PDA

View Full Version : TEV on long liquid line, liquid hammer







Peter_1
10-02-2010, 07:50 PM
I have a case with Alco TEV's (the XWB serie) where a contractor installed a liquid line of 1/2", 25 tot 30 m (75 to 90 ft) long and where in a timespan of 1.5 tot 2 years 5 TEV's failed. Capacity 20 kW (+/- 68200 BTU)
Freezer application and evaporators going each time in pump down when defrosting or temperature reached.
HP kept at 15 bar (215 psi)

The TEV's were each time closed when erratic and only opened again after removing the thermostatic element from their housing. They apparently functioning normal after reassembling.

TEV's (external equalization not connected)where tested in a lab where the bulb was submerged in glycol from -30C and heated to 20C (-22F till 60F)
They measured needle displacement and the displacement of the erratic ones was the same as a new TEV. The new TEV was also measured as a reference.

My theory is that the TEV's encountered a serious liquid hammer when opening the SV which injured somehow the diaphragm and/or the needle assembly of the TEV. I should expect then a complete opening of the valve but they were each time completely closed.

I would appreciate any comment on this issue.

taz24
10-02-2010, 08:08 PM
I have a case with Alco TEV's (the XWB serie) where a contractor installed a liquid line of 1/2", 25 tot 30 m (75 to 90 ft) long and where in a timespan of 1.5 tot 2 years 5 TEV's failed. Capacity 20 kW (+/- 68200 BTU)
Freezer application and evaporators going each time in pump down when defrosting or temperature reached.
HP kept at 15 bar (215 psi)

The TEV's were each time closed when erratic and only opened again after removing the thermostatic element from their housing. They apparently functioning normal after reassembling.

TEV's (external equalization not connected)where tested in a lab where the bulb was submerged in glycol from -30C and heated to 20C (-22F till 60F)
They measured needle displacement and the displacement of the erratic ones was the same as a new TEV. The new TEV was also measured as a reference.

My theory is that the TEV's encountered a serious liquid hammer when opening the SV which injured somehow the diaphragm and/or the needle assembly of the TEV. I should expect then a complete opening of the valve but they were each time completely closed.

I would appreciate any comment on this issue.


Hello Peter.

I have seen this happen before.

Is the liquid subcooled by chance ( or subcooled beyound normal ambient temps )

In one of the situations I know of it was with subcooled R22 on low temp cases and
the valves were about 2.5mtr from the TEV. The actual valves were not damaged but
the flares were knocked clean off the valve body.

We ended up by moving all the solenoid valves nearer to the TEV's (about 40 or 50 cm away)
and then we stopped having problems.

On another site the liquid hammer kept knocking the innards of out of the liquid / suction heat exchangers
out and the liquid was then entering directly into the suction line and flooding back to the comps :confused:.

Again the valves were moved and the heat exchangers removed.

taz

.

chemi-cool
10-02-2010, 08:37 PM
How far from the TXV is the SV located?

hyperion
10-02-2010, 09:21 PM
If the solenoid is local to the condensing unit, then it should be moved to be as near to the TEV as is practical to reduce the effect of the liquid slug when the solenoid opens. This should also slightly improve the pump down time as the effective liquid line to the TEV will be shorter.

Peter_1
10-02-2010, 09:57 PM
...
Is the liquid subcooled by chance ( or subcooled beyound normal ambient temps )

...
There's indeed a possibility of subcooling because a PHE is installed in the discharge to preheat water and this water can be as cold as 12C.
Liquid lines are +/- 25 to 30 m long, so in my opinion very long.

I was asking myself why this can close or block the valve completely where it seems to work again afterwards very well.

Superfridge
11-02-2010, 07:30 AM
Maybe charge migration? Check if valve power assembly is cross charged.

Peter_1
11-02-2010, 08:00 AM
2 which were blocked were tested afterwards in a lab and they functioned again for 100%. So no charge migration.

lawrence1
11-02-2010, 09:02 PM
Peter1,
I would agree with taz24 and move the solenoid to as close as possible to the tx valve.
Regards

al
11-02-2010, 09:35 PM
I also have a vague memory of some older cabinet installation instructions stating that a tee piece with a vertical section of a few inches be fitted just after the solenoid, something to do with vapour rather than liquid though......

alec

taz24
12-02-2010, 08:41 AM
I also have a vague memory of some older cabinet installation instructions stating that a tee piece with a vertical section of a few inches be fitted just after the solenoid, something to do with vapour rather than liquid though......

alec


Do you know I've seen those tee thingies on lots of
well freezers and I always thought it was the tee point for the hot gas bypass if fitted.

