Blowdown of Vessels and Pipelines: S.M. Richardson & G. Saville
Blowdown of Vessels and Pipelines: S.M. Richardson & G. Saville
INTRODUCTION
In the case of vessels, blowdown leads to a hazard because of the very low
temperatures generated within the fluid in the vessel. This leads to a reduction in
the temperature of the vessel walls and possibly to a temperature below the
ductilebrittle transition temperature of the steel from which the vessel is fabricated.
It can also lead to the formation of hydrates in cases when free water is present in
the vessel or to the formation of liquid condensate which can get carried over into
a flare or vent system. For blowdown from the top of a slugcatcher of inside
diameter 4 m, wall thickness 150 mm and length 50 m containing mainly methane
at an initial pressure of 110 bara and temperature of 278 K (5 C) through an
orifice of equivalent diameter 50 mm, the minimum gas temperature is about 228 K
(45 C) and the minimum temperature of the inside of the vessel wall in contact
with gas is about 243 K (30 C).
In the case of pipelines, the hazard from blowdown arises not only because of
the low temperatures that can arise in the pipe walls but also because of the large
total efflux and high efflux rates that arise when the very large inventory in a
typical line is blown down. For a 40 km long 0.4191 m (16.5 in) bore gas line
containing mainly methane initially at 120 bara and 283 K (10 C), the initial efflux
rate following a fullbore rupture is about 3 te/s and a total of about 1000 te can
escape if the line is not fitted with subsea isolation valves.
195
I CHEM E SYMPOSIUM SERIES No. 130
Motivated by the need to predict hydrocarbon and vessel and line wall
temperatures and also efflux rate, composition and phase, a computer program called
BLOWDOWN has been developed by the authors for the simulation of blowdown
[14]. Because the physical processes occurring during blowdown (fluid mechanics,
heat and mass transfer and thermodynamics) are generally extremely complex,
development of a complete description and therefore model of blowdown is
effectively impossible. The main objective of BLOWDOWN is, instead, to simulate
all physically significant effects: all insignificant effects are eliminated. Clearly, a
great deal of judgement is required in order to be able to select physically
significant effects. This judgement is based upon extensive experimental work which
has been carried out concurrently with the modelling. The experiments have proved
to be invaluable in this sense, as well as providing validation for predictions made
using the model. A short description of the main features of BLOWDOWN (more
details are given elsewhere [2,4]) and of its experimental validation (more details are
again given elsewhere [3,4]) follows, together with two case studies which illustrate
typical applications of the program.
Space discretisation
A vessel is assumed to comprise three distinct zones (see Figure 1):
a top zone of gaseous hydrocarbon (including evaporated water and suspended
liquid droplets which have condensed from the gas);
a middle zone of liquid hydrocarbon (including dissolved water and gas bubbles
which have evaporated from the liquid);
a bottom zone of free water (including dissolved hydrocarbons).
A pipeline (whether on its own or as part of the blowdown system for a vessel) is
divided axially into a number of discrete elements, the size of each of which is
varied dynamically during the calculation in such a way that changes in physical
properties along the element can be neglected.
Time discretisation
The blowdown is broken down into a sequence of discrete pressure steps. Steps of
specified pressure decrement, rather than time steps, are used because pressure is a
much more relevant parameter in a thermodynamic sense.
Fluid mechanics
196
I CHEM E SYMPOSIUM SERIES No. 130
Heat transfer
In a vessel, heat transfer is by:
forced/natural convection in the top zone;
nucleate/film boiling in the middle zone;
natural convection in the bottom zone;
transient conduction through the wall (including any insulation);
forced/natural convection to the air or sea surrounding the vessel.
In a pipeline, heat transfer is by:
forced convection to the fluid (or else heat transfer is ignored and the flow
assumed to be either isothermal or adiabatic: experimental evidence is usually
consistent with an assumption of isothermal flow as far as blowdown times and
rates of efflux are concerned);
transient conduction through the wall (including any insulation);
forced/natural convection to the air or sea surrounding the line.
