Flare

Dear all:
i am going to adjust&optimize fuel gas which is using as sweeping gas for flare
tailpipe in our gas refinery.
the problem is that two different velocity has been founded in the contractor
documents for fuel gas velocity.(0.2 & 0.02 ft/sec).

1) would you please advise me how can I find right velocity ?

introducing standard or other document are highly appreciated.

As per Fluor's process manual,continuous sweep gas velocity for elevated flare
would be as follow:

- Normal operation: 0.10 ft/s (0.03 m/s)

in stack/tip
- Upset condition: 3.3 ft/s (1.0 m/s) in
stack/tip

Thus,you can use this recommended practice to calculate required gas velocity
in tail pipe and to know which value is correct.

Fallah

Useful data can be found at http://www.cheresour...n-flare-header/. According to

post by gvdlans there (8 Nov 06), purge gas velocity (depending on gas nature
and header diameter) can be calculated (after H Husa & API 521) as per attached
purge.xls (for example).
Nevertheless the thread continues by referring to BP standard RP44-3 (design
guidelines for relief disposal system), that can be now downloaded
from http://igs.nigc.ir/igs/BP/RP44-3.PDF. Recommended safety factors of 2 to 5
(the latter for densities close to air), to cover high wind conditions, indicate
velocities of fuel gas purging close to 0.2 ft/s, not 0.02 ft/s.
Besides it is understood that purge velocities below 0.5 ft/s are not common
(BP/RP-44-3, 7.1.6); also that Fluor practices may increase purge rate 33 fold in
an upset condition (strong wind?), as indicated by Fallah.
So it is concluded that between a purge velocity of 0.2 ft/s and 0.02 ft/s, correct
is the value of 0.2 ft/s. But I have the feeling that even this value is low. The last
time I dealt with purging (1990), I was instructed to assume a velocity of fuel gas
of 1 m/s = 3.3 ft/s. A lot may have changed since then; hopefully a member with
experience will check this post (providing raw data at least) and give clearer
information.

Attached Files: Purge

dear All:
for more clarification:as all knows different stream from different units
connected to flare header in the refinery,and two purge gas streams is used :first
connected to end point of header in each unit which I named tail pipe flare purge
gas and second which injected in the stack inlet.
I think all recommendation is related to second one (stack inlet purge gas).

my question:
1) is tailpipe flare purge gas calculated as per stack inlet purge gas?
2) what is the function of tailpipe flare purge gas?

1) As said, my knowledge on flaring is limited and experts' help would be

welcomed. Meanwhile I have not heard of purge gas injected into flare
stack.http://www.gasflare..../Flare_Type.pdf.
This could explain the low velocities reported for stack (0.02 - 0.2 m/s)
versus higher purge velocities heard for tail pipe (previous post), if stack
received injected purge gas only. Nevertheless purge gas in stack must be
the sum of the two (tailpipe+injected) (plus some water vapors if seal
drum is used).
In the few cases examined by me in the past, tailpipe flare purge gas
continued into the stack, passing from the water seal drum. So stack inlet
purge gas=0 (as additional flow).
2)Tailpipe flare purge gas creates a positive flow into the flare to prevent
air back flow from tip under conditions that could occur (which could
create combustible mixtures with the gas of the header).
After a peak of hot discharge, vacuum can be created in tail pipe when gas
flow becomes (almost) 0. Tail pipe flare purge gas flow prevents air
ingression. It is an additional protection to the water seal drum (or other
seals).
Apart from this, air can intrude from valves after a hot discharge, since
valves with guaranteed no external air ingression under internal vacuum
can be hardly found (if at all). So tailpipe flare purge gas prevents it, even
in case of no discharge to flare.

I have heard of controlled purge flow after a hot discharge, but not seen in
design or at field. Also that said controlled purge flow can be local, which
indicates a sophisticated purge system and instrumentation.

