O U Mohamed Ismail (R670215011) .Protected
O U Mohamed Ismail (R670215011) .Protected
A PROJECT REPORT
Submitted by
MOHAMED ISMAIL
(R670215011)
In the partial fulfillment of the requirement for the award of the degree of
MASTER OF TECHNOLOGY
in
CHEMICAL ENGINEERING
with specialization in
PROCESS DESIGN ENGINEERING
Under the guidance of
Dr. Murali Pujari
Mr. Vidya Bhusan
Deputy General Manager Mr. Bomma ramanjaneyulu
Assistant Professor (SG)
Engineers India Limited
Department of Chemical Engineering
I hereby declare that this submission is of my own, and to the best of my knowledge
and belief that this document does not contain any content previously published by
other person, or material which has been accepted for the award of any other Degree or
Diploma of the University or other Institute of Higher Learning, except when the
acknowledgement is made in the text.
Mohamed Ismail
R670215011
CERTIFICATE
This is to certify that the thesis titled DESIGNING OF FLARE SYSTEM submit-
ted by MOHAMED ISMAIL (R670215011), to the University of Petroleum & Energy
Studies, for the award of the degree of MASTER OF TECHNOLOGY in Chemical Engi-
neering With Specialization in Process Design Engineering is a bonafide record of project
work carried out by him/her/them under my/our supervision.
Date:
i
ACKNOWLEDGEMENTS
First and foremost, praise and gratitude go to The Almighty, who has blessed me and
gave the opportunity, strength, courage and patience.
I thank Mr. Vidya Bhusan, Mr. Varun Vijay (Engineers India Limited, Gurgoan), Dr.
Murali Pujari and Mr. Bomma ramanjaneyulu (Assistant Professor, UPES, Dehradun,
India) for supervising my M. Tech. thesis. I’m thankful to Dr. Santosh Kumar Gupta,
Dr. Parichay Kumar Das (Distinguished Professor, Chemical Engineering Department,
UPES, India), Dr. P. Vijay (Assistant Professor, UPES, India) and Mr. Umar Afzal(Engineers
India Limited, Gurgoan) for clarifying doubts.
Secondly, I owe a lot to the management of Engineers India Limited for their kind
approval to undergo internship in their organization. I would like to mention the magna-
nimity of various people who helped me and the immense amount of gratitude I owe to
each one of them for their valuable guidance.
Last but not the least, my deep gratitude and appreciation go to my parents and sister,
for their sincere prayers, support and encouragement.
Sincerely
Mohamed Ismail
ii
NOMENCLATURE
Abbreviations
• CV - Control Valve
• HC - Hydrocarbon
• O2 - Oxygen
• N2 - Nitrogen
• NA - Not Available
iii
ABSTRACT
Process safety is a disciplined framework for managing the integrity of operating sys-
tems and processes handling hazardous substances. It is achieved by applying good design
principles, engineering and operating & maintenance practices. Flare systems plays an im-
portant role in the safety of Oil and Gas installations by serving as outlets for emergency
pressure relief in case of process upsets. Accurate design of the flare system plays a key
role in containing possible process safety hazards on the oil and gas installation, petro-
chemical industries, especially oil and gas offshore platforms. This Project is focused on
designing the a flare network for a particular plant and optimizing the design procedure
of the flare system. This project will give the optimized procedure involved in detailed
design engineering for selection and sizing of Pressure relieve devices (PRD) and design-
ing of Flare headers, Knock out drum, water seal drum, Flare stack, Flare seal, Flare tip,
Flame Front Generator and documentation of Flare system. The Various relief scenario is
considered like fire case, CV failure case, reflux failure case, electricity failure case, etc.
The simulation tool used in this project is Aspen Techs Flare system analyzer (known as
FLARENET),which is a steady state simulation tool.
Keywords: Process safety, Flare system, Pressure relief device, Knockout Drum, MAWP,
Water seal drum, Flare seal, Aspen Tech, Steady state, FLARENET, Hydrocarbon
iv
TABLE OF CONTENTS
CERTIFICATE i
ACKNOWLEDGEMENTS ii
ABSTRACT iv
1 INTRODUCTION 1
2 LITERATURE SURVEY 6
v
3.1.2 Load calculation for control valve failure case . . . . . . . . . . 12
3.3.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
vi
5 LIQUID SEAL DRUM 35
6 FLARE STACK 38
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
8 DISPOSAL TO FLARE 51
9 CONCLUSIONS 54
9.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
REFERENCES 64
vii
LIST OF TABLES
viii
LIST OF FIGURES
ix
6.5 Sizing a Flare Stack . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
x
CHAPTER 1
INTRODUCTION
Across the global oil & gas industry, considerable effort has been focused on the pre-
vention of major incidents. For the oil & gas industry the emphasis of process safety is
to prevent unplanned releases which could result in a major incident. Process safety is a
disciplined framework for managing the integrity of operating systems and process han-
dling hazardous substances(C.L. Beyler, 2002). It is achieved by applying good design
principles, engineering and operating & maintenance practices. It deals with prevention
and control of events that have potential to release hazardous materials and energy. Such
incidents can result in toxic exposures, fires or explosions and could ultimately result in
serious incidents including fatalities, injuries, property damage, lost production or envi-
ronmental damage.
As a major safety requirement at oil and gas installations such as refineries and process
facilities, a flare system is usually installed to relieve built up pressure that may occur
during shut down, start up or due to process system failure, reducing other safety hazards
associated with process emergencies. Accurate design of the flare system plays a key role
in containing possible process safety hazards on the oil and gas installation, especially oil
and gas offshore platforms(CCPS, 1993).
The flare system is the single largest pipe network in an oil & gas processing plant. It
serves as a relief system for depressurizing different process and production units in cases
of shut down or unexpected cases of hazardous process emergencies, by collecting excess
fluid through relief devices and a pipe network and disposing of it to the required outlet.
