Brine Pumping Problems - Hycool 20

[Brian_UK]

Question regarding the pumping of a secondary refrigerant namely Hycool 20, potassium formate for -20°C.

Basis:
We have a brine chiller, Frigabon, with four compressors/evaporators serving gravity coolers in a chilled working area and a +2°C cold room. Set point is -7°C for the brine and is achieved and held without problem.

The pump set is based on two Grundfos inverter driven pumps with a low loss header between the two circuits.

Circuit 1 has the pump at a fixed speed and circulates the brine around the chiller and the low loss header.

Circuit 2 has the pump controlled by a differential pressure sensor, across the pump, to maintain constant pressure. It takes the brine from the low loss header, pumps it around the load circuit and returns it to the header.

The secondary load side has seven separate circuits fed via two port controls valves. There is a permanently open bypass circuit across the secondary header circuit to maintain some circulation when the valves are closed.

This bypass also passes water as the various valves close.

Without taking the insulation off the pipes I am guessing that the secondary header pipe is either 3" or 4" diameter (75 or 100mm) and the bypass is only 3/4" or 1" (20 or 25mm).

When the brine is going through the bypass line there is considerable noise being generated where it passes through the regulating valve orifice.

System static pressure with pumps off is 1.5bar; with pumps running the suction pressure at the low loss header would be around 0.8bar. Discharge pressures are roughly 1.0bar above this. System is a closed system with small expansion vessel.

Problem:
At certain flow conditions such as with one control valve open and the remainder closed, brine rushing through the bypass with a lot of noise, the system suction pressure will fall to around 0.2bar to 0.4bar and hold at that for some time.

This will cause the brine system low pressure switch to trip and shut down the system. Obviously not good for the the end product.

Thoughts:
I don't know enough about brines but from what I have read they can form a slurry if the temperature gets near their lower limits and also that they should not pass through too small restrictions.

Do these brines act like our regular refrigerants when throttled through a small orifice such as my bypass valve and pipe?

If they do then is it possible for my -20°C brine to turned into a partial slurry because the operating temperature of -7°C is having a pressure/temperature drop through the orifice?

If I assume that a slurry is being formed then is this physically reducing the system brine volume and thereby reducing the system static pressure?

[Magoo]

Hi Brian_UK,
the specific density changes with brine as solutions drop in temp., and there thermal conductivity changes as well. so with a bypass low load situation the brine solution changes, and concentrates, less heatexchange going on, chiller conditions change. The noise is because the solutions is getting denser/ lumpy.
from memory Carrier say a chiller has a rating drop with a glycol solution rated at -15'C solution mix of approximately 15%. Lower the solution temp and rating drop is fairly linear. All to do with solution spec density.

[MikeHolm]

At first glance, part of the problem might be a NPSH issue. If you could raise the static pressure by .5bar you might find the problem reduce somewhat. There should not be a problem on the glycol side. The other thing to look at is the location of the expansion tank. It should be a bit bigger and should be on the suction side of the pumps. Any pump head developed will be added to the static and the suction head pressure should be at the expansion tank pre-pressure (which should be increased to the new static pressure of the system).

If the exp tank is on the discharge side the pump pressure will be subtracted from the static and you will get closer to cavitation and you will have increase noise and the LP switch will be affected.

Hope this helps.

[mad fridgie]

You can produce slurry using very low pressures "Vacuum ice", but you are not in this area. However you are very likely to have very high velocities through the one valve, which can cause very small pockets of a very deep vacuum, and cause the fluid to boil. When the vapor bubble implodes is causes a bang. You have lots of these little implosions, hence the noise around the valve,

[MikeHolm]

As MF says, Cavitation, implosions.

[Brian_UK]

Thanks guys, some more information to take on board.

The problem with the restriction through the bypass makes sense to me.

I am still having trouble getting my head around the drop in working pressure. The low pressure reading is taken about half way up the low loss header which is a vertical tube with the pump connections at top and bottom sides of the tube.

'Getting lumpy' and 'denser' I can see but is it the action of getting lumpy that could be reducing the water content of the brine and therefore the physical size of the fluid. I'm not explaining this very well, sorry.

I may need to change the pump 2 from a maintained pressure running mode to a controlled pump curve mode to see if this reduces the velocity at low loads.

[install monkey]

can u fit a stat to stop the pumps when brine reaches a low setpoint ie when all zone valves are at setpoint the secondary brine temp will plummet if a stat stops the pumps this will be a cheaper option than replacing pumps for a constant speed .or can the inverter pumps slow down when output pressure rises?

[Brian_UK]

In theory the secondary brine pumps, load circuit, should shut down after 5 minutes of all of the valves being shut. In fact both the primary and secondary are shut down which then shuts the chiller off on a flow switch failure mode.

