Plastic pipe degradation 1

[mmenarry]

Hi folks,
I have a third-party water treatment package, multimedia filters followed by water softeners. Total volume treated is 180 cubic metres an hour.
I'm experiencing a strange problem:
The plastic pipework (ABS, PVC and u-PVC) seems to be becoming brittle. I'm experiencing failures of pieces such as flanges and short sections of pipe. The manufacturer of the pipe is currently testing to see if there is a pipe fault in manufacture, but this seems unlikely, as it would be affecting multiple batches of multiple pipe sizes (from 1/2" to 8&quot
Fissure cracks (by which I mean long, up to 2 to 3 feet on the 8" pipework) seem to be the biggest problem, and the flanges seem to develop from a single point.
I'm thinking that as this is affecting all sizes of pipework, both lengths and fittings, that it may be something removing the plasticizers from the material itself.
Incoming water is normal potable water, dosed on site with chlorine (sodium hypochlorite) to about 1.5 ppm, filtered in the multi media filters, chlorine removed with sodium bisulphite, softened in exchange resin, and then re-dosed with chlorine. This problem is showing itself on pipework carrying water raw, after chlorine dosing and after bisulphite dosing.
Any ideas, anyone?
Michael

 

[hydrae]

Michael

is the point of hypo injection using a quill?

hydrae

 

[KRSServices]

Micheal,

How old is the system, and what are it's operating ranges, particularly flow? Do you have variable surge pressures (such as water hammer) or are you operating at or near the maximum working pressures of the plant? From your description, the failures may be a result of fatigue. I had a small plant that kept the operator busy with breakage until he evaluated the sytem and removed excessive vibration from the pumps and we upgraded the treatment process to accomodate higher process flows. We were nearing our design limits. One other addition was to upgrade the pumps to soft staart systems to mitigate the effects of water hammer on the delivery systems. Your flanges have to be aligned perfectly otherwise the forces of stress and vibration will cause an eventual fatigue failure.



KRS Services

http://www.krs-services.com

 

[mmenarry]

Hydrae,
Injection is low flow (max 1.5 litres/min of 16% Cl solution) through a 1/2" pipe ported directly into the main header (8&quot after the softener stage. Not sure what you mean by a "quill", the failures are both before and after this point of injection.
KRS,
The plant is practically brand new in operating terms. The plant was installed and commmissioned, and has been running for about a year. Because the site is still being built, it has been operated on a tempoarary basis when needed, filling a large tank.
There has been no (noticable) hammer in the system. Pressure fluctuations are present, especially when the filters run a backwash. At present, we start and stop the filters & softeners in sequence, but the design of the plant under automatic operation is for all to start at the same time together. Normal fluctuations at start-up (all together) is from 4 to 7.5 bar (60-110 PSI). System has been tested to 10 bar (147 PSI). Under running conditions, there are no pressure fluctuations.
The only thing I can think of is water lying in the pipe-work for extended periods of no flow (sometimes 2 or 3 days), somehow leaching plasticizers from the material.
I just would not have expected fatigue to show up to such an extent on a plant this new.
Thanks for your help, I'll be monitoring it closely. I recently checked all (and I mean ALL!) the torque settings on the plant, but there appeared to be no problems there. I will examine alignment and see how good this is.
Michael

 

[hydrae]

michael

A quill is used to force the chemical into the center of the process stream instead of entering at the pipe wall. The 1/2" pipe would extend into the 8" header at least one inch so the concentrated chemical does not directly contact the header. The end of the 1/2" is cut at an angle facing down stream so it looks like a writing quill.

Another item is, if you are using a saddle at the injection point some saddles oval the pipe when they are overtightened which will cause a lenghtwise crack.
hydrae

 

[mmenarry]

No, no quill, Hydrae. It does not seem to be a problem with the concentrated chemicals, the exact same specification pipework is used to convey the undiluted chemical to the injection point.
The thing that has me confused is that this pipework is normally indestructable! The walls on the 8" runs are 12mm thick, you could hit one of these pipe with a sledge and it would just bounce off. When one damaged section was removed after repairs had been carried out (8&quot, it was dropped from 1 metre. It literally smashed into about a dozen pieces. All of a sudden, this pipework seems to have gotten very brittle.
Michael

 

[KRSServices]

I have been in the water distribution and treatment industry for over 15 years and have not heard of instances whereby the plasticiziers are leaching out. Vibrations, exacerbated by unsuported pipe or fluctuating pressures, can cause a pipe to fatigue over time. When operating, are the process flow rates near the design limit? Are the pumps operating without undue vibration? Do any of the pipes "shake, rattle or roll" when the pumps come on or shut down or during the backwash process? The cracking of the pipe could be the result excessive working pressures and vibration on the pipe.

