2010-ST-31-eng

8/13/2019 2010-ST-31-eng

http://slidepdf.com/reader/full/2010-st-31-eng 1/39

XIX International Gas Convent ion AVPG 2010, May 24th - 26th Caracas, Venezuela

Page 1

REDESIGN OF FLARE SYSTEMS AT OFF-SHORE PLATFORMS

AS A WAY OF MINIMIZING PROCESS RISKS

PABLO GRAMAJO

Engineering Office – Flargent S.A.

8/13/2019 2010-ST-31-eng



Page 2

SUMMARY

Offshore platforms are characterized by the compact design of its facilities of

reception and processing of oil and natural gas, in order to use as much of the limited

available space as possible. Also, it is fundamental to bear in mind the personnel and

facilities security.

Thus, any system of final disposition and flaring of residual gases should have

restrictions both of space and maximum allowed values for thermal radiation and

noise emission; the selection of models and correlations for calculation must be very

careful, in order to obtain a safe design that does not imply a great amount of

overdesign that could result in an uneconomical investment; at the same time it

should allow a maximization of production.

This work is focused in the verification of 6 existing platforms, in which due to an

increase in production it was necessary to analyze the maximum caudal of gas that

can be flared.

The analyzed facilities include high pressure flare systems (sonic flares), low

pressure flare systems (subsonic flares), multiple stages of simultaneous flaring with

stacks in parallel, effect of watershields, etc. Also, it has been considered continuous

and emergency gas emissions.

From the obtained results, recommendations have been made in order to, if

necessary, adapt the flare system in order to fulfill the limit values recommended byregulations and accepted industry practice for thermal radiation, noise, pressure drop

and exit gas velocity.

8/13/2019 2010-ST-31-eng



Page 3

INTRODUCTION

Amongst process units, the platforms offshore for gas extraction and treatment stand

out because there must be performed a high economic investment in a limited space

(in comparison with other units of equivalent production); as a result its components,

being these a part of gas extraction installations at submarine bed or one of different

skids for processing the raw gas, must be confined in an area of few square meters

(distributed in one or several levels).

On the other hand the installations must be able to face adverse climatological and

marine conditions, and that must be taken into account for any structure to bemounted in its interior; the isolation respect to other facilities, being it far away from

the coast, implies very strict accident control measures to be followed.

In resume, the principal characteristics of these installations are the following ones:

• High production in limited space

• Need to resist adverse climatological and marine conditions

• Isolation and need to adopt strict accident control measures

8/13/2019 2010-ST-31-eng



Page 4

Figure 1 – P-52 Platform (Petrobras)

As in any process unit, the operation in offshore platforms is associated with the need

to eliminate gaseous vents, being these continuous or from an emergency event.

The elimination of gaseous effluent must take into account the following restrictions:

• Minimization of the risk for the personnel and the facilities

• Minimization of environmental impact

• Use of the available space in the facilities

• Economic considerations: the elimination process should not be onerous from

an economic point of view, simultaneously it must be compatible with a

maximization of the production.

8/13/2019 2010-ST-31-eng



Page 5

From the previous paragraphs it can be seen that the design of a flare system for an

offshore platform should be performed as much precise as possible, for avoiding

great overdesigns and at the same time for guaranteeing the safety of personnel and

facilities.

FLARE SYSTEMS IN OFFSHORE PLATFORMS

Roughly, a flare system consists of the following elements:

• Flare collectors: pipes that receives the gas from the primary discharge

element (safety or relief valve, blowdown valve, etc.) and extends up to the

elements of final disposal.

• Knock-out drum: vessel for the separation from the gaseous flow of any liquid

that could be retained.

• Flare: final element of the system; on its end the flaring gas is burned, in order

to liberate to the environment carbon dioxide and water as final products; its

length and total height must be enough to minimize the effects of thermal

radiation and noise at the process facility level.

According to the gas pressure in the flare collector, flare systems classifies in:

High pressure flare systems: the gas arrives to the flare base with a pressure

between 5 and 10 kg/cm2g, in the top it reaches sonic flow (Ma = 1, being Ma the

Mach number, ratio between gas velocity and sound velocity at the same pressure

and temperature conditions.