We have got the BMS guy to shut the chiller down first and then the pumps which get rid of that annoying automated call-out function.

Both pumps have inverter drives but the chiller circuit pump (no.1) simply runs at a set speed, we maintain 1bar across the chiller.

Why do you say that the brine temp will plummet?

[install monkey]

no cooling rejected to the evaps so the secondary circuit will absorb cooling from the primary circuit but with all the coolers satisfied theres no heat to absorb

[mad fridgie]

I would suggest that your expansion tank is to small or in the wrong position, and it also may be being effected by the cold temps.
Why.
It would seem that you are maintaining temperature, so it would follow that there is little change density/volume. so any small temp changes in control and changes in density should be absorbed by the expansion vessel, but when you are under load, a good proprortion of the actual fluid will be over your set point, greater volume. When you have no load "bypass" all the fluid will be close to setpoint, reduced volume, and it would seem that your expansion tank can not handle it, so reduced pressure occurs.
Do you have the expansion tank remote mounted, well it should be will a good length of pipe even with a heater tape

[MikeHolm]

If I see the schematic correctly, the exp tank should be on the low loss header and as MF says keep it good distance away. Also, when you change the tank make sure the diaphram can handle the cold temps long term. Some are better than others, EPDM is worse than SBR at those temps but the difference may only be an issue in the very long term.

Given the nearly constant pumping, I would have thought that there would not be any lumpy issues in the brine, we don't typically find them in the 40% propylene glycol/water that I am used to. How is this brine different? I still think an increase in static would be helpful.

[Brian_UK]

I would suggest that your expansion tank is to small or in the wrong position, and it also may be being effected by the cold temps.
<snip>

Yes, I understand now, thank you.

If I see the schematic correctly, the exp tank should be on the low loss header and as MF says keep it good distance away.<snip> I still think an increase in static would be helpful.

That's correct, there is a flexible hose from the header centre line to two 12litre(?) vessels, I think they are set at a cold fill pressure of 1.5bar.

Thanks all for your help and advice, it is really useful.

[MikeHolm]

I forgot to mention that,if you increase the pressure, to take a bike pump and boost the tank pressure to equal the static. Don't know the liquid volume but I would think there should be at least a 50l tank given the header size.

[Tradewinds]

Hi Brian,
I tend to agree with Mike about raising the pressure.

1. Is the pressure in the bellows correct? We use to isolate the expansion tank. Drain it, then read the pressure inside the bellows with a car tyre guage and match it to the manufacturers data. Then pump it up with OFN to the correct setting and re fill the tank.

2. Have you eliminated the possibility of air in the system. This would be an easy fix and will effect both flow rates and cause noise across valves etc. Vent from the highest point in the system - both with pumps running and with pumps off to allow air to migrate to top of pipework.

3. Is their enought glycol in the system?

4. Are the flow rates correct?

Is this a new installation? If not, was it working ok in previous years? If so then have a look at the points above before changing anything.

[Brian_UK]

Hi Tradewinds, thanks for joining in.

1. As far as I know, yes it is correct and I use the same method.

2. System is well vented, no air bubbles in the system. There is an air separator in the top pipe which I have removed the AAV from and looked into the isolating valve, full of fluid.

3. Quantity of brine will need to be increased to force a rise in system pressure. Note it is not a glycol solution.

4. I am assuming that flows are correct, The pump set is designed by the chiller manufacturer. I have a pump remote controller on order so that I can interrogate the pump to find it's settings.

Newish installation, about 15 months old now and no it has had problems from day three; sound familiar?

[MikeHolm]

Someone messing with it??? or there are some other changes such as bladder on an exp tank is toast or the impeller is not in the shape it was at the beginning (cavitation can do that over time as can other deposits on the impeller).

[PaulZ]

Hi Brian
You have said you have -20 brine, is this the coldest you can run or is it the freezing point?
If you are operating at -7 and depending on the temp drop across the chiller probably about 3 degrees your brine could be too strong. Freezing point should be about 4 degrees colder then the coldest temp out of the chiller.
We have found the stronger the concentration the harder it is to pump around the system and this can cause problems with the suction on the pump. Basically the brine is too thick. We mainly use glycols and alcohol brines, presume the same rules apply to potassium formate.
Paul

[Brian_UK]

Hi PaulZ

Good point, -20 is the coldest that the brine can run and -7 is the set point of the chiller.

Also for added info.. chiller is rated at 200kW @ -8°C and the circulation pump is at 58cu.m/h.

I need to recalculate the expansion vessel sizing as I think it is undersized at present.

