Vacuum Rise Test Misconceptions

[DTLarca]

I think I have now been able to upload the images...

Pre-Commissioning Tasks

Prior to commissioning a refrigeration or air conditioning system we carry out three key tests on the pipework. A strength pressure test, a tightness pressure test and a vacuum rise test. The tightness test is to prove the system does not have leaks and the vacuum rise test is to prove the system is moisture free.

The tradition is to carry out the vacuum rise test only at the end of the vacuuming period. Commonly 24hrs vacuuming is applied to smaller systems. The first 4 hours to reach the target Torr level and the other 20 to remove the small quantities of pipe and component impregnated moisture.

A recent disappointing experience has motivated me to ask a few questions. The answers so far have made me think the traditional sequence of method followed by myself, and those I have spoken to, should probably change.


My Question

If we vacuumed a newly installed system to 0.5 Torr, over a period of 24 hours, is it possible to then see no rise in the vacuum within 30 minutes, after the vacuum pump has been isolated, even if there in fact is still free water in the pipework?


Time and Heat

At 0°C and 1 atmosphere liquid water has an arbitrary heat content, enthalpy, of 0kJ/kg and then at 0.5 Torr and -27°C on the sublimation line, where vapour can deposit as ice crystals, the enthalpy is -388kJ/kg. At 0°C and 0.5 Torr water can only exist stably as a vapour shown as point A in figure 2. Vapour at 0.5 Torr has an enthalpy of 2451kJ/kg. As such, on starting a vacuum pump, the 0°C water mass only has enough heat to vapourise 16% (388/2451x100) of itself taking it to point B. The other 84% of the heat required has to be conducted and radiated from the pipework. A major time constraint is the saturated vapour’s specific volume down at 0.5 Torr where it would take a 5CFM vacuum pump nearly 11 days to remove the volume of vapour produced by each kg of ice. The idea is to instead keep up the heat supply which keeps up the vapour pressure and therefore vapour density where, with proficiency, can result in a full kg of water being removed in a single working day.


Vacuum Ballast Valve

Ice originally in contact with the pipe walls would sublimate leaving a large remaining portion reliant purely on radiant energy for its heat source. Ice subliming under vacuum of course by definition does not pass through a liquid phase on taking heat so there would be no liquid pools ever bridging the thermal gap between suspended ice and the lower pipe walls. Ideally the vacuum pump gas ballast valve would not allow pressures to fall below 4.6 Torr in order that freezing is avoided in addition to reducing moisture condensation in the pump oil. Unfortunately, with most modern vacuum pumps, the difference in achievable Torr with and without the gas ballast valve open is a mere 0.2Torr, not sufficient to prevent the water freezing. Instead a difference of about 4Torr is required.


Vacuum Rise Test Illusion

Upon closing the service gauges for a vacuum rise test if, during that motion, the vacuum happens to rise just 0.3 Torr from point B to point C from 0.5 to 0.8 Torr, as is commonly seen when free water exists but is as commonly ignored, then the sublimation line moves 5K higher above the ice temperature to point D instantly subcooling the remaining ice. This means it is quite possible that any quantity of ice still in the piping could avoid detection by a mere 30 minute vacuum rise test and even for as long as 120 minutes under winter conditions.


Proposed Method Modifications

To ensure ice quantities formed do not also hide below a transitional sublimation temperature it seems that a 30 minute vacuum rise test should instead be carried out earlier after the first say 4hrs of vacuuming immediately after the first instant say 2 Torr is reached. Even with free water in the system it can take 20 minutes for the vacuum rise test to reveal its presence despite a prior vacuum period of merely 2 to 4 hours. If the initial 30 minute rise test detects moisture then another rise test done patiently over perhaps 24hrs should be carried out after each subsequent vacuuming session. Until the system passes any follow up vacuum rise test vacuuming should be done with the gas ballast valve open and the oil changed frequently. Each subsequent vacuum rise test need only be done over 24 hours if the pressure attained during each vacuuming session was below the triple point of water, approximately 4.6 Torr, under which all remaining water is likely to have frozen and so escape the detection by the traditional test method.