Flare purge rates

In order to minimise flaring how do we calculate the minimum purge rate required for the flare header? As I
understand the purge is just to ensure no flashback into the header and also to prevent a flame out of the
pilot.What data,and calculations are reqd for this? Thanks

Some of the data you require for these calculations are:

- average molecular weight of purge gas
- whether or not any "seals" are present in the system i.e. molecular seal, liquid seal or velocity seal.
- Size of the flare header
- presence of hydrogen (that changes everything as the burning velocity is very very fast)
- type of flare tip (pipe, coanda or other proprietary make)
- minimum recommended purge from the flare tip manufacturer.

Hello, if you read the API-RP-521, they show you this equation:

Qg = (0.005xd^3)/M^0.565

Where:
Qg: Gas flow (mmscfd)
d: diameter of the flare (in)
M: Molecular weight

I2P
Relative to the purge, basically, if you need to have the flare available all the time, you need to keep it "alive"
by constantly putting gas to it in the main header.
The question is only .. how much is enough?
 
The flare vendor will give you a recommendation about the minimum flow but you really need to find this out
for yourself as it applies to your gas and your flare tip.
 
The minimum flammable flow you need is that which will just  keep a small visible flame at the tip and prevent
burn-back.  When this flow goes too small to flame will try to disappear into the top of the tip.  It may make
smoke because the flame inside the tip does not have enough air.  If you have smoke suppression (air or
steam?) turn it down (or almost off) whilst you try to find the minimum so that it doesn't overpower the
flame.  Then when you find the minimum put back just enough smoke suppression flow to clear up the flame
and still allow it to be visible.  You may need a bit more gas to stop it going back into the top of the tip.  If
smoke appears when you have no flame or if the smoke is a single column coming from the center of the tip,
that is a sign of burn-back.  Drop me an email at flareman_xs@netzero.net and I'll send you a paper which
has a formula to give a first order estimate of the burn-back rate for a generic tip design.
The rate will probably be something like 0.2 fps based on the internal flow area.
 
Purge rates are often quoted at much lower rates than this.  Those are the rates needed to prevent air from
getting back into the riser, when there is no flame.  My paper also covers a formula for this.
 
Sometimes, purge rates on plant are specified by the contractor on the basis of sweeping or sweetening the
flare sub-headers.  In these cases, the sub header rates may be based on a Reynolds number condition (say
2000 Re) to keep a turbulent flow.  These flows often produce a much greater requirement than that actually
needed by the flare itself.
 
Additional factors which influence the amount of purge used are
 
- the use of an instrumented purge input.  Often this is set to keep a positive pressure at some point in the
headers.  I cannot argue against the logic for this as it is a plant decision intended to ensure that all leakages
are outward.  However, it does consume a large amount of purge gas when the gas MW is less than air
because the vertical height of the stack sets the system pressure due to the weight of gas in the riser, which is
probably already negative at the bottom of the flare when referenced against atmosphere in a PSL or PT.
 
- supplementary purges which allow for system shrinkage.  If there are relief cases which heat up the header
and then cool rapidly, or dramatic ambient temperature changes, or gases with a high probability of
condensation in the header, the contractor may have determined the need for a supplementary purge intended
to overcome the consequences of shrinkage.
 
These are things which need to be reviewed by the plant process engineers and addressed individually.
 
When all these things are considered you will turn up with a number of different rates for purge.  Then, you
have to decide which ones are the most important based on how you run the plant day-to-day.  In a large
number of cases, the day-to-day flows from vents etc. exceed the purge rates.  So you may be able to settle
on a permanent purge rate based on keeping air out of the system for no flame, and handle all of the other
cases by operational procedure.  This means watching the flame for burn-back and adjusting the flow if
needed, monitoring for temperature changes and sweeping hot gases from the system before they condense,
etc. according to the specific condition.
 