Eeeeeeeee int life great.
You learn somthing new each day.

Logicaly a little vapour lock above the liquid will absorb any hammer in the liquid itself
and because of the nature of the tee it will almost always have vapour inside it???

Thanks Alec.

taz

.

Peter_1
12-02-2010, 09:14 AM
Removing the SV as close as possible to the TEV was also what I recommended but can this liquid hammer provoke the described phenomenas, a complete closure of the TEV?

Chef
12-02-2010, 10:24 AM
Removing the SV as close as possible to the TEV was also what I recommended but can this liquid hammer provoke the described phenomenas, a complete closure of the TEV?

Peter - Have you thought about the fact that if you suddenly close the TEV you will get a P wave ie positive pressure caused by the waterhammer upstream.

But if a line valve upstream closes you are just as likely to get an N wave - ie negative pressure on the upstream side of the TEV.

Maybe that would make some differance to how you view the hammer effect.

Chef

Peter_1
12-02-2010, 11:46 AM
Chef, I can't follow you on this. Can you enlighten this a little bit please?

Chef
12-02-2010, 12:35 PM
I am assuming you have a valve upstream of the TEV but you must correct me if that assumption is wrong.

If the valve upstream suddenly closes then the the pressure upstream of that valve will possibly increase due to liquid hammer but downstream the pressure will suddenly reduce, It means in the scenario I discuss here that the TEV will see a sudden drop in pressure. So the P wave is upstream and the N wave is downstream of the valve - so the TEV may see a negative pulse of pressure.

As the TEV is generally governed by 3 forces, the bulb pressure, the spring, and the evaporator pressure it seems strange that any change in the upstream pressure would effect it much.

However the portion exposed to the upstream looks like a classic poppet valve and this has its own characteristics, the back face which is fully exposed to the upstream pressure is larger than the needle portion in the orifice. So changes in upstream pressure can easily effect the valve - although they are small surfaces in comparison to the diaphram the hammer effect can introduce very large pressures for short periods. It is possible this hammer - be it a positive or negative pressure pulse, is trying to open or close the TEV.

The reason it stays closed is yet to be discussed.

Maybe a P or N wave gets stuck inside the pipe between the valve and TEV and that stretch of line is sealed of and just holds the TEV shut? Only a theory.

Chef

Peter_1
18-02-2010, 07:28 AM
This is a pictures of the 2 TEV valves

mad fridgie
18-02-2010, 08:57 AM
I tend to think that you actually have a fouling problem. I do not not believe hammer will cause a problem, as the internal gause filter is likely to disperse the pressure waves (not to say the waves do not break pipes especially around short radius elbows)
I think you may have a built up of a very fine particulate, when you remove the guts you disturbe the debries (falls out, not seen)

Chef
18-02-2010, 09:24 AM
Peter - This looks like an electrical schematic, is it the correct picture of the TXV valve? Someting like a cutaway or parts schematic would be great.

Chef

Peter_1
18-02-2010, 09:36 AM
sorry, uploaded the wrong picture, corrected it now
Mad Fridgie, certainly not a fooling problem.

mad fridgie
18-02-2010, 09:51 AM
sorry, uploaded the wrong picture, corrected it now
Mad Fridgie, certainly not a fooling problem.
You said that during testing that equalising line was not connected, on site is the equalising line fitted before or after the bulb, if before and it is seeping it will close the valve prematurely. Have you refited the the old cages, or just bench tested?

Chef
18-02-2010, 11:08 AM
There seem to be a number of scenarios that may come into play here.

1st one
The SV closes and the 30m pipe of full of liquid is fed into the evap leaving just gas in the pipe. This is highly likely as the flow rate will fall dramatically from the 0.148Kg/s in normal operation to just a few tens of grams/s being driven by the higher temp liquid line and lower evap pressure.

The evap balances out (pressure and temp) and the TXV valve closes.

The SV now opens and liquid begins to flow from a reservoir, at presumably 15 bar, and assuming the 30m liquid line is full of gas at the evap pressure it will provide little resistance to the liquid coming down the pipe (some may even condense but in the period available for the liquid to arrive it would need time to be studied more)

The full pressure of the upline liquid is now on the poppet valve as shown in your diagram. This plus a possible extra impulse of around 10-11 bar (or as low as 7-8 bar) from hammer may jam the valve closed and stop it opening from the classic forces of bulb, spring and evap pressure. Here you may see a pulse of 25 to 26 bar on the valve end which is quite a significant impulse. The valve entry geometry may even increase this pulse!