The light hydrocarbons and other gases are released by combustion into the atmosphere
while the heavier hydrocarbon, liquids are let out through drains and are often pumped
back into the separation system. Flare knock out drums are installed in the units to remove
liquids from the relieved vapors. These liquids are routed to the units closed blow down
system. Vapors from the units flare knock out drum leave the units battery limit and join
the OSBL flare header. OSBL flare headers from various units in a complex are combined
and routed to centralized flare systems. Depending upon the quality and quantity of flare
loads and the physical dimensions of the complex, one or more centralized flare systems
may be installed.The centralized flare system typically consists of an OSBL Flare Knock
Out Drum, a Water Seal Drum, a Flare Stack, a Flare Seal, a Flare Tip and a Flame Front
Generator(CCPS, 1993).
There are various causes for overpressure in a process unit. Some of the major causes
are reproduced here below(CCPS, 1993 ; API, 2014):
1. Blocked Outlet:
Inadvertent closure of a valve at the outlet of equipments, instruments or piping
may subject them to overpressure. Example: Blocking the outlet of a reciprocating
compressor or pump subjects the intervening equipments, instruments and piping to
overpressure.
2
2. Inadvertent Valve Opening:
Inadvertent opening of a valve at the inlet of equipments, instruments or piping may
subject them to overpressure. Example: Opening of valves designed for the reduc-
tion in steam pressure may subject the downstream system to overpressure.
4. Utility failure
Loss of utility to equipments or instruments can lead to overpressure. Some of the
possible utility failures and the equipments that get affected due to the same are as
follows:
3
(c) Instrument Air:
Control valves for various process control functions Possibility of sudden in-
strument air loss to units is mitigated by the installation of a instrument air
surge vessel of sufficient size to allow either rectification of instrument air loss
(e.g. restarting of spare instrument air compressor) or safe shut-down of com-
plex.
(d) Steam:
Turbine for pumps, compressors, blowers, combustion air fans or electric gen-
erators; Reboilers; Equipments using direct injection of steam (e.g. Strippers);
ejectors and eductors
(e) Fuel:
Boilers, Reboilers, Engine drivers for pumps or electric generators, Gas tur-
bines.
5. Loss of Fans:
Loss of fans due to electrical or mechanical failure can lead to loss of cooling in
air cooled heat exchangers thus causing overpressure. Loss of fans of cooling tow-
ers can cause high cooling water temperature which could lead to loss of cooling in
units, causing overpressure.
4
pressurization of this column can occur. Example: Vacuum column downstream of
a Atmospheric column in a Crude Distillation Unit.
7. Reflux Failure:
Loss of reflux in a column due to reflux pump or reflux control valve failure can
cause overpressure.
5
CHAPTER 2
LITERATURE SURVEY
Process safety is a disciplined framework for managing the integrity of operating sys-
tems and processes that handle hazardous substances(API 520, 2015). It relies on good de-
sign principles, engineering, operating and maintenance practices. In recent years, major
incidents in both the upstream and downstream industries have highlighted the importance
of having these robust processes and systems in place.
Flare and disposal system plays an important role to prevent major incidents and it is
part of process safety design of a plant. As seen in figure the Swiss cheese model, hazards
are prevented/contained by multiple protective barriers. Barriers may have weaknesses or
holes. When holes align hazard energy is released, resulting in the potential for harm. Bar-
riers may be physical engineered containment or behavioral controls dependent on people.
Holes can be latent/incipient, or actively opened by people. Flare and disposal system is
one of the major prevention barriers for the safety and integrity of the operating assets(API
520, 2015).
• method of enhancing mixing at the flare tip (i.e., steam-assisted, airassisted, pressure-
assisted, or non-assisted).
Figure 2.1: Swiss Cheese Model
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. Further, the products
of combustion can be dispersed above working areas to reduce the effects of noise, heat,
smoke, and objectionable odors.
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. This mixture on ignition establishes
a stable flame zone around the gas core above the burner tip. Cracking can occur with the
formation of small hot particles of carbon that give the flame its characteristic luminosity.
If there is an oxygen deficiency and if the carbon particles are cooled to below their ignition
temperature, smoking occurs. 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(API 520, 2015).
Steam-assisted flares are single burner tips, elevated above ground level for safety rea-
sons, that burn the vented gas in essentially a diffusion flame. To ensure an adequate air
supply and good mixing, this type of flare system injects steam into the combustion zone
to promote turbulence for mixing and to induce air into the flame.
7
2.2.2 Air Assisted flares
Some flares use forced air to provide the combustion air and the mixing required for
smokeless operation. These flares are built with a spider shaped burner with many small
gas orifices. The principal advantage of the air-assisted flares is that they can be used where
steam is not available. This type of flare is not generally economical when the gas volume
is large. The air assist flare is not usally used on large flares.
The enclosed flare’s burner heads are inside a shell that is internally insulated. This
shell reduces noise, luminosity, and heat radiation and provides wind protection. A high
nozzle pressure drop is usually adequate to provide the mixing necessary for smokeless
operation and air or steam assist is not required. The height must be adequate for creating
enough draft to supply sufficient air for smokeless combustion and for dispersion of the
thermal plume. These flares are always at ground level. Enclosed flares generally have
less capacity than open elevated flares and are used to combust continuous streams. Stable
combustion can be obtained with lower Btu content stream gases (50 to 60 Btu/scf).
Flare, vent and blow down system are very critical systems in oil & gas plant. Initial
system design for a typical topside facility is for maximum relief from the largest source
for a particular relief scenario decided during design phase of the plant. As the time goes,
subsequent modification projects, subsea tie-in to the existing topside facility makes flare
system vulnerable. Some times each and individual project estimates the additional relief
loads they will put into the existing flare system and compare with the available capacity
in the flare system. In most of the cases the new sources are added to the existing flare
system without any modification or upgrade of current system. Again, building a new
flare system (which includes tail pipes, main header, KO drum and flare stack) requires
8
Figure 2.2: Flare stuctures
heavy investment and typically in an offshore installation where there is restriction on total
allowed weight on top side equipment makes it no feasible.
