June
2016
HYDROCARBON
ENGINEERING
30
which are connected to the flare, it may be prudent to
specify the maximum back pressure at the process unit
battery limit when sizing the outside battery limits (OSBL)
headers and sub-headers from various units in the complex.
This would enable the unit engineer to size his or her
network independently and later integrate the same, along
with the OSBL relief network. This approach is slightly
conservative, yet effective when there are several different
entities, or offices, involved in the design of a complex relief
network, which involves many process units. For the
preliminary sizing of the inside battery limits (ISBL) relief
network, a maximum back pressure of 1.5 kg/cm
2
g at the
unit battery limit could be used (assuming that balanced
bellows PSVs are used at lower set pressures, and
accounting for the pressure losses for the liquid seal drum,
gas seal and flare tip). Further optimisation can be targeted
once all the relief valve data and pipe lengths, based on the
final piping isometrics, have been entered in the flare model
for the complete system.
n
n
Actual length of flare headers and sub-headers: the sizing
of the relief network should be based on the actual
routing and configuration of the PSV laterals, sub-headers
and main flare header. One of the most important factors,
which has a pronounced impact on the back pressure, is
the piping expansion loop configuration, which is based on
the design temperature of the flare header. Each piping
expansion loop would result in four additional 90˚ bends,
some additional straight length, and the exact
configuration of the expansion loops on the main
header/sub-headers, which could lead to a huge increase
in the built up back pressure, especially in cases involving
high pipe diameters. This needs careful attention when
designing a new relief network. The normal practice is to
specify a single design temperature for the entire length of
the OSBL flare header(s); however, sometimes this results
in a highly conservative expansion loop configuration,
leading to increased back pressure and cost. To optimise
the number of expansion loops, a flare header
temperature profile can be generated based on the actual
heat lost from the hot flare header to the ambient. Based
on this temperature profile, a lower flare header design
temperature can be specified for those OSBL header
sections farther away from the process unit(s), which
would lead to a smaller number of piping expansion loops,
thus lowering overall cost. Close coordination with the
piping stress engineer is required to make this exercise
effective.
n
n
Allowance for the pressure drop in the liquid seal drum
(where provided): the liquid seal drum should, generally, be
included in the scope of the flare vendor, as most of the flare
vendors provide their own specialised liquid seal drum
designs (including the specification of the proprietary drum
internals, e.g., anti-slosh baffles, etc). Generally, around
0.2 - 0.25 kg/cm
2
is considered adequate for the liquid seal
drum when sizing a new relief network.
Flare knockout drum and pumps
An adequately sized OSBL flare knockout drum should be
provided in order to prevent the carry over of liquid droplets
to the flare stack. The sizing of the knockout drum should
consider all the major relieving scenarios in the complex. In
situations involving extremely high peak flare loads, the
possibility of providing two parallel knockout drums, with
symmetrical inlet piping, could be explored. On a large
diameter horizontal knockout drum, the inlet header may be
Table 1.
Relief network checklist
1
Check that the PSV/control valve/BDV/XV layout in the
model is as per the final plot plan/3D piping model.
2
Check that the routing of all individual PSV outlet laterals,
flare headers and sub-headers is as per the actual (final)
piping isometrics.
3
Check that inner diameter (ID), length and required fittings
for all PSV outlet laterals/flare sub-headers/flare headers
are based on the final piping isometrics.
4
Check that the correct pipe schedule (based on the actual
selected pipe class) has been selected for all PSV laterals,
sub-headers and headers.
5
Check that the flare knockout drum(s) have been modelled
correctly with respect to type (vertical/horizontal), location
and dimensions.
6
Check that the flare stack height and diameter are based on
final vendor data.
7
In case a liquid seal drum is being used, check that an
adequate pressure drop has been included for the same.
8
Check that the flare tip diameter has been input based on
final vendor data.
9
Check that the flare tip pressure drop curves have been
input based on final vendor data.
10
Check that all the relevant relieving scenarios in the plant or
complex have been built in the model.
11
For each relieving scenario check that all the active PSVs
have been correctly selected.
12
For each relief valve check that the following data have
been input based on the final relief valve datasheet:
(a) Set pressure.
(b) Normal (kg/hr).
(c) Molecular weight (usually calculated based on the
component-wise composition data entered).
(d) Relieving temperature and pressure.
13
Check that the following data for each relief valve have
been based on the final relief valve vendor datasheet:
(a) Relief valve type.
(b) Orifice designation.
(c) Outlet flange diameter.
(d) Relief valve rated flow.
14
Check that the system back pressure has been entered
based on the overall project design data.
15
Check that the following data have been entered as per the
overall project design data:
(a) Atmospheric pressure.
(b) Ambient temperature.
(c) Wind velocity.
16
Check that the following constraints have been entered as
per overall project design data:
(a) Mach number constraints (these may be different for
headers and sub-headers).
(b) Liquid and vapour velocities (if applicable).
(c) Noise data.
(d) Constraints on rho*v2 (if applicable).
17
Check that an appropriate thermodynamic model, e.g., Peng
Robinson method, has been selected for VLE and enthalpy.
18
For two phase systems check that an appropriate method,
e.g., Beggs and Brill (Homogeneous) method, has been
selected for pressure drop calculations. For single phase
systems, the more rigorous adiabatic pressure drop method
should be used.
19
Check that there are no Mach number/choked flow
violations in any of the pipe segments.
Find out
Serving the tan
TSA16UpdatedAdv
