How to limit throttling/Joule-Thompson effect? 3

[j.doe]

Hi, I've designed a stainless steel cylinder cell to measure the electrical conductivity of R134a in liquid phase (under pressure). I'm a mechanical engineer and not a chemical engineer so i was wondering if you guys can kindly help me out with this. Initially, the procedure we were hoping to implement was to fill from a valve at the bottom (see picture attached) and wait before opening the top valves slightly to see whether liquid is coming out to ensure the refrigerant is in liquid phase before carrying out the experiment (measuring current from electrode inside the cell). The problem is, when opening the top valve, the flash evaporation caused the whole cell to cool significantly (roughly 50C temperature drop, calculated using Clausius–Clapeyron equation) which affected our readings because we want to measure at 25C. The saturation pressure of 134a is 665kPa, it gets released to atmospheric conditions.

I've been thinking about adding a pressure sensor and thermocouple between the top valves (see pic) to use the pressure reading as an indicator of liquid phase, instead of opening it from the top. Then, I would release the liquid from the same bottom valve I filled cell from into an empty cylinder with a relatively long pipe in between, to ensure the cooling effect does not reach the cell. Is there anything you think i can do further to reduce the throttling effect in my system? There must be a way of removing the liquid with a gradual decrease in pressure that stops throttling from occurring when opening a valve right?

I was thinking of covering the valves and pipes at the bottom and top of the cell with heat tapes (set to 25C) as well to limit the throttling. The problem with this though is that a temperature gradient could develop maybe causing bubbles to rise to the top of the cell which would affect our reading (we need uniform liquid conditions). Our cell is within a stainless steel cylinder enclosure to hold it together while being pressured. There is an air gap between the inside of the enclosure and the outside of the cylinder electrode (see pic) so the electrode can't be heated from the outside.

https://files.engineering.com/getfi...99-a95d-a87696065fb8&file=initial_set_up.jpeg

https://files.engineering.com/getfi...6-9c36-6e39175a5a51&file=proposed_set_up.jpeg

 

[j.doe]

@Compositepro Thanks a lot for clearing that up. Sorry, didn't understand what the term meant from your previous posts.

@georgeverghese Yeah i agree, will probably adopt that approach.

@chicopee How do you propose to reduce the rate of discharge from the top valve? I've got two ball valves at the top which i basically manually open and close, do you think i need a different type of valve or something? You're right, i studied biomedical engineering in my undergrad but have come across a lot of ppl who don't know what that is, so didn't want to confuse ppl here when reading my post and claimed to be a ME cos i thought that was the next closest subject. We studied bio instead of thermodynamics, but that was literally the only difference in my uni. Sorry.

 

[IRstuff]

Hypothetically, if the pressure is sufficiently high enough, its equilibrium state will still be liquid, even at 20C

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[Latexman]

A ball valve is NOT the greatest throttling valve. Try a globe valve or a needle valve.

Good Luck,
Latexman

 

[j.doe]

Thank you all so much. You guys are the best!!

 

[georgeverghese]

The additional kit you will need to enable subcooling is nothing to be scoffed at - you can create a tear refrigerant stream from this R134a and let it down to a much lower pressure to chill / subcool the test stream. Talk to a senior process or mechanical engineer to help with configuring the subcooling section.

 

[TiCl4]

"You need to pressurize the cell to at least 665 kPa with air or nitrogen before introducing liquid R134a into the cell."

What's the point of that? As you say, the pure component vapor pressure of R134A from the diagram is ~665 kPa at 25 C. If you put liquid R134A into a closed vessel that has 665 kPa due to N2 or air pressure, the total cell pressure will equilibrate to ~1,330 kPa at 25C. Liquid refrigerant will still vaporize until the partial pressure of the refrigerant is 665 kPa in the headspace. Liquid vaporization is controlled solely by the partial pressure of that component - injecting an inert gas to increase total system pressure will have NO effect on the vaporization of the refrigerant. If you try to bleed the vessel while simultaneously injecting liquid in order to maintain pressure at 665 kPa, you will be dropping the partial pressure of R134a below 665kPa because the purge stream will be comprised of air/N2 and R134a. This will continue until all the air/N2 is purged - pressurizing the vessel beforehand will actually lengthen the bleed process and induce more vessel cooling, meaning the OP will still have the same cooling issue he experienced when trying to fill the vessel when starting at ambient conditions.