In all cases, standard correlations are used to determine heat transfer coefficients
[914].
Orifice
Because transit times through the orifice are comparable with times for nucleation
and growth of gas bubbles in a volatile liquid, nonequilibrium flashing flow is
assumed when just volatile liquid is fed to the orifice. Otherwise, the flow
approaching the orifice is assumed to be in thermodynamic and phase equilibrium.
Choking
When choking occurs, whether for a gas, liquid or twophase (gasliquid) flow, the
mass flow rate is a maximum.
Balance equations
In order to close the system of equations describing blowdown, mass, energy and
(for pipelines) momentum balances are performed. The reason why momentum
balances are required for pipelines and not vessels is that vessels are assumed to be
at a spatially uniform pressure; pipelines are not.
Blowdown experiments have mainly been conducted using two different vessels,
oriented vertically or horizontally and containing a range of different fluids with
blowdown from the top, bottom or side through chokes of various different sizes:
vessel A (flat ends): length 1.524 m, inside diameter 0.273 m and wall thickness
25 mm;
vessel B (torispherical ends): tantotan length 2.250 m, inside diameter 1.130 m
and wall thickness 59 mm.
Vessel A was oriented either vertically or horizontally, vessel B was always oriented
vertically. Transducers were used to measure the pressure, with an estimated
accuracy of 0.2 bar. Barewire thermocouples were used to measure the
temperature of the fluid within the vessels and also of the inside and outside vessel
walls, with an estimated accuracy of 0.5 K and a response time of order 0.1 s.
197
I CHEM E SYMPOSIUM SERIES No. 130
Predictions made using the BLOWDOWN program have been compared with
all of the validatory experiments. One point that must be borne in mind when
comparing the experimental results and the BLOWDOWN predictions is that the
program contains no disposable parameters. There has, therefore, been no adjustment
of parameters in order to ensure good agreement between the experimental
measurements and the predictions. In order to give some feel for the agreement
between the experimental results and the BLOWDOWN predictions, two
representative examples will be examined.
In the first example, vessel A was oriented vertically and nitrogen blown down
from the top through a choke of equivalent diameter 6.35 mm. The experimental
results are summarised in Figure 2, which gives the variation with time of the
pressure in the vessel, and Figure 3, which gives the variations with time of the
bulk gas and inside wall temperatures. Also shown on the figures are predicted
variations made using BLOWDOWN. There is clearly excellent agreement between
the measured and predicted pressures in the vessel. This illustrates the more general
result that it is relatively easy to predict the pressure decay rate using any
physically reasonable assumptions. Clearly, there is also good agreement between the
measured and predicted bulk gas and inside wall temperatures. Interpolation and
smoothing of the spatial variations in the gas temperature give the isotherms shown
on the lefthand side of Figure 4. These indicate the presence and hence importance
of natural convection within the vessel, schematic streamlines for which are shown
on the righthand side of Figure 4.
In the second example, vessel B was oriented vertically and 66.5 mole%
methane, 3.5 mole% ethane and 30.0 mole% propane blown down from the top
through a choke of equivalent diameter 10 mm. The experimental results are
summarised in Figure 5, which gives the variation with time of the pressure in the
vessel, Figure 6, which gives the variations with time of the bulk gas and bulk
liquid temperatures, and Figure 7, which gives the variations with time of the
temperatures of the inside of the wall in contact with the gas and with the liquid.
Also shown on the figures are the predicted variations made using BLOWDOWN.
There is again excellent agreement between the measured and predicted pressures in
the vessel. There is also reasonable, and usually good, agreement between the
measured and predicted bulk gas, bulk liquid and inside wall temperatures.
It follows from the comparisons reported here and from others reported
elsewhere [3] that the BLOWDOWN program predicts results which are in close
agreement with experimental measurements: minimum average bulk fluid and wall
temperatures can be predicted with an estimated typical uncertainty of 3 K. This
agreement permits confidence to be placed in the predictions made using
BLOWDOWN.
