Water Main Leakage Test - Calculate theoretically? 4

[winterrules]

I recently witnessed a pressure test followed by the leakage test. Pressure dropped from 161psi to 159psi for 5,710 ft of 8-inch DI pipe. Is there a way for me to verify the leakage (7.5 gal) theoretically?

I'm suspicious of the leakage number because they pulled the suction line out of the water tank used to pump the line up....and I suspect air got in the suction line and messed with the test. Thanks for any help you can provide!

 

[LSPSCAT]

See the link below to a reference book from EJ Prescott. Great field reference guide for various questions especially when dealing with old equipment. The link includes a formula to calculate the allowable leakage while pressure testing.

http://www.ejprescott.com/reference/WaterMainFlushPresR-23.pdf

(4.34 Gallons) for your example.

The pressure test criteria does allow you an allowable which can be calculated in the above formula. At the end of the test you are allowed to pump an additional 4.34 gallons back into the line to bring it back to the test pressure. Typically in the northeastern U.S. I have been required to hold between 150 psi - 200 psi anywhere from 1 hour to 2 hours and everwhere in between. Once completed, if I have dropped below the test pressure, I am allowed to pump in the allowable leakage, calculated above, back into the line to get back to the test pressure. If it does not pump back up, you fail. If you pumped in an additional 7.5 gal then it appears to me that you failed. Keep in mind everywhere I have done a pressure test the rules and regulations have been different! Same basic idea but everyone seems to have there own take.

 

[LSPSCAT]

Additionally be suspicious if the contractor sees an instantaneous jump on the gauge and sucked air as you describe as the system can become air locked and is pointless after that. Contractors try everything from a "secret" buried valve or curbstop to pass if a line has become troublesome. Pay careful attention to the setup and make sure you understand the location of every valve, hydrant, and blowoff.

 

[winterrules]

Thanks for the input. I should have pointed out that it was a 2hr test that passed the pressure test portion, and seems to have passed the leakage test performed immediately following the pressure test. I am involved with the testing quite heavily and am doing my best at keeping an eye on him whenever he leaves my sight.

What I'm looking for is basically a ballpark way to estimate the leakage as a check. I'm a relatively new engineer and I like to have a way to check results on things that I haven't built up enough experience to verify. In college water was always treated as incompressible, so I'm not sure how to calculate it. In theory it seems like there should be a simple "ideal gas law" type method of calculating the water needed to bump "x" volume at atm pressure to "y" volume at "z" pressure, but I can't seem to find it.

 

[LSPSCAT]

To my knowledge the pressure test is the leakage test? No? If it leaks the pressure decreases...therefore it has leaked, the allowable leakage is calculated as above. Not sure what you mean by a supplementary leakage test?

 

[winterrules]

They are conducted concurrently, but they're considered separate tests. You pump up to 155psi and hold it (within 5psi) for 2hrs. Then they pump in the volume of water needed to get back to the starting pressure. The allowable leakage calculated with that formula is a max leakage allowed and the amount pumped back in must stay at or below that.

For example: Pump 3,000ft of 8-inch water main to 155psi at Noon, monitor until 2pm. Let's say it ends up at 152psi....it passes the pressure test. Then you have to pump water into the system until the pressure gets back to 155psi. Whatever volume (at atm pressure) is pumped back in is your leakage that must be equal to or less than the calculated allowable leakage.

 

[LSPSCAT]

Agreed.

 

[psnyder]

Winterrules, that is correct in that reference must be made to AWWA C-600 (Installation of DIP), Section 5.2.1.4 Testing allowance. See the section below, but I would purchase the installation standards from AWWA for the various types of pipes.

