Professional Documents
Culture Documents
P
rocess modelling errors and failure Computer models are
to design vacuum unit equipment invaluable tools used Ejector
as an integrated system has caused during CPD to establish system
yield loss, poor gasoil quality, and bottlenecks and define
unscheduled shutdowns. The vacuum scope. But the process
unit charge pump, fired heater, transfer flow sheet models must
line, column internals, and ejector sys- represent true operation Diesel
tem must be evaluated and designed and account for the
together so that operating temperature many equipment non- LVGO
and pressure can be optimised to meet ideal situations encoun-
economic goals. Revamps need to push tered in a refinery
major equipment to its intrinsic limits vacuum unit. If not, Crude
to minimise investment. inaccurate calculated
MVGO
Real equipment performance should process stream data will
be the basis of a revamp, not office- result in short run- Crude
based assumptions or cursory reviews of lengths and low gasoil HVGO
the original equipment manufacturer’s product yields.
data sheets. Even though vacuum unit One such example is
equipment is often highly constrained the influence of transfer
by existing equipment, an experienced line phase separation on Column
revamp engineer can often manipulate predicted heater outlet
Transfer
heater outlet temperature, column flash temperature and the cal- ATB line
zone pressure, coil steam injection, or culated wash section liq-
vacuum bottoms stripping to achieve uid flow rates. Unlike Fired
revamp yield and reliability targets. some theoretical mod- heater
The challenge is to accomplish these elling concerns that have Stripping
steam
objectives while meeting stringent few real penalties, failure
investment criteria or Capex restric- to account for transfer Fuel gas Vacuum
tower
tions. Vacuum unit heater-inlet- line phase separation bottoms
Crude Crude
through-ejector-outlet (Figure 1) must will lower the gasoil
be considered a single system when a product yield and result
practical, operable and cost-effective in a very low wash oil Figure 1 Single system, ejector through heater inlet
revamp is to be implemented. A success- flow rate.
ful revamp will meet all these standards. Wash bed packing requires adequate built based on a low capital investment
Revamps should always start with a wetting in the middle of the bed to strategy and not best practices method-
thorough test run gathering all neces- avoid stagnant zones where coking is ology. Poorly designed equipment may
sary data on current unit and equip- initiated. In reality, accounting for operate well at low temperature and
ment performance. Next, the field data transfer line phase separation raises the high pressure, but when severity is
is used to calibrate a baseline process wash oil flow rate by 200–300 per cent increased, the run-length or yield expec-
model and quantify current equipment over conventional modelling practices tations are not met.
performance. The calibrated model will that assume liquid and vapour leaving Many poorly designed vacuum
be the basis of the revamp calculations. the transfer line are in equilibrium. This heaters have been built because the
Only then can the revamp engineer one oversight alone has caused many selection criteria were based mainly on
apply equipment design know-how to unscheduled shutdowns after less than a low price. Low-cost heaters have high
identify all under-utilised equipment year of operation. Process modelling heat flux imbalances among the passes
and exploit it to minimise investment. plays a significant role in meeting over- that significantly reduce the maximum
Without this approach during the all refinery economics. reliable heater firing.
conceptual process design (CPD), all Rigorous equipment models need to Often, engineering approaches are
major cost bottlenecks may not be fully account for the various designs that are based solely on superficial evaluations
defined. This leads to scope growth and encountered when revamping. Many such as average radiant heat flux or ven-
cost escalation during later stages of vacuum unit fired heaters, transfer lines, dor data sheet review. These methods
engineering. columns and ejectors were designed and minimise engineering cost but often
85
P T Q SUMMER 2002
w w w. e p t q . c o m
REVAMPS and SHUTDOWNS
86
P T Q SUMMER 2002
REVAMPS and SHUTDOWNS
88
P T Q SUMMER 2002
REVAMPS and SHUTDOWNS
Pass #2
Outlet
Oil residence time
Pass #1
Secondary
steam
FI injection
0.2 wt%
Primary
steam
0.1 wt% injection
0.05 wt% Control valve
pressure drop
Figure 5 Cracked gas production: oil film temperature vs Figure 6 Minimised pressure drop: multiple pressure
residence time locations
These heaters can operate at average being able to determine the time and temperature or oil residence must be
2
heat flux rates above 10 000Btu/hr-ft -°F temperature relation in each tube and reduced to prevent short heater run-
without excessive coking. Therefore, the the quantity of gas the ejector must be lengths.