In all these cases you will continue to have permanent pilot burners.  Each will consume fuel.  Modern pilots
only use about 50,000 - 70,000 Btu/h each, though it is a lot of gas on a yearly basis.  If your pilots are the
older design and consume, say 200,000 Btu/h each (a big visible flame), then you may wish to think about
modernizing.
 
The 50,000 Btu pilot gives a fairly small flame which is often difficult to see against a bright sky.  That often
makes operators turn up the flow for their confidence level.  Most modern pilots are fitted with thermocouple
to register the flame and, in many cases also start an automatic relight sequence if the  flame goes out.  This
overcomes the low confidence issues.

 
As an aside ... thank you to Nosey for your kind remarks  .  I do what I can!!  The web site is suffering from
lack of attention right now but I'll get back to it with more things soon.

David

Safe gas flare system operation,

flow and control
Part 1: Operate this critical mechanical equipment safely.

Amin Almasi
Jan 25th, 2018
Image courtesy of Amin Almasi
Flaring is a combustion control process for gases (such as flammable
gases and hydrocarbons) in which the gases are piped to a remote,
usually elevated, location and burned in an open flame in the open air
using a specially designed burner tip (often flare stack tip), auxiliary fuel
and steam or air to promote mixing for nearly complete destruction.
Completeness of combustion in a flare is governed by flame temperature,
residence time in the combustion zone, turbulent mixing of components
to complete the oxidation reaction, and available oxygen for free radical
formation. Elevating the flare can prevent potentially dangerous
conditions at ground level where the open flame (i.e., an ignition source)
is located near a process unit or operating facilities. Further, the
products of combustion can be dispersed above working areas to reduce
the effects of heat, smoke, noise and objectionable odors.

A flare is a critical mechanical piece of equipment intended for the safe,

reliable and efficient discharge and combustion of gases and materials
from pressure-relieving and vapor-depressurizing systems. Since it is
critical to the safety of an operating plant, a flare should be continuously
available with high reliability and capable of its intended performance
through all operating plant emergency conditions, including a site-wide
general power failure or whole plant emergency trip. The flare and
related mechanical components should be designed to operate and
properly perform for the specified service conditions for an overhaul-to-
overhaul period, say five to seven years, without the need for an outage
of the operating facility. If a flare fails (for instance, because of a tip
failure), whole systems that rely on it for protection should usually be
tripped for some days (or even a week or more) for the flare repair,
which can result in considerable financial losses.

In most flares, combustion occurs by means of a diffusion flame. A

diffusion flame is one in which air diffuses across the boundary of the
fuel and combustion product stream toward the center of the fuel flow,
forming the envelope of a combustible gas mixture around a core of fuel
gas. On ignition, this mixture establishes a stable flame zone around the
gas core above the burner tip. This inner gas core is heated by diffusion
of hot combustion products from the flame zone. Cracking can occur
with the formation of small, hot particles of carbon that give the flame its
characteristic luminosity. If oxygen is deficient and if the carbon
particles are cooled to below their ignition temperature, smoking occurs.
In large diffusion flames, combustion product vortices can form around
burning portions of the gas and shut off the supply of oxygen. This
localized instability causes flame flickering, which can be accompanied
by soot formation. As in all combustion processes, an adequate air
supply and good mixing are required to complete combustion and
minimize smoke. The various flare designs differ primarily in their
accomplishment of mixing.
Practical notes

Flares can be used to control waste gas flows, and they can handle
fluctuations in gas concentration, flow rate, heating value and inert
content. Flaring is appropriate for continuous, batch and variable flow
vent and relief stream applications. The majority of plants have existing
flare systems that are designed to relieve emergency upsets that require
the release of large volumes of gas. These large-diameter flares designed
to handle emergency releases can also be used to control vent streams
from various process operations.

Consideration of vent and relief stream flow rates and available

pressures should be given for a flare design. Normally, emergency relief
flare systems are operated at a small percentage of total capacity (total
capacity is supposed to be for total plant shutdown) and at negligible
pressure. To consider the effect of controlling an additional vent stream,
the maximum gas velocity, system pressure and ground-level heat
radiation during an emergency release should be evaluated.

