Converting leakage rates (mg/s) to ppmv concentrations.

[A8yssUK]

Hello everybody,
I work with the ESA (European Sealing Association) and as part of my work with them I also liaise with the FSA (Fluid Sealing Association of America).
The ESA/FSA are jointly working on a new test standard for spiral wound flange gaskets. (SWG)
I have a problem perhaps you guys can help with,

Basically we have a bolted flange system, correctly installed, pressurized and heat cycled and we measure the "leakage" from the system with a sniffer device (either methane or helium), at multiple stages. The system is encased so that any leak should be taken into the sniffer device, but isn't fully vacuumed out so that a flow rate past the outer seal face is mantained.
The measured output is in ppmv or parts per million by volume. This is a normal "leakage" measurement in the industry on plants..

Now I am European and I work in a lab So I use normalised mass transport values for leakage rates, normally mg/s/m (miligrams per second per meter of seal length) as used in EN1591 / EN13555

I want to devise a method to convert between the two measurements, however this is seemingly quite difficult even when you make some quite large assumptions.

For the lab test device if i assume all the leakage from the device is sampled through the sniffer and not lost, i should be able to convert to a mass rate right..??

ppmv > mg/m^3 is easy (

https://en.wikipedia.org/wiki/Air_pollutant_concentrations)

so i know how many mg I have for the reading by then need a sampling rate of the sniffer to change it to a rate.

The sniffer device is rated for EPA method 21 useage, and section 6.4 of the EPA method 21 states the sniffer should sample between 0.1 and 30 L/min.. a 30x spread!
therefore,

mass rate (mg/s) = concentration (mg/m^3) * sample rate (m^3/s)

So for a gasket with a "length" (or inner circumference) of 1m if it is giving a reading of 1 ppmv Helium then the leakage rate is between 2.7e-7 and 8.32e-6 mg/m/s

In Europe The EN1591 states/will state that a minimum tight seal has a leakage rate of 0.01 mg/m/s. Now inverting the equation, a 0.01 mg/s leakrate measured by an EPA method 21 capable sniffer would give a value between 1202.0 and 36061 ppmv which is a massive spread!!

Redoing but with methane (16.032u) as the test gas, 0.01 mg/m/s I get between 300 and 8996 ppmv

OK, firstly is there anything wrong with my math/assumptions ?

Now I don't have a lot of experience with sniffer devices, but can they not be controlled to give a tighter range of sample rate ??

Anybody else know of any other standards that may contains sampling rates that sniffer devices conform too ?

Thank you

Edit : I've also now located ISO 15848-1 Annex B section 5.1.1.1 d) which states a flow rate of 0.5 - 1.5 L/min which makes the spread not so bad..

 

[chicopee]

There is no way that I can see about converting ppmv to mg/(m^3-s) unless you have some flow rate (L^3/s)(from the leaks.

 

[A8yssUK]

some workings and I have tried to make it easier to follow for discussion,

1 PPMv = 1 ml/m^3 i.e. there are a million ml in a m^3 so 1 part per million by volume is 1ml in a m^3 ( The metric system is wonderful )

At standard temperature and pressure ( STP 20°c, 101.325kPa ) 1ml of helium weighs 0.1636 mg or 1ml of methane weighs 0.6572 mg (Using Ideal gas constant / laws)

So if a helium sniffer sampled at 1 m^3 per second and read 1 PPMv the mg/s value would be 0.1636 mg/s and methane 0.6572 mg/s.

However sniffers run at much lower sample rates ISO15848-1 (valve testing standard) annex B sets the rate between 0.5 and 1.5 L/min, so lets say 1 L/min which is (1/1000)/60 = 1.66e-5 m^3/s

So if we assume all the leakage is going straight to the sniffer and not being lost/collected anywhere (Due to our test being in a lab and the partial encasing of the test device)

To give a steady reading of 1 PPMv of helium at a sample rate of 1 L/min then every second 0.1636 * 1.66e-5 = 2.73e-6 mg of helium must be going through the detector.

So for a 1 L/min sample rate 1PPMv = 2.73e-6 mg/s of helium or 1.10e-5 mg/s of methane being sampled.

Now the test gasket is a 3" class 600 spiral which has an inner diameter of 101.6mm, which equates to an inner circumfrence or "seal length" of pi*101.6 = 319.2 mm = 0.3192 m

So if the seal length was 1m the leakage rate per meter of seal length would be 2.73e-6 / 0.3192 = 8.55e-6 mg/m/s He or 3.45e-5 mg/m/s CH4

The value however depends highly on the sniffers sampling rate!!

 

[Compositepro]

So, why does this surprise you? If you understand how the test works, of course it depends on the sample flow rate (sampling rate is a term usually used to describe the frequency of taking samples or measurements). Measuring this sample flow rate is part of calibrating the test instrument.

 

[A8yssUK]

What surprises me is that I cannot really find much data about how accurate the flow measurement on helium/methane sniffers are, data I have found seems to imply that multiple devices from multiple manufacturers all seem to have different (sometimes widely) different sample flow rates, some cheap devices may not even have a constant flow rate,some expensive devices do allow you to set a rate, but I doubt these are readily available on plants.

Anyway more research points to many sources stating that EPA Method 21 style measurements are not good enough for quantifying leaks only detecting them, and new technologies are under review the most relevant being optical imaging monitoring techniques. So perhaps our testing method is a little old school and we should incorporate some newer technologies.

 

[A8yssUK]

Just an update,

I located how the EPA change the PPMv values into kg/hr emission rates, which are obviously easy to convert to mg/s. To cut a long story short they use a correlation chart from measurements made on plants in the late 80's early 90's. As you would expect that is a terrible way to measure/convert anything.... anyway that is what is used.

Info,
EPA calculation document for PPMv screening values (SV) to kg/hr emission rates (ER)

https://www3.epa.gov/ttnchie1/efdocs/equiplks.pdf

Table 2-10 is most relevant as it gives the equation for flanges in the oil and gas industry.*

Here is proof the correlation values from 1995 are still being used in calculations,

https://www3.epa.gov/ttn/chief/efpac/protocol/Protocol Report 2015.pdf

Table 2-2

Here is some info on a new technique they are testing involving gas detecting cameras that can quantify leakages (qOGI) i.e. directly measure emission rates in kg/hr

https://www.isa.org/participate-in-...opments-in-epa-ldar-enforcement-chris-lehmann