[Tesla]

There appears to be no misconception here - as in the field we do things a little different. Firstly we generally go by how deep the vacuum is (microns) as opposed to timing it dependent on the estimated size of the system and the pump. We also use different processes for the evacuation. We may use the triple evac method, the deep vac method, or/and incorporate a desiccant (chemical or mechanical) to remove the moisture, along with other gases which suck up moisture like nitrogen and refrigerant. Sure we take into account the amount of time taken but our main measure is the actual vac. We may be using the vac as a secondary gas tightness check.
Technically when there are other chemicals in the system - do the straight out water pressure/temperature relationships apply for calculation? Then there is the measurement - a big factor considering it will differ from the pump end to the other. Another factor is the oil (boiling point...) It's just not an exact science and calculable to last extent in a dynamic real world situation.
We do learn these scientific calculations but lets face it laws are made to be broken and scientists as us technicians are sometimes not 100% correct.
We welcome further discussion as this is an important topic - especially if there is improvement required in our process or calculation.

[DTLarca]

There appears to be no misconception here - as in the field we do things a little different. Firstly we generally go by how deep the vacuum is (microns) as opposed to timing it dependent on the estimated size of the system and the pump.

Let's say it's the largest VRF system possible with 3 outdoor units piped in parallel.

Let's say it's a new installation.

Let's say you are pulling a vacuum overnight in 0°C ambient conditions.

Let's say there happens to be 2kg of free water in the pipework because you were on the roof of a hotel installing 16 such systems with 2 other crews helping you and over the last few days there was heavy snow and sleet and one of the crew didn't realise that he had left open to atmosphere one of the many pipe ends.

[nevgee]

[quote=DTLarca;212349]Sorry, the site's image uploader does not work but the gist of the idea can still be grasped.


My Question

If we vacuumed a newly installed system to 0.5 Torr, over a period of 24 hours, is it possible to then see no rise in the vacuum within 30 minutes, after the vacuum pump has been isolated, even if there in fact is still free water in the pipework?


Simply ..... Yes I believe so only because I have experienced simillar situations, where further treatment / vacuum was needed.

You call it FREE water but do you meen water as ice?

I can follow your reasoning moderately, (I am a little dim) But I wonder that you might be driling down too deep theoretically for a practical apllication.

Tessla does make a reasonably comment that in practice events don't necessarily follow the rules. Physical laws may be fine in theory but the outside world applies many other variables that don't take part in following the rules.

As you state various properties of water may exist in the deep vac condition. Sublimation being one. However, should you not also be considering the physical state of the other components in the system. Oil for instance will have changes brought about it's physical state. how does that perform under the circumstances. Will it creat a partial pressure to causes background interference?

If moisture exists within the oil, will any amount of vacuum have influence on its presence. Will it boil off from the oil? will the depth of oil cause sufficient pressure to hold the moisture regardless of the amount of vacuum and heat applied?


I think you've raised a proverbial can of worms .. :D

[DTLarca]

Simply ..... Yes I believe so only because I have experienced simillar situations, where further treatment / vacuum was needed.

You call it FREE water but do you meen water as ice?

So would you agree that the vacuum pressure rise test should take place very soon after the vacuum process starts - as soon as, perhaps, the pressure reaches 2 Torr?

Free water very quickly turns to ice under vacuum.

I have finally been able to add the pictures to the original post.