The whole flare system should be evaluated if the gas stream pressure is
sufficient to overcome the flare system pressure, since it is not
economical nor technically possible to provide gas mover systems
(compressors or blowers) to a flare system. Other considerations should
also be respected such as maximum gas velocity limits or groundlevel
heat radiation limits for the flare system and flare stack location, height
and details. A flare system should be optimized while considering effects
and factors such as pressures, flows, hydraulic effects, flare stack height,
radiation limits, temperature limits, material selection for all
components and optimal cost of the whole system. To ensure an
adequate air supply and good mixing, some flare systems inject steam
into the combustion zone to promote turbulence for mixing and to
induce air into the flame.

Operation & design

The flammability limits of flared gases influence ignition stability and
flame extinction. The "stoichiometric" composition is a chemically
correct mix of air and flammable gas capable of perfect combustion with
no unused fuel or air. The flammability limits are defined as the
stoichiometric composition limits (maximum and minimum) of an
oxygenfuel mixture that will burn indefinitely at given conditions of
temperature and pressure without further ignition. In other words, gases
should be within their flammability limits to burn. When flammability
limits are narrow, the interior of the flame may have insufficient air for
the mixture to burn. Gas fuels with wide limits of flammability are
therefore easier to combust.

For most vent streams, the heating value also affects flame stability,
emissions and flame structure. A lower heating value produces a cooler
flame that does not favor combustion kinetics and is more easily
extinguished. The lower flame temperature also reduces buoyant forces,
which reduces mixing. The density of the vent stream affects the
structure and stability of the flame through the effect on buoyancy and
mixing. Gas velocities in many flare systems are extremely low;
therefore, most of the flame structure is developed through buoyant
forces as a result of combustion.

Lighter gases tend to burn better. In addition to burner tip design, the
density directly affects the minimum purge gas required to prevent
flashback, with lighter gases requiring more purge. Poor mixing at the
flare tip is the primary cause of flare smoking when burning a given
material or gas. Streams with high carbon-to-hydrogen mole ratio (say
greater than 0.34) have a greater tendency to smoke and require better
mixing for smokeless flaring. For this reason, if steam injection is
required and selected, one generic steam-to-gas (gas being flared) ratio is
not necessarily appropriate for all vent and relief streams. The required
steam rate is dependent on the carbon-to-hydrogen ratio of the gas being
flared. A high ratio requires more steam to prevent a smoking flare.
Vent and relief streams are sent from the facility release point to the flare
location through the gas collection flare header. The flare piping is
designed to minimize pressure drop. Ducting is not used because it is
more prone to air leaks. Valves should be kept to an absolute minimum
— usually, no valve is allowed expect one or two check valves in a whole
flare system. Piping layout is designed to avoid any potential dead legs
and liquid traps. The piping is equipped for purging so that explosive
mixtures do not occur in the flare system on startup or during operation.
An important consideration in the material selection of flare and piping
design is the possibility of low temperatures as the result of relief and
blowdown of high-pressure gases. For instance, as a rough indication, in
a unit with gas pressure above 90 Barg, it could be expected that some
low temperatures be achieved in case of blowdown or relief. Different
scenarios should be established and simulated to find the lowest possible
temperature. High-pressure gas inventories, if assumed at an ambient
low temperature (say between 2°C and 10°C), can produce low
temperatures of around -30°C to -50°C in flare piping. A flare system
design temperature of -50°C or -60°C is not unusual for a high-pressure
facility.

Liquids that may be in the vent stream gas or that may condense out in
the collection header and transfer lines are removed by a "knockout
drum." The knockout drum is typically either a horizontal or vertical
vessel located close to the base of the flare stack. Some traditional flare
designs used a vertical vessel located inside the base of the flare stack as
a knockout drum. Liquid in the flare stream can extinguish the flame or
cause irregular combustion and smoking. In addition, flaring liquids can
generate a spray of burning materials that could reach ground level and
create a safety hazard. For a flare system designed to handle emergency
process upsets, a knockout drum should be sized for worst-case
conditions (for instance, total plant shutdown), and it is usually quite
large.