Many cases it is gaseous/vapor phase fluid goes through the flare system. In certain cases
the flow could be two phase with both liquid and gas phase present. Then it necessitates a
detail study of the flow regimes, velocities of different phases, reaction forces and change
in fluid property along the flare header(API 520, 2015).
9
CHAPTER 3
For each of the overpressure scenarios, individual relieving rates need to be determined
in order to size the safety valves that need to be installed to protect against overpressure.
Generally liquid and vapour relief loads are determined by estimating net energy inputs.
Energy inputs can be as heat input leading to an increase in pressure or it can be a direct
pressure input from a higher pressure source. In determining the relief loads, the proba-
bility of two unrelated failures occurring simultaneously and normally does not need to be
considered. The effect of pressure and temperature need to be considered to calculate the
relief load. This means that relieving loads need to be considered at the set pressure of the
safety device and not the normal operating pressure.
The location of the PRD devices is based on the nature of the vessel, operating conditions
of the vessel,etc. The failure case scenario is to be analysed for that particular vessel like
fire case, CV failure case, etc. The flare load for each PRD devices is calculated based on
the components inside the vessel, failure case scenario, geometry of the vessel, height and
thickness of the vessel, design pressure and set pressure of the vessel. After determining
the individual relieving rates of a unit for all possible relief scenarios, the same are tabu-
lated in the form of a Unit Flare Load Summary. A tabulation of flare loads from all units
connected to a flare system is then prepared. This is called an Overall Flare Load Summary.
The highest load out of all the loads in a flare load summary is called the Controlling Load
or Governing case and determines the size of the relevant flare system facilities. (API521,
2014 ; API 520, 2015)
3.1.1 Load calculation for fire case
If the entire plant or a small portion catches fire, the components in the vessels in that
fire area absorbs heat and it will pressurize the vessel. The fluid in that vessel should be
relieved to depressurize and protect the vessel. The relieving load from each vessel is
calculated from wetted area of the vessel. The wetted area depends on the geometry and
orientation of the vessel. If the vessel located above 7.6 m from the base, there is no need
to fix the PRV to that vessel for fire case. The wetted area to be calculated upto 7.6 m from
the base.
h
πDi2 h
1 4 Di − 0.5 + B
= − 0.5 B + 1 + ln √
8 Di 4 2− 3
Where B
s 2
h
= 1 + 12 − 0.5
Di
Di = Internal Diameter
h = Height
!
(h − r)
= Di π − arccos L
r
Where L = Length
" #
πDi2
1 2 + 2
= 2 + ln √
8 4 2− 3
11
• Wetted Area of vertical Cylinder
= πDi h
Q = C1 F A0.82
ws
v
k + 1
u
u
2 k−1
u
t
= Cd A kρ0 P0
k+1
Where Cd = Discharge coefficient
A = Cross sectional Area
cp
k= cv
If the control valve is failed to close, the fluid flow from the valve is maximum and it
will pressurize the vessel next to it. The fluid from the control valve is to be relieved to
depressurize and protect the vessel. The maximum flow of the control valve is calculated.
12
The choked flow is the relieving load for control valve failure case.
Mass flow rate due to choked flow
v
k + 1
u
u
2 k−1
u
t
= Cd A kρ0 P0
k+1
If any valve is failed to open or blocked, the vessel before to the valve will be pressur-
ized. The maximum flow from the valve or the pipe is the relieving load for blocked outlet
case.
If the condenser, heat exchanger or cooling fans in a column is failed to operate due to
electricity failure or pump failure, the column will be pressurized. The reflux stream will
be failed to cool the stream from the tray next to it. The vapour load from that tray is the
relieving load for that column. Probably this load will be the governing case among the
other failure case.
The pressure loss between the vessel and PRD devices is calculated to design the di-
ameter and thickness of the pipe connecting vessel. The Governing case load for vessel
is used to determine the pressure loss. The diameter of the connecting pipe is designed in
such a way that pressure loss in the line is less than 3% of the set pressure. The Darcy
equation is used to find the pressure losses in the line.
Darcy Equation,
13
∆P ρ v2
= fD
L 2D
The Pressure relief valve is used to control the pressure in a particular system which
can build up for process failure like fire, instrument or equipment failure. The relief valve
is designed to open at a set pressure to protect the system. The pressure is relieved by
allowing the pressurised fluid to flow from the system (API 520, 2015).
3.3.1 General
The process in the particular system is operating under a certain pressure called oper-
ating pressure. The design temperature of that system is determined by the design rules
of the pressure design code. The design temperature is based on the minimum permis-
sible thickness and characteristic of each component. The Maximum allowable working
pressure(MAWP) is the maximum pressure that the weekest component of the system can
handle at a designed temperature. This is based on the design codes, fabrication of the
vessel, piping, etc. The MAWP doesn’t remain constant throughout the life of the system.
It will reduce due to corrosion, wear and fatique. The design Pressure is the maximum
pressure that the system that can be exposed to. Design pressure should be less than or
equal to MAWP. The set pressure is the inlet pressure at which that PRV set to open to
protect the vessel. The set pressure should be equal to or less than MAWP. Accumulation
is pressure above the MAWP of the vessel. Accumulation is expressed as a percentage of
MAWP. Overpressure is pressure above the set pressure of the PRV. The overpressure is
expressed as a percentage of set pressure. The relieving pressure is equal to the set pressure
14
plus the overpressure. The temperature of the flowing fluid at relieving conditions can be
higher or lower than the operating temperature.
