from the KV40 and KV80 values in the above table and then see if it holds or not?

It actually takes only a minute or two. All you have to do is to stick the values in:

https://www.widman.biz/English/Calculators/Operational.html

The answer is, yes, the extrapolation works. You get similar results to what Shannow got for the interpolation.

I'm not too worried about ASTM D341 being applicable to finished oils with VII and DI or not. It's one of the better tools we have. The data here seems to confirm it.
 
Originally Posted by Gokhan
Originally Posted by Gokhan
Originally Posted by Garak
For the extrapolated values, that won't hold.
OK, rather than opposing everything without doing any calculation yourself, why don't you extrapolate the KV100 from the KV40 and KV80 values in the above table and then see if it holds or not?

It actually takes only a minute or two. All you have to do is to stick the values in:

https://www.widman.biz/English/Calculators/Operational.html

The answer is, yes, the extrapolation works. You get similar results to what Shannow got for the interpolation.

I'm not too worried about ASTM D341 being applicable to finished oils with VII and DI or not. It's one of the better tools we have. The data here seems to confirm it.

This study was base oil and VII only. The chart holds because VI improvers are hydrocarbons, thus ASTM D341 applies. You can't use this data to prove it works for fully formulated oils, which are not hydrocarbon fluids. ASTM D341 specifically warns that non hydrocarbons do not respond linearly. We don't have a clue what the effect of the multitude of combinations of non hydrocarbon components used in any specific oil fully formulated oil might have on an extrapolated number.

Without measured KV150 numbers you can't even begin the process of empirically validating your equations.

Ed
 
Originally Posted by edhackett
Originally Posted by Gokhan
Originally Posted by Gokhan
Originally Posted by Garak
For the extrapolated values, that won't hold.
OK, rather than opposing everything without doing any calculation yourself, why don't you extrapolate the KV100 from the KV40 and KV80 values in the above table and then see if it holds or not?
It actually takes only a minute or two. All you have to do is to stick the values in:

https://www.widman.biz/English/Calculators/Operational.html

The answer is, yes, the extrapolation works. You get similar results to what Shannow got for the interpolation.

I'm not too worried about ASTM D341 being applicable to finished oils with VII and DI or not. It's one of the better tools we have. The data here seems to confirm it.
This study was base oil and VII only. The chart holds because VI improvers are hydrocarbons, thus ASTM D341 applies. You can't use this data to prove it works for fully formulated oils, which are not hydrocarbon fluids. ASTM D341 specifically warns that non hydrocarbons do not respond linearly. We don't have a clue what the effect of the multitude of combinations of non hydrocarbon components used in any specific oil fully formulated oil might have on an extrapolated number.

Without measured KV150 numbers you can't even begin the process of empirically validating your equations.

Ed

Hi Ed, you're overinterpreting things.

Exact wordings from ASTM D341:

"It should also be emphasized that fluids other than hydrocarbons will usually not plot as a straight line on these charts."

The way I would interpret it would be "Don't expect it to work it for any fluid" rather than your interpretation "We have this nice oil tool but don't expect it to work if the oil is a finished oil."

I would worry more about the VIIs' effect, which greatly modify the viscosity. The fit function isn't supposed to work well when there are hydrocarbon phase transitions, which is what the VIIs go through in order to modify the viscosity. Therefore, what you said about the hydrocarbons always fitting nicely is also an overinterpretation. Now, you're worried about the DI pack, which has a much smaller effect on the viscosity and should easily be absorbed into the coefficients of the fit function?

Also, if ASTM D341 is completely useless for finished oils, why in the world do they bother documenting the viscosity index (VI) in the PDSs for virtually every finished oil? The viscosity index is based on ASTM D341.

Yes, like every approximation, there is an error in ASTM D341, which varies with the oil and additives. However, to say that it's useless for anything other than pure (neat) base oils is a stretch.
 
phase transitions ???

Please explain what you mean by phase transitions in the performance of Viscosity Modifiers...

edit...obviously within the range of 0-150C which is what we are talking about...
 
Last edited:
Originally Posted by Shannow
phase transitions ???

Please explain what you mean by phase transitions in the performance of Viscosity Modifiers...

edit...obviously within the range of 0-150C which is what we are talking about...

(Quasi) phase transition as in the VII molecules changing their shapes...
 
I just want to emphasize that VII coil size expansion with temperature is not necessary to achieve significant elevation of viscosity index (VI).
The results in this study - How Polymers Behave as Viscosity Index Improvers in Lubricating Oils, show that coil size expansion with temperature is not necessary to achieve significant elevation of viscosity index (VI), but polymers which do expand with temperature have higher VI contributions than those that do not.
As you will see from the study, polymer coil size in solution is relatively invariant with temperature for OCP VIIs. Olefin copolymers thicken base oils by about the same proportion, regardless of temperature. Situation is different for PMA VIIs because they do expand when temperature increases.

You can find the study here - https://www.researchgate.net/public...sity_Index_Improvers_in_Lubricating_Oils
It is also available in .pdf format.
 
Last edited:
Originally Posted by emod
I just want to emphasize that VII coil size expansion with temperature is not necessary to achieve significant elevation of viscosity index (VI).
The results in this study - How Polymers Behave as Viscosity Index Improvers in Lubricating Oils, show that coil size expansion with temperature is not necessary to achieve significant elevation of viscosity index (VI), but polymers which do expand with temperature have higher VI contributions than those that do not.
As you will see from the study, polymer coil size in solution is relatively invariant with temperature for OCP VIIs. Olefin copolymers thicken base oils by about the same proportion, regardless of temperature. Situation is different for PMA VIIs because they do expand when temperature increases.

