Reactor effluent air cooler safety through design

Quality-controlled replacement of carbon steel with Duplex 2205 for revamps can
increase the service life and reliability of the REAC in the high-pressure
Eric Lin and Peter Risse Chevron Lummus Global
in hydroprocessing units, the reactor effluent cooler (REAC) is one of the
most vital pieces of equipment, and any hindrance to its smooth operation
immediately impacts the whole high-pressure loop. Older REAC designs
used carbon steel, but these required low concentra- tions of ammonium
bisulphide and/or polysulphide sulphide injection together with frequent
and thorough inspections. As feeds in most heavy oil hydroprocessing
service have become more laden with sulphur and nitrogen, the
concentrations of ammonium bisulphide with economic levels of water
injection have risen to a point where carbon steel tubes have routinely
been substituted by alloy tubing.
Duplex 2205 and Alloy 825 are used, with the former being very
popular because it is relatively less expensive. Initially, there were
several prob- lems associated with Duplex 2205, which were a result
of poor fabrication techniques. These included a rapid cooling rate
associated with thick header boxes, which could result in high ferrite
and thus poor corrosion resistance; welding of thick tubes to tube
sheets with joint leaks; and lack of control of welding, resulting in
high hard- ness and thus susceptibility to sulphide stress cracking.
Many of these initial problems seem to have been overcome.
However, problems persist around the REAC, primarily because of a
lack of attention to detail and not adhering to licensor specifications. In
this article, we will illustrate real REAC problems from recent projects
and consider the remedies that were recommended

Background
In the hydroprocessing industry, the REAC is one of the most
important components in the high-pressure recycle gas loop. It
typically provides the final cooling solution before separating the
vapour (recycle gas) from the oil effluent and the sour water. The
outlet temperature directly impacts recycle gas molecular weight as
larger hydrocarbon molecules dropout of the vapour phase. The
same mechanism also affects the hydrogen partial pressure, which
directly impacts reactor catalyst life. The impor- tance of this piece
of high-pressure equipment cannot be overstated. However,
operating under high pressures and low temperatures can bring a
host of issues into the equation, not the least of which includes
ammonium bisulphide (NH4 HS) and ammonium chloride (NH4Cl)
precipitation (leading to pressure drop build-up), corrosion and/or
erosion-corrosion. Some refiners have also experienced cracking in
REAC welds, lead- ing to fires, and loss of both time and money.
Tube metallurgy has a huge impact on a REAC’s life expectancy.
Materials currently used in REAC systems include carbon steel, Type
400 series stainless steels, Type 300 series stainless steels, duplex
stainless steel Alloys 3RE60 and 2205, Alloy 800, Alloy 825, Alloy 625
and Alloy C-276. Early refiners used carbon steel and this was found
to be effective with NH4HS concentrations up to 3 wt%. As sour
opportunity crudes became available, the resulting NH4HS crept up
into the double digits. In cases where the resulting NH4 HS was
between 3-8 wt%, polysulphide injection was a very cost-effective
solution for the continued use of carbon steel (especially for revamps).
However, over time, the persistent plugging/oper- ation problems,
frequency of inspection and the foul smell of the polysulphide liquid
greatly diminished its use. Increasingly sour feeds demanded more
corrosion-resistant materials such as nickel-based Alloys 625, 800 and
825. Capable of handling up to 15 wt% NH4 HS, Alloy 825’s corrosion
resistance is normally matched by an equally hefty price tag. As the
cost of materials continued to rise, a more economical alternative
with comparable protection against corrosion was sought.
Duplex stainless steels are often successfully used in these systems
because they offer advan- tages from both the ferritic and austenitic
stainless steel families. They are often cost effec- tive due to their
higher strength and reduced alloy element content compared to other
higher alloys (up to one-third the cost of Alloy 825). However, since
these materials consist of dual- phase microstructure, heat-treating,
fabrication and welding techniques need to be carefully reviewed and
monitored to assure that the balanced microstructure is not
compromised. In the past, Duplex 3RE60 was used, but it had infe-
rior corrosion resistance and toughness at the welds (it is no longer
available). The most commonly used grade today is Duplex 2205.
Early implementations of duplex REACs failed to show significant
reliability improvements, and a few units failed by hydrogen
embrittlement cracking or sulphide stress cracking (SSC).
Advancements in steel manufacturing have mini- mised
microstructural deterioration during fabrication. Weld procedures
and practices have been developed to assure balanced ferrite and
austenite content, thereby improving reliability. API TR 938C
provides guidance on materials and fabrication practices to achieve
good corrosion resistance in duplex stainless steels.