If this is indeed true it may mean the poppet valve gets rammed into the seat and sticks a little so requiring a larger force from the diaphragm to open it.

It is very difficult to calculate all this as we dont really know the state of the liquid line in between the SV and TXV nor the stiction of the valve into the seat.

But it would appear the force on the valve poppet is about 3.1Kg closing force whilst the opening force is about 4Kg per deg C of SH.
Note these forces are non dimensional as I dont have detailed dimensions on the valve but they can be used as a comparison to each other.
It looks unlikely that the upstream pressure will hold the valve closed against normal SH rising as the system pumps down unless it is stuck.

Scenario 2 involves looking at the SH and evap pressures and the amount these can 'distance themselves' during pump down to provide sufficient force on the diaphragm to actually open the valve.

For this detailed knowledge of the evap configuration would be required.

Hope this helps at the moment.

Chef

mad fridgie
18-02-2010, 07:53 PM
If we look at the practical aspect of the valve, during pump down the internal evap pressure is very low and the suction temperature (at bulb position is high) The valve will be be fully open.
When the SV is opened, the first thing that is going to happen is that the liquid is going to flash in the liquid line (liquid line at evap pressure) The vapour will enter the valve first at a lower pressure than the stored liquid pressure (pressure drop through the liquid line but with a vapour pressure drop) If there is an internal leak into the equalising chamber and the equalising line is small (unable to release the pressure) then ineffect the valve "thinks" the evap pressure has risen to a high level, the bulb pressure will be equal to the room temperature. The valve will close very quickly (giving the impression of no superheat)
So is your equalising line a capilary or a 1/4 line.
The problem will only arise if the seal to the equalising chamber becomes damaged (over time).
Are the pressure pulses causing the damage?
You could install a small bleed solenoid in parralell with you main SV with time delay on your main SV, this would reduce any sudden surges

Peter_1
19-02-2010, 06:33 AM
The eq. line is indeed a 1/4"

What I proposed as a final solution is moving the SV near the TEV (on the ceiling of the freezer) or a bypass SV of 1/4" over the 1/2" one with a time delay over the 1/2" or open the 1/2" first 1", then wait 5", open it again 2" wait 5", and so on, so opening it gradually.

But I'm almost sure that there must be a relation between blocking of all these valves and the long distance between SV and TEV.

Pinpointing the real cause is extreme important because I have to point someone who's responsible for the loss of all the unfrozen goods (220,000 euros!)

mad fridgie
19-02-2010, 09:05 AM
If you have 1/4" equaliser line then pressure should not build up in the equalising chamber, so that is that theory out of the window. Bear with me!
There must be thousands of systems with longer pipe runs and SVs at the unit. I am not a user of Alco valves, so do not know how good they are.
you mentioned you had a de-superheater (this will not give sub-cooling, it will drop the head pressure) only if the water is cold.
De-superheaters can cause the discharge pressure to rise especially when the water is hot. If we look at the long pipe runs in conjuction with the pressure increase caused by the desuperheater, I suspect that you must have very high discharge temperatures. The high temps could be causing oil to break down. After a period your valve could be fouling with wax at freezer temperatures, when you remove the cage and bring back to room temp the wax reverts back to a liquid. it would be normal to see oil in the internals of a refrigerant system.
This may be a build up over a period of time, not an instantaneouse blockage.
The original description indicates fouling,

Chef
19-02-2010, 12:00 PM
If we look at the practical aspect of the valve, during pump down the internal evap pressure is very low and the suction temperature (at bulb position is high) The valve will be be fully open.

When the SV is opened, the first thing that is going to happen is that the liquid is going to flash in the liquid line (liquid line at evap pressure)

Mad Fridgie - I was assuming the valve and compressor started/opened at the same time but that may not be realistic, even so if they do operate simultaneously then the action at the SV end will all be over in less than a couple of seconds and the compressor may not yet have started a useful pump down, not in 2 seconds.

Your second point is most valid and the flash gas and also the resident gas in the liquid line between the SV and the TXV will act as a big spring defusing the hammer effect. It is a most difficult to analyse it as there is the flash gas, resident gas, acceleration of liquid down the line caused by the upstream pressure, pipe friction and the spring effect of the trapped gas.

