Chemical Engineering June 2012 - 54

TABLE 1. PRE- AND POST-SHUTDOWN OPERATING CONDITIONS
Pre-shutdown Post-shutdown
Engineering Practice
Feed from the deethanizer,
thousand lb/h
brought to minimum governor speed.
Steam to the reboiler was started
10 h later. The depropanizer was operated
in the liquid re-injection mode for
the following 12 h. Throughout this period,
several moves were made on the
reflux and reboil rate. Since these operations
were unsuccessful in bringing
the tower bottom within specification,
the operation mode was transitioned
to the condensate stripper scheme.
Furthermore, roughly 5,000 lb/h of offspec
C4's from storage was fed to the
depropanizer through the external
C4's feed point to help load the tower.
The tower was still performing poorly
and process engineering support was
then requested by operations.
Depropanizer troubleshooting.
Prior process engineering work [2]
done on startup conditions highlighted
the risks associated with operating
high-capacity trays below their minimum
liquid limit. There was concern
that the tray load might not be sufficient
to seal the downcomers. However,
an initial high-level review of the
operating conditions did not reveal
anything abnormal.
As described in Table 1, the tower
operating conditions were not significantly
different from the pre-shutdown
ones. Plant data also suggested
that the upper section was working
fine, since reflux changes and overhead
composition were varying as expected.
The stripping section, however,
was performing poorly.
Simulation work and a gamma scan
on the tower were then conducted.
The Operations Dept. was asked to
maintain stable conditions, regardless
of performance, for the duration required
to conduct the gamma scan on
the tower. The intent of the scan was
to determine if any mechanical anomalies
were present inside the tower
and to assess the tray-loading profile.
The gamma scan results (Figure 3)
indicated no mechanical anomalies and
no sign of flooding. Vapor-liquid disengagement
looked better compared to a
previous scan done at a normal plant
rate. Many of the high-capacity trays
showed a clear vapor space equivalent
to the vapor space above the reboiler
return and the bottom tray. Scanning
across the tray could not provide any
information with regard to the downFeed
from the condensate
stripper, thousand lb/h
External C4's, thousand lb/h
Reflux rate, thousand lb/h
LP steam to reboiler, thousand
lb/h
Tower pressure, psig
112.8
63.0
98.9
24.2
123
Tower pressure drop (in. H2O) 111
(at time of Gamma Scan)
105.6
72.8
5 (off-spec C4s)
94.3
26.3
122
112
TABLE 2. PROBLEM ANALYSIS
Deviation: Depropanizer bottom stream C3s outside specification
Is
Is not
What Object:
Depropanizer
bottom
stream
Defects:
C3s outside
product specification
Where
Depropanizer
stripping
section
Object:
Deethanizer
Defects:
Bottom
specification
Depropanizer
rectifying
section
*
Feedrates
* Properties
* Process
conditions
* Composition
*
Trays
* Tray type
* Composition
*
Hydraulic
load
* Temperature
* Properties
When Startup
Pre-shutdown
*
Path to lowrate
operation
*
Low
rates in
C3 area
* Stripping
section tray
downcomer
unseal
* Stripping
section tray
weeping
comer operation. Based on the work
done by Urbanski and others [3], it is
assumed that tray weeping cannot be
determined by the gamma scan.
The tower was simulated with inputs
reflecting operating conditions
sustained at the time of the scan. The
intent of the simulation was first to
ensure that the trays were operating
within their hydraulic capabilities.
The operating point relative to flooding
was calculated using the Kister
and Haas correlation [4] for the sieve
tray section.
The operating point for the highcapacity
trays was plotted on the tray
operating window obtained from the
tray vendor. The simulation results
estimated the sieve tray operation at
72-73% of flood, which compare to
68-70% of flood for the pre-shutdown
operation. The high-capacity tray operating
points are shown on Figure 4
and compared against loads calculated
50 CHEMICAL ENGINEERING WWW.CHE.COM JUNE 2012
for pre-shutdown operation. As for
the sieve tray, the startup conditions
showed loads slightly higher than preshutdown.
Tray
loading calculations were consistent
with the gamma scan results,
which supported the conclusion that
the trays were not flooded but rather
were operating at low rates. Some of
the key information gathered was
structured into a problem analysis matrix,
as shown in Table 2. The concept
of problem analysis is to look at what,
where and when the problem is - and
is not - and to look at differences and
changes, as well. It was clear from
Table 2 that the main change was the
low feedrates in the C3 area. A review
of previous startups concluded that
the plant had never been brought up
to steady-state conditions at such low
rates. A light feed slate similar to preshutdown
conditions explained the
lower rates to the C3 and C4 area. This
* Low
rates in
C3 area
* High capacity
tray turndown
not
met
* Low
rates in
C3 area
Differences Changes Probable
cause
http://WWW.CHE.COM
Chemical Engineering June 2012

Table of Contents for the Digital Edition of Chemical Engineering June 2012