ATB pump is a critical component of capable of handling. Rigorous tube-by- Transfer Line
meeting both the HVGO product cut- tube rating allows the designer to eval- Heater outlet (Figure 7) pressure is set by
point and run-length. uate oil residence time and peak oil the transfer line design. Transfer line
Fired heater film temperature in each tube (Figure pressure drop and two-phase critical
The fired heater supplies the energy 5). Coke forms because the oil film velocity both influence this pressure.
needed to raise the oil temperature high inside the heater tube reaches temper- Pressure drop is based on transfer line
enough to meet the gasoil yield target ature where the oil is no longer ther- configuration. Two-phase critical veloci-
and supply sufficient wash oil flow to mally stable. High temperature will ty is much lower than the sonic velocity
keep the wash section packing from cok- not cause excessive coking at low resi- of the gas phase alone. Hence, many
ing. Heater outlet temperature and pres- dence time. designers will calculate a low percentage
sure are important variables. Heater Coil steam lowers oil residence time. of sonic velocity and predict a much
outlet pressure depends on the transfer The revamp engineer must decide lower heater outlet pressure than would
line design. The maximum allowable oil where to inject the coil steam and how be feasible. Calculation methods for crit-
temperature is based on the rate of coke much steam is needed to suppress cok- ical velocity are complex.
laydown inside the tubes and the target- ing while balancing feed pump Accurate field measurements of many
ed heater run-length. hydraulic limits. Picking the wrong transfer line pressures and temperatures
Coke insulates the oil inside the tube location will cause coking in the tubes and subsequent calculations using a
from the hot flue gas. Therefore, the upstream of the coil steam injection model with critical velocity calculations
outside tube metal temperatures (TMT) point or high-pressure drop which lim- have confirmed that critical velocity is
must increase to transfer heat through its charge rate (Figure 6). The influence real. Therefore, heater outlet pressure is
the insulating coke layer. Tube metallur- of heat flux imbalances and other non- a function of both hydraulic losses and
gy, tube thickness, and tube life will ideal conditions should not be over- two-phase critical velocity.
determine the maximum TMT before looked because many low-cost heaters Transfer lines usually have long hori-
the heater must be decoked. have very poor radiant section heat dis- zontal runs with line diameters of
Targeted heater run-length will deter- tribution. Heaters designed with the
mine the acceptable rate of coke forma- individual tube passes stacked have
tion. Coking rate can only be inferred poor heat distribution.
by measuring the TMT with either skin Thorough evaluation of the heater
thermocouples or infrared thermogra- and the transfer line is necessary to
phy. When coke is produced, so is establish gasoil product yield and heater Flash
zone
cracked gas, and it can be directly mea- run-length. A rigorous heater and trans- ATB
sured from the ejector hotwell. As coke fer line model will calculate pressure
formation increases, so does the amount drop and limit operating pressure to the
of cracked gas. Gas leaving the hot well two-phase critical velocity limit. These
will be primarily cracked gas unless the models must make the calculations
unit has excessive air leakage. backwards from the flash zone through Fired Transfer Column
Low cracked gas rate will be 0.05 wt% the heater. heater line
of feed. High cracked gas rate will be After each run, peak oil film tempera-
0.2–0.3 wt%. Once the gas rate exceeds tures and oil residence times in each Absolute pressure, mmHg
these values, short heater run-lengths tube are evaluated to determine if rapid
should be expected. coke lay down will occur. If high coking Figure 7 Heater outlet pressure influence
The key to vacuum heater design is rates are predicted, then heater outlet on product yields
89
P T Q SUMMER 2002
REVAMPS and SHUTDOWNS
optimisation. Exceed-
Vapour/liquid ing vacuum column Ejector
phase separation capacity causes high system
vacuum tower bottoms
(VTB) entrainment
C.W.