The three basic types of pressure relief valves are conventional, balanced and pilot
operated.
The conventional valve is used in where the backpressure is less than 10% of the
setpressure. The conventional PRV is a self actuated spring loaded PRV that is designed
to open at a predetermined set pressure and protect the system from excess pressure by
relieving the fluid from that system. The basic elements are an inlet nozzle connected to
the system to be protected, a movable disk that controls flow through the nozzle, and a
15
spring that controls the position of the disc(API 520, 2015). Under normal conditions the
disc is seated on the nozzle preventing flow through the nozzle. When the inlet pressure is
below the set pressure, the disc remains seated on the nozzle in the closed position. When
the inlet pressure exceeds set pressure, the pressure force on the disc overcomes the spring
force and the valve opens. When the inlet pressure is reduced to the closing pressure, the
valve recloses.
A balanced PRV is a spring-loaded PRV that incorporates a bellows on the valve disc to
minimize the effects of backpressure on the valve.For balanced PRV the allowable back-
pressure is 10 - 50% of the set pressure. When the backpressure is constant, the spring
load in conventional valve can be reduced to compensate for the effect of backpressure on
set pressure, and a balanced valve is not required. But the backpressure(superimposed +
built up) is not constant always. The superimposed backpressure may be variable(API 520,
2015; Peter Smith et al., 2003). In a balanced PRV, bellows is attached to the disk holder
with an effective bellows area is equal to or greater than seating area of the disc. This
isolates the seating area from the variable superimposed back pressure. Thus the variable
back pressure would not affect the PRV opening pressure.
In pilot operated setup, main relief is combined with and controlled by a smaller self
actuated pilot valve.This relief valve valve uses the process fluid (circulated through a pilot
valve) to apply the closing force on the valve disc. The pilot valve is a small safety valve
with a spring. The main valve does not have a spring but is controlled by the process fluid
from pilot valve. The piston on the valve disc is designed to have a larger area on the top
than on the bottom. Up to the set pressure, the top and bottom areas are exposed to the
same inlet operating pressure(Adam Bader, et al., 2011). Because of the larger area on
the top of the piston, the net force holds the piston tightly against the main valve. As the
16
Figure 3.2: Conventional PRV
operating pressure increases, the net seating force increases and tends to make the valve
tighter. At the set pressure the resulting net force is now upward causing the piston to lift
and process flow is established through the main valve.This arrangement allows operation
of pilot operated valves with a very narrow margin between set pressure of the relief valve
and operating pressure of the protected equipment. The lift of the main valve piston is not
17
Figure 3.3: Effects of backpressure on convetional PRV
affected by built up backpressure. Allowable back pressure is typically more than 50% of
the set pressure.
The selection of PSV among Conventional, Balanced and pilot operated PRV is de-
pends on the pressure drop in inlet line of PRV, backpressure, set pressure, number of
relief valves, failure scenario, nature of fluid, inlet and discharge piping, tail pipe sizing,
space available, cost feasibility and backpressure in flare header while process failure.
The conventional valves requires the backpressure is less than 10% of the set pressure.
The balanced valves allows the use of smaller tail pipes and relief headers because of the
larger pressure drops allowed, as a result of higher allowable backpressure (10-50%). The
balanced valve is more expensive than conventional valves; however, the total cost of the
use of balanced valves plus the smaller header system may be lower(Adam Bader, et al.,
2011). The bellows should be checked periodically for leakage because A leaking bellow
does not provide backpressure compensation. If the super imposed backpressure is con-
stant, the spring load in conventional valve can be reduced to compensate for the effect
of backpressure on set pressure, and a balanced valve is not required. The superimposed
backpressure may be constant or variable. If it is variable, the balanced valve is preferable.
When the pressure drop in inlet line of PRV exceeds 3% of set pressure, the pilot operated
valve should be used.
The conventional valve is used in non corrosive services and the balanced valve is used
18
Figure 3.4: Balanced PRV
in corrosive services because the bellows isolates the disk from corrosive fluid. The pilot
operated valves should be considered for clean services within their temperature limita-
tions. The pilot operated valves cannot be used in multiphase fluids. For multiphase fluids,
balanced valve is preferrable.
19
Figure 3.5: Pilot operated PRV
20
3.3.4 Orifice area calculation
The Orifice area needs to be calculated in order to have proper amount flow of the
process fluid. The disc valve in the PRV held in the closed position by the spring. When
the system pressure reaches the desired opening pressure. The pressure force of the pro-
cess fluid pass through the inlet and then it is acting over Area A1 equals the force of the
spring, and the disc will lift and allow fluid to flow out through the outlet(Adam Bader, et
al., 2011). When pressure in the system returns to a safe level, the valve will return to the
closed position. The orifice area has been calculated from the set of formulae given in the
Appendix A. Some certain area has been standardized in API 526 (Flange Steel Pressure
Relief Valves) and designated into certain alphabetic as shown on Table. The actual orifice
area should be equal to or greater than calculated orifice area.
r
W TZ
A=
CKd P1 Kb Kc M
or
√
2.676V T ZM
A=
CKd P1 Kb Kc
or
√
14.41V T ZGv
A=
CKd P1 Kb Kc
Where
A = Required discharge area in mm2
W = Flow in kg/hr
v
(k + 1)
u
u
u 2 (k − 1)
C = 0.03948 k
t
k+1
21
P1 = Upstreamm relieving pressure in kPa
Kb = Capacity correction factor due to backpressure
Kc = Combination correction factor
T = Relieving temperature in K
Z = Compressibility factor
M = Molecular Weight
V = Flow in Nm3 /min
Gv = Specific gravity
The orifice area has been calculated based on the actual relieving load. If the actual
orifice area is greater than calculated orifice area, the rated relieving load has to be calcu-
lated(Adam Bader, et al., 2011). The formulae to calculate the rated relieving rate is given
in Appendix A. The tail pipe and flare header to be sized based on the governing case of
rated flow.