This will take some time, but as a quick take it seems the liquid might accelerate to up to 20 or 30m/s before the other factors start to slow the pulse down and the pressure could be from anything from 5bar over supply pressure (with lots of gas effect) to 194bar (if the liquid actually slams into the TXV.)
The latter is very doubtful so the best estimate at the moment is around 30 or 40 bar with a pulse duration of just .2 seconds, not really enough to hold the valve closed I think.

The long run is not ideal and getting the SV closer to the TXV may be a solution but how much damage is done to the TXV each time it sees this pulse is a question.

So far there is no reason to suspect hammer is the major cause.

Chef

Chef
19-02-2010, 12:19 PM
The eq. line is indeed a 1/4"

What I proposed as a final solution is moving the SV near the TEV (on the ceiling of the freezer) or a bypass SV of 1/4" over the 1/2" one with a time delay over the 1/2" or open the 1/2" first 1", then wait 5", open it again 2" wait 5", and so on, so opening it gradually.

But I'm almost sure that there must be a relation between blocking of all these valves and the long distance between SV and TEV.

Pinpointing the real cause is extreme important because I have to point someone who's responsible for the loss of all the unfrozen goods (220,000 euros!)

Maybe you could use a motorised valve which would open slower - very common in the steam and petrochem industries to avoid large pressure pulses.

Mad does have a good point about the wax and oil breakdown - I had to replace a valve just recently blocked by dodgy oil.

Do you want me to continue to develop some analysis of the liquid line and the SV effects as it it opens?

Chef

BESC5240
19-02-2010, 01:24 PM
2 which were blocked were tested afterwards in a lab and they functioned again for 100%. So no charge migration.

Hi Peter,
charge migration is completely reversible. So a valve blocked due to charge migration (bulb charge completely migrated to the thermostaic element) will function again for 100% once the charge has re-migrated to the bulb. Submerging the bulb in the -30C glycol bath is enough to re-migrate the charge.

And because charge migration can be initiated and amplified by the pumping effect of fluctuating pressure difference (due to the unstable feed towards the TEV), I tend do go for charge migration as the cause for the blocked valves.

Best regards
Bruno

mad fridgie
19-02-2010, 08:09 PM
The eq. line is indeed a 1/4"

What I proposed as a final solution is moving the SV near the TEV (on the ceiling of the freezer) or a bypass SV of 1/4" over the 1/2" one with a time delay over the 1/2" or open the 1/2" first 1", then wait 5", open it again 2" wait 5", and so on, so opening it gradually.

But I'm almost sure that there must be a relation between blocking of all these valves and the long distance between SV and TEV.

Pinpointing the real cause is extreme important because I have to point someone who's responsible for the loss of all the unfrozen goods (220,000 euros!)
If you open and close the valve then all you will do increase the number of pulses, as Chef says the pulse will last 0.2 of a second.
A couple of other options, that may assist, is installing a drier in the room just prior to the TEV, This will aid in moisture removal, and could solidify any wax that maybe in the system.
Reduce the orifice size and install a capillary in parralell with the TEV (sized for part load). Again this is a bandaid approach, but would keep the system running if TEV failed.
The client should fit a remote monitored alarm.
In the good old days, we always fitted 2 systems each at 66% capacity, so if one failed the other would be able to limp through until the break down was fixed.
Your client has to take responsiibilty for his product, if he does not have the protection to cover break downs (which will occur) then it is his fault. A bit harsh I know, but there isa bit of difference between a small fidge with a few euro in it and 220,000euro.

Peter_1
19-02-2010, 09:49 PM
....

Do you want me to continue to develop some analysis of the liquid line and the SV effects as it it opens?

Chef

If it's not too much work for you, then I should appreciate your efforts.

mad fridgie
20-02-2010, 01:07 AM
If we look at the pump down, your suction temp at the compressor, could reach close to ambient (during pump down) plus the head pressure could rise (if you are flooding the condensor and depend upon the reciever size and system charge) if this is the case your discharge temp (at point of compression) could reaching close to 150C. Even though pure oil should not break down at these temps, the slightest impurity (which most system have) may cause it fracture at these temps.

Magoo
20-02-2010, 02:40 AM
Peter 1
quick and easy, install a small by-pass solenoid valve around the main SV to increase down stream pressure, fit a timer to open the main SV, pipe hamer gone.

Chef
20-02-2010, 06:37 AM
If it's not too much work for you, then I should appreciate your efforts.

I think I will leave out all the equations at this time (unless you want them) and just describe the process.

So to calculate this we need to look at the refrigerant accelerating down the 30m pipe as the SV is opened.