[Tesla]

Hi DTLarca
In the event of the let's say (if) scenario you mentioned, most technicians from experience would notice that the required vac would not be reached in time. If a large pump had been used compared to the volume of the system - we would notice it would rise too much during a vac hold test. As nevgee has correctly noted there are many other considerations like depth of oil (not to mention surface tension). Story - I once had a 1600 kW chiller to vac and after realising my pump was too small I purchased the biggest I could. Then it had to run for 24 Hours to run in - in all it took 8 days to reach an acceptable vac with minor leaks taken to account.
Asides from that we technicians have intuition - a sort of touchy feely tool we use like seeing frost/condensation on a part of the system that we would notice then apply heat to that part to speed up the vac process. We do learn as part of our training the calculations you use along with other affecting points not to mention the experience passed on from our seniors. I have also worked on chillers that have burst evap tubes half filling the chiller with water and DX systems which sucked water in directly where other methods of dehydration were used.

[DTLarca]

If a large pump had been used compared to the volume of the system - we would notice it would rise too much during a vac hold test.

So we are assuming a new VRV system - therefore no oil exists in the newly installed pipework.

The engineers attach their vac and go home for the night.

Next morning they see the vac has got down to 0.7 torr and since 2 torr is for them acceptable they now do a vac rise test.

After 2 hrs they do not experience a vac rise and so assume there cannot be moisture in the system.

Two days later they call me because in heat mode the units are tripping on LP.

In this video I am thinking to myself "This has just got to be ice - it cannot be anything else - only water can do this" and at the same time I know the guys did a vac rise test for 2 hrs because it was witnessed by a site Foreman who on two previous occasions I had shown him how to witness a vac rise test.

In the end I explained to them that the icy stuff just has to be water that must have turned to ice and subcooled after they isolated the vac pump. The bubbling could only have been trapped R410A escaping.

http://www.youtube.com/watch?v=tlQ5gpbgilo

I was hired in as an independent to guide all pre-commissioning procedures. The guys on site - subbies to the installation company - had already successfully done several other vac rise tests with me. There is a longer story to the matter but essentially they or we were in the end let down by our blind reliance on the vac rise test principles which up till this point in my carrer have not ever stated when a vac rise test is best carried out - that was the point of my article.

[mad fridgie]

Yes this can happen, and the principle of sublimation is used in freeze drying.
If i go back to my early years in the trade, I was told by the old dogs, that you could not rely on vacuuming a system (for moisture removal) if part of the system was below 4C. ( at that stage I never questioned there expertise, and did believe that water did not vapourise below that temperature, knowing that to be incorrect now) So perhaps commissioning documentation should say the vacuum for moisture removal should be applied to system when temperature are above 5C.

[nevgee]

Lets get back to basics. ..... no theoretical wizardry no sadle sore razzamataz
The system is an isolated pipework installation. There are no other components or valved arrangements ... it is only a pipe work circuit with all condensers, condensing units, compressors, evaporators isolated and valved off.
The pipework is insulted, is isolated from all other components, has been purged, pressure tested and assumed to be leak tight and clean.
Then a vacuum pump is applied and after a period over night vacuuming down to 0.7 torr, a rise test after 2 hrs shows no rise in pressure. The system is set into operaton then after two days of running an lp trip occurs?

[nevgee]

Why then do you assume there is a problem with moisture in the system?

[Tesla]

I checked out the video and have seen this stuff before at a TEV blocked with frozen moisture and oil. Which leads to a few questions.
Was the vacuum measured correctly?
Where does the time of 2 hours come from for the vac hold test? I always thought it should be over night but I have mainly worked on bigger systems. I also think if the ballast was left open until 1500 microns was reached and then closed it helps remove the moisture without freezing but in your case you had very low ambient temps.

[DTLarca]

Lets get back to basics. ..... no theoretical wizardry no sadle sore razzamataz
The system is an isolated pipework installation. There are no other components or valved arrangements ... it is only a pipe work circuit with all condensers, condensing units, compressors, evaporators isolated and valved off.
The pipework is insulted, is isolated from all other components, has been purged, pressure tested and assumed to be leak tight and clean.
Then a vacuum pump is applied and after a period over night vacuuming down to 0.7 torr, a rise test after 2 hrs shows no rise in pressure. The system is set into operaton then after two days of running an lp trip occurs?