Case studies
Many of the reported incidents involving Duplex 2205 REAC failure
have a common thread: pres- sures greater than 1000 psig and a
NH4HS concentration of 6 wt% or greater. The following two recent
case studies highlight these conditions.
A licensee in Asia (“Company A”) had a carbon steel REAC running
for a number of years. An expansion of the unit required a REAC
metal- lurgy upgrade to Duplex 2205. The design pressure of the
REAC was ~2400 psig and the NH4 HS concentration was expected
to be 5 wt% during the modelling phase. A new Duplex 2205 REAC
was installed and ran without issue for about two years, then fire
erupted due to REAC failure, and nearby equipment and piping
were damaged. Cracks were observed on the weld joints between
 the top plate and tube sheet, as well as the bottom plate and tube
sheet of the floating header. Fin tubes were deformed and the
walkway was nearly unusable. An investiga- tion was commissioned
to find the root cause of the REAC failure. The investigation
discovered that during the REAC fabrication,Charpy impact testing
was conducted at 0ºC instead of -40ºC (as specified by CLG). This
oversight led to inad- equate toughness and low ductility for the
welds as well as less than favourable microstructural phase balance.
Hardness values in the heat-af- fected zones (HAZ) and at the weld of
failed specimens were higher than those recommended (in the range
of 313-359 HV10 vs 310 HV10 maximum). As a result, CLG
concluded that REAC failure in this case was due to sulphide stress
cracking (SSC).
Another licensee based in the US (“Company B”) had a two-stage
hydrocracker and hydro- treater ready for startup. The unit required a
Duplex 2205 REAC. The design pressure of the REAC was ~2400 psig
and the NH4HS concentra- tion was expected to be 8 wt% based on
modelling. During initial startup, a decision was made to skip the
high-pressure tightness testing. A fire erupted from cracks formed in
the REAC outlet temperature indicator thermowells in two separate
trains. These cracks were found to be a result of improper
manufacturing practices (of the failed duplex material) and would have
been detected during a high-pressure tightness test. By not following
standard operation practices, the unit startup was delayed.

REAC safety through design

Issues that plague the industry’s REACs can be mitigated or
eliminated altogether by consider- ing “Safety Through Design” . In
preparing a basic engineering package for licensees, CLG highlights
certain areas and design considera- tions that are considered strong
recommendations. These considerations include:
• Using balanced, symmetrical piping
• Limiting tube velocities based on metallurgy, NH4HS concentration
and H2S partial pressure
• Specifying fabrication and welding guidelines for Duplex 2205 above
and beyond those speci- fiedin API TR 938C
• Using single-point water injection instead of multi-point water
injection.
Symmetrical piping for REAC inlet tubes is critical for reactor
effluent. It is impossible to implement any type of flow control for two-
phase flow, so the use of symmetrical piping will ideally provide
even distribution of gas and liquid phases as it enters the air cooler.
The danger of uneven phase distribution is that tubes favouring
vapour will have more NH4HS depos- its, and tubes favouring liquid
will have more of the injection water. These deposits will eventu-
ally plug the tubes and increase the rate of corrosion. Consequently,
unplugged tubes will see higher than acceptable velocities,
increasing the rate of erosion/corrosion. It is possible for piping to be
symmetrical and still be unbal- anced. Balanced headers require
additional splits, but will prevent flow that follows the “path of least
resistance” .
During basic engineering, CLG specifies tube velocity limitations
based on the tube metallurgy selected along with NH4HS
concentration:
• For carbon steel tubes and piping with less than 3 wt% NH4HS, 10-
20 ft/s is allowed, with 15 ft/s preferred
• For 2205 duplex stainless steel tubes and piping with 3-12 wt%
NH4HS, 10-30 ft/s is allowed, with 25 ft/s preferred
• For Alloy 825 tubes and piping with up to 15 wt% NH4HS, 10-40
ft/s is allowed, with 35 ft/s preferred.
In all cases, tube velocities falling below 10 ft/s become quite
hazardous, as phase separation and corrosion can result.
For the industrial use of Duplex 2205, API TR 938C gives a great
starting point for fabrication and welding guidelines. CLG has crafted
a stand- ard specification that addresses certain areas in more
depth, and creates greater accountability for both the fabricator and
the welder. For example, CLG specifies a maximum hardness rating
of 310 HV10 versus API’s requirement of 320 HV10. There is also
tighter control of weld consumable and filler metals. Qualifications for
Weld Procedure Specifications (WPS) and Procedure Qualification
Record (PQR) are also much more stringent. As the two case studies
illustrate, Duplex 2205 REAC failure can be attributed to not
respecting the guidelines specific to duplex material.
One final design consideration is whether to use single- or multi-
point water injection before the REAC. With the numerous valves
involved in a multi-point water injection system, water management is
difficult. Even with symmetrical piping, a multi-point water injection
system may have some tubes favouring more water than others. This
scenario can be dangerous, as some tube temperatures fall below
the NH4HS precipi- tation temperature. Single-point water injection
ensures that the global water requirement is met from the start, well
before the effluent splits into their separate headers (the injection
point is typically 10 pipe diameters upstream of the split). Another
benefit of single-point injection is that the recycle gas compressor
spill back line can be positioned downstream of the injection point so
that it is washed with injection water. This setup minimises NH4Cl
and NH4HS forma- tion in the dead-leg piping.

Summary
The combination of high pressures and corrosive environments can
spell disaster for REACs in the refining industry. Fortunately,
measures such as using balanced symmetrical inlet piping, limiting
tube velocities based on metallurgy and NH4HS concentration,
working with suppliers and weld- ers who have good-quality control
with Duplex 2205, and the use of single-point water injection can
increase the service life and reliability of the most important air
cooler in the high-pressure loop.

Eric W Lin is a Principal Process Engineer with Chevron Lummus Global in

Bloomfield, New Jersey. With more than 15 years of experience in the refining
industry, he holds a BS degree in chemical engineering from Columbia University.
Peter J Risse is a Senior Staff Materials Engineer with Chevron Lummus Global in
Richmond, California. With over 30 years of experience in the refining industry, he
holds a BS degree in chemical engineering from University of Rhode Island and a
MS degree in materials science from University of California at Berkeley.