8/13/2019 2010-ST-31-eng



Page 6

Low pressure flare systems: the gas arrives to the flare base with a pressure

between 0 and 1 kg/cm2g; the flare operates at subsonic gas velocity (Ma <1, this

value generally varies between 0,2 and 0,8).

An offshore platform typically possesses both flare systems (having separated

collectors, manifolds and flares for high or low pressure discharges), being these able

to operate simultaneously.

The high or low pressure flare can have one or several stages in parallel (generally

up to four stages). As gas flow increases, stages are successively enabled, up to

design flow (maximum flow), in which all stages are operatives.

Figure 2 – Scheme of a multistage flare control system

8/13/2019 2010-ST-31-eng



Page 7

DESIGN AND VERIFICATION OF A FLARE SYSTEM

The conceptual design of a flare system implies determining or establishing, for the

maximum flaring flow, the following:

• Diameter of the flare stack and the tip (in parallel multistage flares, it is

necessary to determine the diameter of each stage), so that flaring is

performed in a suitable hydrodynamic condition.

• Flare length / total height, for guaranteeing maximum values of thermal

radiation and noise in the facilities, below the permissible levels.

• Dimensions (diameter and length) of knockout drum (if there is more than one

flare system, it should be dimensioned a high pressure and a low pressure

knockout drum)

• Flare collector diameter (high and low pressure system), in order to being

compatible with available pressure drops given by the resulting backpressures

at safety or blowdown valves.

• Materials to be used (according to the characteristics of flared gas and the

pressure and temperature conditions)

• Ignition system (direct or front flame system ignition)

• Other characteristics (air or water vapor assistance against smoke formation,

water shields, etc.)

8/13/2019 2010-ST-31-eng



Page 8

The verification of an existing flare system implies to determine if its dimensions are

appropriate for the treatment of the maximum discharge flow (design flow). This work

is focused in the verification of existing flares.

Both in design and verification there must be taken into accout the restrictions related

to the gas discharge process and to the safety of personnel and facilities:

Maximum thermal radiation level: maximum value attainable at process facilities

grade and in any place where there could be continuous or eventual human

presence; this value determines the flare length and its minimum height.

Maximum pressure drop: this value depends on the available backpressure at flare

collector, which determines collector and flare diameter.

Mach number: the API Standard 521 recommends maximum and minimum values

for the Mach number in subsonic flares, in order to mantain a stable flame on the flare

top end.

Maximum noise level: maximum value of the attainable sonorous pressure level in

any place where there could be continuous or eventual human presence.

SCOPE AND DESCRIPTION OF THE PERFORMED WORK

To a greater platform production correspond a greater need for flaring; the scope of

the performed work has been to determine, for existing platforms, the maximum gas

flow that could be flared taking into account restrictions of thermal radiation and noise

level, pressure drop and gas velocity; the more suitable physical models and

calculation software have been selected in order to obtain values with a reasonable

precision level.

8/13/2019 2010-ST-31-eng



Page 9

This study has been made for 6 platforms (FPSO MLS, P-35, P-40, P-52, P-53 and P-

54), with different flare system characteristics (diameter, height, stage quantity, etc.).

The thermal radiation and noise level calculations have been made for 10 receptor

points on each platform; each point was selected by its closeness to the flare, its

personnel occupation level or its importance in the gas extraction and treatment

process.