Vapour into the HVGO prod-
uct. Minimising col-
Diesel
umn flash zone
Liquid pressure reduces the
heater outlet tempera- LVGO
ture. Higher column
operating pressure
Figure 8 Process modelling: transfer line allows more gasoil Crude
phase separation production without
MVGO
exceeding the column
36–84in prior to entering the column capacity factor, but it Crude
flash zone. These large diameter horizon- requires higher heater HVGO
tal runs cause the liquid and vapour outlet temperature. Efficiency of wash section
phases to separate. Calculated phase Lower column oper-
Deentrainment tray
regimes are either stratified or stratified ating pressure reduces
wavy (Figure 8). Stratified phases cause heater outlet tempera- Vapour horn
the liquid and vapour to have poor mass ture. However, lower
and energy exchange across the inter- column operating pres-
face. Hence, liquid and vapour contact- sure requires a larger Stripping
ing is poor. Thus, phase separation causes column cross-sectional trays
vapour to flow through the top of the area to produce a given
horizontal portion of the transfer line. amount of gasoil.
This vapour becomes superheated due Vacuum column
to pressure reduction as the vapour diameter is often one
approaches the column flash zone. of the major limits Vacuum
tower
Vapour and liquid entering the flash that the revamp engi- bottoms
zone of a vacuum column are not in neer must circumvent.
phase equilibrium. Vacuum column Ejector system inlet Figure 9 Vacuum column design
wash section coking causes unscheduled pressure and heater
shutdowns. There are several potential outlet temperature must be optimised. Vapour and liquid feed enters the col-
causes. However, failure to account for Once the overhead pressure is selected, umn at velocities as high as
phase separation in the transfer line is the pressure drop through the column 380–400ft/sec. The vapour phase con-
one of the most common. Coked wash determines the flash zone pressure. The tains small droplets of VTB that have
section packing has occurred in nearly flash zone pressure (oil partial pres- been generated in the transfer line.
100 per cent of columns operating sure) and the heater outlet temperature Droplet size is too small to allow set-
above 750°F where the wash rate was and pressure determine the amount of tling in the transfer line because the
calculated assuming equilibrium. oil vaporised. velocity is too high. Ideally, vapour
Wash oil flow rate to the packing Proper design of vacuum column leaving the horn should enter the wash
must be sufficient to keep the packing wash sections, flash zone, and strip- section uniformly distributed across
wetted, otherwise high oil residence ping sections are necessary to meet the column cross-section and it should
time stagnant zones are created in the yield, quality, and reliability targets. be free of entrainment because entrain-
middle of the packed bed. Transfer line Stripping section performance influ- ment contains high metals, concarbon
phase separation increases the required ence heater outlet temperature. Higher and asphaltenes.
amount of wash oil flow rate. Super- stripping steam rate and improved effi- The wash zone removes entrained
heated vapour entering the wash section ciency reduce the heater outlet temper- residue from the flash zone vapour and
vapourises a portion of the liquid flow- ature. Flash zone internals should provides some fractionation of HVGO
ing down through the packing. separate much of the VTB from flash product. The fractionation require-
Once the vapour superheat is zone vapour and help distribute ments depend on crude oil type, gasoil
removed, oil condensation occurs from vapour to the wash section. Wash sec- yield target, and product quality. Typi-
the vapour. The bottom of the wash sec- tion internals include liquid and cally, a wash zone consists of a packed
tion is kept wetted by flash zone vapour distributors above and below bed, spray header or gravity flow
entrainment. The top of the packing is the bed, respectively, and the packing. trough type liquid distributors, and a
wetted by the wash oil flow rate. The Packing type and bed depth determine collector tray below the packing. Cok-
wash oil flow rate must be high enough efficiency. ing in the wash zone occurs either in
to ensure that the middle of the packing Figure 9 shows the various parts of a the packing itself or on the collector
remains wetted. Otherwise coke forms vacuum column. tray under the packed bed.
in the middle of the packed bed. Bed Flash zone and wash section Revamps may produce high gasoil
depth has a significant influence on cok- As flash zone operating temperature yields for a 6–18 month period; how-
ing because as bed depth increases there and capacity factor increase, design of ever, wash oil flow rate is not suffi-
is more middle section depth that must these internals becomes increasingly cient to operate for four to five years.
be wetted. important. Design mistakes have The collector tray below the wash bed
Vacuum column reduced HVGO product yield and can also coke due to high liquid resi-
Vacuum column diameter often drives caused coking and unscheduled shut- dence time. High liquid residence
the system pressure and temperature down of numerous vacuum units. time results from either the tray
90
P T Q SUMMER 2002
REVAMPS and SHUTDOWNS
91
P T Q SUMMER 2002