198
I CHEM E SYMPOSIUM SERIES No. 130
There has been relatively little practical investigation of very rapid blowdown of
large lines containing hydrocarbons. There were, of course, the accidental blowdowns
following the sequential, probably fullbore, rupture of the three gas lines connected
to Piper Alpha, one of which, that between Piper and MCP-01, was monitored
sufficiently well to give useful validatory information. The blowdown of this line is
discussed in the second case study.
The measured and predicted variations with time of the pressure and
temperature at the intact end of the line and of the inventory of the line are given
in Figures 8, 9 and 10, respectively. There is clearly good agreement between the
values measured experimentally and those predicted using BLOWDOWN. Alteration
of the assumed composition to 90 mole% propane and 10 mole% butane changes the
predicted pressure and temperature at the intact end of the line by less than
0.5 bar and 1 K, respectively, and the inventory of the line by less than 1%. The
only experimental feature not predicted by BLOWDOWN is the small pressure
undershoot near the start of the blowdown (see Figure 8). This undershoot is
determined by the usually irrelevant dynamics of the initial expansion wave.
The first case study concerns blowdown of a high pressure gascondensate separator.
The carbon steel vessel is taken to be vertical and of tantotan length 10 m,
inside diameter 4 m, wall thickness 120 mm and to have hemispherical ends of
thickness 80 mm. The choke is of 50 mm equivalent diameter and blowdown to
atmosphere is either from the top or from the bottom. The hydrocarbon feed to the
vessel is of the following composition: methane: 75 mole%, ethane: 5 mole%,
propane: 3 mole%, butane: 4 mole%, pentane: 4 mole%, hexane: 4 mole%, heptane:
3 mole%, octane: 2 mole%. Liquid hydrocarbon condensate initially fills 20 vol% and
the coexisting gas phase 80 vol% of the vessel. The initial pressure is 100 bara and
the initial temperature 283 K (10 C). The vessel is assumed to be standing in
freelyconvecting air at a temperature of 278 K (5 C).
The predicted variations with time of the pressure in and inventory of the
vessel are shown in Figures 11 and 12, respectively. In both figures, lines marked T
refer to blowdown from the top and lines marked B to blowdown from the bottom.
The main features of the blowdowns are as follows:
blowdown to essentially atmospheric pressure takes about 3000 s, irrespective of
whether blowdown is from the top or bottom of the vessel;
the initial fluid inventory in the vessel is 31.6 te: when the vessel has been blown
down, the remaining inventory in the vessel is about 16.3 te when blowdown is
from the top of the vessel and about 0.1 te when it is from the bottom;
the maximum efflux rate through the choke is about 0.027 te/s (110 MMSCFD)
199
I CHEM E SYMPOSIUM SERIES No. 130
when blowdown is from the top of the vessel and about 0.154 te/s (420 MMSCFD)
when it is from the bottom;
the minimum bulk gas temperature is about 222 K (51 C) when blowdown is
from the top of the vessel and about 220 K (53 C) when it is from the bottom;
the minimum inside wall temperature is about 267 K (6 C) when blowdown is
from the top of the vessel and about 261 K (12 C) when it is from the bottom.
The principal distinction between blowdown from the top of the vessel and
blowdown from the bottom is the rate of decrease of inventory in the vessel.
Although vessels are usually blown down from the top at present, there is an
increasing interest (especially after events on Piper Alpha) in blowdown of vessels
containing significant quantities of flammable liquid from the bottom since doing so
minimises the inventory which might otherwise feed a fire [16]. Of course, blowdown
from the bottom also leads to a flow mainly of liquid out of the vessel which could
not normally be fed directly to a flare or vent. (In this case study, however, it
turns out that the liquid condensate is sufficiently volatile that it flashes to form
mainly gas downstream of the choke which could be fed directly to a flare or vent.