The following premises have been taken into account:

• The flare system simulation software Flaresim, version 2.1, distributed by the

company Softbits, have been used for calculation procedure

• Events of continuous and emergency discharge have been taken into account

• The possibility of simultaneous flaring of both the high and the low pressure

flare system has been considered

• In calculations of thermal radiation level, the contribution of solar radiation has

been considered

• Both the presence and absence of watershield has been considered

• Calculations have been performed for different meteorological conditions (wind

speeds and directions)

• Air or water vapour assistance against smoke formation has not been

considered

The sequence for the work development is the following:

8/13/2019 2010-ST-31-eng



Page 10

• Selection of a thermal radiation emission model

• Establishment of the most suitable correlation for calculation of F parameter

(fraction of the total heat that is emited from the flare by radiation)

• Selection of the noise emission model

• Establishment of limit values to be considered in the verification

• Selection of the receptor points to be analyzed in each platform

• Establishment of the meteorological conditions to be considered

• Input of the necessary information in the flare simulator

• Performing of the calculation runs for different flaring events and conditions; if

it would be necessary, to realize several determinations using a trial and error

process up to obtain the final thermal radiation or noise level value

• Presentation of results

• Conclusions and recommendations

MODEL FOR THERMAL RADIATION EMISSION

The different available models varies in complexity, according to how conservative

would be the obtained result; each model proposes a determinate hypothesis about

flame behavior and shape, resulting from the gas combustion.

8/13/2019 2010-ST-31-eng



Page 11

The models that were considered are the following ones:

• Models of thermal radiation punctual source

• Model of Integrated Punctual Source (IPS)

• Model of Integrated Diffuse Source (IDS)

• Model of Integrated Mixed Source (IMS)

a) Models of thermal radiation punctual source

The simplest and more commonly used model is that proposed in the API Standard

521 (Pressure-relieving and Depressure Systems), which considers that all heat is

emited from a punctual source (the flame average point).

The calculation is based on the Hajek and Ludwig equation; considering the minimal

distance ( D ) from the flame emission point up to the object where thermal radiation

level ( K , energy for unit of time and area) is calculated:

2..4

..

D

QF K

π

τ = (1)

Where:

Q : Total heat resulting from gas combustion

8/13/2019 2010-ST-31-eng



Page 12

This value depends on the gas flow (W ) and on its low heating value ( LHV ):

W LHV Q .= (2)

F : Fraction of total energy emited from the flare that is transmitted by radiation;

10 << F .

This is an empirical parameter that depends principally on the flared gas composition,

its flow, the used tip design (subsonic or sonic) and the smoke presence.

τ : Fraction of total energy transmitted by radiation not absorbed by environment

(transmisivity); 10 << τ

This parameter takes into account the heat absorption capacity of surrounding air. For

conservative calculations it is considered 1=τ .

In its basic form, this model considers that the emisor point is on the flame base. A

more realistic approach considers that this point is on the flame half; to apply thishypothesis it is necessary to complement the model with a method that makes

possible the determination of the flame length. To accomplish that, the API Standard

521 presents graphs that correlate information of flame length obtained from field

tests.

The Brzustowski and Sommer method (1973) constitutes a variant of the API method

that takes into account the angle between the normal line to the surface on which the

thermal radiation level is estimated, and the straight vision line to the flame center;

this method is mentioned in the API Standard 521 as an alternative to determine the

flame center location.

8/13/2019 2010-ST-31-eng



Page 13

b) Integrated Punctual Source Model (IPS)

In this model the flame is splitted into elements, each one being a punctual source;

the thermal radiation emited by the flame is the summation of each punctual source

contribution.

∑=

=n

i i

i

D

l

L

QF K

12

...4

..

π

τ (3)

Where:

ni K,2,1= : Flame element (element 1, element 2, and so on up to element n )

L : Total flame lenght

il : Lenght of flame element i ; ∑=

=n

i

il L1

i D : Distance from the flame element up to the object where thermal radiation level is

calculated

K : Thermal radiation level (energy for unit of time and area)

Q : Total heat resulting from combustion of flared gas (formula (2))

F : Fraction of total heat emited by the flare that is transmitted by radiation



8/13/2019 2010-ST-31-eng



Page 14

To apply this method, the following hypotheses should be fulfilled:

• The flame emits radiation uniformly in all its extension

• The flame is long in comparison to its width, so it can be considered a linear

source

From these hypotheses can be seen that the flame is transparent to thermal radiation

and that a flame element does not interfere with each other.

c) Model of Integrated Diffuse Source (IDS)

This model assumes that flame is opaque, so thermal radiation emission is entirely

from its surface.