5.2.1.4 Testing allowance. Testing allowance shall be defined as the maximum
quantity of makeup water that is added into a pipeline undergoing hydrostatic
pressure testing, or any valved section thereof, in order to maintain pressure within
±5 psi (34.5 kPa) of the specified test pressure (after the pipeline has been filled with
water and the air has been expelled*). No pipe installation will be accepted if the
quantity of makeup water is greater than that determined by the following formula:
In inch–pound units,
(Eq 1)
L=[S*D*(P)^1/2)]\148,000

Where:
L = testing allowance (makeup water), in gallons per hour
S = length of pipe tested, in feet
D = nominal diameter of the pipe, in inches
P = average test pressure during the hydrostatic test, in pounds per square
inch (gauge)




PAUL S SNYDER, P.E.

 

[winterrules]

Shouldn't that be over 133,200 for those units?

 

[rconner]

I fear a reader to this thread might come out of it with some misunderstanding of the specific test it appears was performed, as well as the complexity in general of hydrostatic testing issues. To begin with my reading of the formula as well as tables of C600 tells me that when make-up water must be pumped back into an 8” ductile iron pipeline 5,710 feet long tested at say 160 psi, the ALLOWABLE “testing allowance” for a two hour test per this standard should not exceed:

L=(2 hrs) (5,710)(8)160^.5/148,000 or 7.81 gallons. [It appears the contractor in this case for whatever reason pumped only 7.5 gallons back in, that would appear to be a “pass” at least per the AWWA standard .]

With all due respect also to some of the verbiage and terminology in references in this thread, as well as in some of the responses, I feel compelled to make the following additional comments related to the further, original question in general of "calculating leakage". In my opinion, while it is possible to do all manner of calculations e.g. with “bulk modulus of compressibility” of water and elastic stretching of pipes etc. it is generally not really practical to “calculate leakage” in buried pipeline tests (except maybe in some very HIGHLY controlled test conditions of vessels and limited pipelines). This is due to the following realities, some that may really be common to installation of pipelines of all materials:

1. Despite best intentions/design, most rigorous or expensive installation/construction, and inspections etc., pipelines are not necessarily laid precisely with regard to either the horizontal or e.g. vertical lines on plans etc.
2. Similar to the realities/imperfections of “1.”, air valves and/or other air release mechanisms etc. are not necessarily always installed precisely at the top or apex of all local, vertical crests that may happen by design, or inadvertently in the installation of pipelines.
3. Due to the combination of “1.” and “2.”, some air inevitably becomes trapped in pipelines.
4. As most venerable standards and specifications also require that pipelines be filled, e.g. in preparation for hydrostatic testing “slowly” [defined in some as at velocities <1 fps (0.3mps)], it is also a reality (at least when high velocity flushing is not accomplished prior to hydrostatic testing) that that any unintentionally trapped and unvented air will simply not be removed (or “scavenged by sufficiently high flow velocity) at the time the (in effect combination air and water containing) pipeline is pressurized in hydrostatic testing.
5. While trapped air may or admittedly may not cause any sort of noticeable problems in subsequent hydrostatic text results, it is well known that it CAN wreak literal havoc in others. In other words, pipelines that contain trapped air can appear in various fashions to sometimes fail hydrostatic testing criteria (incidentally either raising or falling in pressure, or requiring excessive make-up water!), even though they really have no leakage!
6. For this reason, it is also stated in many long-standing specifications that the installer should remove most or all air (a requirement that appears, at least in many circumstances, sometimes easier said than done and some incongruous with the realities of 1-4!)
7. Unlike water, that as you note is relatively “incompressible”, air is instead highly compressible, and air is also highly volume or pressure (if contained as in a closed hydrostatic test) responsive to changes in temperature in accordance with “Boyle’s”, “Charles’” and “Gay Lussacs’” et al laws/principles. Thus if (indeed probably instead “when”, in the case of much testing that is of rather long duration) there are any temperature changes this will be accompanied by inevitable changes in pressure etc. and/or disproportionate requirements for make-up water etc., that do not have anything to do with “leakage”.
8. It may be tempting to just assume that pressurized air is just like pressurized water (the old “pressure is pressure” argument), and therefore will not meaningful affect test results. However, this argument appears to neglect yet another phenomena associated with air/water mixes, that being that when there is no separating membrane air is also somewhat free also go into and out of “solution”, and this in turn is probably also influenced by various drivers/variables of pressure and changing temperature etc. (not to mention the locations of at least any large columns of air e.g. from the pressurizing end etc.). While much at least surface source water is already “saturated” with air, this may not be true of all filling water sources, and also other conditions. In this regard I saw a writing many years ago (I think from a manufacturer of air valves), “A typical 1-mi (2-km) long pipeline of any diameter that has been supposedly vented of air will, in most instances, still contain enough dissolved air to completely fill over 100 ft (30 m) of the pipe.” Likewise, some water can soak into cement mortar linings of some types of pipe, also over time/pressure.
9. Lastly, it is also possible that there can be slight movement, settlement, or some extension of pipelines due to installation and/or Bourdon effects etc., again that has nothing necessarily to do with leakage but that would change the test volume and manifest itself in at least some loss of pressure during a long-term test.