The problem would then be the very high flow rate which would probably
necessitate use of a choke with a smaller equivalent diameter, at least during the
early part of the blowdown.) There are, therefore, problems associated with the
disposal of significant quantities of liquid: storage in another vessel such as a flare
knockout drum can just alter the location of the hazard.
The second case study concerns the probable [16] fullbore blowdown of the gas line
between the Piper Alpha and MCP-01 platforms on the evening of 6th/7th July
1988. It is based on evidence [17] given by the authors to the Public Inquiry into
the tragedy on Piper. The line was of the following dimensions:
subsea line length: 53804 m;
subsea line bore: 0.4191 m;
riser height (from sea bed to topsides): 160 m;
riser bore: 0.4064 m;
roughness lengthscale (assumed): 0.063 mm;
orifice diameter (assumed fullbore at top of riser on Piper): 0.4191 m.
The gas in the line had the following composition: methane: 73.60 mole%, ethane:
13.40 mole%, propane: 7.40 mole%, butane: 1.40 mole%, pentane: 0.15 mole%,
hexane: 0.02 mole%, nitrogen: 4.03 mole%. The initial pressure in the line was
117 bara and the initial temperature was 283 K (10 C), assuming that the gas
adopted a typical summertime sea temperature. Hence the initial inventory in the
line was 1280 te (51 MMSCF).
The measured and predicted variations with time of the pressure at the intact
(MCP01) end of the line are given in Figure 13. There is clearly good agreement
between the values logged at MCP01 and those predicted using BLOWDOWN.
It is of interest to investigate the effect on the total efflux from the line of
fitting an SSIV (subsea isolation valve). In addition to the base case:
case 1: no SSIV in the line;
already discussed and corresponding to the actual case at Piper, two cases have
been investigated:
case 2: an SSIV at a distance of 400 m along the seabed from the base of the
riser at Piper;
case 3: an SSIV at a distance of 200 m along the seabed from the base of the
riser at Piper.
The SSIV is assumed to start to close 60 s after rupture of the line, since it would
200
I CHEM E SYMPOSIUM SERIES No. 130
take about this length of time for the rupture to be detected and for valve closure
to be initiated. It is assumed that there is a linear reduction of flow area with time
while closing the SSIV and that closure takes 20 s, which is based on the
ruleofthumb that 1 in of closure takes about 1 s.
The predicted variation with time of the total mass efflux from the line is
given in Figure 14: line 1 refers to case 1, line 2 to case 2 and line 3 to case 3.
The main features of the blowdown are as follows:
in all three cases, immediately after rupture of the line, the flow is choked and
the mass flow rate out of the line is about 3 te/s;
in case 1 (no SSIV), about 1270 te (that is virtually all of the gas) has left the
line when blowdown effectively ceases about 25000 s (7 hr) after rupture of the line;
in case 2 (SSIV located 400 m from riser), about 37 te has left the line when
blowdown effectively ceases about 90 s after rupture of the line;
in case 3 (SSIV located 200 m from riser), about 36 te has left the line when
blowdown effectively ceases about 90 s after rupture of the line.
Clearly, fitting an SSIV, whether 400 m (case 2) or 200 m (case 3) from the riser
drastically reduces the amount of gas leaving the line and hence the threat to the
platform compared with fitting no SSIV (case 1). Moreover, fitting an SSIV close to
the platform (200 m from the riser; case 3), with possible risk of damage to the
SSIV from debris falling from the platform and from the anchors of ships, does not
give a large reduction in the amount of gas leaving the line compared with fitting
it some distance away (400 m from the riser; case 2).
CONCLUSION
BLOWDOWN has now been used by many oil and gas companies for the
simulation of future and existing depressurisation systems. Applications have included
a large number of individual vessels on offshore platforms and onshore installations, a
number of multiple vessels connected to common blowdown headers, subsea
pipelines and also accident investigation. BLOWDOWN can, therefore, be used
confidently as an engineering tool for the design of depressurisation systems.