∑=

=n

i i

ii

D

senl

L

QF K

122

)(..

.

.. β

π

τ (4)

Where:

ni K,2,1= : Flame element (element 1, element 2, and so on up to element n )

L : Total flame lenght

il : Lenght of flame element i ; ∑==

n

i

il L1

i β : Angle between the tangent to the flame element i , and the straight vision line

from the flame center for element i up to the object where the thermal radiation

level is calculated

8/13/2019 2010-ST-31-eng



Page 15

i D : Distance from the flame element i up to the object where thermal radiation level

is calculated

K : Thermal radiation level (energy for unit of time and area)

Q : Total heat resulting from combustion of flared gas (formula (2))

F : Fraction of total heat emited by the flare that is transmitted by radiation



d) Integrated Mixed Source Model (IMS)

In this model it is applied a linear combination of results from IPS and IDS models

previously depicted.

IDS IPS IMS K aK aK ).1(. −+= (5)

Where:

IMS K : Thermal radiation level resulting from applying the Integrated Mixed Source

method

IPS K : Thermal radiation level from Integrated Punctual Source Method

IDS K : Thermal radiation level from Integrated Diffuse Source Method

8/13/2019 2010-ST-31-eng



Page 16

a : Constant ( 10 << a ), whose value is obtained from correlating results obtained with

the IPS and IDS methods

e) Comparison between presented methods

The punctual source methods (of which the API and Brzustowski and Sommer

methods have been presented) has demonstrated to provide reasonable results at

relatively big distances from the flare; for this and for its simplicity they are extensively

used in the design of flare systems on onshore facilities.

On the other hand, for distances near to flare base these methods do notappropriately predict radiation levels; therefore, they are not the best option to apply

on offshore platforms, as one of they characteristics is the little availability of space

and the closeness of process facilities to the area for final disposition of residual

gases.

For results obtained with multiple source models, field experience has demonstrated

that these methods provide reasonable values of thermal radiation levels of, with the

following features:

• Integrated Punctual Source (IPS) and Integrated Diffuse Source (IDS) models

predict similar values for thermal radiation levels at great distances from flare

• The Integrated Punctual Source (IPS) model tends to overestimate thermal

radiation levels nearby the flare

• The Integrated Diffuse Source (IDS) model tends to underestimate thermal

radiation levels nearby the flare

8/13/2019 2010-ST-31-eng



Page 17

The target of the Integrated Mixed Source (IMS) model is to combine the two previous

methods, IPS and IDS, in order to obtain a better prediction in areas nearby to the

flare base.

From the previous paragraphs, the Integrated Mixed Source (IMS) model results the

most attractive, amongst the presented models, for developing the study at the 6

offshore platforms.

The calculation software determines the τ parameter according to the distance from

the flame center up to the object where thermal radiation level is calculated, and to

the relative ambient humidity. For the used correlation, the parameter value usuallyvaries between 0,8 and 0,9.

Taking into account the absence of additional information that could allow a more

precise modeling, for simulating how a watershield (waterfall for blocking and

diminishing the amount of radiated heat from the flare) mitigates thermal radiation

levels it has been established a value of 0,3 for the τ parameter.

The n value has been fixed in 20 flame elements, since using major values would

increase calculation time without any significant improvement in the calculated value.

The flame inclination is calculated solving exit gas velocity, wind velocity and flame

buoyancy vectors, while its length is calculated considering the total amount of heat

resulting from the combustion of the flared gas and of the tip type used.

