Exploded GRP Water Tanks

Maybe this is the sort of tank you should have...You are obviously exceeding the loading capabilities of the design for the age and conditions that are present...

Is this the type design you have? What sort of internal support structure do you have?

Thank you for the feeback. Below is more information regarding the issue.

The structure looks more or less the same except that the tank are each 4 meters in height with the lower panels being 1 meter each and the top panel being 2 meters.

The tank the recetnly exploded is a collection tank for grey water and therefore the water entry is purely dependent on the usage of the water in the building and there are therefore no level controls to regulate the water level in the tank.

The panels in both cases where the tanks have exploded have failed at the join located 1 meter above the bottom of the tank. At this point, there is disconuity in the external reinforcement channels which are continous the rest of the three meters to the top. The tank also has an open vent at the top in addition to the access manhole to the tank.

Hope this information helps.

Can you be more specific on the construction of the vessel?

By your description thus far it seems that it is constructed of annular rings about a metre high and these are bonded (glued) together to construct a vessel of 4 rings and thus 4metres high.

What is the diameter?

Being greywater it'll have an enclosed top, or lid, so as to allow control of odors.

Are the components of the tank filament spun?

Is it a hand lay up moulded set up built around a male mould? Or a female mould?

Does it have any annular re-inforcement, at or near the obvious joins in the segments? These may be the steel rings you mention.

Is it, in fact, cylindrical? Or square? Made up of modular panels fixed together by various means, and re-inforced by steel beams (girts)?

More info please.

Stu.

At this point, there is disconuity in the external reinforcement channels which are continous the rest of the three meters to the top.

I reckon your main contractor has these ass about. Should be continuous at the lower level with the extensions at the top or at least staggered around the tank perimeter. What you have here is line of weakness all the way around.

Do you have a copy of the assembly manual?

I agree with you. It sounds upside down.

Note that if tank is filled to overflow level, then it becomes (potentially) a flow restrictor on the pumped fluid. The difference between air departing the vents and water departing the vents could add significant pressure to the tank and the weakest seams fail.

Trapped air is compressible and so a smaller vent can balance large water inflow UNTIL you try to push the water through.

The fact that the tank provider could so easily diagnose the problem (failure cause) and that you are not contesting it also indicates it is highly likely to be the case.

such failure would be seen as tearing and failure of apparently "good" material rather than a fatigue failure where the materials would appear rayed or degraded.

Do you have proper level controls for the tank?

Do they have a reliable backup?

Have they been tested and verified to be effective?

Are you filling the tank to capacity (overflow) and then shutting off teh pumps?

If this has happened multiple times and only for your group as customers of this suppleir, then it is something about your filling and usage process.

Do not be tempted to put in larger vents. The water shock (hammer effect) that will happen as the water hits the full/surcharge condition will still cause failure.

A repeated failure of this type I find very strange.

Even operating the tank at over flow level is part of the tank design and shouldn't cause failure. The over flow is there to prevent overfilling.

I believe there is an Australian Standard (certainly I had to prove to mining clients once) that a tank couldn't possibly over-pressure (by 10kPag greater than atmosphere). The common rule of thumb I use is to make the overflow line diameter 150% of the incoming line - therefore not possible to have any significant pressure drop (and I would assume tank manufacturers / system designers would also do something similar).

Having said that if the overflow / vent is not properly designed, perhaps overpressure caused it. But I would like to see the full system design to be certain of this (P&ID, Process flow diagram, mass balances, tank deisgn specifications, etc).

Regarding other tanks onsite, as I can only think this is either a system design problem, or a manufacturer problem. And assuming same desing engineer and same manufacturer for all tanks, then there is definite doubt over the other tanks onsite.

Suggest bringing in an experienced process design form for an audit. Any of the engineering consultancies will be able to find the cause of this. If it then is tank manufacturer issue, then it is money well spent in ammunition against the tank manufacturer (warranty or not - bad publicity hurts).

Good luck,

Anthony

Thank you for the information. The tank is for collecting grey water and does not therefore have level control as water flows freely from all points in the building.

The defect liability period for this tank only expired two weeks ago which was one year since handover from the main contractor and the tank should be therefore be relatively new. Overflow of the tank as highlighted in my mail is only based on hearsay and no supporting evidence from the contractor. I however want to establish the technical aspects behind the tank's structure and construction which could have been linked to the explosion before addresing the entire issue.

The explosion of the first tank happened a short period before handover last year and was therefore still with the main contractor and was rebuild by them.

Obtain a copy of the witnessed hydraulic test record/certificate for the tank. Find out how and why this test level/pressure (same thing - P=ρgh and all that) could have been exceeded. Something is wrong at process design stage for this to have been able to happen.

are the tanks subject to wide temerature changes? if so. do the tanks have relief valves?

The tanks are installed inside the building and hence not subject to wide temperature change.

