Claus waste heat boiler economics

Part 2: mechanical considerations

The design of a cost effective waste heat boiler faces many, often opposing
factors affecting the performance and reliability of the exchanger

Nathan A Hatcher, Clayton E Jones, Simon A Weiland, Steven M Fulk and Matthew D Bailey
Optimized Gas Treating, Inc.

T
he Claus waste heat boiler lectively increase the peak heat
Flow rate and composition of AAG
(WHB) runs under quite flux at the front of the boiler
and SWAG feed streams
harsh operating conditions, well above predictions from
has serious reliability challenges, Amine AG + SWS AG models that ignore or discount
and is one of the most fragile TGU recycle some (or all) of these factors.1
equipment items in the sulphur Flow rate, std. m3/h
H2S mol%
140
88.2
26
33.1
Greatly elevated tube wall tem-
recovery unit (SRU). It not only CO2 mol% 6.4 – peratures well downstream
provides heat recovery from the NH3 mol% – 40.9 of the area of protection pro-
thermal section, but it also affects vided by ceramic ferrules for
the unit’s hydrogen balance and Table 1 the higher mass velocity cases is
COS levels through recombina- demonstrated, lending theoret-
tion reactions. ical support to documented failures in the indus-
Part 1 of this two-part series (see PTQ, Q1 2019) try. In this article, tube wall temperatures, and
discussed general process considerations includ- heat flux predictions from the model are exam-
ing the effects of tube length, pressure drop, and ined down the length of the tubes along with the
COS/H2 reactions on sulphur recovery. One of implications of sulphidic corrosion and the result-
the main concerns was exothermic recombination ing effect on boiler tube life and SRU reliability
reactions and how they affect SRU performance economics examined with this new information.
in terms of hydrogen make, COS creation, and
sulphur recovery: Case study
The following case study is a continuation from
H2 + ½ S2 ⇌ H2S Part 1 and is a typical 125 lt/d sulphur plant (see
CO + ½ S2 ⇌ COS Figure 1) with two converter stages processing
both amine acid gas (AAG) and sour water acid
Part 2 focuses on determining the heat flux and gas (SWAG). Table 1 shows the conditions of
tube wall temperature profiles along the length these two acid gas feed streams. The WHB was
of the boiler. This is aimed at understanding the sized in Part 1 by fixing the process-side mass
particularly important area near the critical tube- flux, the tube size, and the process outlet tem-
to-tubesheet joint where WHB mechanical failures perature. The tube count and tube length were
frequently occur. Also relevant are boiler tube cor- adjusted to meet target specifications. The results
rosion rates and subsequent boiler failure together were obtained using the kinetic heat transfer and
with the cost of its mitigation. The recombination chemical reaction rate based SulphurPro SRU
reactions are exothermic. They increase both the simulator within ProTreat Version 6.4.
process fluid and tube wall temperatures, as do Table 2 shows values for other design param-
the species shifting between the S2, S6 and S8 allo- eters assumed for this application of the model.
tropes of sulphur. Radiative heat transfer coupled The boiler produces 350 psig saturated steam
with the exothermic recombination reactions col- from preheated boiler feed water.