8/13/2019 2010-ST-31-eng



Page 18

CORRELATION FOR CALCULATING THE F PARAMETER

The F value depends on the properties of the flared gas, on the gas flow regime in

the flare and on the tip constructive characteristics. Its estimation is fundamental for

obtaining reasonable values of thermal radiation levels.

a) Gas natural correlation

This correlation was developed specially for a natural gas flow of molecular weight

equal to 19; the F value depends on the exit gas velocity an the flare top end.

b) Kent method

This method, proposed in 1964, relates the F values to the gas low heating value;

the used formula is:

900

.2,0 LHV

F = (13)

Where:

PCI : Gas low heating value in BTU/Sm3 (standard conditions being at 14,7 psia

and 60 °F)

The gas low heating value (for hydrocarbons) is correlated with its molecular weight:

100.50 += MW LHV (14)

For a gas mixture:

8/13/2019 2010-ST-31-eng



Page 19

∑=

=n

i

ii LHV y LHV 1

. (15)

Where:

MW : Gas molecular weight

i LHV : Gas low heating value of i component i in a gas mixture

i y : Gas molar fraction of i component in a gas mixture

,

The F values obtained varies from 0,2 for methane; 0,33 for propane; up to 0,55

for other hydrocarbons. The author provides neither experimental validation for

this method nor its interval of application; nevertheless, other authors report its

satisfactory application for flare systems design.

c) Tan method

This method, proposed in 1967, relates the F parameter to the flared gas

molecular weight:

MW F .048,0= (16)

Where:

MW : Gas molecular weight

8/13/2019 2010-ST-31-eng



Page 20

This coterrelation throws the following F values: 0,2 for methane; 0,33 for

propane and 0,4 for major molecular weight hydrocarbons. The author provides

neither experimental validation for this method nor its interval of application.

d) Recommended values from API Standard 521

The API Standard 521 presents a table with F values for hydrogen, methane,

butane and natural gas (95 % of methane), obtained experimentally for different

burner diameters.

e) Cook method

In 1987 Cook proposed a method based in the assumption that the flame emit

radiation uniformly from its surface. The used equations are the following:

Q

PF = (17)

f A E P .= (18)

H mQ .= (19)

From (17, (18) and (19) is obtained:

H m

A E F

f

.

.= (20)

Where:

8/13/2019 2010-ST-31-eng



Page 21

:P Total energy transmitted by radiation from the flame (W)

:Q Total heat liberated by gas combustion (W)

E : Emissive power (W/m2)

f A : Flame area (m2)

m : Gas mass flow (kg/s)

H : Heat of combustion (J/kg)

From adjustment with experimental data an average value of the emmisive power

( E ) of 239000 W/m2 has been obtained. Using this method the F values obtained

varies from 0,017 to 0,344.

f) General Pipe method

This method is based on the adjustment of results obtained with the Kent, Tan, Cook

and Natural Gas correlations, in a range of exit gas velocity and gas molecular

weight.

g) ‘High Efficiency’ method

This method uses a proprietary interrelation of Flaresim software, which takes into

account the tip type (subsonic, sonic, etc.), the exit gas velocity, its molecular weight

and the component hydrocarbons grade of saturation.

8/13/2019 2010-ST-31-eng



Page 22

h) Comparison between methods to obtain the F value

Amongst the presented methods, the ‘High Efficiency‘method obtains the less

conservating values for the F parameter, while the Natural Gas method proposes the

highest values.

For the purposes of this work, the ‘High Efficiency‘method has been selected, since:

• The tips used in the platforms are of recent installation and a correct design

has been considered for the different cases of flaring

• The flared gases are in most cases paraffinic hydrocarbons with low molecular

weight, therefore its burning do not generate smoke.

MODEL OF NOISE EMISSION

The noise generated in a flare during the gas flowing and burning can be subdivided

in two components:

Combustion Noise: produced by combustion of gases in the flare top end

Jet noise: produced by the gas discharge

Each noise component will have a major influence in the total noise level value

depending on the flare type: in the low pressure flares (subsonics) the noise from gas

combustion prevails; in the high pressure flares (sonics) the jet noise prevails.

The following methods have been considered for noise level calculations:

8/13/2019 2010-ST-31-eng



Page 23

a) API method

This method is the simplest one; it is exposed in the API Standard 521 (Pressure-

relieving and Depressure Systems). It takes into account only the contribution of the

jet noise.

