Leak Testing

[Samarjit Sen]

We have just installed a refrigeration plant with copper pipe lines. The plant after installation was tested for leakage by putting Nitrogen at 300 psi and the same was retained for 48 hours. There was no leakages nor any pressure drops. All the joints were tested with soap water

After about a month it was found that there was a shortage of refrigerant in the system. Again before filling the system with R 22 refrigerant we tested for leakage with Nitrogen. After 2 days of proper working again there was a shortge of leakage. I personally visited the site and had both the sides ie. the high side and the low side tested seperately with lekage with Nitrogen at 300 psi. Kept the pressure overnight. There was no leakage. Charged the system with refrigerant and the plant operated absolutely fine for two days. Now again there is shortage of refrigerant and the temperature is not being attained.

The compressors are semi hermetic of a renowned make, the Evaporators are also of renowned make.

The unit is to maintain a temperature of - 35 Deg C inside the room which it is performing so long there is no leakage of refrigerant.

Could some one advice me and guide me as to what other tests could be carried out so as to detect the leakages. Should the pressure of Nitrogen be increased and if so to what extent so as not to damage any component.

Please advice soon It is not possible to shut down the plant for a long time, as this is an Ice Cream Hardening Room.

[mgtet]

I would check the relief valves if you have any, but it sounds to me like you have a hot leak. I would pressure test as high as you can, at least upto the high pressure swith setting on the high side and then whatever the maximum test pressure is on the low side. If you still can't find the leak try leak testing whilst the equipment is running, paying particular attention to the condensor.

[The Viking]

Sounds like if you have a leak that reacts on temperature (i.e. metal expanding/contracting thereby opening up as the temperature changes) these are hard to find, typical example would be a flare.
You should be able to find it with a good quality electronic sniffer (there is a substantial leak somewhere).
Another option would be to put UV dye in the system and wait until it goes down again and then check it with an UV lamp.

BTW, you have checked the running pressures and compared with any pressure-relief/safety valves ratings?

[NH3LVR]

I would check the relief valves if you have any

I was prepared to make this reply when I saw this post from mgtet. He is, of course correct to suspect the relief valves.
This may not apply to your situation, but I just returned from a job plagued with leaking relief valves. There is a sensor (NH3) on the relief line. After dealing with seeping valves for the last year and a half we installed rupture disks under all of them.
It is common to remove the relief valves in industrial applications while pressure testing. Every thing is fine until you reinstall the valves. Rubber gloves, balloons and other items can be installed on the outlet of the valves to check for leakage. I suggest this as they will allow you to test over a time period. Relief valve leaks seem to come and go, so testing them now does not mean they will not leak in an hour.

[US Iceman]

If the entire refrigerant charge is being lost this quickly you have major problems. A simple and quick inspection of the piping looking for oil stains will tell pretty quickly if the pipe is leaking this badly.

My guess is the relief valves and this quote is the closest to the solution and problem I think.

BTW, you have checked the running pressures and compared with any pressure-relief/safety valves ratings?

Relief valves will typically start to weep at about 90% of their set pressure. At or above this percentage they will simply pop open. If the operating pressures routinely approach this percentage of the relief valve set pressure, the valve spring can get weak and start to leak at lower pressures.

What is your design condensing temperature and what refrigerant are you using? Also, what is the relief valve set pressure?

If you are using air-cooled condensers and a high pressure refrigerant in India I think the relief valve settings may have to be quite high. Perhaps 400 psi!

[Samarjit Sen]

Dear US Iceman,

The plant has water cooled condensr of the PHE type. The refrigerant is R 22. The reli.ef valve is set at 400 psi.

I am instructing my technicians to check the relief valves, the caps of service valves, flare joints and remove the complete insulation over the suction line and check the brazing joints.

I am also asking them to increase the Nitrogen pressure to 350 psi before doing the leak testing.

I shall revert back to the RE with the results and seek further advice.

With best wishes,

[US Iceman]

In an attempt to understand your issues I looked at the sequence of events.

Problem 1

...tested for leakage by putting Nitrogen at 300 psi and the same was retained for 48 hours. There was no leakages nor any pressure drops.

Problem 2

After about a month it was found that there was a shortage of refrigerant in the system.

Problem 3

After 2 days of proper working again there was a shortage of leakage.

Problem 4

Charged the system with refrigerant and the plant operated absolutely fine for two days. Now again there is shortage of refrigerant and the temperature is not being attained.