Mass flow rate due to choked flow
22
v
k + 1
u
u
2 k−1
u
t
= Cd A kρ0 P0
k+1
23
3.4 Tail pipe sizing
Tail pipe is the outlet line of the PRV. Maximum load for each PRV is the governing
case of that PRV. The outlet line is sized accorinding to that governing case load. The
allowable mach number in the tail pipe is usually fixed to size the tail pipe. It depends
upon the MOC and support of that pipe. Usually it is 0.7 in tail pipe. The pressure drop
in the tail pipe also calculated(Peter Smith et al., 2003). The tail pipe should join the flare
header at the angle 45 degrees to minimise the pressure drop. There are various method to
calculate the pressure drop.
Under segregation, usually two flare headers are provided to collect the flare releases
from the various safety valves. These are generally called LP and HP flare headers. All
safety valves with a low value for the maximum allowable back pressure are connected
to the LP flare header, whereas safety valves which can take a higher back pressure are
connected to the HP flare header. As flare headers are long uninsulated pipes carrying
condensable vapors. The condensate needs to be separated from vapors before reaching
the flare. For this purpose, flare headers are built with no pockets and continuously sloping
(typical slope- 1:500) towards the knock out drum. Typically in a straight segment, expan-
sion loops are installed at a center to center distance of 80-100 m for the impact of thermal
stress(Scandpower, 2004; Dr.A. Alizadeh et al., 2007). The allowable mach number in the
flare header is based on the flare support. It is usually 0.5 in header. The flare header is
sized according to governing case scenario. The governing case scenario is maximum load
of each governing case. For example if 5526 kg/hr load for fire case and 15625 kg/hr load
for reflux failure case, the governing case scenario is reflux failure case. The flare header
is need not to be sized for double jeoperdy or two failure scenario at a time. The pressure
drop in the flare header is calculated by various methods in the literature.
24
3.6 Pressure drop calculation
The pressure drop is need to be calculated from tail pipe to flare tip. The pressure in the
entire flare network should be greater than atmosphere(API 521, 2014). The flare fluid will
move towards the flare tip because of the positive pressure in the network. The pressure
drop is calculated from one of the following methods.
Where
G = Mass flow
a = Cross sectional area of the pipe
P1 = Upstream Pressure
P2 = Downstream Pressure
R = Universal gas constant
f = Moody friction factor
φ = Internal diameter
L = Equivalent length
T = Temperature
M = Molecular weight
Where
25
γ = Ratio of specific heats
The allowable backpressure in outlet of PRV is based on the type of PRV. The actual
backpressure should not exceed the allowable backpressure. The actual backpressure in
the tail pipe and header for particular failure scenario is calculated by
fl 1 P1
= 2
[1 − ( PP12 )2 ] − ln( )2
d M a1 P2
and
fl 1 P1 P1
= 2
[ ][1 − ( PP12 )2 ] − ln( )2
d M a 2 P2 P2
Where
f = Moody friction factor
l = Equivalent length
d = Pipe inside diameter
M a1 = Mach number at pipe inlet
M a2 = Mach number at pipe outlet
P1 = Upstream pressure in kPa
P2 = Downstream pressure in kPa
0.5
−5 G ZT
M a2 = 3.23 × 10
P2 d2 M
26
From the combination of Mach number and ratio of upstream and downstream pres-
sure, the diameter of the pipe is calculated after some iterations.
27
CHAPTER 4
Flare systems generally require a flare knockout drum to separate liquid from gas in
the flare system and to hold the maximum amount of liquid that can be relieved during an
emergency situation.
The cost of drum design can influence the choice between a horizontal and a vertical
drum. If a large liquid storage capacity is desired and the vapor flow is high, a horizontal
drum is often more economical. Also, the pressure drop across horizontal drums is gener-
ally the lowest of all the designs. Vertical knockout drums are typically used if the liquid
load is low or limited plot space is available. They are well suited for incorporating into
the base of the flare stack.
The various types are
• Horizontal drum with the vapor entering at the top end the vessel and exiting at the
another end of the vessel with no internal baffling.
• Vertical drum with the vapor inlet nozzle entering the vessel radially and exiting at
top of the vessel in vertical axis.
• Horizontal drum with the vapor entering at each end exiting at center of the drum.
• Horizontal drum with the vapor entering at the center and exiting at each end.
The function of the knockout drum is to provide residence time for liquid discharges
and to limit the size of droplets directed to the flare burner. Large liquid droplets and liquid
loading can cause smoke.The phenomenon is called as burning rain, occurs when a liquid
hydrocarbon droplet does not burn completely within the flare flame envelope and the rate
of burning is lower than the rate of settling of the liquid droplet. Liquid droplets 300µm
and larger may drop out of the gas stream at less than 2 m/s. If liquids are not drained
from the system, flare flows with gas velocities exceeding about 3 m/s or 4 m/s can entrain
liquid droplets up to 1000µm in size. Liquid droplets exceeding 1000 µm can readily lead
29
to burning rain regardless of flare type. Burning rain can occur at smaller droplet sizes for
some flare types(Daniel A.Crowl et al., 2013).
Sizing a knockout drum is generally a trial-and-error process. The first step is to de-
termine the drum size. The KOD sizing procedure for simple horizontal configuration is
described below
1. The governing load has been calculated for the failure case scenarios.
2. Calculated the fluid equilibrium. Usually 10% of the total load is liquid.
4. The pressure of the KOD is calculated by back calculation from flare tip. The pres-
sure at the flare tip is considered as slightly greater than atmoshpheric pressure. Then
the pressure drop is calculated for flare stack, liquid seal drum, molecular seal drum,
fittings and joints and other components. Then the pressure at the oulet and inlet of
KOD is calculated.