It is based on the following equation, in which the noise level is calculated as the

sound pressure level in decibels at a distance of 30 m from the gas discharge point to

the ambient:

)..5,0log(.102

30 cW L L m+= (6)

Where:

30 L : Sound pressure level at 30 m from the atmospheric discharge point, in decibels

L : Sound pressure level (in decibels), obtained from a graph of the norm API 521,

which correlates this value of L with the quotient between the pressure up-stream

from the end of the flare and the atmospheric pressure

mW : Mass flow of gas (kg/s)

c : Sound velocity in the flared gas (m/s)

For other distances, in addition to the 30 m considered in the equation (6), the sound

pressure level is calculated this way:

)30/log(.2030 r L L p −= (7)

Where:

8/13/2019 2010-ST-31-eng



Page 24

p L : sound pressure level (in decibels) at a distance r

r : Distance from the source of noise (flare top end) (m)

b) Spectrum method

This method takes into account both combustion and jet contribution noise level. The

noise level is expressed as sound pressure level ( SPL ), in the following way:

⎟⎟ ⎠

⎞⎜⎜⎝

⎛ =2

0

2

log.10P

PSPL (8)

Where:

P : Sound pressure

0P : Reference sound pressure (2.10-6 N/m2)

The sound pressure level can be expressed in decibels A (dB (A)), this is a weighted

scale that take into account the difference in sensibility of human audition at different

sound frequencies (in the range of human audible frequencies, the contribution of

average frequencies to the weighted total value is grater than in the case of lower or

higher frequencies). The noise frequency spectrum is divided in several bands of

octaves, from 63 Hz to 8000 Hz.

Combustion noise: it depends principally on the total heat emission in the flare tip

and on the tip design. The calculation is based on typical curves, depending on the tip

type (subsonic, sonic, etc.).

8/13/2019 2010-ST-31-eng



Page 25

For example, the following typical curve can be used:

Figura 1 – Typical curve to determine the combustion contribution to sound pressure

level (Spectrum method)

Figure 1 establishes combustion sound pressure levels at a distance of 20 ft from the

sound source and for a total emited heat of 81 MMBTU/h

At other distances and for other values of emited heat, the sound pressure level is

corrected applying the following equation:

8/13/2019 2010-ST-31-eng



Page 26

ASPL D

QSPLSPL −⎟

⎠

⎞⎜⎝

⎛ +⎟

⎠

⎞⎜⎝

⎛ +=

20log.20

10.1,8log.10

71 (9)

Where:

1SPL : Sound pressure level obtained from Figure 1 (dB(A))

ASPL : Sound pressure level substracted from the total value due to atmospheric

attenuation (dB(A))

Q : Total heat emited by combustion of the flared gas (MMBTU/h)

D : Distance from the flame average point up to the position where the sound

pressure level is calculated (ft)

In the flare simulation software it is necessary to select a standard curve for

determining the combustion component of noise, or to input values of total heat

emited from flare versus sound pressure levels, for each used band of frequency.

Jet noise: it arises from the expansion of the gas flow at the flare end; its value

depends on the expanded gas kinetic energy and on its acoustic efficiency. The

following equation applies:

2

.

..

2u

V PWL ρ

η =

(10)

ASPL DPWLSPL −−−= 49,0log.20 (11)

where:

8/13/2019 2010-ST-31-eng



Page 27

PWL : Sound power level of the noise source

η : Acoustic efficiency

V : Volumetric flow of flared gas

ρ : Density of flared gas, downstream from the flare top end

u : Velocity of flared gas, downstream from the flare top end

D : Distance from the flame average point up to the position where the sound

pressure level is calculated (ft)

ASPL : Sound pressure level substracted from the total value due to atmospheric

attenuation (dB (A))

To determine the acoustic efficiency it is necessary to take into account the expandedgas velocity and its flow regime (subsonic or sonic).

For subsonic flow, the acoustic efficiency is obtained from Figure 2, where it is

correlated with the ratio between gas velocity and sound velocity, both measured

downstream from the flare top end.

