8 replies to this topic

[KPFM]

DOW's TEG properties differ from most simulation properties. Would appreciate any insight on which to use for sizing the heat exchangers and reboiler for a TEG dehy?

[MTumack]

I'd go with the properties of the TEG from the manufacturer who you are buying your Glycol from.

 

Buying from DOW? Use DOW's Data they got from testing their product in the lab. Buying from someone else? Use their empirical data instead.

 

You could probably buy 1000 different versions of TEG all with slightly different properties if you looked hard enough. Might as well size with data that comes from Glycol created with said manufacturing process and controls.

[latexman]

How big are these differences?  Give us an example on a few properties.  I'm wondering if it's enough to worry about or not, because Engineering is usually not that exact of a science.  That's why we use safety factors. 

[breizh]

KPFM ,

is it this document your reference ?

I will follow advices from both latexman and MTumack .

 

Breizh

[RockDock]

It depends. Are you sure your simulation is correct? The two should match pretty closely. I have almost always matched their data when simulating the process. You are using ProMax to simulate it, right? That should be a given, but I've been surprised by a number of threads on this site with engineers using the wrong tools.

 

The point of simulation is to make sure the vendor gives you what you need. If you simulate one thing and then accept something else, you might as well not have done any simulation in the first place. Dow's job is to sell the solvent and their expertise. Not that they will be unethical, but they are promoting their products and services.  I think Ronald Reagan's expression, "trust, but verify" applies here.

[KPFM]

Breizh,

 

I am using the Dow reference you attached. The differences I have noticed so far are specific heats and viscosities. I am using VMGSim and here are a couple of examples:

 

1. At 135F for 99.686%TEG:

 

Dow: Cp=0.567 Btu/lb-F & cP = 11.26

VMG: Cp=0.559 Btu/lb-F & cP = 8.32

 

2. At 267.2F for 99.686%TEG:

 

Dow: Cp=0.669 Btu/lb-F & cP = 1.40

VMG: Cp=0.605 Btu/lb-F & cP = 1.51

[breizh]

KPFM ,

 

I will not worry much about the differences and will take latexman's advice into consideration (SF).

Good luck,

Breizh

[MrShorty]

A couple of additional thoughts, if I may. In many ways, I am from the "outside looking in", since I have very little practical design experience. However, I do get a lot of questions from clients like, "Simulator says A, manufacturer says B, literature, if any, says C, D, and E -- how do we decide what to use in design.

 

1) Just to muddy the water further, I checked your values for those 2 properties against DIPPR's database for pure TEG. At 267.2 F, I get 0.597 +/- 1% BTU/lb/R for Cp, and 2.39 +/- 10% cP for viscosity. At 135 I get 0.544 +/- 1% BTU/lb/R for Cp, and 11.9 +/- 10% cP for viscosity. I don't know how much one would expect that 0.341% impurity (water?, mass basis or mole basis?) to change these from the pure TEG values, but it doesn't seem like it should be very much.

 

2) I don't want to say that DIPPR's are better values, because they may not be. One thing I like about DIPPR's values is that, through the database, I can look up the literature sources DIPPR used for their correlations. I believe that most simulator programs also allow you to look up references for the data they are using, and I would expect most manufacturers to have some kind of bibliography. While it is not always true 100% of the time, I generally prefer values that I can trace through the calculation/regression steps to the raw, measured data. When a value comes "undocumented" I tend to treat it with more skepticism.

 

3) Another part is evaluating the applicability of the equations used. For most pure component properties, these equations may not be very complex. When it comes to treating that 1/3 % impurity, one might check the validity of the equations and mixing rules used to account for the impurity. Do these have a good basis?

 

4) As noted by others, one also tends to evaluate how much difference the error makes in the overall design, and build in sufficient safety factors to make sure that the design will work.

[mvp]

The differences are between a pure product and commercial product; secondly, if it is a heat transfer fluid, there are additives (like inhibitors) in every commercial product. It is a good practice to use the properties provided by manufacturer