[MikeHolm]

Having a 1 bar drop over the pump with a 1.5 static is quite tight when the density of the material changes so much from the fill density. I will bet the bladder in the small tanks (I assume they are bladder) will be right up to the port most of the time driving suction pressure lower.

[Brian_UK]

Agreed Mike, I have had the expansion vessel manufacturer size up a vessel today. The recommended is 60 litre size.

I have shown the client this thread today as it was easier than trying to explain each piece of advice given.

So, I think the synopsis is that we fit a larger expansion vessel (60l), pre-charge it to 2.0bar, top up the brine charge to provide a static pressure of 2.0bar.

Pre-charge the expansion vessel to 1.5bar (?)

[MikeHolm]

Great... and get out the ol bike pump or nitrogen cylinder to charge it up the the new static pressure. Let us know how it works out. I learn so much from you guys so it's nice to be able to use my knowledge in a good way.

[Tradewinds]

Hi Mike, I have a question for you. I was under the impression that we should only use pressurization units on sealed systems and that pressure vessels are only for open systems. ie gravity fed in multi storey buildings. Is that correct?

Is that the case here as Brian had said 'System is a closed system with small expansion vessel'.
Should the closed system have a pressurization unit to help maintain the pressure?
I'm just curious and may be wrong.

[MikeHolm]

Almost all the systems I see here now are closed loop, pressurized systems (with the exception of some cooling tower systems). In the old days we had vented heating, which I assume to you is an open pipe in the attic or through the roof so the system pressure is the weight of the liquid. There would be a tank above the highest radiator to accept liquid expansion.

In the pressurized systems we used to have tanks without bladders pressurized to approx 1bar (for a small building) but the problem with these is that over many cycles the air in the tank would become entrained in the water and the tank would get "water logged", and the air would get to the highest point in the system to get bled off. With an auto water feed set to 1bar and no expansion capacity the pressure relief valve (usually set to 2 bar) would blow and the system would be messed up. Pressure would go way down and the feed would open up to 1bar again.

Almost all systems over here now use the bladder tanks to keep the separation between liquid and air so we then have a much wider pressure operating range. I cannot think of a gravity system installed here post 1940 or 1950 and if they were, they were probably converted to pressurized systems sometime since.

The non bladder exp tank had its port on the bottom so the air pocket stayed at the top. I assume this is the same as the "Pressure vessel" you use? A pressurized system must have a make up system expecially if there is a gas or oil boiler involved where a lack of water with no makeup can result in a big hole in the roof. I'm not sure what your local code calls for but in any pressureized closed loop systems we need a water feed, backflow preventer (to keep any chemicals out of the water supply), a low water cutoff to shut off the boiler and a properly sized exp tank.

[Tradewinds]

Thanks Mike.
Just to clarify what I meant:

Pressurized Vessel - Steel vessel with bladder inside
Pressurization Unit - Has a small pump with makeup tank

I guess if a bladder has perrished then you could have a similar problem.

[Chris Jennings]

Hi Brian

Not sure if i can help at all, we are the global distributor for Hycool, please give me a call (xxxx 450982) we may know the system your are working on and be able to offer a solution.

Cheers

Chris

[MikeHolm]

Thanks Mike.
Just to clarify what I meant:

Pressurized Vessel - Steel vessel with bladder inside
Pressurization Unit - Has a small pump with makeup tank

I guess if a bladder has perrished then you could have a similar problem.

Right, sorry about that. We seem to use some different terms sometimes. We would use a pressurization unit only on the closed loop system and then usually if there was a particular chemical needed for the system but a pressure vessel, whether a bladder type of just a tank with an air pocket would always be needed.

I've done work in many 20-30 story buildings where the boilers and other equipment is in the basement and the expansion tanks are sometimes with the boiler or sometimes on one of the upper floors. My experience with open systems in tall buildings is limited to turning it into a closed system with new boilers and BAS and most of these systems will only have a water make up using mains pressure (which in Toronto is 4bar). For high buildings the boilers are often on a middle floor and we use booster pumps for the domestic water anyway.

The UK still seems to install open systems....is this correct?

[Brian_UK]

Tradewinds, i understand your thoughts on having a pressurisation unit and it would be nice to have one on this system but it is not a good option.

The brine mixture is made up of the chemical salts and inert water, distilled for want of a word, so it is of prime importance not to mix it with oxygen. A pressurisation unit would normally need an open header tank to feed the pump so cannot be used.

I know that there are some systems which adjust the bladder air side pressure using compressors and vents but I think that would be a bit of overkill.

Mike, what are your thoughts on setting the expansion vessel pre-charge pressure to? I get so fixated on heating system expansion that working it backwards for chilled water makes my eyes water a bit.