If you were to take the vessel at atmospheric conditions and pump liquid R134A into it until it is 80% full, you vessel would be approximately 665 kPa + 500 kPa = 1165 kPa (compressing air isothermally from 100% volume at 100kPa to 20% volume). OP, if your vessel has an MAWP greater than 1,200 kPa, that is the easiest way to go.

The approach described by IRstuff is the correct one. Pull you vessel down to as low of a vacuum as you can. Then inject R134A until 80% full. No need to bleed the system, and depending on the level of vacuum you obtained, you should be between 665-800 kPa in the vessel. You can estimate the % full of the vessel by the final vessel pressure at 25C. For an isothermal compression of an ideal gas, P1V1 = P2V2. If you initial vessel is 5 kPa, your vessel will be ~80% full at 665 kPa + (5 kPa *100%Volume/20%Volume) = 690 kPa. Or, you could just weigh the vessel before and after adding to determine the amount added.

If air contamination is an issue, then you will need to bleed pressure down to 665 kPa at 25 C. There is no way to avoid the cooling effect during this bleed - you will need to keep it in a water bath and wait for temperature to recover to 25 C before taking measurements.

Edit: Corrected bad terminology Re Latexman's and Compositepro's corrections.

 

[Compositepro]

Adding nitrogen pressure is pointless. It does nothing to affect the equilibrium vapor pressure of the liquid. It does prevent boiling, however. Boiling occurs when the liquid vapor pressure is greater than the ambient pressure. When vapor pressure is less than ambient pressure, evaporation is what happens.

 

[Latexman]

Liquid boiling is controlled solely by the partial vapor pressure of that component - injecting an inert gas to increase total system pressure will have NO effect on the boiling point of the refrigerant.

So, water boils at the exact same point (temperature) no matter the altitude?

What is a "partial vapor pressure"? Vapor pressure of a component is a function of temperature only. There's no concentration term in the Antoine equation. I think you are confused with partial pressure.





Good Luck,
Latexman

 

[Latexman]

Adding nitrogen adds one degree of freedom. As long as the total pressure is more than the saturation pressure, one can control temperature and pressure. With only R134a present, those are forever tied together. It adds controllability.

Good Luck,
Latexman

 

[Compositepro]

What would be the point of controlling or changing total pressure independently of temperature for this experiment?

 

[IRstuff]

The issue is nitrogen contamination of the R134 test sample. At the pressures required to keep the R134 liquid, there's going to be a high percentage of nitrogen dissolved in the liquid R134. That potentially messes with the bulk electrical conductivity measurements.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[Latexman]

True, I hadn't thought much about dissolved nitrogen in the R134a. Anyone know the solubility at around 1 Mpa? I didn't find any free data using Google on this.

Good Luck,
Latexman

 

[IRstuff]

btw, j.doe

How are you dealing with the overall question of R134 purity? Seems odd to me that you'd purify R134 into one container and then have to transfer it to another, risking additional contamination at several interfaces. How do you even know you're not contaminating the R134 in the primary container already? Every valve is a potential source of organic contamination.

There may even be contaminants that aren't even tested for. We had the misfortune to discover that stainless steel typically contains trace amounts of boron, which they don't test for, but wreaks havoc in certain semiconductor materials. Since almost all process piping is stainless steel, that undetectable amount of boron in the steel essentially tanked our development project.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[Latexman]

I thought the OP was having difficulty controlling the R134a in the cell. He said it would flash off and not be at 25 C where it was needed. That's why I suggested to add nitrogen or air; to disconnect the temperature/pressure relationship of the vapor/liquid R134a in the cell.