In next month’s February issue, Part 2 of this article will discuss design
considerations in more detail.

Amin Almasi is a senior rotating machinery and equipment consultant.

He is a chartered professional engineer of Engineers
Australia and IMechE. Almasi is an active member of Engineers
Australia, IMechE, ASME and SPE and has authored more than 100
papers and articles dealing with rotating equipment, condition
monitoring, offshore, subsea and reliability.

Optimising sweep gas flow in a flare header

How to calculate the minimum flow of flare header sweep gas needed to maintain safe conditions

SATISH MATHUR
Bechtel India

Download Complete Article

Viewed : 21015

Article Summary

The purging of flare systems prevents the ingress of air through the open stack, which can create an inflammable
mixture of hydrocarbons and oxygen inside the system. Stack purge, together with purge reduction seals, is generally
used to prevent this, in addition to a make-up gas for sweeping the flare headers. Some estimates of header sweep
gas quantities are suggested by Duggan, Simpson and others.1

With rising energy costs, there is a need to minimise the amount of gas used for sweeping without compromising the
safety aspect, which is that there should be no air ingress in the system. This article develops the correlations for
estimating the header sweeping gas requirement needed to compensate for the shrinkage effect due to ambient
cooling or rapid cooling of the gases after a hot gas release. It also provides guidelines to facilitate a decision on
investing in instrumented intermittent purge systems, which can provide further savings in the quantity of sweep gas.
The practice of using rule-of-thumb flare header sweeping gas rates could lead to substantial annual utility costs in
large flare networks and is therefore not a desirable practice.

A typical flare network consists of several sub-headers that collect the discharge gases from relief valves, process
flaring control valves, emergency depressurising systems and equipment depressurising for shutdown purposes. In
large complexes, the sub-headers may terminate in unit knockout drums for liquid removal before joining the main
header leading to the common flare stack.

The flare stack can have a water seal to prevent any flashback from the stack to headers (except in cryogenic
applications), a seal on the stack to prevent infiltration of air and a dedicated continuous flare stack purge. The
quantity of stack purge gas required is dependent on the size of the flare tip, the composition of the purge gas, the
composition of the waste gases and the design of the seal.2 In most cases, the seal purge gas quantity is specified by
the seal supplier.
In addition, the unit area flare headers and sub-headers are provided with a sweeping gas supply in most
installations, although some flare systems with a water seal and stack designed for internal explosion may not require
continuous header purge.3 This is to compensate for a shrinkage effect due to ambient cooling or rapid cooling of the
gases after a hot gas release in the flare header network.

This article deals with header sweeping gas quantity, which is independent of the stack purge flow due to their
different objectives, as explained above. Natural gas available from the plant fuel gas system is often used for
purging, both at the stack as well as in the sub-headers, with a backup gas supply source for ensuring uninterrupted
gas availability.

Basis
Under normal operation, when the plant is in steady condition and if the valves are not passing, there should be no
flow of waste gas into the flare headers. In this condition, the stagnant gases in the flare headers can be subjected to
volume shrinkage due to cooling. The factors influencing shrinkage are:
• Temperature of gas in the header after stoppage of flaring
• Ambient temperature
• Wind speed
• Flare pipe surface area
• Heat capacity of flare pipe metal
• Thermal conductivity, viscosity and density of gas.