This involves momentum and is powered by the upstream pressure, P1, and the downstream pressure, P2.

But as the slug of liquid passes along the tube to the TXV it compresses the gas in the pipe and also flash gas is added so P2 continually rises. This is treated as an isentropic process at the moment.

This acts like a big spring slowing the rush of liquid down. When P2=P1 the liquid will not accelerate any more but will now have Kinetic energy as it is moving at a good speed. ie 0.5*m*v*V

This kinetic energy continues to compress the gas until P2 is higher than P1 and slug of liquid is slowed down to zero m/s. At that point the higher P2 begins to reverse the flow in the pipe. But also the higher pressure of P2 will now cause some gas to re-condense.

All of these variables have been accounted for and the resulting plot is shown below.

The distance curve is the face of the liquid plug versus time down the tube (red curve) and so at about 0.6 second it is 10m down the pipe from the SV valve and at a velocity of nearly 20m/s. (blue line)

Looking further in time at about 1.9 seconds you will see the P2 reaches its maximum of 93 bar and liquid face has reached about 25m down the pipe. At this point the liquid stops and then starts to reverse as the P2 pushes it back down the pipe (hope you do not have a non return valve in the line somewhere!!!!!)

The initial conditions were P1=15bar, density=1000Kg/m3 viscosity 11E6pa.s P2=1bar speed of sound=550m/s and time steps of 10milliseconds per iteration.


3365

It is easy to change the variables if you want other conditions.

Chef

mad fridgie
20-02-2010, 06:48 AM
I think I will leave out all the equations at this time (unless you want them) and just describe the process.

So to calculate this we need to look at the refrigerant accelerating down the 30m pipe as the SV is opened.

This involves momentum and is powered by the upstream pressure, P1, and the downstream pressure, P2.

But as the slug of liquid passes along the tube to the TXV it compresses the gas in the pipe and also flash gas is added so P2 continually rises. This is treated as an isentropic process at the moment.

This acts like a big spring slowing the rush of liquid down. When P2=P1 the liquid will not accelerate any more but will now have Kinetic energy as it is moving at a good speed. ie 0.5*m*v*V

This kinetic energy continues to compress the gas until P2 is higher than P1 and slug of liquid is slowed down to zero m/s. At that point the higher P2 begins to reverse the flow in the pipe. But also the higher pressure of P2 will now cause some gas to re-condense.

All of these variables have been accounted for and the resulting plot is shown below.

The distance curve is the face of the liquid plug versus time down the tube (red curve) and so at about 0.6 second it is 10m down the pipe from the SV valve and at a velocity of nearly 20m/s. (blue line)

Looking further in time at about 1.9 seconds you will see the P2 reaches its maximum of 93 bar and liquid face has reached about 25m down the pipe. At this point the liquid stops and then starts to reverse as the P2 pushes it back down the pipe (hope you do not have a non return valve in the line somewhere!!!!!)

The initial conditions were P1=15bar, density=1000Kg/m3 viscosity 11E6pa.s P2=1bar speed of sound=550m/s and time steps of 10milliseconds per iteration.


3365

It is easy to change the variables if you want other conditions.

Chef
Hi chef, not going to argue with the maths, are calcs based upon a dead head? Have you allowed for the TEV being open (normal operating conditions, before the fault occurs) If not was influence does this have.
Would this not reduce, peak pressure and thus reduce reverse pressure wave pressure.
Always love to see your analitical brain in overdrive.
" I should have paid more attention at school, well more attention to the teachers, not the girls"

Peter_1
20-02-2010, 07:18 AM
...
Always love to see your analitical brain in overdrive.
" I should have paid more attention at school, well more attention to the teachers, not the girls"

I couldn't agree more

Peter_1
20-02-2010, 07:29 AM
If we had charge migration like BESC5240 - what could cause charge migraton Bescie? - and we had therefore a blocked valve, then this must be the condition to have a theoretical pressure of +90 bar. Am I right?

TEV valve is installed on the ceiling of the freezer, bulb (isolated with Armaflex) is attached on the suction line outside the ceiling, so in a temperature of let's say 20C.

This is for me an interesting problem

Peter_1
20-02-2010, 07:36 AM
If we had charge migration like BESC5240 what could cause charge migraton Bescie? - and we had therefore a blocked valve, then this must be the condition to have a theoretical pressure of +90 bar. Am I right?

TEV valve is installed on the ceiling of the freezer, so in a temperature of let's say 20C.