The first two things I notice are:
The system holds pressure during a leak test.
Sometime shortly after being started, the leak almost immediately starts within two days.

This points to an operational issue with the system, OR, how the system is being ran or used.

At any time during operation of the system does it start operation with a hot or warm room or high heat load? The pull down load could exceed the condensers ability to keep the condensing pressure below the relief valve set pressure.

Is the water flow being restricted at some time (temporarily) during operation?

Are the water supply temperatures to the condenser low enough? High entering water temperature could cause this also.

One thing I did not see mentioned, so I will ask instead of assuming this took place. Was the system evacuated prior to charging? Any air left in the piping will raise the discharge pressures due to non-condensable gas.

[Samarjit Sen]

Hi US Iceman,

Let me answer your quaries properly.

1. The plant when being restarted ha a room temperature of anything between 0 Deg. C to -12 Deg. C.

2. The water from the cooling tower is at about 32 to 24 Deg. C.

3. The Discharge pressure being maitained is about 160 psi, whereas the set pressure is 300 or 350 psi.

4. Yes prior to charging we evacuated the system for about 4 to 6 hours and also held the same for 24 hrs.

One thing has been noticed that we had retained our technician at site for 4 days after recommissioning the plant. During that period there was no proble. After we withdrew our staff from site within a period of 3 days the leaks again occurred. We told the owners that we are having doubts that maybe some body in his staff is creating this problem. The owner agreed to put a security guard in the plant room. Even then the leakage ocurred.

[The Viking]

Back to the temperature related issue then I'm afraid.
(see previous post)

[Samarjit Sen]

It is not a temperature related issue. I had stated the temperatures as US Iceman desired to get more details. The idea of stating the temperatures is that to ascertain wether the leakages are occuring due to high temperature.

All your opinions and advices shall help me to sort out this problem.

With best wishes,

[lana]

We told the owners that we are having doubts that maybe some body in his staff is creating this problem. The owner agreed to put a security guard in the plant room. Even then the leakage ocurred.

This happened to me once. Someone tampered with the electric control unit and the electric defrost was turned on for whole night .
Fortunately we proved our case and that someone was sacked.
What people would not do for money .
Good luck

[The Viking]

It is not a temperature related issue. I had stated the temperatures as US Iceman desired to get more details. The idea of stating the temperatures is that to ascertain wether the leakages are occuring due to high temperature.

All your opinions and advices shall help me to sort out this problem.

With best wishes,

As any refrigerant plant operates it's components changes temperature (otherwise you wouldn't bother running it). When metals changes temperature, they do expand and contract, this sometimes causes leaks. Normally this type of leaks are found around where different metals joins, like in a flare. This change in temperature together with moisture (condensation) can even shear/crack flare-nuts.

The problem with these type of leaks is that they are hard to find as when you come to pressure test they might have sealed themselves due to the temperature being closer to room temperature again (as the system now aren't working).
Good quality electronic leak-detector or dye is normally the only way of finding these leaks.

[US Iceman]

The Viking raises a valid concept. These types of leaks are difficult to find.

The issue I have the most problem with on this is the entire charge is being released within two days. That is a significant leak.

Of course, a cracked flare on a high pressure gas line would just about do this too.

[Samarjit Sen]

We have to attend to the plant. But we can not get a shut down for more than 2 days, as at this time of the year the sale of Ice Cream is very high and the production is going on day and night. The owners have kept R 22 in stock and every 2-3 days 4 to 5 kg is being charged to top the gas level.

I would appreciate that before I again send my technicians to attend to the plant, if you could advice me as to how should I go about for tracing the leakage.

What I had in mind is to devide the plant in two sections ie the high side which will be the compressor, discharge line, and the condenser and the low side which will be the liquid line, cooling coil and the suction line up to the compressor. We propose to pressurise these area upto 350 psi with nitrogen and retain the same for 24 hours while testing for leaks with soap and water.

This time we shall give more importance to the flare joints and the pressure relief valves.

Would you suggest a pressure of 350 or 400 psi of nitrogen for testing purpose ?

Please corerect me if my proposal is wrong or incase you have something else in mind, which will give better results.

[The Viking]

Before you go for a shut-down, I would try less intrusive methods.

If you are loosing 4-5kg in 2 days, then an electronic sniffer must be able to find it whilst the machine is running (these are normally rated to a couple of hundred gram/year)

Should this not be possible then add some Ultra Violet dye to the system, this will blend with the oil and after a couple of week you will find a yellow (or green, or red, depending on manufacturer) stain around where the leak is, when you light it up with an UV lamp.