These were Toshiba 3-Pipe heat recovery systems. The outdoor unit isolation valves are not opened until the the strength pressure test, the leak pressure drop test and the moisture indicating vacuum rise test are all complete.

The unit was put into operation on a Friday. They called me on the Monday.

[DTLarca]

I checked out the video and have seen this stuff before at a TEV blocked with frozen moisture and oil. Which leads to a few questions.
Was the vacuum measured correctly?
Where does the time of 2 hours come from for the vac hold test? I always thought it should be over night but I have mainly worked on bigger systems. I also think if the ballast was left open until 1500 microns was reached and then closed it helps remove the moisture without freezing but in your case you had very low ambient temps.

The vacuum was measured with 2 different electronic Torr gauages - one of them known to be calibrated correctly for the project.

In the UK they talk of a 30 minute vacuum pressure rise test.

The ambients were low during the events unfolded - 0°C to -5°C - lots of snow in between.

[al]

So is adding heat the only way to guarantee this doesn't occur....very thought provoking.

I had a York RTU heatpump that burst a condenser in winter, roof was at -5 all week, vac was only achieved by gently applying oxy/ace, vac gauge showed 2 torr but i'd witnessed water ingress to the coil after repair!

BTW how do you fare with warranty now?

Alec

[Gingerair]

No offence to your engineers but i agree with Tesla, an 'experienced' engineer may have asked the right questions & 'sniffed' out a possible problem during the initial vacuum. It can be as important to see how a vacuum initially pulls down as a vac-rise test is once it has reached a suitable level.
It's not good practice to just put a vacuum on then skedaddle, especially when you're ambient conditions would have meant that you were up against it anyway.
If the water in the pipework was already ice before applying a vacuum you'd either need a mind-reader or an honest installer who might be good enough to let you know about a possible risk...

[DTLarca]

So is adding heat the only way to guarantee this doesn't occur....very thought provoking.

I had a York RTU heatpump that burst a condenser in winter, roof was at -5 all week, vac was only achieved by gently applying oxy/ace, vac gauge showed 2 torr but i'd witnessed water ingress to the coil after repair!

BTW how do you fare with warranty now?

Alec

Of course another question is "how accurate are my instruments?".

See if the table attached is of any use to you. Of course you might have a better one.

[DTLarca]

No offence to your engineers


Not my engineers - they were subbies hired by the company my customer hired me, just for the duration of the project, to oversee.


but i agree with Tesla, an 'experienced' engineer may have asked the right questions & 'sniffed' out a possible problem during the initial vacuum. It can be as important to see how a vacuum initially pulls down as a vac-rise test is once it has reached a suitable level.

These were firstly installers who admitted to me they ordinarily do not even carry out a vac rise test. Secondly I failed the vac on suspected moisture at the first witness. The engineer said he thought his oil needed replacement again. I went in the next day again to witness the vac rise and failed it again. This time the engineer said he did not leave it on vac all night because he was afraid his vac pump would be stolen. So I left them with my own pump which they left on the system for two nights and nearly 3 days. I could not make the next vac rise test witnessing but I left my gauge with the site Foreman and got him to demonstrate to me he knew how to carry out a vac rise test. Then I talked to him over the mobile each time he checked the Torr reading on the day so to the best of my knowledge he was doing everything right.


It's not good practice to just put a vacuum on then skedaddle, especially when you're ambient conditions would have meant that you were up against it anyway.
If the water in the pipework was already ice before applying a vacuum you'd either need a mind-reader or an honest installer who might be good enough to let you know about a possible risk...

I agree - and the article above hopefully invokes such in others.

[jimmy wilkes]

triple evacuation....ledgendery!!!

[Gingerair]

Hi DTLarca
I do sympathise with your situation, if there was a large volume of water present it would probably take more than 3 days on vac to remove..
Suppose part of your problem was identifying how much water had entered your system, as this would determine the steps taken to remove it.