Chef
20-02-2010, 07:41 AM
Hi chef, not going to argue with the maths, are calcs based upon a dead head? Have you allowed for the TEV being open (normal operating conditions, before the fault occurs) If not was influence does this have.
Would this not reduce, peak pressure and thus reduce reverse pressure wave pressure.
Always love to see your analitical brain in overdrive.
" I should have paid more attention at school, well more attention to the teachers, not the girls"

When you say dead head I presume you mean the upstream pressure is constant at 15bar and if so yes this is what I assumed.

Also assumed the TXV was closed but if it was open as you say it may reduce the gas pressure P2. But correct me if I am wrong - if the temps and pressures in the evap equalise will the valve be open or closed? I assumed it would display low or no SH and so close.

As for the chicks, well the 'Le Grande Fromage' is cooking dinner so I have to do this in between appreciating the aroma and the tenderness of the meat.

Overdrive (read overload) will arrive when Peter wants surface tension, pipe bulk modulus, bends and reflected waves included.

Chef

mad fridgie
20-02-2010, 07:43 AM
Do you know what oil was originally used. Had a quick look, pour point temp might be an issue.

Chef
20-02-2010, 07:49 AM
Peter has just answered this TXV open or shut question as the valve and bulb are at 20C so then it must be open. Will redo the calcs and include gas feed through the valve.

Chef

mad fridgie
20-02-2010, 07:58 AM
When you say dead head I presume you mean the upstream pressure is constant at 15bar and if so yes this is what I assumed.

Also assumed the TXV was closed but if it was open as you say it may reduce the gas pressure P2. But correct me if I am wrong - if the temps and pressures in the evap equalise will the valve be open or closed? I assumed it would display low or no SH and so close.

As for the chicks, well the 'Le Grande Fromage' is cooking dinner so I have to do this in between appreciating the aroma and the tenderness of the meat.

Overdrive (read overload) will arrive when Peter wants surface tension, pipe bulk modulus, bends and reflected waves included.

Chef
Sorry for not explaining, dead head, i meant closed TEV "down stream"
The TEV at this point will be fully open(ish), evap pressure will be around 0Barg (pump down) suction temp (at bulb postion) could be at 20C, external to chiller area (new info) so superheat would be nearly 65C (off the top of my head) valve would be open.
They evap would act like an pressure accumulator.
The vapour passing through the valve is compressorable, so absorbing energy whilst flowing at a lesser rate than the presssure energy absorbed.
One would presume that flow would be limited to the "choke point"

mad fridgie
20-02-2010, 09:02 AM
If we look at the TEV in the off position, 65C superheat, there must be great pressure on the needle/shaft/orffice look at the area of the therm bellows, and the area of the needle. i would expect the pressure excerted is 100s of bar, so regardless of the pulse pressure, I can not see the valve being pushed into the body causing it to seize and stop flow.
can not see the charge reversing as the ambient (where the body is ) is always higher than the suction temp.
I still say it is fouling!

Peter_1
20-02-2010, 01:06 PM
Problem is that I haven't dismantled these valves. The technician did it and they were send straight to the lab.
I haven't seen neither the tests in the lab

mad fridgie
20-02-2010, 07:03 PM
Problem is that I haven't dismantled these valves. The technician did it and they were send straight to the lab.
I haven't seen neither the tests in the lab
Even more difficult.
My suggestion, for what it is worth.
Have the oil tested,
Do not pump down.
Control SV with compressor.

Chef
21-02-2010, 05:30 AM
Mad Fridgie asked to see what effect the valve being open had on the transient.

Well looking at 3 cases - high, medium amd low flash rates the valve being open had no effect on the high flash rate at all because there is so much gas available and pressures are low plus the flow through the valve is small.

For the medium case it did show a slight change as the valve bleeds away gas and so lowers the line pressure and allows the liquid to get further down the line before it it is slowed so in fact it makes it a bit worse than a closed valve.

For the Low flash rate the flow out of the valve is very influential. As the pressure rises so the flow through the valve is high and this makes the liquid even closer to the valve and some pretty high impacts are shown in the graphs below. This Low flash rate could be due to some large subcooling.

The graphs go off the roof if subcooling is increased much more and could be enough to pop the valve shut, even with the high bulb temp.

The main problem is how much flash gas do you get in around 200milliseconds, I dont know of any work done in that area but would love to be pointed to it if there is? The other unknown is the condensation rate which is assumed as the reverse of the flash rate.