5. The viscosity, Density, Specific gravity and Molecular weight of the fluid is calcu-
lated.
30
0.13 × 108 ρv (pd)3 (ρl − ρv )
C(Re)2 =
µ2
In USC units,
10. The Drag Co-efficient, C is determined from the drag co-efficient graph.
11. The residence time for the fluid from low liquid level to High liquid level in KOD is
30 minutes. The liquid holdup for 30 minutes is determined.
12. The Cross sectional area for liquid and vapour flow is determined by assuming the
length and diameter of the drum.
13. Check L/D ratio. The L/D ratio should be less than 6 and greater than 5.
14. The vertical height of the liquid level in horizontal drum is calculated based on the
geometry of the vessel.
15. The dropout velocity of the stream is calculated using this equation,
r
gD(ρ1 − ρv )
uc = 1.15
ρv C
31
hv 1
θ = ( 100 )( 100 )
In USC units,
θ = ( h12v )( 100
1
)
uv = ( 7.34
N
)( A1v )
In USC units,
uv = ( 260
N
)( A1v )
Lmin = uv θN
The assumed length should be greater than or equal to the minimum required length.
The snapshot of this calculation can be seen in the figure. The model calculation for KOD
Sizing is in Appendix. In vertical vessel, the vapour velocity is equal to dropout velocity.
The required cross sectional area is determined by dividing volumetric flow rate by dropout
velocity.
32
Figure 4.2: Determination of Drag Co-efficient
33
Figure 4.3: Snapshot of Horizontal KOD Calculation
34
CHAPTER 5
The purpose of the liquid seal drum in flare system is to prevent flashback originating
from the flare tip from propagating back through the flare system and to maintain positive
pressure to ensure no air leakage into the flare system. The iquid seal also used to prevent
air ingress into flare system during sudden temperature changes or condensation of flare
gas.
To prevent air entry, it is necessary that the seal dip-leg height and the density and amount
of seal liquid within the drum be sufficient to prevent the seal from being broken as a result
of the vacuum formed in the flare header.The physical dip-leg height is measured from the
top opening of the seal head or end piece to the bottom of the horizontal section of the flare
header piping immediately upstream of the inlet leg(Daniel A.Crowl et al., 2013).The seal
drum should be designed to provide the volume of liquid to fill the vertical seal leg up to
the specified vacuum. Experience has shown that a minimum dip-leg height of 3 m (10 ft)
above the liquid level is effective in minimizing the ingress of air into the flare header from
flare stacks for typical refining applications.
Liquid seals typically use water as the seal medium, however, other fluids are possi-
ble. Fluid selection requires consideration of freeze protection in cold climates, hydrocar-
bon/water separation, implications of carryover, compatibility with the relief stream, cost,
availability, and disposal(Daniel A.Crowl et al., 2013). In facilities that have cryogenic
products released into the flare header, consideration should be given to the effect of the
cold material on the seal medium. Water seals are not recommended if there is a risk of ob-
structing the flare system due to an ice plug. Alternate sealing fluids such as a glycol/water
mixture or other means to prevent freezing can be required.
Consideration should be given to providing a continuous flow of seal fluid (typical for
water seals), which allows for the continuous skimming of hydrocarbons as well as main-
taining liquid level. Proper liquid seal drum operation is dependent upon maintaining the
design liquid level in the seal. Routine surveillance and hydrocarbon skimming, if ap-
plicable, are required to ensure proper seal operation. Seal drums that overflow to open
sewer should be evaluated as to whether condensed flammable and toxic vapors can be
discharged and the need to provide suitable containment and mitigation systems.
Sizing depends on whether the drum is horizontal or vertical. The minimum diameter
of a vertical seal drum is determined by the total seal liquid volume.
r
H
D=d ( + 1)
h
For a horizontal drum this criteria typically does not control the drum diameter because
the length of the drum can easily be adjusted to get sufficient seal liquid volume. The
volume of seal fluid in a horizontal seal drum shall be adequate to fill the dip-leg(Daniel
A.Crowl et al., 2013). Optimization of the L and w dimensions can be done subsequently.
L.w = π4 d2 Hh
102p
h=
ρ
The area for the gas flow above the liquid level should be at least three times the inlet
pipe cross-sectional area to prevent surges of gas flow to the flare. The diameter of a verti-
cal drum based upon avoiding pulsing is determined by providing an area ratio 1:3.
D2 − d2 = 3d2
36
D = 2d
The typical horizontal liquid seal drum is shown in figure 5.1.
Key
1 - to flare
2 - flare header
3 - try cocks used to check for hydrocarbons
4 - vent
5 - to sewer
6 - water supply
7 - submerged weir welded on end of flare line
8 - water level
9 - baffle
10 - drain
37
CHAPTER 6
FLARE STACK
Flare system design and plant equipment layout should minimize the need for operator
attendance and equipment installed in locations of high radiant heat intensity. The design
of towers or other elevated structures exposed to flare radiation should consider radiation
effects on the ability to safely egress.If personnel exposure to radiant heat exceeds the
set of guidelines, then shielding or other protection should be considered. It is often most
effective to accomplish this by locating ladders and platforms on a side away from the flare.
Personnel are commonly protected from high thermal radiation intensity by restricting
access to any area where the thermal radiation can exceed 6.31kW/m2 . The boundary
of a restricted access area can be marked with signage warning of the potential thermal
radiation exposure hazard(Daniel A.Crowl et al., 2013). There are practical differences
between laboratory tests and full-scale field exposure. Heat radiation is frequently the
controlling factor in the spacing of equipment such as elevated and ground flares. The table
presents recommended design total radiation levels for personnel at grade or on adjacent
platforms. The extent and use of personal protective equipment can be considered as a
practical way of extending the times of exposure beyond those listed. Each company may
select the radiation level to which personnel can be exposed, either for a short duration or
continuously. The wind and ambient temperature can influence the amount of radiation a
person can withstand.