8/13/2019 2010-ST-31-eng



Page 28

Figura 2 – Acoustic efficiency for subsonic flow

The adimensional parameter B is obtained from the following equation:

2

. ⎟⎟ ⎠

⎞⎜⎜⎝

⎛ =

∞∞ T

T B

ρ

ρ (12)

Where:

ρ : Flared gas density, downstream from the flare top end

8/13/2019 2010-ST-31-eng



Page 29

∞ ρ : Flared gas density at athmosferical conditions

T : Flared gas temperature, downstream from the flare top end

∞T : Flared gas temperature at atmospherical conditions

For sonic flow, the acoustic efficiency is obtained from Figure 3, where it is correlated

with the ratio between the pressures upstream and downstream from the flare top

end:

Figure 3 – Acoustic efficiency for sonic flow

8/13/2019 2010-ST-31-eng



Page 30

VALUES LIMIT FOR PARAMETERS

The limit values for each parameter involved in the calculation are the following:

a) Limits of thermal radiation

• Emergency flaring (maximum): 4737 W/m2 (1500 BTU/h.ft2)

• Continuous flaring (maximum): 1577 W/m2 (500 BTU/h.ft2)

• Solar radiation level: 790 W/m2 (250 BTU/h.ft

2)

The proposed limits are based on values from the API Standard 521, where the

thermal radiation levels that incide in human skin are related to times estimated for

reaching the pain threshold.

At a radiation level of as much as 1500 BTU/h.ft2 it is considered as factible to attend

to an emergency situation during two to three minutes, with personnel that does not

possess any special protection but that is provided with suitable work clothes and

basic safety elements. At a level of as much as 500 BTU/h.ft 2 it is possible the

permanent presence of personnel without any special protection from thermal

radiation but with suitable work clothes and basic safety elements.

The given solar radiation level is also based on values proposed by the API Standard

521.

b) Limits of Mach number (subsonic flares)

The following values have been adopted:

8/13/2019 2010-ST-31-eng



Page 31

• Emergency flaring (maximum): 0,7

• Continuous flaring (maximum): 0,3

The API Standard 521 recommends, for low pressure systems design, to use a Mach

number of 0,5; nevertheless, it also mentions that major values could be acceptable if

the flare tip has a suitable design.

For continuous flarings, it is mentioned for the Mach number a value of 0,2; however

a value of 0,3 has been adopted as it is satisfactory according to the industrialpractice.

c) Limits of pressure drop in the flare

These values depend on the available backpressure downstream from safety or

blowdown valves, therefore they have been taken from the design specifications of

each platform. In most cases, the available pressure values in the stack base are the

following:

• High pressure flare systems (maximum): 490 kPaabs (5 kg/cm2abs)

• Low pressure flare systems (maximum): 115 kPaabs (1,17 kg/cm2abs)

d) Limits of sound pressure level

• Emergency flaring (maximum): 100 dB(A)

• Continuous flaring (maximum): 90 dB(A)

8/13/2019 2010-ST-31-eng



Page 32

• Environmental sound pressure level: 60 dB(A)

The extreme values have been established from the flare system design

specifications of the platforms.

METEOROLOGICAL CONDITIONS TO BE CONSIDERED

The following meteorological conditions have been taken into account:

a) Wind speed

• No wind presence

• 8,2 m/s

b) Wind direction

• The flare flame approaches the process

• The flare flame moves away from the process

• Another wind direction, according to the considered platform

ANALYZED RECEPTOR POINTS ON EACH PLATFORM

The points of interest for observing thermal radiation and noise levels during a gas

flaring event can be different according to the platform that is considered, but in

general the following ones have been selected:

8/13/2019 2010-ST-31-eng



Page 33

• Flare boom base

• Operation modules: skids or equipment nearer to the flare base

• Operation modules: skids or equipment nearer to the flare top end (top end of

towers, etc.)

