Professional Documents
Culture Documents
Prelease PDF
Prelease PDF
Prelease PDF
Heat Flux
ric ratio of air-to-fuel (Figure 1).
Figure 1 is general for both a jet and a pool Global fuel-controlled
fire; the difference being a higher flux for the Global-
jet fire. For a pool fire, the API fire (4) is illus- ventilation- Open-fuel-controlled
trated as the lower, dashed line to the right. controlled pool fire
Note that in the equation API RP 521 uses, in-
creasing the area reduces the flux. The dashed
lines represent the average heat flux. Howev-
er, when studying the total volume of a fire, 1.0
any point on the continuous curve will be Stochiometric Ratio
found. A ventilation-controlled fire is to the
left of the peak heat-flux in Figure 1 at a stoi- Figure 1. The heat flux from a fire and its relation to the stoichiometric ratio.
chiometry of < 1. The fuel-controlled fire is to
the right, i.e., the stoichiometry is > 1. and more soot attaches to the surface. For more on absorp-
tivity and emissivity, see Ref. 5.
The fire equation By combining the suggested highest or lowest typical val-
The heat flux absorbed by a segment from a fire, qabsorbed ues into the fire equation, the heat fluxes toward a cold seg-
(kW/m2), can be modeled as: ment are found (Table 4 (6)). Typical heat fluxes measured in
large-scale jet fire and pool-fire tests are within the maximum
(
qabsorbed = segment fire T 4 fire segment T 4 segment + ) and minimum values in Table 4. Norsok (7) recommends
using the initial incident heat fluxes as specified in Table 5.
(
h Tgas Tsegment ) (1)
Sizing procedures
The fire-relief and depressurization calculations determine:
The absorbed heat flux will be reduced with increasing size of the relief valves and depressurization orifices
segment temperature, and a steady-state segment tempera- requirements for passive fire protection (PFP)
ture will be reached when the heat influx from the fire size of the pipes downstream from the relief and de-
equals the heat outflux from the segment. pressurization valves (if any)
The view factor, which is not included in Eq. 1, is a
scaling factor for the radiative terms. The view factor is Table 1. Typical flame emissivities for global
1.0. It equals 1.0 when the segments that absorb radiation and local fires.
see nothing but an optical, thick flame. Calculation of Type of Fire Global Fire Local Fire
view factors is difficult and a conservative assumption in- fire fire
volves use of a view factor of 1.0, which results in Eq. 1. Ventilation-controlled pool fire 0.60.75 0.70.9
The incident heat flux is calculated by setting segment = Fuel-controlled pool fire 0.60.75 0.70.8
1.0 and disregarding the segment emissivity term. The Jet fire 0.50.75 0.60.75
initial incident heat flux from a fire is
calculated by setting segment = 1.0 and
Table 2. Typical temperatures and convective heat-transfer
Tsegment equal to the normal operating tem- coefficients for a global fire.
perature (of the cold segment).
Type of Fire Tfire, C 2 Tgas, C h, W/m K
Input to the fire equation Ventilation-controlled pool fire 1,0001,050 850950 1530
Fuel-controlled pool fire 9501,000 800900 1530
The different terms in the fire equation Jet fire 1,0001,150 9501,050 50125
are combined to achieve the required ini-
tial heat fluxes. Tables 1, 2 and 3 suggest
values to be used. The segment absorptivi- Table 3. Typical temperatures and convective heat-transfer
coefficients for a local fire.
ty and emissivity in Eq. 1 are normally
equal and depend upon the nature of the Type of Fire Tfire, C Tgas, C h, W/m2K
surface. Typical values are 0.70.9. A Ventilation-controlled pool fire 1,0501,125 Equal to Tfire 2030
value of about 0.8 is typical for oxidized Fuel-controlled pool fire 1,0001,050 Equal to Tfire 2030
Jet fire 1,0001,150 Equal to Tfire 100150
surfaces. The value will change as more
Table 4. Minimum and maximum heat fluxes calculated by Eq. 1 for the suggested values of the
input parameters toward a cold segment.
Type of Fire Maximum Minimum Maximum Minimum
Incident Flux, Incident Flux, Absorbed Flux, Absorbed Flux,
kW/m2 kW/m2 kW/m2 kW/m2
Global jet fire 326 121 306 98
Local jet fire 367 187 347 160
Global fuel-controlled pool fire 138 88 127 65
Local fuel-controlled pool fire 187 124 171 92
Global ventilation-controlled pool fire 158 101 145 75
Local ventilation-controlled pool fire 228 142 208 105
Step 1:
Reduce the
Estimate the size of all orifices and calculate the pressure profile and
size of the
flare rates for all segments. Use the fire with the largest heat
orifice.
input (kW). No PFP in this initial iteration.
