Pilot Gas Calculation for Flare

[OilBoiler]

Hey everyone,

I'm new in the flare world. Currently, we have a flare that is designed for using propane as the pilot gas, but is too expensive for us right now, so we'd like to change it to natural gas that our refinery already has. This is a flare for our docks, so it operates more on a batch basis. I know the MW, mole%, volume, temperature, and pressure of the vapors to be burned by the flare. I would like to know that if based on this information, I could calculate how much natural gas I need? Are some type of combustion calculations needed? I could find out how much propane we use right now and I'm actually waiting on that and I could find the amount of nat. gas needed based on the HHV equivalences, but I just would like to learn if there is any other way to find this. Thanks!

 

[NeedAHoliday]

I do not work specifically with flares but I am involved in specifying burner duties for other processes, so I will attempt to explain the theory with respect to combustion calculations as they are the same essentially.

The easiest way is to compare the calorific content of NG against propane for the set point being used with your current propane burner(s).

The calculations you speak of can certainly be established by taking the mass flow rate, the specific heat of your gas (based upon consituents) and the dT of your stream from pre-combustion to destruction temperature i.e. mCpdT.
Within the dT you can account for the calorific content of the compounds being destroyed as this will have a regain effect on the kW being used.

This will allow you to determine volume flow of NG based upon calorific content vs kW required.

Remember to account for any vaporisation effects etc as this will effect your burner power as well as your efficiency of combustion.

A point to check is the burner itself as I have found that generally some burners will have altered characteristics of operation if you change fuel types. You may find that the pressure to the burner can be decreased very slightly when using NG over propane but not enough that I would worry about altering set points, and your flame length will most likely reduce, also the amount of combustion air will alter proportionally to the variation in fuel flow.

I know this is very bsic but with the limited information provided in your query it is hard to formulate a more defined answer.

Cheers.

P.S. Anyone who works specifically in flares please feel free to add or query post, as I siad this is a little out of my field.

 

[TD2K]

I'd talk to the flare vendor as it might require a change to your pilots. The pilots I've seen typically have an orifice which is used to mix atmospheric air with the fuel and then delivers it to the pilots where it's burnt (very crudely, analogous to your barbeque). MW is nearly 3x times higher in MW so the flow through the orifice with NG is going to be much less than with propane. You may be able to compensate for this depending on the pressure of th NG you can supply but I'd talk to my flare vendor first.

 

[Flareman]

OilBoiler

As a general rule, flare pilots will be designed for a specific heat release. This is commonly around 50,000 - 75,000 Btu/h but could be as much as 250,000 Btu/h. The usually way to relate comparable requirements for different gases based on jet flow is through the

Wobbe Number ........ which is

Calorific value (btu/scf)
-------------------------
sqrt{specific gravity}


Then

sqrt{Pressure ratio} is proportional to Wobbe number
(for pressures less than critical)

This is a little simplistic if you have a venturi involved because, if you reduce the MW of a hydrocarbons you usually need a lesser air/gas ratio for combustion, which would normally suggest reducing the gas pressure at the venturi, whereas the Wobbe formula says increase the pressure at the jet. Many pilot designs will compensate for this apparent problem because they are not designed critically and both gases may fall within an acceptable operating range for the venturi.

Natural gas and Propane are about as far apart as one could realistically expect within a single venturi-based design with no changes.

Usually, the pilot capacity has little to do with the gases being burned in the flare itself because the pilot simply provides a stable flame to continuously re-ignite the main flame. Unless the main flame gases are very lean or almost non-flammable it is probably not an issue.

Having said all that, I agree with TD2K that checking with the flare vendor is the most appropriate course because he understands (we hope) the design which he provided and should know its capabilities.

 

[Flarespec]

Nothing that has been written so far is wrong but the Wobbe thing is a bit of a red herring.

The chief objective in pilot operation is to supply a pilot gas/air mixture to the pilot nozzle at a sufficiently high pressure that a stable flame is sustained at all times, including occasions when there might not be any secondary air available continuously at the pilot nozzle, or "wind" conditions at the nozzle that will produce an axial vector that can disturb the desired mixture.

The fact is that, in this case, a certain pilot topography exists - and the design of the whole pilot is relevant*. As it is quite likely that the operation of the flare prevents quick changes, I read the problem as what can you do with what you've got in place.

A reliable flare manufacturer should have supplied equipment for "optimal" operation within the original design constraints. Thus any change in gas will represent a change from this optimal design. But what represents "optimal" is very broad and is not amenable to verifiable quantification. It is quite likely that you can just change the gas and, probably, its pressure to achieve successful operation. But the aim, I would argue, is to achieve a balance within the pilot to get a gas/air mixture that is approximately of the same degree of aeration as the original design but using the "flammable range" as the measure of this function (rather than the straight v/v analysis). This involves more than the chemistry-based Wobbe assessment. The venturi action is a momentum process and the velocity of the entraining raw pilot gas is as important element in developing a "good" mixture as is the mass (i.e. quantity) value as derived by the Wobbe calculation. (It also is relevant to point out that changing the gas also changes the flammability limits).

On the other hand, you have no real confirmation that the operation on propane is "optimal", nor would you know that following Flareman's advice is correct until you observe a pattern of pilot failure that is markedly different from your previous experience. My advice is to follow Flareman's suggestion and keep a close watch on the pilot. As was said in my youth, "Suck it and see!".

(Incidentally, just putting in the same heat release to the pilot will make life easier with any regulatory authority to whom you may have to report!)

* The gas orifice size and inlet shape, the position of this orifice w.r.t. the venturi, the shape and internal finish of the venturi, the length, size and shape (e.g. are there any bends in it?) of the air/gas mixture tube, the arrangement and area of the burner nozzle and , maybe, the arrangement for keeping the wind away from the pilot flame.

 

[ProchemPL]

You should use API521 "Guide for Pressure-Relieving and Depressuring Systems"

 

[TD2K]

Where in API 521 does it go into pilot gas calculations?