Method
Heat transfer from the hot flare gases to ambient is by the following means:
• Inside the pipe: by natural convection (gases inside the pipe are considered stagnant when not flaring)
• Across the pipe wall: by conduction
• Outside the pipe: by natural convection, forced convection and radiation.
Theoretical analysis of natural convection heat transfer inside enclosed surfaces is provided in the Nusselt equation:4

NNu = a (NGr NPr)m

where
NNu = hiL/k = Nusselt number
NGr = L3ρ2GβΔt/μ2 = Grashof number
NPr = cμ/k = Prandtl number.
The factors “a” and “m” for horizontal cylinders are given in the reference. Using these, hi, the heat transfer coefficient
for natural convection inside the pipe, is computed.
When a heated flare pipe surface is exposed to flowing air, the convective heat transfer outside the pipe is a
combination of forced and free convection. For this mixed convection condition, Churchill recommends the following
equation for computing the heat transfer coefficient h:5

(Nu - δ)j = (Nuf - δ)j + (Nun - δ)j

The forced convection Nusselt number for the horizontal pipe of diameter D is given by Incropera and Dewitt5 as
follows:

Where ReD = VD/ν = Reynolds number.

The natural convection Nusselt number for the horizontal pipe of diameter D is given by Churchill and Chu5 as
follows:

Where RaD = g.β.ρ.cp(Δt)D3/(ν.kf) = Rayleigh number.

The overall Nusselt number Nu is computed using j = 4 and δ = 0.3. Once the Nusselt number is computed, the heat
transfer coefficient external to the pipe is computed as:

Continuous purge gas flow in flare network

3 Min Read
 Share This!

Flare/Vent network in oil refineries, oil and gas processing plants, all the relief
devices that can emit hydrocarbons are connected to a flare/vent network. The
flare/vent network ends up in a flare or vent stack which continuously emits the
purge gas hydrocarbons to atmosphere. A vent stack does not burn the
hydrocarbons and simply disposes them to the atmosphere at a safe location.
Hence gas flow emitted by a vent stack is typically small. When the gas flows in to
be continuously disposed are quite high, the gas needs to be combusted before
disposing it to the atmosphere. This consumption is achieved in a flare stack. The
hydrocarbons are continuously combusted at the flare stack tip before disposing to
atmosphere.

Air Ingress

Typically the flare/vent network along with the vent KO drum operates at a low
pressure close to atmospheric pressure. This pressure is actually the built-up back
pressure due to the continuous flared or purged gas flow in the vent stack.
Pressure at the vent stack tip is atmospheric and back-pressure in the knock out
drum is atmospheric pressure plus the frictional pressure drop from continuous
venting/purging of the gas.

If the flow of gas to vent/flare stack stops for some reason, there is a possibility of
air ingress into the flare stack and into the vent KO drum, vent network. This can
result in an explosive mixture of air and hydrocarbons in the vent/flare network,
which can be catastrophic.

Continuous purging of hydrocarbon gas to avoid air ingress

One of the ways to avoid air ingress into the flare stack, vent KO drum, flare
network and subsequent catastrophic consequences, is to continuously purge a
small flow rate of hydrocarbon gases. This continuous hydrocarbon gas purge flow
in the flare network helps to build up some positive backpressure at the vent knock
out drum. This is the normal operating pressure for the vent KO drum when no
other sources (relief valves, blowdown valves etc.) are emitting any relief flow. Thus
continuous purge gas flow maintains a small positive backpressure in the vent KO
drum, vent header and helps to keep the air out.

It is important for the source of the continuous purge gas flow to be located at the
farthest end of the flare/vent network from the flare stack. This way the constant
positive back pressure can be felt throughout the flare network, at all the relieving
sources.

API standard 521 describes the equation to determine minimum purge gas flow
requirement for a flare stack of given diameter and given composition of the
continuous purge gas flow.

Another ways to avoid ingress into the flare stack, flare KO drum, flare network are
- liquid seal, flame arresters etc. Although flame arresters are not frequently used
due to possibility of obstruction in the flare stack.

Estimation Of Purge Gas Requirement In

Flare Header
I am currently working on modification to a Process oil and gas platform. We are
adding launchers and receivers. Main drive behind this project is to stop flaring
and divert gas to onshore. Currently, the gas is flared. The gas is 80 mol%
methane 8% ethane and 5% propane.