System idle, pumps off, brine temperature say 10°C.
Isolate expansion vessels from circuit.
Pump in some more brine into the circuit until static pressure is 2.0bar.
Set pre-charge pressure of expansion vessel to 1.5bar.
Open expansion vessel to circuit.
Run chiller down to -7°C and monitor pressures.

[Brian_UK]

Hi Brian

Not sure if i can help at all, we are the global distributor for Hycool, please give me a call (XXXX 450982) we may know the system your are working on and be able to offer a solution.

Cheers

Chris

Thanks Chris, I'm sure that you will know the system and I will call you tomorrow, I will also email you tonight with my details. Brian (I've edited your phone number to reduce the threat of junk calls to you from the scammers)

[mad fridgie]

Set up your pressures when at normal working conditions -7C, this way you have an expansion tank, if you do it when the systems is warm, then you have a contraction tank

[Brian_UK]

Set up your pressures when at normal working conditions -7C, this way you have an expansion tank, if you do it when the systems is warm, then you have a contraction tank

DOH, that is so blindingly obvious, I think I need a beer. Hangs head in shame and slinks off, stage right.

[MikeHolm]

The beer is the obvious solution the the entire problem. All issues become clearer with beer (hic).

MF gets to it before i do. You are lucky here because your normal operating spread is so small so although the tank may seem big for the system, you should try solar systems. Not only do we need to account for an operating range of -30C (in some climates) to 150C+ but our tanks must accept all the fluid as it is pushed out of a solar panel when the system turns to steam. As an example, I have a resi solar system with 2 flat plate panels (5m2 total), system volume of 20L and the nominal heat output is 4kw.....the exp tank is 18L.

If we sized your tank at ambient (20C) and it did not have the buffer built in we could so sub-atmospheric it is the easiest time for air to seep into the system through a gasket and cause other problems.

If you can cool the medium as much as possible, even down to freezing when you put it in, it wil help in getting the pressure right. I have an easy formula I use at my shop which I can look up and if I find it under the piles on my desk, I'll post it. I also have a crazy German formula but it is 2 lines long.

Without the formula, and knowing the suggested 60L tank size, I would get the medium cool, pump it up to the precharge pressure and then bump up the tank to 2bar. The tank should be big enough to allow the bladder to distend towards the port without compromising the expansion/contraction capabilities.

[Brian_UK]

Had a good day for getting information today.

Spoke with an expansion vessel supplier who had said go for the 60l vessel, this size was selected because his calculation said around 25l would be OK but his next size up was 60l,

Anyway, rechecked site and found that we have two 24l vessels fitted giving 48l, so should be OK.

Also spoke with Chris (post #25 above) who supply the brine fluid who confirmed our general thoughts that the pressure was too low. Possible reason for the fall off in pressure since original install is the slow entrapment of air by the auto-air-separator which normally occurs after the initial fill.. Low temperature operation slows this process down also.

Whereas I have a standing pressure of 1.5bar he recommends 2.0 to 2.5bar.

So just got to get a filling pump and some more fluid and we should be home and dry.

[MikeHolm]

Great, and don't forget the bike pump and a tall cold mug of ale.

[Tradewinds]

Excellent news Brian.

And thanks for sharing the result with us.

[Brian_UK]

It will take a few days to get the materials and tools together but I will post the results.

Once again thanks to all for your help.

Virtual doughnuts (donuts) all round.

[Brian_UK]

OK News Hounds, been back to site today with a small pump, connecting hoses/fittings plus a small pressure gauge as the fill point is at the low level of the system and the pumps are one floor up.

Taking Chris's recommendation of a 2.5bar fill pressure I actually ended up with 2.6bar; this was with the fluid at around -6°C.

The slight overpressure was due to my pump discharge hose beginning to spray as the hose clip wasn't tight enough. Between trying to tighten that up (why I didn't shut the pump and valves off I don't know) and dropping the screwdriver the system ended up 0.1bar over. I think I can live with that though.

Restarted the pumps and the common suction pressure is now 2.5bar.

Reinstated the low fluid level pressure switch and put everything back into automatic operation. Just hope everything is OK as I hate to be the one to generate a weekend call out, especially when I'm not the one on-call.

So, once again, thanks for all the input; it has been an interesting teach-in for me.

[Brian_UK]

Went to site today for another matter and checked on the system.

Still running and no faults recorded on the chiller - sweet.

Think we've cracked it.

[MikeHolm]

Most cool (pun intended), where's the virtual beer? The Fullers vintage ale is calling my name.

[Brian_UK]

How about an i-beer?

[Tradewinds]

Classic.

Please stop, your making me thirsty!

[Brian_UK]

Nice YouTube ibeer link..
http://www.youtube.com/watch?v=8b9PH55EtJI

[MikeHolm]

Damn, I have a CrackBerry