Good Luck,
Latexman

 

[TiCl4]

Thank you both for the corrections. "As iron sharpens iron"...

As to N2 solubility, I've found the following reference for the Henry's Law parameter as being 0.0013 for N2 in R134a (Table 3).

Solubility of nitrogen in one-component refrigerants: Prediction by PC-SAFT EoS and a correlation of Henry’s law constants November 2011International Journal of Refrigeration 34(8):2109-2117

https://www.researchgate.net/public...oS_and_a_correlation_of_Henry's_law_constants

Latexman: Even if N2 were added, opening the bleed valve would bleed of a N2/R134a mixture, causing additional R134a to vaporize to maintain the vapor pressure. He would still have the same cooling issue, no?

 

[IRstuff]

Seems to me that the OP has several problems

> JT expansion of R134 into 2nd container -- that may be completely unavoidable, given the high pressures involved; BUT, it's a transient thing, so the question is whether the gas will condense back into liquid after equilibration. It seems to me that in the steady state, if there was sufficient pressure in the 1st container and the 2nd container is small compared to the 1st, then you'd be back to the saturated gas and liquid in the second container.

> Not even sure how big a deal this is anyway. My generation of EEs, at least, has taken something like Dust-off and inverted the can to spray liquid freon onto ICs to cool them down, so getting liquid Freon out of a pressurized container is achievable in the transient case. This suggests that an internal bladder (another contamination source) might be the best approach, or super-cooling the 2nd container with LN2, or somesuch, to keep the Freon from flashing.

> Why not put the R134 into the 2nd container from wherever came from to get into the 1st container; that would eliminate a truckload of potential contaminants.

> Contamination -- possibly unavoidable as well; the OP has not stated how the R134 came into being in the first place and is attempting to make measurements that are ostensibly better than previous. One would think that would put contamination at the top of the list, even ahead of an JT considerations. High purity R134 is only around 99.5%. Some claims of 99.99%, but who knows... I mea, if it's just to be used as a refrigerant, why does one even need 99.99%?
TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[Latexman]

Good find, but that's not Henry's Law parameters. It's a standard deviation of Henry's Law parameters between an EOS and a correlation.

I don't think that article lists any Henry's Law parameters; not of R134a anyway. It does give parameters to an equation that allows one to calculate R134a's Henry's Law parameters.

Good Luck,
Latexman

 

[Latexman]

I'll concede adding nitrogen may affect the measurement, not sure how much, but let's focus on doing this with only R134a. Agreed?

Good Luck,
Latexman

 

[IRstuff]

The OP is essentially duplicating the test setup of the paper I referenced earlier, but, in the paper, the calculation of the bulk resistivity of the R134 is predicated on complete filling of the test (2nd) container to a known volume, so having any gas phase in the 2nd container is a non-starter. This would mean that the OP needs to either bleed the gas off the second container, or allow the gas to get back to the 1st container.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[j.doe]

yeah my supervisor just told me the same thing about not using nitrogen, so i'll just be using 134a like you guys said. Thanks for the detailed process explanation @TiCl4. To answer your questions about purity and contaminants @IRstuff, the first container containing the 134a is too large to keep inside my lab, it's a pharmaceutical grade cylinder with ≤6ppm moisture content (moisture is the main impurity we're trying to limit). The company supplies the 134a to pharma companies for use as an inhaler propellant, and this whole project is to do with understanding the electrostatic properties of formulations within inhalers (not sure i should be disclosing this stuff lol). I guess we are risking additional contamination when transferring to the smaller cylinder but they taught me how to limit such contamination when doing the transfer (by opening the valves first to purge the valves before connecting the 2nd cylinder). You're right contamination is at the top of my list, and i think that's the main reason why i got such higher values for resistivity in the preliminary tests compared to previous studies.