3366

3367

3368

Also checking the pulse with various lengths of pipe is seems that an SV to TXV distance of less than 2 m is ideal.

Chef

Chef
21-02-2010, 05:42 AM
If we look at the TEV in the off position, 65C superheat, there must be great pressure on the needle/shaft/orffice look at the area of the therm bellows, and the area of the needle. i would expect the pressure excerted is 100s of bar, so regardless of the pulse pressure, I can not see the valve being pushed into the body causing it to seize and stop flow.
can not see the charge reversing as the ambient (where the body is ) is always higher than the suction temp.
I still say it is fouling!

With such a high SH I tend to agree that it's unlikely the valve gets shut buy a pressure pulse, it would need to be in the top hundreds of bar to do so and then it has to 'stick' shut against many hundreds of kilos of opening force. It does seem unlikely.

But you seem to be in a 'think out of the box' mode so any other ideas? It certainly is a curious one.

Peter - what is the distance from the reservoir to the SV valve as thsi may be an effect too?

Chef

mad fridgie
21-02-2010, 06:05 AM
Mad Fridgie asked to see what effect the valve being open had on the transient.

Well looking at 3 cases - high, medium amd low flash rates the valve being open had no effect on the high flash rate at all because there is so much gas available and pressures are low plus the flow through the valve is small.

For the medium case it did show a slight change as the valve bleeds away gas and so lowers the line pressure and allows the liquid to get further down the line before it it is slowed so in fact it makes it a bit worse than a closed valve.

For the Low flash rate the flow out of the valve is very influential. As the pressure rises so the flow through the valve is high and this makes the liquid even closer to the valve and some pretty high impacts are shown in the graphs below. This Low flash rate could be due to some large subcooling.

The graphs go off the roof if subcooling is increased much more and could be enough to pop the valve shut, even with the high bulb temp.

The main problem is how much flash gas do you get in around 200milliseconds, I dont know of any work done in that area but would love to be pointed to it if there is? The other unknown is the condensation rate which is assumed as the reverse of the flash rate.

3366

3367

3368

Also checking the pulse with various lengths of pipe is seems that an SV to TXV distance of less than 2 m is ideal.

Chef
Very Interesting.
If the unit is off the reciever will be at equalibrium, so apart for the static head of liquid within the reciever and the line between the reciever and the SV, there will be no sub-cooling. this would reduce the likelyhood of low flash rate.
What would happen if we had a slug of liquid,( before the wave) followed by vapour, then liquid.
Would the pressure pulse increase the hydrolic pressure within the original slug,

Chef
21-02-2010, 06:23 AM
What would happen if we had a slug of liquid,( before the wave) followed by vapour, then liquid.
Would the pressure pulse increase the hydrolic pressure within the original slug,
I take the point about the SC being low as equilibrium has set in, but the rate of flashing is still unknown and as it all happens in 100 to 200 milliseconds it is a pretty small period to deal with, turn a bottle of R12 upside down and open the valve - liquid everywhere and does not just go 'flash' so its a big query in this. Oh I remember R12 is no longer available - OK forget that test and pretend.

As for the slug/gas/slug difficult to see how it arises but the answer is you may likely get 2 pulses. If the liquid hits the TXV the the pulse is rho*C*V where density, speed of sound and velocity come into play. If it is a liquid slug running into a wall of high pressure gas it is just V*V so a much lower pulse.

More likely the plug of liquid would disperse at the speeds in operation though.

Chef

mad fridgie
21-02-2010, 07:31 AM
I am trying to get a feel for the system.
Question for Pete
When you arrived at site, did you find the unit cycling on the LP switch "Off for a while, on for a short while"

How did your remove the TEV,
as a total valve, or as a cage/orifice only, or cage and thermostatic element.?
Did you pump the system down (closing liquid reciever outlet hand valve)
Did you get freeze burn on your hands?
Did you add refrigerant after changing the valve?
Has oil ever been added to the system (which oil)?
When you arrive is the head of the compressor "hot"?
Did you change the drier each time.
Have you has more or less problems with alco, over other valves (I have not used Alco sorry)

Please indicate any short cuts you may have taken

Peter_1
21-02-2010, 08:32 AM
....
Peter - what is the distance from the reservoir to the SV valve as thsi may be an effect too?