KRS Services

http://www.krs-services.com

 

[hydrae]

Michael
is there a source of a hydrocarbon, ozone, or UV near by?
Hydrae

 

[mmenarry]

KRS,
I'll knock the plasticizer leaching idea on the head then so. Maybe I'm being too kind towards the pipe-work and assuming it should be less brittle than it actually is. If that's the case, then pressure fluctuations could well be the root cause. Pipe work is very securely placed, "Unistrut" type clamps used on vertical sections, and large u-bolts on horizontal sections. No damping is used anywhere on the system. As far as working flows vs max flows are concerned, the current flow though each filter or softener is 60 cubic metres per hour, max capable is in the region of 170 cubic metres per hour. However, during a backwash sequence, butterfly valves can go through closed/open/closed sequence in approx 20 seconds, with the associated pressure changes. Perhaps the pipework is so securely clamped that any hammer is "fixed", i.e. present but movement restricted to such an extent that it is not visible. I'll loosen the clamps slightly and see if vibration is present. On the flange failures, the failures would seem to be the "classic" metal fatigue type, a number of "rings" around a point on the edge, followed by a full failure the rest of the thickness. (I have a couple of pictures, not great)
Hydrae,
No, no HCs, solvents, no ozone, UV or other oxidsers.

It could just be exactly as KRS has suggested, fatigue. These failures are all happening one after the other, it may just be coincidence that the fatigue failures are all happening together rather than brittleness. Fatigue could certainly cause the same failures as brittleness, I just assumed that the pipework would be more fatigue resistant than it was. I'll get the on-site engineers to examine the way in which the whole pipework is fixed in place, and see if we can reduce and pressure fluctuations or hammer knock first, and if it's present, try and engineer out the effects through damping.
Thanks for your help guys, I could have been looking for spurious leachate for weeks!
Michael

 

[jdarco]

Is the pipe outside? Pipe exposed to strong sunlight can become brittle, although I think it would take more time than your system has been in place.

 

[mmenarry]

jdarco,
No, no significant sunlight exposure, pipework is either indoors or in a pipe-rack. Also, it's ireland. We rarely get sunlight!
Michael

 

[JedClampett]

Is the pipe coated?

 

[BobPE]

I would have to agree with KRSServices, plastic is a very poor piping material in dealing with transients. Look at you pumping systems espicially with the cracks you are describing. Getting an engineer to look at it is a great idea, but don't use the same one that designed it, try a second opinion, it will cost you the same and you wont get an engineer trying to hide anything.

KRS

Plasticizer leaching is a big problem with plastic pipe and may be a problem here, although I doubt it. Levels of plasticizers found in drinking water is kinda alarming, its in a part per trillion area, but levels are high enough to cause concern as endoctine disruptors in current research.

BobPE

 

[mmenarry]

JedClampett, No, no coatings.

Thanks Guys,
We're looking at the whole pipe-work systematically now, and we're finding leaks starting all over the place. Looks like fatigue is the cause, aggravated by the pipework being so securely clamped that all vibrtations and water-knock is being absorbed by the piping itself.
It's looking like replacing all the pipe-work and proper vibration dampers/bellows will have to be done. It's a pity that it has happened so soon, but in hindsight it's probably best we experienced this problem before the site was finished.

If I can tie down any exact details, I'll post back with the information. (Manufacturer's investigation, etc.)
Michael

 

[HDW]

What is the temperature?
What is the MAWP?
Are there detergents in the water? - I have already seen damages on water wetted thermopastics because of stabilisator wash out caused by detergents.
Are there free radicals in the water ? They may eat up the stabilizer.
Furthermore, I suggest to install some electrinic manometers with a memory for the maximum pressure measured - This may give a hint whether MAWP is exceeded or not.
Is there the possibility of condensation above the water level in semi filled pipes? Condensating water may sometimes be very agressive. If yes, is it feasible the change the process in such a way that the pipes are alway cimplete filled?
To check out, whether there is wash out of stabilizers, this may be a procedure:
Take small material samples from (1) the wetted surface at a point of failure, (1) from middle wall thickness of same pipe length, where undamaged (3) outside, where undamaged, and (4) from a retained piece of same pipe and go to a plastic laboratory to measure the OIT (Oxidation Initiation Time) in air by mean of Differential Scanning Calorimetry (DCS). test temperature should be such that OIT of undamaged samples is between 8 and 48 hours. If OIT where wetted is significantely less, there will be washout or utilisation of the stabilizer.
Harald

 

[mmenarry]

Normal working temp is about 7.5 degrees C, between 5 and 15 degrees.
MAWP is of the order of 12 bar. 8.5 bar is never exceeded.
No detergents are present. Thanks for the test method for the stabilizers, though.
Regards,
Michael

 

[stanier]

Sorry I had missed this debate before.

Thermoplastic pipe has a coefficient of expansion 10-20 times that of carbon steel. If you lock it in then you will get fatigue damage due to diurnal temperature fluctuations.

Vibration of piping could also be a cause particularly with the supports as described as they provide notch sensitivity. Notches provide the sites for cracks to initiate from and then they continue to propogate.

One as pect that may need consideration if the chlorine is gaseous and the unit doesnot operate continuously then there could be a large concentration build up when the unit is switched off. I experienced this with a ss header and it turned black overnight.


Sharing knowledge is a way to immortality

 

[mmenarry]

Update:
Got some reports back on the pipe-work. "Suspect" pipe was failing at about 2.2kN in a crush test, good pipe was taking 7.7kN before failing.

Also got what could only be decribed as a report in the loosest sense of the word. Plenty of "maybe", "possibly" and other wishy-washy vocabulary. This report claimed damage due to solvents in the MEK used for the joint welding, based on the reporter's "database" on the subject. High on pictures, poor on actual content. No conclusive detail, though.

So I now know I need to replace much of the pipe-work. Without a root cause, though, I'm looking to replace with steel. If I ever find the definate cause, I'll update y'all.
Michael