We have a 24" flare header to the scrubber. A 20" line from scrubber to sub sea
flare. Gas flow rate is 72 mmscfd.

Once flaring is stopped we need to provide a continuous purge.

I am looking for a way to estimate purge rate with the 80% methane stream.
Purge gas is at 65 psia and 94 Deg F. I read in API-521 that there is an article by
H. W. Husa, "HOW TO COMPUTE SAFE PURGE RATES", published in Hydrocarbon
Processing and Petroleum Refiner -1964.
Any inputs regarding this will be useful.

i can be reached at goldsuresh@hotmail.com

regards
K Suresh

Following I copied from a discussion on eng-tips.com (post was by Flareman),

see http://www.eng-tips....d.cfm?qid=15647

There is a paper by Husa which basically says that the oxygen profile which
establishes in a vertical open pipe to atmosphere varies as an exponential
relationship of the pipe diameter, purge rate and the distance down the pipe.
This formula can be expressed as

O2% = 21 * exp{- U * L /0.0036/Fb/(D^1.46)}

U = purge velociy - fps
L = distance from top - ft
Fb = Gas buoyancy factor
D = diameter - inches

If the purge is a single component, then

Fb = exp{0.065*(28.96 - MW)}
If the gas is multicomponent, Husa's formula becomes a little top heavy and I
recommend using
Fb = 6.25*[ 1 - 0.75*((MW/28.96)^1.5)]
MW = Mol wt of purge gas (mixture)

In the same discussion thread, there was a quote from API RP 521:

"Alternatively, the continuous introduction of purge gas can be used to prevent

flashback. Studies [13] have shown that a safe condition exists in situations that
involve hydrocarbon-air mixtures if a positive flow of oxygen-free gas is
maintained, allowing the oxygen concentration to be no greater than 6 percent
at a point 25 feet (7.6 meters) from the flare tip."

So you can find required purge velocity U by entering O2% = 6 and L = 25 feet, as
well as the applicable MW and D for your particular situation into the above
mentioned formulas.

I may have a hardcopy of Husa's paper somewhere, but this will take time to
find... From what I remember it did not contain much more than the above
mentioned formulas.

Please check out appendix C of following

document: http://igs.nigc.ir/i...d/BP/RP44-3.DOC
The method as it is described in this appendix is similar to what I used in the
past, apart from the Ki values for heavier-than-air gases.

After some digging I did find the 1977 Husa's paper, being a presentation to the
fire/safety engineering subcommittee of the API, Fall 1977 meeting. It contains
the same formulas as included in the BP standard (see the link in my previous
post), but he used Imperial units instead of metric. He gave the following K
values:
Gas/K
Hydrogen/+5.783
Helium/+5.078
Methane/+2.328
Nitrogen/+1.067 (no wind)
Nitrogen/+1.707 (wind)
Ethane/-1.067
Propane/-2.651
CO2/-2.651
Butane+/-6.586

So the K values for heavier-than-air gases are negative and different from the
ones given in the BP standard. It's hard to say which K values are better,
although it seems like BP has further developed the method. As indicated in the
BP standard: "In view of the uncertainties involved in purging a flare, the
calculated purge rates should be multiplied by factors ranging from 2 for light
gases, to about 5 for gases similar in density to air in high wind conditions."

How do they get the following values from the BP RP 44-3 equation
Ethane / +0.94
Propane / +0.38
Carb.diox / +0.38
Butan+ / +0.15

Shouldn't these be the same as for the HUSA formula except with oppsite signs?

-jha-

Perhaps someone could confirm that using this formula to compare a minumum
flowrate of a fuel gas as purge vs nitrogen purge results in a lower flowrate
requirement for nitrogen than a fuel gas containing 50% Methane 30% Ethane
15% Propane and 5% Heavier than propane.