You can find the study here - https://www.researchgate.net/public...sity_Index_Improvers_in_Lubricating_Oils
It is also available in .pdf format.

Thanks for this! This is an excellent, high-quality serious recent research article from the Applied Science Department of Warren Buffet's Lubrizol additive company! I've added it to my library.

Quickly glancing at it, one of its conclusions seems to be that the OCP VII has very similar viscosity - temperature relationship to that of pure (neat) base oils.

Given that the OCP VII is by far the most common VII these days because it doesn't leave nearly as much turbocharger deposits, this explains why the ASTM D341 viscosity - temperature calculator for the finished oils and my base-oil-viscosity calculator work so well.
 
Originally Posted by emod
I just want to emphasize that VII coil size expansion with temperature is not necessary to achieve significant elevation of viscosity index (VI).
The results in this study - How Polymers Behave as Viscosity Index Improvers in Lubricating Oils, show that coil size expansion with temperature is not necessary to achieve significant elevation of viscosity index (VI), but polymers which do expand with temperature have higher VI contributions than those that do not.
As you will see from the study, polymer coil size in solution is relatively invariant with temperature for OCP VIIs. Olefin copolymers thicken base oils by about the same proportion, regardless of temperature. Situation is different for PMA VIIs because they do expand when temperature increases.

You can find the study here - https://www.researchgate.net/public...ity_Index_Improvers_in_Lubricating_Oils.
It is also available in .pdf format.


Here's a directly accessible study that shows the same thing. In fact, it seems that OCP polymers actually contract with heat. They also explore if VI can be predicted by measuring the polymer behavior. The short answer is no, because as usual, it's more complicated than that.

Viscosity Modifiers: A Fundamental Study

Ed

P.S. Gokan, I'll address your last reply to me later today. I replied late and didn't finish my thoughts.
 
Another conclusion of the Lubrizol article is that:

"Olefin copolymers (OCP) thicken base oils by about the same proportion, regardless of temperature."

This addresses Shannow's long-standing concern regarding the relationship between the viscosity index (VI) and VII content or, more precisely, the lack of such a simple relationship in my A_Harman-index and base-oil-viscosity calculations. One would naïvely expect to be able to deduce the VII content from the VI of the finished oil but this is not the case in reality.

KV40 and KV100 are multiplied by the same number for a given VII content. This results in the viscosity index (VI) for thinner base oils increased more for a given OCP VII content than for thicker base oils.

In other words, the viscosity-boost rate of OCP VII doesn't depend on the base-oil viscosity much but the viscosity-index-boost rate greatly increases with decreasing base-oil viscosity.

You can use the viscosity-index calculator to see what happens to the viscosity index (VI), for example for a 20% OCP VII boost simultaneously in both KV40 and KV100 for different starting KV100 values but same starting VI values.

https://www.widman.biz/English/Calculators/calc-visc-index.html

Once again, thanks for this great reference!
 
Originally Posted by Gokhan
OK, rather than opposing everything without doing any calculation yourself, why don't you extrapolate the KV100 from the KV40 and KV80 values in the above table and then see if it holds or not?

How am I supposed to verify an extrapolated KV without the equipment to run a KV test in the first place? What I'm concerned about here are not KV40 and KV100 values, since those points are the ones in which we actually have the most confidence when using calculators and we routinely have those values available, as much as that isn't as ideal as raw data. Even interpolating, I'm not all that concerned about KV80, though a measurement is better.

My point is that if we're extrapolating a KV150 without anyone having measured a KV150, we've got a lot of nothing. KV150 figures in here and is used for calculations without the number having been experimentally verified. The way a proper error analysis is conducted, as you're well aware, is comparing the prediction to a measured value, and the measured value itself will have an uncertainty. Right now, we're dealing with a prediction without testing that prediction. We're not grabbing an accepted value out of the CRC Handbook here.
 
Originally Posted by Garak
Originally Posted by Gokhan
OK, rather than opposing everything without doing any calculation yourself, why don't you extrapolate the KV100 from the KV40 and KV80 values in the above table and then see if it holds or not?
How am I supposed to verify an extrapolated KV without the equipment to run a KV test in the first place? What I'm concerned about here are not KV40 and KV100 values, since those points are the ones in which we actually have the most confidence when using calculators and we routinely have those values available, as much as that isn't as ideal as raw data. Even interpolating, I'm not all that concerned about KV80, though a measurement is better.

My point is that if we're extrapolating a KV150 without anyone having measured a KV150, we've got a lot of nothing. KV150 figures in here and is used for calculations without the number having been experimentally verified. The way a proper error analysis is conducted, as you're well aware, is comparing the prediction to a measured value, and the measured value itself will have an uncertainty. Right now, we're dealing with a prediction without testing that prediction. We're not grabbing an accepted value out of the CRC Handbook here.

What I meant was that you can extrapolate the KV100 from the experimental KV40 and KV80 values in Shannow's table and then see if it matches the experimental KV100 in his table within the error bars.
 
Oh yes, I understand what you mean, but my issue isn't so much there. There can be problems with interpolations, but at least we do have a lot of KV40 and KV100 values in reality from data sheets and VOAs. When it comes to the KV150 business, that's where I really wonder. Out of interest's sake, I might play with some of those numbers you suggested if I can get a few free minutes.
 
Status
Not open for further replies.
Back
Top