I think many of these practical realities were not lost on at least the formulators of venerable standards that have existed for decades, e.g. ANSI/AWWA C600 for the installation and testing of iron pipelines. In this regard, if after more than 2 hours a particular closed off small diameter pipeline still holds basically a 160 psi level of pressure, while only losing a pound or two of water pressure, I suspect many Owners or installers of pipelines of many materials would probably be pretty happy with such an apparently tight line. As psnyder notes/quotes above, the C600 standard indeed states, “Testing allowance shall be defined as the maximum quantity of makeup water that is added into a pipeline undergoing hydrostatic pressure testing… IN ORDER TO MAINTAIN PRESSURE within ±5 psi (34.5 kPa) of the specified test pressure (after the pipeline has been filled with water and the air has been expelled*).” While I guess I understand your perception of a second required test (you call “leakage” test) and I guess you may or may not have any very slight seepage in this mile plus length of buried piping based on your results of that second test, see my items 1-8), if literal “C600” is the required testing specification it would appear in your case that based on the original testing ABSOLUTELY NO WATER WAS REQUIRED to be pumped back “in order to maintain...” (of course I added the caps in the quote above) Thus, this standard in and of itself does not really state that the line must be pressurized a second time e.g. for a “leakage test”, unless the line pressure is falling below the required test pressure minus 5 psi (in your case say 160 psi -5psi =155 psi?); then, and only then, would you be specifically directed to pump it back up and measure make-up water.

Now with regard my statement that I suspected many folks might be happy with a small line, at least pretty much full of water, holding pressure this tightly, I recently had occasion not long ago to myself to hydrostatically test a small exposed pipeline of predominantly flexible plastic piping (that was supposed to be “50-psi rated”) about 45’ long that had a bunch of fittings and valves etc. My first pressure test after slow filling to 35-40 psi test pressure failed miserably, with the pressure dropping off very rapidly once the line was closed off. I pumped it up again and the same thing happened. As I could not see any drips, but I did suspect there was some air trapped at locations of the piping, I eventually decided to “soap” all of my connections as well as my end ball valve closures (and indeed I found an air leak in one threaded connection). Once I fixed this problem, I pumped up and valved off this pipeline (that I then knew had no leaks) up yet again this time to 35 psi, and still this pipeline lost 5.3 psi in 2 hours. As I knew I had a good bit of air entrapped in this line, I decided to run yet another test, this time removing the air with high velocity water flow before I valved the line off. This time my closed and obviously then drip-tight little pipeline held at least a little better, but still lost pressure to a 30.4 psi gauge reading after 2 hours (I guess due to inelastic swelling and/or stretching of this plastic pipe!) One thing that ran through my mind is what would I have known, or suspected, if this were a buried pipeline and I could not as easily do the diagnostic and remedial work? In looking at manufacturers’ field hydrostatic testing guidelines for example for polyethylene pipes often perceived to be alleged “leak-free”, I have found that the make-up water allowance for testing such pipelines is in reality many times greater than that for ductile iron water pipelines (in the case at least of buried piping, how much of that much more liberal make-up water allowance they know is the result of some sort of precise swelling or stretching of the pipes, and how much might be due to air effects or leaking out of imperfect fusions or mechanical connections etc.?, would appear to be somewhat of a mystery).