www.digitalrefining.com/article/1002299 PTQ Q2 2019 1

 Figure 1 Case study sulphur recovery unit flowsheet

Heat flux, tube wall temperature, and corrosion high tube wall temperatures. Tube wall tempera-
implications ture together with the concentration of H2S pres-
Heat flux is a very important factor in the relia- ent correlate directly with the sulphidic corrosion
bility, life cycle cost, and the safe long term oper- rate.2 Figures 2, 3 and 4 show the heat flux and
ation of the WHB. Heat flux varies greatly along tube wall temperature profiles along the length
the tube length; it is highest at the process inlet of the exchanger tubes for 1.5in, 2in, and 3in
and decreases as the process gas cools. In terms diameter tubes, over a range of mass fluxes. Tube
of reliability, it is our experience that failures lengths have been determined so that the process
tend to become more common with heat fluxes gas just reaches 550°F (288°C) at the exit from the
exceeding 50 000 Btu/ft2·h. At elevated heat flux, tubes.
problems with steam blanketing on the utility Each curve shows a wave approximately one-
side are more common. Locally high heat flux third of the way along the boiler tubes. This is
values can result in local steam blanketing which, caused by the sulphur species shifting from S2
in turn, can result in locally higher tube wall tem- to S6 and S8 (sulphur redistribution), which are
peratures and increased corrosion rates. A host exothermic polymerisation reactions. Note that
of other factors such as tube pitch and orientation the heat flux considered here does not account
also play into the likelihood of steam blanketing; for thermal protection provided by ferrules or
these are discussed elsewhere.2 The life cycle cost for the enhanced heat transfer effect of eddies at
of a WHB depends greatly on the corrosion rate the ferrule exits. Eddies in the process gas flow as
of the boiler tubes caused by high heat flux and it exits the ferrules can cause the heat flux to be
amplified several times for a short
Design parameters
distance.2 As the mass velocity is
increased at a constant boiler tube
size, the peak heat flux and tube
Parameter
wall temperature both increase at
Outside convective heat transfer coefficient, Btu/h·ft ·°F
2
150
Emissivity for radiative heat transfer, unitless 0.9 the front of the tubes. This is where
Inside fouling resistance, h·ft2·°F/Btu 0.008 almost all failures from sulphidic
Outside fouling resistance, h·ft2·°F/Btu 0.002 corrosion occur. Tube wall temper-
Tube material Carbon steel
ature ties directly into the sulphidic
Tube wall thickness, inches 0.1085
corrosion rate. As the mass flux
through the boiler tubes increases,
Table 2 the tube wall temperature increases

2 PTQ Q2 2019 www.digitalrefining.com/article/1002299

by approximately 130°F (72°C),
750
which translates directly into

Tube wall temperature, ºF

2
higher sulphidic corrosion rates. 3
700

The predicted corrosion 4 650

rates shown in Figure 5 for two 50 5 600

Heat flux, BTU/ft2.h ×103

mass fluxes in a 2in tube were 40
550
determined from our digital 500
correlation of the well-known 30
450
Couper-Gorman curves, repro-
20
duced from the correlation here
as Figure 6. 10
At constant mass flux,
Figures 2-4 show that larger 0
0 5 10 15 20 25 30
boiler tubes tend to produce Distance along tube, ft
higher tube wall tempera-
tures. Corrosion concerns tend Figure 2 Heat flux (solid lines) and tube wall temperature (dashed lines)
to increase at tube wall tem- profiles along 1.5in OD boiler tubes with mass flux (lb/ft2·h) as parameter
peratures greater than 600-
650°F (315-343°C). For this
study, the tube-wall temper- 750

Tube wall temperature, ºF

2
ature at the front of the boiler 3 700

tubes is generally approach- 50 4 650

Heat flux, BTU/ft2.h ×103

5
ing this temperature range 600
40
for mass fluxes between 550

2 and 3 lb/ft2·s. The service 30 500

life of the exchanger is directly 450

determined by the corrosion 20

rate, and the life cycle cost is a 10

strong function of service life.
Looking at how sulphidic cor- 0
0 5 10 15 20 25 30 35 40
rosion occurs through the length Distance along tube, ft
of the exchanger, the higher
mass flux case has significant Figure 3 Heat flux (solid lines) and tube wall temperature (dashed lines)
corrosion rates through the first profiles along 2in OD boiler tubes with mass flux (lb/ft2·h) as parameter
five feet or so of the tubes. Figure
5 shows how the corrosion Tube wall temperature, ºF
trends through a 2in outside 2 750
3
diameter tube at 2 and 5 lb/ft2·s 700
4
mass flux. The corrosion rate 50 5 650
increases by almost a factor of
Heat flux, BTU/ft2.h ×103