In case of
(Step 1) any of the "ORs"
Is the flare system
capacity utilized (when No (Step 1) Evaluate to increase the blowdown rate,
adding all of the simultaneous preferably for the most hazardous blowdown section.
blowdown rates
together)?
Step 2:
Add insulation if required. Calculate the process segment pressure
profile. Use the fire with the largest heat input (kW).
Yes Tip: Do several calculations with varying amounts of fire insulation.
(Step 1)
No Is the blowdown Yes Step 3:
rate less than Calculate the wall-temperature profile for all pipes and equipment.
maximum Use the local fire with the highest heat flux (kW/m2).
l-dP/dt l
Step 4:
Use the temperature profiles from Step 3 to calculate the rupture pressure
for all pipes and equipment. Compare with the pressure profile from
Step 2 (Step 1 in the first iteration).
Step 6:
Acceptance criteria: Decide which pipe/equipment to fire insulate
- Pipe rupture pressure Step 5: or
- Equipment rupture pressure Are the acceptance No Increase orifice diameter if available
- Released fluid at rupture criteria for rupture capacity in the flare system, or reduce system
- Time to rupture met? volume by relocation of sectionalization valves
- No rupture or increase the flare system capacity or change
material quality or increase wall thickness.
Yes
Step 7:
Calculate the minimum design temperature (low-temperature design
temperature) of the blowdown section and the flare system tail pipe.
Is the
minimum design No Start depressurization at a higher
Figure 2. Follow this sizing procedure temperature temperature (or change material).
to design depressurization orifices. acceptable?
Yes
pressurization orifices, using the capacity of the flare sys- within a segment, i.e., the temperature profiles from the
tem in the calculations. A recommended first estimate is an first iteration can be kept throughout the whole iteration
orifice diameter that takes the pressure below the unaccept- procedure. A final update of the temperature profiles
able rupture pressure within the typical time to rupture. The must be performed prior to the last iteration.
initial pressure should be the highest normal operating pres- Step 4: Calculate whether or not rupture occurs. De-
sure or an equalization pressure (settle-out pressure) for a termine the stress or strain that all pipes and equipment are
compression segment. A global fire should be used. The typi- exposed to for the temperatures and pressures seen during
cal time-to-rupture can be set at that interval it takes to reach a the depressurization (Calculated in Steps 1 or 2, and Step
600800C wall-temperature, depending upon the wall thick- 3) and determine whether the segment will rupture.
ness. A value of 510 min is typical for a dry wall exposed to Two failure (rupture) criteria are often used: the maximum
a medium-heat-flux jet-fire, with no depressurization. stress or maximum strain (% elongation). The maximum
One way of improving the safety of the plant is increase stress criterion is usually the UTS. Rupture strain is a matter
the rate of depressurization, as the hazardous aspects of the of definition. Strain calculations require finite-element mod-
segment increase. The total blowdown rate can be kept un- eling of the system, which is usually not performed during
changed by increasing the depressurization times of the this step-wise method. Such calculations should be per-
less hazardous segments. A segment containing large formed for verification purposes during the final design.
amounts of light liquids (e.g., condensate or liquefied The suggested approach is to calculate the stress
petroleum gas (LPG)), those that will result in boiling-liq- from the internal pressure and add extra stress (margins)
uid expanding-vapor explosions (BLEVEs) are regarded as when calculating the longitudinal stress. The stresses of
a particularly hazardous section. In any case, there may be importance for a pipe are the hoop stress caused by in-
limitations on maximum pressure gradients for items such ternal pressure, and the longitudinal stress. The longitu-
as compressors or gaskets. dinal stress is the sum of axial stresses due to pressure;
Step 2: Add insulation, if required, and simulate the the weight of the pipe, valves, fittings, branch pipes,
pressure profile during depressurization when expos- etc; stress due to reaction forces exerted on the pipe by
ing the segment to the global fire. For the first iteration pressure; and stress due to thermal elongation of the
only, omit this step and go to Step 3. A global fire will add pipe. The equivalent stress (von Mises) is the stress to
heat to the fire-exposed area without PFP. The initial pres- be compared with the temperature-dependent UTS to
sure in this calculation should be equal to the highest nor- determine whether rupture occurs. The hoop stress,
mal operating or settle-out pressure. Credit for insulation hoop, is equal to:
should be given only for PFP. Piping and equipment with
insulation used for purposes other than for PFP should be Pressure Outer dia.
hoop = (2)
regarded as uninsulated. 2 Wall thickness
Unrealistic backpressure in the flare system may result
in a too-rapid simulated depressurization. The orifice back-
pressure should be based on the time-dependent simultane- The longitudinal stress, long, is given by:
ous depressurization rates.