Let me know if I misunderstand any of the conditions of your particular test.

 

[Trirdsavani]

Hi,
We are doing hydro testing of C900 10” Raw water pipe form Deep well to water reservoir. We are continuously losing from 175 to 165PSI and it holds forever 165 PSI. 175PSI is test pressure, which is 1.5 time
Please share you thoughts.
Thanks for your time.

 

[rconner]

This new pvc pipeline test that has piggybacked on this thread would appear to not be meeting the requirements of new ANSI/AWWA C605-05, that includes testing requirements for pvc pipelines. Arguably some like the earlier DIP standard this appears to require that (per my reading) pressure fall to no less than the specified test pressure minus a maximum of 5 psi. It appears in your case the pressure should be maintained to no less than 170 psi, if the starting pressure is say 175 psi. As to why this is occurring, with no more information supplied, is anyone’s guess. While it may or may not have anything to do with the problems you are experiencing, you may want to clarify exactly what pipe (DR etc.) you are using per “C900”, as pressure ratings etc. have not long ago changed (in accordance with the history etc. at

http://eng-tips.com/viewthread.cfm?qid=227976

) and there have been many thicknesses etc. of pipe per this standard. If you have run multiple tests with exactly the same pressure drop performance however, one would think you may well have a leak somewhere in this closed system (one would think it has to be either that, or the system keeps expanding with exactly the same ending resistance to said movement!)

 

[Trirdsavani]

Hi rconner,

Thanks, it was interesting to read your past threads too.

It is DR18 pipe which we are testing at 175 PSI. Working pressure of pipe is 116 PSI.Is there any ways to identify pressure drop cause by air inside the pipe and loss of water. If DR 18 pipe is rated for 235PSI and fittings as compatible to that, any chances to expand pipe and increase in makeup water (which is count as water losses during hydro test).

Thanks,

 

[rconner]

In partial answer to your first question regarding the pvc pipeline testing problem, in a normal installation you very well might not have enough accurate information to "identify pressure drop cause by air inside the pipe" (exactly how much air is inside the line, what is the original and any changing temperature and saturation levels of air and water etc. etc.) I think that is why I believe standards attempt to define based on experience etc. reasonable amounts of a "testing allowance" (not necessarily that this is small leakage, but also to some extent to account reasonably for many variables).

With regard to your second question, that is an interesting one, but while the long-term modulus of pvc (with a slow rate of load/stress application like a hydrostatic test) is much less than that most often published e.g. for rapid loading ring testing of pvc per ASTM D2412, you would think that the founders of new C605 standard would have accounted for any creeping pipe expansion in their requirements for holding pressure (and make-up water testing allowance). If so, that is not likely an answer to the testing failure, either.

 

[rickwitt]

When is it proper to use values 133,200 and 148,000 in the equation?

 

[rconner]

Essentially, I think both have been "proper" at different times in history. I believe 35 or more years ago the predominant testing allowance for at least iron pipelines was based on a formula that had 133,200 in the denominator. The current testing formula requirement for modern ductile iron pipe systems installed per ANSI/AWWA C600 requirements has 148,000 in the denominator. With all else being equal, the current (more rigorous) testing allowance would thus be about 133,200/148,000 or 90% of that accepted many years ago. [For whatever it is worth, I think you will also find that for many common testing parameters/requirements, the “10 gallons per inch per mile per 24 hours” rule of thumb that I understand has been promoted for both welded and rubber-gasketed steel pipelines, when pro-rated to whatever the required testing duration of 2 hours or more is for those lines, yields a practical testing allowance also that is pretty close to what is now required by ANSI/AWWA C600.]

 

[rickwitt]

Thanks. Would the same denominator be used for PVC pipe as ductile iron?

 

[rconner]

Yes {it appears they claim so per recent AWWA C605).

 

[rickwitt]

@rconner thanks.