600
four at the higher mass flux. 40
550

30 500
Life cycle cost and economics 450
There are two ways that Claus 20
WHBs cost the operator:
• Capital cost incurred when the 10
plant is built: ‘pay me now’
• Operating and maintenance 0
0 5 10 15 20 25 30 35 40
costs that are incurred: ‘pay me Distance along tube, ft
later’.
Capital cost is relatively Figure 4 Heat flux (solid lines) and tube wall temperature (dashed lines) pro-
straightforward to estimate. For files along 3in OD boiler tubes with mass flux (lb/ft2·h) as parameter

www.digitalrefining.com/article/1002299 PTQ Q2 2019 3

 25 Assumed cost parameters
5 lb/ft2.s
Corrosion rate, mpy 20 2 lb/ft2.s Corrosion allowance, in 0.125
Production advantage, $/lt 1000
15 Outage duration, days 14
Sulphur production, lt/d 125 (design rate)
10 CEPCI basis, year 2016

5 Table 3
0
0 5 10 15 20 25 30 35guishing parameter that affects
40 45
Cumulative tube length, ft O&M costs is replacement fre-
quency, and whether a planned
Figure 5 Sulphidic corrosion rates predicted at mass fluxes of 2 and 5 lb/ft2·h or unplanned outage was taken.
through 2in OD boiler tubes The replacement cost was esti-
mated to be a new exchanger
the purpose of this article, the capital costs were purchased in the future year plus a variable
calculated based on ratioed in-house budget- ‘outage’ cost component. If a particular hydro-
ary equipment quotes, then applying an overall carbon producing plant happens to be sulphur
installation factor of five to place the estimates on constrained, then an outage can result in quite
a US Gulf Coast battery limits installed cost basis. significant lost profit opportunity based on an
Consideration of a steam drum was included in advantaged feedstock margin. In many cases,
the cost estimate. the O&M cost may be considerably less if, for
Normal operating and maintenance (O&M) example, only refractory work or retubing were
costs that run day to day were assumed to even conducted. The results here are intended to
out across the various cases. The main distin- be illustrative, and it is the relative difference
between the lifecycle costs that
is important, not the absolute
10
1 mpy
values.
2 mpy Table 3 summarises the
3 mpy assumed cost parameters.
5 mpy The case of 1.5in outside
10 mpy
diameter boiler tubes and a
15 mpy
1
20 mpy
mass flux of 2 lb/ft2·s was arbi-
25 mpy trarily selected as the base case
30 mpy for doing relative comparisons
30 mpy between various designs within
50 mpy
the three tube size, four mass
0.1 flux design matrix. The cost cal-
culations included:
• Initial installed cost
• Annualised lost profit oppor-
tunity for downtime
0.01 • Equipment replacement cost
• O&M annualised cost
• O&M relative cost vs the base
mol % H2S

No corrosion case
• TIC (total installed cost) sav-
0.001 ings vs the base case
500 600 700 800 900 1000 1100 • Payback period relative to the
Temperature, ºF base case.
Table 4 shows the O&M cost
Figure 6 Couper-Gorman sulphidic corrosion curves for carbon steel vs the base case, and Table 5