201
I CHEM E SYMPOSIUM SERIES No. 130
ACKNOWLEDGEMENT
Some of this work was supported by Shell U.K. Exploration & Production.
REFERENCES
1. Haque, M.A., Richardson, S.M., Saville, G. & Chamberlain, G., 1990, Rapid
depressurization of pressure vessels, J. Loss Prevention Process Industries, 3,
47.
2. Haque, M.A., Richardson, S.M. &; Saville, G., 1992, Blowdown of pressure
vessels I: Computer model, Trans. I. Chem. E. part B: Process Safety
Environmental Protection, 70, 39.
3. Haque, M.A., Richardson, S.M., Saville, G., Chamberlain, G. & Shirvill, L.,
1992, Blowdown of pressure vessels - II: Experimental validation of computer
model and case studies, Trans. I. Chem. E. part B: Process Safety Environmental
Protection, 70, 1017.
4. Richardson, S.M. &; Saville, G., 1991, Blowdown of pipelines, Offshore Europe
91, (Society of Petroleum Engineers, Aberdeen), paper SPE 23070.
5. Richardson, S.M., 1989, Fluid Mechanics, (Hemisphere, New York), page 259.
6. Hetsroni, G. (editor), 1982, Handbook of Multiphase Systems, (Hemisphere,
Washington DC), chapter 2.
7. Rowlinson, J.S. & Watson, I.D., 1969, The prediction of the thermodynamic
properties of fluids and fluid mixtures - I: the principle of corresponding states
and its extensions, Chem. Eng. Sci., 24, 15651574.
8. Saville, G. & Szczepanski, R., 1982, Methanebased equations of state for a
corresponding states reference substance, Chem. Eng. Sci., 37, 719725.
9. Perry, R.H. & Chilton, C.H. (editors), 1973, Chemical Engineers, Handbook,
5th edition (McGraw-Hill, New York), pages 1010 to 1015.
10. Rohsenow, W.M., 1952, A method for correlating heattransfer data for surface
boiling of liquids, Trans. A.S.M.E., 74, 969976.
11. Jordan, D.P., 1968, Film and transition boiling, Advances Heat Transfer, 5,
55128.
12. Lienhard, J.H. & Dhir, V.K., 1973, Hydrodynamic prediction of peak
poolboiling heat fluxes from finite bodies, Trans. A.S.M.E. J. Heat Transfer,
95, 152158.
13. Berenson, P.J., 1961, Filmboiling heat transfer from a horizontal surface,
Trans. A.S.M.E. J. Heat Transfer, 83, 351358.
14. Churchill, S.W. & Chu, H.H.S., 1975, Correlating equations for laminar and
turbulent free convection from a horizontal cylinder, Int. J. Heat Mass
Transfer, 18, 10491053.
15. Cowley, L.T. &; Tarn, V.H.Y.: Consequences of pressurised LPG releases: the
Isle of Grain fullscale experiments, presented at GASTECH 88 (13th
International LNG/LPG Conference), Kuala Lumpur, 1988.
16. Cullen, W.D., 1990, The Public Inquiry into the Piper Alpha Disaster, (HMSO,
London), pages 134 & 393.
17. Richardson, S.M. & Saville, G.: Analysis of flows in Piper Alpha gas import
and export lines, Transcript of Days 90/91 of Piper Alpha Public Inquiry,
1989.
202
I CHEM E SYMPOSIUM SERIES No. 130
203
I CHEM E SYMPOSIUM SERIES No. 130
204
I CHEM E SYMPOSIUM SERIES No. 130
205
I CHEM E SYMPOSIUM SERIES No. 130
206
I CHEM E SYMPOSIUM SERIES No. 130
207
I CHEM E SYMPOSIUM SERIES No. 130
208
I CHEM E SYMPOSIUM SERIES No. 130
Figure 14 Total mass efflux out of ruptured (Piper) end of line: line 1
refers to case 1, line 2 to case 2 and line 3 to case 3
209