If a depressurization segment is 100% fire-insulated, long = 1/2hoop + x (3)
then the integrity of the insulation and supports usually de-
termines the maximum allowable depressurization time, The equivalent stress is given by:
which is typically 3060 min. Account for the integrity of
the insulation by extending the depressurization time for a
Von _ Mises = hoop 2 + axial 2 hoop axial (4)
100% fire-insulated section. The reduced depressurization
rate for this section is used to allow for the increase of the
rate from a most-hazardous depressurization-section. A re- The term x in Eq. 3 represents all stress except for that
duced depressurization rate may increase the fire duration, set up by the pressure. A piping engineer should be con-
if a leak in this section is the source of the fire. sulted when determining the value of x.
Step 3: Simulate the temperature profile for all It is recommended that the UTS by reduced by 20% or
piping and equipment in the depressurization seg- more, depending on the reliability of the UTS data. The
ment when exposed to the local peak-heat flux. A jet 20% is a safety margin. Reduce the wall thickness by ac-
fire is normally used in these calculations, but the local counting for the mill tolerance. It must be assumed that the
load for a pool fire should be used if the segment will lower mill tolerance is delivered. Reduce the strength by
not be exposed to a jet fire. All piping means all pipes including the weld factor. Again, a piping engineer should
with different diameters, pressure classes and/or materi- be consulted.
al qualities. The temperature profile for one particular Step 5: Check the rupture pressure against the ac-
pipe usually is rather insensitive to pressure changes ceptance criteria. If all piping and equipment in the de-
Antisurge Valve
The real geometry of any type of process segment (e.g., the above system drawn in continous
lines) is transformed into a hypothetical cylindrical vessel (the vessel below in continuous line).
The hypothetical vessel is used in calculation of system pressure during depressurization and relief.
The diameter and length are increased until the volume equals the volume of the original system
(use the dominating diameter of the original system). The liquid level is adjusted to match the
liquid volume of the original system, "Add or subtract" wet and dry areas to match the wet and
dry area of the original system. Set a wall thickness equal to the dominating wall thickness of
Figure 4. Modeling the original system and adjust the weight until the weight matches the original sysem weight.
the real process segment Hence, we have a hypothetical segment where the system area and weight do not fit the
hypothetical cylinder, but do fit the original system.
as a hypothetical one
simplifies sizing proce-
dures. PSV
Flare System
Blowdown
EV Flare System
Blowdown
Orifice
Length
Di
will increase as the lighter components evaporate, and Hypothetical segment used for system pressure
the wall will eventually reach the rupture temperature, The hypothetical segment is modeled with the real sys-
even if it is liquid-filled. tem volume, system outer-area, with and without PFP, sys-
tem inside-area in contact with the gas, system inside-area
Modeling process segments in contact with the liquid, and system weight (of piping
Here, we present an approach to modeling complex and equipment).
depressurization and relief segments. The method models The hypothetical segment is modeled as a cylinder.
the complete complex geometry by creating one hypo- This shape is specified with a diameter equal to the most-
thetical segment that represents the total system volume dominating (volume) diameter of any item of piping or
and heat-transfer areas, and several sub-segments that equipment in the original segment. The associated wall
represent the real geometries of the segment. thickness for this diameter should be used. The length of
The hypothetical segment is used for calculation of the the cylinder is set such that its volume equals the volume
system pressure during depressurization or relief. The sub- of the original segment. The liquid level is adjusted to ob-
segments are used for calculation of the temperature re- tain the actual liquid volume. The hypothetical cylinder
sponse of each piece of piping and equipment within a pro- now represents the correct system volume. However, the
cess segment. The sub-segments are modeled with the outer area of this cylinder must be corrected to the real-
same geometric information that is required for a wall-tem- system outer area by subtracting or adding gas and liquid
perature-response calculation namely, the sub-segment area to the segment. Therefore, heat transfer with the sur-
wall thickness, segment or pipe diameter and inside fluid. roundings (fire, ambient air) will be modeled over the real