• Operation modules: other points of interest (for example, chemicals product

deposits)

• Perforation tower platform

• Top of perforation tower

• Crane operation cabins

• Platform extreme points, on the area where the flare system is located

8/13/2019 2010-ST-31-eng



Page 34

Figura 4 – Observation points for thermal radiation and noise levels (P-53 Platform -

Petrobras)

9

4

7

3

6 102

5

81

N

1

2

E

6

8/13/2019 2010-ST-31-eng



Page 35

CALCULATION PROCEDURE

First of all the necessary system specifications must be loaded in the simulator, for

each defined case (continous or emergency emissions, different meteorological

conditions, etc.):

• Flared gas information: gas composition or its bulk properties (molecular

weight, low heating value, Cp/Cv), gas temperature

• Meteorological data: wind speed and direction, ambient temperature, ambient

relative humidity, atmospheric pressure, solar radiation level, backgroundsound pressure level, ambient transmisivity (τ )

• Information of each flare stack: ubication, length, angle with horizontal line

• Information of each flare tip: type (subsonic, sonic, etc.), burner number, seal

type (fluidic or molecular), calculation method for the factor F , calculation

method for flame length, standard curve for calculation of sound pressure level

due to combustion (for the Spectrum method), length, angle with horizontal

line, exit diameter, stack diameter, flared gas flow

• Receptor points information: coordinates of each point where the thermal

radiation and noise level is to be calculated

• Calculation options: selection of the calculation method for thermal radiation

levels, selection of the position of the average point in each flame element and

number of these elements (for multiple source models), options for including

major detail in calculations (inclusion of solar radiation, inclusion of cooling due

to wind, etc.), selection of the calculation method for the sound pressure level.

8/13/2019 2010-ST-31-eng



Page 36

As soon as all the necessary information is loaded the calculation runs can proceed.

Since the flares can be of parallel multiple stages, the entire gas flow must be

distributed between the stages; it should be taken into account that, after the calculus

of the pressure drop on each stage, upstream the distribution manifold it is necessary

to obtain the same pressure value.

P1 = P2 = P3 = P4 = P

Therefore, different flow combinations must be proved in the stages up to obtain, with

a trial and error process, the same pressure at every stage beginning.

PRESENTATION OF RESULTS

The results can appear in a table or graphically in isopleths of thermal radiation andsound pressure level, presented on a platform layout.

Stage 1

Stage 2

Stage 3

Stage 4

ΔP1

ΔP2

ΔP3

ΔP4

P1

P2

P3

P4

P

Flared gas

8/13/2019 2010-ST-31-eng



Page 37

Figure 5 – Example of thermal radiation level isopleths – Plant view (P-53 Platform -

Petrobras)

Case 6 - Emergency

Watershield Yes

Wind speed 8.2 m/s

Wind direction Approaching

flame

9

3

2

8/13/2019 2010-ST-31-eng



Page 38

Figura 6 – Example of thermal radiation level isopleths – Elevation view (P-53 Platform -

Petrobras)

Case 6 - Emergency

Watershield Yes

Wind speed 8.2 m/s

Wind direction Approaching

flame

1

2,

8/13/2019 2010-ST-31-eng



CONCLUSIONS AND RECOMMENDATIONS

For each flaring event it was analyzed which is the limitant parameter for maximum

flow: thermal radiation or sound pressure level in the observed platform points, Mach

number (for subsonic flares) or flare pressure drop.

If the analyzed platform operates at present with a maximum flaring flow greater than

the maximum permissible obtained as a calculation result, it is recommended to

implement any of the following actions, according to economical or operative

possibilities:

• To diminish the maximum flaring flow, modifying the process in order to achieve

this goal without compromising the production level

• To adopt a system of watershield in order to diminish radiation levels

• To replace flare tip by another that is suitable for the present levels of operation

BIBLIOGRAPHY

• ANSI/API Standard 521 – Pressure-relieving and Depressuring Systems - Fifth

Edition, January 2007

• Flaresim User Manual

• Heat radiation from flares – Guigard, Kindzierski & Harper – May 2000

• Flare Radiation Estimated - McMurray, R. - Hydrocarbon Processing, Nov.

1982 pp 175 181

2010-ST-31-eng

Documents