Dispersion analysis or study 1

[engr2GW]

Hello,

I am looking for on direction on how to do a dispersion analysis; for PSV relief (or tank hatch). Intent is to incoporate this into spacing of equipment and possible ignition source.
Any rule of thumb, document/standard/code, or general direction to start from?

Thank you.

As much as possible, do it right the first time...

 

[The Obturator]

I'm no expert on dispersion analysis, but with regard to a Pressure-relief valve, are you looking for a flow rate that a PRV will give? or how to size a PRV ?

*** Per ISO-4126, the generic term 'Safety Valve' is used regardless of application or design ***

*** 'Pressure-relief Valve' is the equivalent ASME/API term ***

 

[RVAmeche]

Dispersion analysis is based off the flow rate, velocity, density, weight of fluid, discharge location (elevation, surroundings) and wind conditions. It's not the sort of thing you can guestimate with a rule of thumb.

In the past we've hired consultants to perform dispersion analysis. I would recommend the same approach if you don't have access and expertise in the programs used to perform the analysis.

 

[engr2GW]

Thanks for the reply @The Obturator,
I am familiar with sizing a relief valve, my question is with respect to discharge of flammable vapor/gas in relation to possible ignition source.
NOTE: These are all relief valves that are relieving to the atmosphere and there are no equipment above the relief valve, just open space

As much as possible, do it right the first time...

 

[The Obturator]

OK. You may want to consider information contained in API-521 which provides information on the causes of overpressure and methods of calculating the flow. This incudes fire cases. Probably for what your looking for, is within Section 5 on Disposal Systems and fluid affects etc. That info might help you.

*** Per ISO-4126, the generic term 'Safety Valve' is used regardless of application or design ***

*** 'Pressure-relief Valve' is the equivalent ASME/API term ***

 

[don1980]

Yes, there is a good (and easy to calculate) rule-of-thumb method for analyzing flammable vapor releases from PSVs. Refer to Eq. 55 in API 521. This equation (valid only for vertical releases into an open area) reveals whether or not the Nre is high enough to ensure safe jet-momentum dispersion. If the installation passes this simple test, then one can be confident that the flammable envelope will have the characteristic shape of a candle flame (height will vary depending on mass flow), and the risk of the gas forming an explosive vapor cloud is minimal. The predictable shape of the flammable envelope, which doesn't drop below the release elevation, is due to the high amount of air that's induced into the jet stream exiting the pipe. Air in induced by the low-pressure zone surrounding the pipe exit, caused by the high-velocity gas stream.

The width of the flammable envelope is much smaller than most people would predict based on intuition. The end of the tailpipe obviously needs to be at a higher elevation than any ignition source in the immediate vicinity.

Be aware that there's no similar rule-of-thumb for assessing toxic vapor releases. That's because the specific exposure hazard varies widely from one chemical to another. For this you need to do a rigorous gas dispersion analysis.

 

[georgeverghese]

There are 2 types of dispersion analyses : jet dominated dispersion and Gaussian dispersion. There are rules in the API which tell you if you are in the jet flow regime. I dont remember the rule now, but I recall it is related to the Reynolds no at the release exit.
One program which has routines for both types is PHAST. Others are licensed by Operating Companies. If you are doing this for an Owner - Operator, they may have a preference for their inhouse program to be used.

 

[TiCl4]

ALOHA is a free Gaussian dispersion modeling software, available from the EPA. However, do be aware of the limits of the gaussian modeling on flammability.

 

[pierreick]

https://www.epa.gov/scram/air-quality-dispersion-modeling-alternative-models

Hi,
Consider the resource above to support your work .

Pierre