The flare tip is a proprietary device. The flare tip design is based on the net heating
value, maximum velocity of flare gas, composition of flare gas, steam rate, auxillary fuel
Figure 6.1: Recommended Design Thermal Radiation
rate and ignition systems(Adam Bader, et al., 2011). The flare tip is a complex design.
The flare tip diameter is calculated from maximum allowable velocity and total flow rate
including auxillary fuel rate.
r
Qtot
Minimum flare tip Diameter, Dt = 1.95
Vmax
The flare tip diameter is rounded upto the next commercially available size. The minumum
diameter is 1 inch. Larger sizes are available in 2 inch increments from 2 to 24 inches and
in 6-inch increments above 24 inches. The maximum size commercially available is 90
inches.
The height of a flare is based on the ground level limitations of thermal radiation inten-
sity, luminosity, noise, height of surrounding structures, and the dispersion of the exhaust
gases. In addition, consideration must also be given for plume dispersion. Industrial flares
are normally sized for a maximum heat intensity of 1500-2000 Btu/hr-ft2 when flaring at
39
their maximum design rates. The intensity of solar radiation is in the range of 250-330
Btu/hr-ft 2 . The flare height also depends on wind effect and safely dispersion of flare gas.
However the wind effect is reduced by flare support structure like self supported, derrick
supported and guy supported flare. The safely dispersion is acheived by plume dispersion
model. If the flare stack height increases, the thermal radiation intensity and the noise de-
creases in the ground level. Usually upto 90 m , there should be no hydrocarbon processing
unit and anyother flammable material. This 90 m may vary depends on the height of the
flare stack.
Heat transfer from flare tip is propagated through three mechanisms: conduction, convec-
tion, and radiation. Thermal radiation may be either absorbed, reflected, or transmitted.
Since the atmosphere is not a perfect vacuum, a fraction of the heat radiated is not transmit-
ted due to atmospheric absorption due to humidity(Adam Bader, et al., 2011). In general,
the fraction of heat radiated increases as the stack diameter increases. The heat release
from the flare tip is calculated from the flare gas flow rate and net heating value. From
the heat release and fraction of radiation, the thermal radiation intensity in ground level is
calculated.
The Flare diameter is determined from Mach number in the flare stack. The recom-
mended Mach number for continuous flare is 0.2 and for emergency flare is 0.5. The
maximum mach number is 0.7. The Flare stack sizing procedure for continuous flare is
described below,
1. The governing load has been calculated for the failure case scenarios.
40
Table 6.2: Fraction of heat radiation
2. The Average molecular weight, Temperature, Heat of Combustion and specific heat
ratio of that stream is calculated.
41
(6.31kW/m2 ) for continuous flare system.
qm ZT 0.5
M a = 3.23 × 10−5 [ ][ ]
p 2 d2 M
9. From the heat liberated, the flame length is calculated by the graph.
10. The vapour volumetric flowrate and the flare tip velocity is calculated.
q
uj = πd2
4
11. The ratio of the wind velocity and flare tip velocity is calculated.
u∞
uj
∆x ∆y
12. From the ratio, the flame perimeters Σ ,Σ are calculated by the graph.
L L
13. The distance from the flame center to the grade level boundary, D is then calculated
by the equation.
r
τFQ
D=
4πK
Where D = minimum distance from the epicenter of the flame to the object being
considered.
τ = fraction of the radiated heat transmitted through the atmosphere.
F = fraction of heat radiated.
Q = Heat released, kW.
42
K = Heat Intensity,kW/m2 .
14. Then distance from the flare and object being considered in the ground is calculated
from the flare perimeters.
The height of the flare stack and flame is calculated.
43
Figure 6.5: Sizing a Flare Stack
44
Figure 6.6: Snapshot of Stack Sizing
45
CHAPTER 7
7.1 Introduction
The design of flare and vent system piping is an important part of the overall system
design for any chemical process. FLARENET has been designed to facilitate the design
and rating of flare and vent system piping throughout the entire design process. The pro-
gram interface uses a flow diagram for direct visualisation of the piping network. This is
supported by detailed tables of all pertinent data and calculated results. FLARENET can
model the piping system topologies most commonly found in flare systems. Multiple relief
scenarios such as ”Plantwide Power Failure”, ”Plantwide Cooling Water Failure” and ”Lo-
calised Fire” cases, as well as the individual relief valve loads can be maintained within a
single file model of the flare system. The following calculations can be done simply from
a consistent data set(Adam Bader, et al., 2011).
• Debottlenecking design of an entire flare system for a single and multiple relief
scenario.
• Rating of an entire flare system for a single and multiple relief scenario.
FLARENET has the option to calculate the pressure profiles using a range of single and
two-phase pressure drop calculation methods. These methods may be used globally through-
out the model or specified at a local level. Robust multiphase thermodynamic models back
up the physical property predictions used by the pressure drop models.
7.2 Data Requirements
Before starting to build Flarenet Model, the data are defined to determine our system.
The basic data is inlet pressure, inlet temperature, allowable backpressure, relieving pres-
sure, type of orifice, allowable mach number, failure case scenarios, length and elevation of
pipe, type of fittings, total mass flow, composition of the stream, methods to calculate pres-
sure loss, properties of stream, properties of pipe and parameters of KOD and flarestack.
The data input of PSV, Control valve, pipe, KOD and flare stack is shown in figure 7.1 7.2
7.3.
47
Figure 7.2: Knockout Drum Editor
The results can be viewed in the form of tables, graphs and charts. Data can be inter-
preted from the graphs. Typical graphs are shown in figure 7.4, 7.5, 7.6.
48
Figure 7.4: Flow Map
49
Figure 7.7: Pressure Profile
Based on the data availabe, the flarenet model has been built. Analysed the results in
the form of tables and graphs. The snapshot of that model is shown in figure 7.8.