Chef
+/- 2 m and vessel is below the SV

Peter_1
21-02-2010, 08:42 AM
..
When you arrived at site, did you find the unit cycling on the LP switch "Off for a while, on for a short while"

How did your remove the TEV,
as a total valve, or as a cage/orifice only, or cage and thermostatic element.?
Did you pump the system down (closing liquid reciever outlet hand valve)
Did you get freeze burn on your hands?
Did you add refrigerant after changing the valve?
Has oil ever been added to the system (which oil)?
When you arrive is the head of the compressor "hot"?
Did you change the drier each time.
Have you has more or less problems with alco, over other valves (I have not used Alco sorry)

Please indicate any short cuts you may have taken
Well, I came in the picture just now and this happened all in 2007. It is a case for the Court. I just have what's in emails, papers, loggings, that's all....
Only the thermostatic element was removed, not by me but in 2007 by the technician.
What happened: Wednesday morning failure of freezer, Saturday evening, call to technician. He told me the pack went continuously in in LP with 3 TEV's on this pack. Because it was already late (10 PM) , his first idea was lack of gas. Added 60 kg of gas and told customer that it should take some time to freeze down the already unfrozen goods. End-user called him back Sunday morning because temperature then was still as high as Saturday evening. Sunday morning, he replaced the 2 Alco TEV's and from then on, you see a steady dropdown of the temperature.
No oil was added.
What could have also be happen is a filterbreakdown. Another poster in a Belgium forum posted a picture of the RE forum with all the drying particles in the TEV body.
For me, an Alco is even superior than a Danfoss, a Danfoss TE5 reacts relative slow.
So when I came in the picture some months ago, all these problems already happened. But after this dramatic failure, it happened once again with both TEV's and it are these TEV's which were examined.

mad fridgie
21-02-2010, 09:03 AM
Well, I came in the picture just now and this happened all in 2007. It is a case for the Court. I just have what's in emails, papers, loggings, that's all....
Only the thermostatic element was removed, not by me but in 2007 by the technician.
What happened: Wednesday morning failure of freezer, Saturday evening, call to technician. He told me the pack went continuously in in LP with 3 TEV's on this pack. Because it was already late (10 PM) , his first idea was lack of gas. Added 60 kg of gas and told customer that it should take some time to freeze down the already unfrozen goods. End-user called him back Sunday morning because temperature then was still as high as Saturday evening. Sunday morning, he replaced the 2 Alco TEV's and from then on, you see a steady dropdown of the temperature.
No oil was added.
What could have also be happen is a filterbreakdown. Another poster in a Belgium forum posted a picture of the RE forum with all the drying particles in the TEV body.
For me, an Alco is even superior than a Danfoss, a Danfoss TE5 reacts relative slow.
So when I came in the picture some months ago, all these problems already happened. But after this dramatic failure, it happened once again with both TEV's and it are these TEV's which were examined.
I would say that your valves are fouled, not totally blocked, if pack was in and in on LP (I presume you mean in and out on LP) This is indicating some flow.
My Gut feeeling is that during pump down you are reaching temps close to -47C, if the wrong LT oil is used,(pour point -46C if pure, more than likely warmer) the valve will start to clog.
As the valve is on the roof (not in the freezer) most of the time the oil will warm and pass through the valve. But if your SV short cycled (presume controlled by thermostat) closed on stat, then for example door was opened, stat would make open SV valve. The oil would not have time to melt, some flow would occur, but due to partial flow very low suction pressures (temps) keeping the oil blocking the valve.
"Same could apply with to very low levels of moisture."
Which is "always" present in these types of systems.
I do not believe that liquid hammer is the problem.
I think this can be proven with an independent oil test. (high end)

mad fridgie
21-02-2010, 09:33 AM
Practically pressure from liquid would be approx 10barg at 20C ambient, so pressure pulses would less.
What Chef calcs did prove was pressure spikes sure can be a problem (breaking of piping) a common problem!
I would hope the client has installed an Alarm system, surely he did not loose product a second time?
There are thousands of system with long pipes runs as you have described.
(I remember Littlewoods systems were always like this 25 years ago, much longer pipe runs)
In court, I do not know how you can prove negligence. It may not be a perfect design, but what is!
On the original call the tech, added 60Kgs of refrigerant (seems like alot), was it removed. maybe there was a leak, some moisture had been drawn in causing the valves to block. No reduction it temp.
Was a leak found, were driers replaced at each service call (technical should always be done)
this you can find on the service invoices.
You can not blame the tech for the first lot of product lose, as already unfrozen before he arrived.

Peter_1
21-02-2010, 09:37 AM
I have 1 week time to give my report and I will add this as an additional advice. Thanks for you ongoing support you both.