Hydrocarbon Processing - June 2022 - 27

Special Focus Process Optimization
F.-J. XU, F.-H. ZHOU and Z.-M. CAO, China National
Air Separation Engineering Co.; Q. YANG and L. ZUBER,
Sulzer Chemtech Ltd., Winterthur, Switzerland; and
J.-J. DONG, Sulzer Chemtech China, Shanghai
Investigations of high pressure drop
observed in ASU columns
The measured pressure drop of columns in air separation
units (ASUs) is often reported as being higher than the
predicted pressure drop. Particularly for high-pressure (HP)
columns, the difference can be up to four times higher, depending
on the operating pressure of the column. To investigate
the difference, the operation data of recently revamped
HP columns were collected; then, vapor and liquid loads as
well as their physical properties were generated using a process
simulation tool. Thereafter, a proprietary hydraulic rating
toola
was utilized to predict the dynamic pressure drop caused
by the column internals. The use of the differential pressure
transmitter and its connecting piping inside the cold box was
also scrutinized. It was deduced from calculations that the
average temperature of gas in the differential pressure transmitter
(PDT) piping located inside the cold box was more
than four times higher than the temperature of gas inside the
column. The resultant different gas static head at the two locations
leads to the deviation between the measured and the
predicted pressure drop.
Since the 1980s, structured packings have progressively replaced
trays, first beginning with the crude argon (CAR) column
and low-pressure (LP) columns. Around 2011, in China,
structured packings eventually became a default option for
HP columns.
It was determined that the pressure drop of columns in the
ASUs was higher than what was predicted by the hydraulic
rating toola
. The extent of the reported deviations was inconsistent-the
typical difference was ~20%, which was initially
believed to be within the accurate parameters of pressure drop
correlations. When the packings-equipped HP columns began
to operate nationwide, the observed deviation jumped.
For a 550-kPa HP column, the measured pressure drop was
three times higher, and for an HP column operated at 900 kPa,
the measured pressure drop was four times higher.
Pressure drop in a column is caused by two factors: dynamic
pressure drop and static pressure drop (commonly called
gas static head). Dynamic pressure drop is the resistance of
column internals, such as packings, to gas flow.1,2
The pressure
drop decreases with declining gas throughput, and reaches zero
if gas flow stops. For a packed column, the dynamic pressure
drop is closely linked to packings capacity-e.g., for a proprietary
structured packing seriesb
, 5.5 mbar/m and 12 mbar/m
Hydrocarbon Processing | JUNE 2022 27
correspond to 90% and 100% capacity, respectively. Therefore,
the dynamic pressure drop measured from a plant is typically
used to monitor packings hydraulic performance. In contrast,
the static pressure drop is caused by the pressure exerted
by gas weight over a certain height.
The gas density in ASU columns can reach 40 kg/m3
, depending
on the operating pressure of the columns, and structured
packings generate much less dynamic pressure drop
than trays. Therefore, the gas static head can contribute a
significant portion to the total pressure drop for ASU packed
columns. The proprietary hydraulic rating tool only calculates
dynamic pressure drop, so the gas static head must be manually
added to the dynamic one to obtain the total pressure
drop of a column. However, the measured pressure drop in
ASU plants matches neither the proprietary hydraulic rating
tool'sa
calculated pressure drop. Rather, it lies between.
Although the deviation does not cause any practical challenges
to column design and operation, a joint investigation
between the authors' companies began in early 2020 with the
objective of benefitting the ASU industry.
Data collection and analysis. Among the various columns
in an ASU, the HP column is the best option for investigation:
the simplest HP column consists of only one feed (e.g., the
compressed air feed) and two products (i.e., liquid nitrogen at
the top and oxygen-enriched liquid at the bottom). The high
pressure drop observed in HP columns will not be masked by
the inaccuracy of any pressure drop correlations.
In the last 2 yr, the authors' companies have successfully revamped
HP columns from tray to structured packings, and the
details of the first project were published elsewhere.3
The operation data of two identical revamped HP columns
in a plant were collected in June 2021. The vapor and liquid
loads and their physical properties were generated using a process
simulation tool, and then the dynamic pressure drop of
packings was calculated using the proprietary hydraulic rating
toola
. The gas static head inside the HP column was calculated
manually using Eq. 1:
∆Pstatic pressure drop in HP
= ρg × g × h
(1)
predicted dynamic pressure drop nor the total manually
Hydrocarbon Processing - June 2022

Table of Contents for the Digital Edition of Hydrocarbon Processing - June 2022