They also have ope vent pipes and the access manhole covers are also not sealed as the rooms are always locked and can be easily lifted which can also help to act as a pressure relieve.

While replacing or repairing the tank, invest in overfill protection/prevention instrumentation and controls immediately.

This facility has been subject to the scrutiny of a HazOp Study, hasn't it? Why was this possibility not considered as part of that study?

It would be a good idea to immediately:

1. Take steps to lower current fluid levels in all tanks you have of this type that are near capacity.

2. Implement operational / procedural limits restricting capacity to each tank of this type.

The other comments are excellent, and I also suspect the tank ruptures are related to overpressure resulting from making vents designed for air deal with pumped liquid....

....but that's just a guess from the luxurious vantage point of not being in your shoes. A thorough investigation needs to be done, until that is accomplished you are essentially flying blind and cannot even begin to effectively devise a long term solution.

If filling the tank can lead to the tank bursting, the possibility of its collapse while being drained must be considered too:

AP,

Whatever the tank is constructed from, when the level reaches "overflow" it is merely full.

If the tank can't stand the loads of being full, then it's a substandard, poorly constructed tank.

No matter what the warranty period, to fail in service means that it had a latent defect, which is squarely the responsibility of the maker.

Sue 'em. Including collateral damage.

Then get product from a tank maker who knows what he is doing.

Cheers,

Stu

(Engineer + 30years in GRP constructions)

Caution: that is why the witnessed test report/certificate is of value. It would determine which party is going to win and which is going to lose - if it ever gets to litigation as a result of its uncovery, that is.

Dear Friend, In case the tank is of depth/height 4m, the ideal configuration is to have 2m ht. X 2 panels or bottom panel of 2m & above panels of 1m each. The reason of panel tank failure occured due to it's liquid column. Since at overflow level, maximum pressure was exerted at 2/3rd depth, a panel of 1m failed as the pressure was exerted at the joint. This would be the only reason causing multiple failures. Other causes mentioned in other replies are also important design considerations but to my mind, the principal cause is using 1m panel in a 4m ht. of tank.

5M Dia and 10M high is a cylindrical tank.Cylindrical tanks are available in all heights and diameters. Panel tanks of GRP are not generally available at 4M height. It appears a manufacturer has stretched the design to unrealistic levels causing failure.

That would explain the 1m extensions....

Maybe these failures are actually a design feature to prevent overflow.

I guess you may consider buckling of the joint owing to pressure of liquid depth as a cause.

Four meter is an unusual height for GRP tanks. Maximum 3m height is advised. The tank should not explode by filling to overflow level. the test pressure for the tank should be higher than the height of the tank. Another important point is temperature of water. GRP is very sensitive to water temperature and its strength will come down drastically as the temperature rises above 20 deg C which is generally the design temperature. If you have higher water temperature then you should set the overflow level lower but you will sacrifice the storage volume.

Seriously now...20ºC?

It doesn't get that cold here.

What have your insurers got to say about the tank failure?

Like others have said, the tank has to be able to survive being filled to the brim.

Interesting (bog actually) how the portion of the tank with the greatest loading is the least reinforced.

Better get a lawyer son.

Dear All, Thank you for your valued feedback. It has been great help. As I am writting to you now, I have just come from clearing water from another plantroom which has been flooded after another grey water tank busted three hours ago.

We have contacted the main contractor and he has reluctantly agreed to call the manufacturer to investigate this failure. Coincidentally, the tank that busted today is the same size as the one that failed last week and it has also failed at the 1 meter height. I hope that the matter will be investigated in details and necessary action taked to reinforce or replace the remaining GRP water tanks.

This forum has been great help.

Reluctant? Scared shitless more likely.

If your main contractor has any nads then he should be giving the tank supplier a bollocking right about now.

One failing is a coincidence, but this many???!!!!

How many tanks are left? Are you bracing them up with timbers and straps?

Please accept my empathetic commiserations mate!

Grey water has a higher SG than clean water right.......hmmm.

Well regardless your next step is to get a good NDE company to do weld examanation as well as taking thickness reading to check for corrosion. While the forum can speculate, there is no substitute for real information.

and then there's the clean up costs.

Once again thank you. Additionally the idea of involving a third party engineer to investigate the entire installation is a great idea. We have an other 5 tanks which are 4 meters high each, 2 tanks which are 2 meters high each and 8 tanks which are three meters high. Your feedback has been great help.

You're very welcome AP #1.

My opinion is: the tank was designed as a 2 meter high module by the manufacturer. Supporting tie rods are located at the mid high point. By stacking 2 modules to create one tank, the juncture between the lower 2 panels becomes a weak point because the 2 meter high tank was designed to a 2 meter head. Stacking the modules increases the load on the bottom module to over 8 times. I'm making an assumption here based on the sketch provided by Solar Eagle. I'm not sure about the 8X, but 4X at least which should be enough for failure to occur.