[The Viking]

Just had another thought... or rather 2....

1. If you lost this amount of refrigerant then there should be some traces of oil there as well, check for oil stains.
2. Does the pipework go through any walls or has an electrician installed cables on to the machinery? Reason I asked; it wouldn't be the firs time someone put a nail or a screw through the pipework! I've seen it in the past, an electrician install the power to our condensor and using self tapping screws to hold the cable clips, put about 10 holes in the pipework. That system also passed the pressure test and Vacuum OK but as soon as we started it up we lost the whole charge in just hours.

[Samarjit Sen]

That is what I thought also. I also tried to find out traces of oil, but there were none. The pipes also do not go through any wall nor any portion of the pipes are plastered. The wirings are seperately laid along the wall.

I am sending my technicians to the site tomorrow and shall revert back with the result. The site is located in another City.

[RAC'ers 98]

Hello Sir!

At some point, it may sound true! that someone or a person is creating a problem here... based on experience... I believe you did all you need to do for the Leak Test and based on advised at previous posts. Putting a company security guard on the plant didn't really assure to catch up for someone doing strange. One piece of advise is to put your own staff on close monitoring for a week or more to really identify whats happening! Then, if something sounds alright, somehow, you can justify it with your customer that you need someone to closely monitor...

GoodLuck!
RAC'ers 98

[Samarjit Sen]

The site is located in another city which is about 1000 km from where me and my office is located. Based on all the advices and the help given to me by my good fellow members of the RE, I have instructed my technicians to carry out the leakage tests accordingly. I hope that this time something is going to work out and I shall revert back to the RE with the outcome.

[Samarjit Sen]

Thank you all for the guidance and the advice. Finally two leaks has been located. One is a cap of the flare nut on the compressor and the other is an internal leak within the PHE Condenser. We applied a pressure of 350 psi of nitrogen when the leaks were found. The cap has been changed and we are changin the phe condenser.

I feel using PHE as a condenser is not a good idea, so I am reverting back to the old Shell & Tube condensers in my future projects.

Thanks all. I have learnt a lot from the various suggestions, which will help me to ensure leak proof installations.

[US Iceman]

The only good results I have had with flare caps is to use a flare washer and a little refrigeration oil on the washer.

Or, use a flare cap with an o-ring in it.

A regular flare cap by itself is not much good.

I never thought of PHE leaking, it must have been an internal leak where the refrigerant went into the water?

[gas_n_go]

Honestly I think your best bet is to hire a reputable outside contractor to come in and help you find it. Leaks can be very tricky to find and sometimes it takes someone with alot of experience to know where to look. No disrespect to you or your techs, Its not just a matter to where to look, its also a matter of how to look. What your describing sounds like its big and should not be too hard to find, but Ild bring in experienced people to do it, so you dont lose anymore refrigerant, or experience unecessary down time.

[Thana]

You may want to try if you have the time... try isolating the pipework from the codenser and pressure teating this seperatly i.e. norrowing down the possibilities..I've found that this works some times

[HOMANJ]

I've have found systems that only leaked when they were hot. Try leak test when system is operating. Also loss of compressor capacity will show up as loss of refrigerant .

[smurphy]

What about a electronic leak detector, also I've had water cooled condensers leak from inside the condenser, disconnect water lines and check with leak detector, good luck

[The MG Pony]

Thank you all for the guidance and the advice. Finally two leaks has been located. One is a cap of the flare nut on the compressor and the other is an internal leak within the PHE Condenser. We applied a pressure of 350 psi of nitrogen when the leaks were found. The cap has been changed and we are changin the phe condenser.

I feel using PHE as a condenser is not a good idea, so I am reverting back to the old Shell & Tube condensers in my future projects.

Thanks all. I have learnt a lot from the various suggestions, which will help me to ensure leak proof installations.

He found it people.

[SteinarN]

Thank you all for the guidance and the advice. Finally two leaks has been located. One is a cap of the flare nut on the compressor and the other is an internal leak within the PHE Condenser. We applied a pressure of 350 psi of nitrogen when the leaks were found. The cap has been changed and we are changin the phe condenser.

I feel using PHE as a condenser is not a good idea, so I am reverting back to the old Shell & Tube condensers in my future projects.

Thanks all. I have learnt a lot from the various suggestions, which will help me to ensure leak proof installations.

What was the make of the PHE condensor?