50
CHAPTER 8
DISPOSAL TO FLARE
The flares are provided with auxillary fuel to combust hydrocarbon vapours when the
flare gas stream falls below the flammability range to sustain a stable flame. The amount
of fuel is calculated based on the flare gas stream net heating value(Adam Bader, et al.,
2011) Typically the natural gas has a net heating value of 1000 Btu/scf. Automatic control
of the auxillary fuel is recommended.
300 − Bv
Required fuel, F = G
Bf − 300
Where
Bv and Bf are the heating value of flare gas and fuel gas respectively in Btu/scf.
The purge gas is used to maintain the positive pressure and to prevent the backfire,
flashback and flame instability. Purge gas is inert gas usually N2 . If there is a possibility
of air formation in the leakage of flare manifold, the purge gas will prevent the formation
of an explosive mixture in the flare system(Adam Bader, et al., 2011). The purge gas re-
quirement is determined by the design of flare seals, which are usually proprietary devices.
Purge flow rate, Fpu = 190.8D3.46 y1 ln( 20.9
O2
)ΣCi0.65 Ki
Gas K
Hydrogen 5.783
Helium 5.078
Methane 2.328
Nitrogen 1.067
Ethane -1.067
Propane -2.651
CO2 -2.651
C4+ -6.586
The steam is used for the smokeless burning of flare gas. The steam requirement de-
pends on the flare gas flow rate, composition of flare gas, steam velocity and flare tip
diameter. Typically 0.01 to 0.6 kg of steam per kg of flare load is required(Adam Bader, et
al., 2011). This ratio is usually estimated from the molecular weight of the gas, the carbon
to hydrogen ratio of the gas and whether the gas is saturated or unsaturated. For example,
olefins such as propylene, require higher steam ratios than would paraffin hydrocarbons to
burn smokelessly. The smokeless flare is a proprietary device. The manufacturer should
52
be consulted about the minumum necassary steam rate for smokeless burning. Usually the
amount of steam required is 0.4 kg per kg of flare load.
Required Steam, S = 0.4G
Where
G = Flare gas flow rate.
53
CHAPTER 9
CONCLUSIONS
9.1 Conclusion
The flare network design is very complicated. Very few EPC companies in India doing
the flare network design because most of the designing systems in flare network is propri-
etary one. This project optimized the flare network design procedure. The flare network in
the old refining and petroleum industries requires huge amount of auxillary fuel and purge
gas. It eliminates heavy radiation to the atmosphere and to the ground level. The carbon
emission from the flare tip is heavy. The chunk cost for the piping and support of the flare
system also heavy. Apart from this the thermal design of Flare tip and ignition systems
also included in the designing of flare system. The supporting structure of the flare system
should be designed before procurement. The expenditure for the entire flare system should
be estimated. After the successful designing and procurement, the commissioning should
be done in that particular plant.
Appendix A
Location No. = xx
Item No. = xx
PSV No. = xx
55
Height of the vessel from the ground = 2 m;
Diameter, Di = 1.1 m;
= 0.866;
Latent Heat = 1695.35 kj/kg; (Calculated from Individual latent heat and weight frac-
tion of its components)
56
A.2 Model calculation to determine line loss
Location No. = xx
Item No. = xx
PSV No. = xx
Diameter = 4 ” ;(In this diameter, the line loss is less than 3% of set pressure);
kg
Density = 0.9874 m3
kg
Total Load = 4555.497 hr
;
3
Flow rate = 4613.424 m
hr
;
Velocity = 158.148 ms ;
57
kg
Line loss = 0.1196 cm 2;
qm ZT 0.5
M a = 3.23 × 10−5 [ ][ ]
p2 d2 M
58
r
45, 360
−5 1 × 422
0.2 = 3.23 × 10 ( )
101.3d2 46.1
d2 = 0.219
d=0.468 m ;
9.46
uj = π×0.4682
= 55m/s;
4
u∞ 8.94
= = 0.162;
uj 55
P 4x
From the graph = 0.85
L
P 4y
= 0.36
L
P
4x = 0.85 × 50 = 42.5m;
P
4y = 0.36 × 50 = 18m;
59
A.7 Model calculation for Flare stack height
r
τFQ
D=
4πK
r
1.0 × 0.3 × 6.3 × 105
D= = 48.9m
4π × 6.3
h0 = h + (0.5
P
4y)
r0 = r − (0.5
P
4x)
D2 = r02 + h02
48.92 = 24.42 + h02
h0 = 42.3m
h = 2.3 − (0.5 × 18) = 33.3m
60
Figure 9.2: Flare Stack Parameters
61
Appendix B
C(Re)2 C C(Re)2 C
10.0 59.00 6,000.0 1.20
20.0 33.00 7,000.0 1.15
30.0 24.00 8,000.0 1.10
40.0 19.00 9,000.0 1.05
50.0 16.00 10,000.0 1.00
60.0 14.00 20,000.0 0.84
70.0 12.00 30,000.0 0.75
80.0 11.00 40,000.0 0.70
90.0 10.00 50,000.0 0.66
100.0 9.50 60,000.0 0.62
200.0 6.00 70,000.0 0.60
300.0 4.70 80,000.0 0.59
400.0 4.00 90,000.0 0.57
500.0 3.50 100,000.0 0.55
600.0 3.20 200,000.0 0.50
700.0 3.00 300,000.0 0.47
800.0 2.80 400,000.0 0.47
900.0 2.70 500,000.0 0.46
1,000.0 2.50 600,000.0 0.46
2,000.0 1.90 700,000.0 0.45
3,000.0 1.60 800,000.0 0.45
4,000.0 1.40 900,000.0 0.45
5,000.0 1.30 1,000,000.0 0.45
62
Figure 9.3: Determination of Drag Coefficient (C)
63