4 PTQ Q2 2019 www.digitalrefining.com/article/1002299

shows the TIC savings vs the
Annualised O&M relative costs ($) vs base case
base case.
At each tube size, as the
2 lb/ft2·s 3 lb/ft2·s 4 lb/ft2·s 5 lb/ft2·s
mass flux increases, the initial 1.5in OD Base case 10 992 54 945 119 418
installed cost decreases (see 2in OD 26 467 50 591 100 243 203 692
Table 5) because the surface 3in OD 58 241 92 336 176 744 276 918
area required with 1.5in tubes
is the lowest of the three tube Table 4
sizes. The replacement cost was
assumed to be the cost of a new Total installed cost (TIC) savings ($) vs base case
exchanger, so this also decreases
with increasing mass flux. The 2 lb/ft2·s 3 lb/ft2·s 4 lb/ft2·s 5 lb/ft2·s
annualised lost profit oppor- 1.5in OD Base case 93 962 131 038 142 055
tunity for downtime is directly 2in OD (226 934) (164 145) (152 986) (162 824)
3in OD (623 465) (624 944) (661 103) (712 166)
related to the corrosion rate, and
Figure 5 shows that the corrosion
rate increases with mass velocity Table 5
and tube size which is directly
connected to the service life of the exchanger. cases would probably require two passes. This
As the mass velocity through the exchanger influences both the initial installation cost as well
increases, the lost profit opportunity from down- as the O&M annualised cost of the exchanger.
time increases because of higher corrosion rates. Scaling cost for two-pass exchangers introduces a
This directly affects the O&M annualised cost level of complexity in the cost analysis that was
shown in Table 4. not explored for this article.
Table 4 shows that as the mass velocity Another factor not fully explored is the pres-
increases, the O&M cost increases; this is the sure drop across the ferrules, although the 1.5in
result of higher corrosion rates, hence the shorter tube size appears to be the best choice, not only
life cycle of the exchanger. Table 5 shows that from an economic standpoint but also for hydro-
1.5in tubes are the most economical choice based gen production and sulphur recovery. The fer-
on TIC savings. As mass flux increases, TIC sav- rules in this size tube may cause high enough
ings increase because tube cross-section is better pressure drop to outweigh the other benefits.
utilised, so less material is required to manufac-
ture the tube bundle. For 2in and 3in tube sizes, Conclusions
TIC savings are negative, meaning that these The intelligent design of a cost effective WHB is
exchangers have higher capital cost. quite challenging. There are many, often oppos-
So, is it preferable to ‘pay me now’ with a ing factors that go into not only the performance
WHB design that has a higher capital cost, or to and reliability of the exchanger, but the econom-
‘pay me later’ with a design that has a higher ics as well. If a WHB is designed without proper
O&M cost over time? In reality, the answer to consideration of all the factors, the design could
this question is not straightforward – one must fall flat in one or more aspects. The SulphurPro
look at both the O&M cost and the TIC. An accu- simulator correctly accounts for all the important
rate simulation tool quantifies the effects of all factors.
aspects of WHB design, revealing right from the It has been shown that WHBs designed with
start what the full implications of various deci- small tubes and low mass fluxes favour safe
sions will be. heat fluxes, the lowest tube wall temperatures,
There are many other factors that enter into and the lowest sulphidic corrosion rate. Each of
the cost analysis; one of them is the number of these favourably affects the lost profit opportu-
exchanger passes. WHBs with a tube length nity from downtime. Additionally, as Table 5 of
greater than 35ft are typically built with two Part 1 showed, small tubes and high mass fluxes
passes because of structural support and plot favour sulphur recovery and hydrogen make.
layout limitations. Table 3 of Part 1 showed that Considering all these competing factors, a bal-
half of the 2in tube cases and all of the 3in tube ance must be struck – one that depends on the

www.digitalrefining.com/article/1002299 PTQ Q2 2019 5

 Optimized Gas. He holds a BS in chemical engineering from the
specific plant’s economic objectives and its per-
University of Oklahoma.
formance requirements. Steven M Fulk is Technical Director with Optimized Gas
Treating. He holds a BS in chemical engineering from Texas A&M
SulphurPro and ProTreat are marks of Optimized Gas Treating, University and a PhD in chemical engineering (2016) from the
Inc. University of Texas at Austin.
Matt Bailey provides sales and marketing support with
References Optimized Gas Treating, Inc. He has over 10 years of industry
1 Karan K, Mehrotra A K, Behie L A, Including radiative heat experience in technical sales, creating safe engineered solutions
transfer and reaction quenching in modeling a Claus plant waste for refineries, chemical plants and engineering companies. He
heat boiler, Ind. Eng. Chem. Res., 33, 1994, 2651–2655. is a registered Engineer in Training with the Texas Board of
2 Martens D H, Tube and tube weld corrosion and tube collapse, Professional Engineers and holds a BS in chemical engineering
Brimstone-STS Sulphur Symposium, Vail, CO, 2011. from Texas A&M University and an MBA from the University
3 http://igs.nigc.ir/STANDS/IPS/e-tp-760.PDF
of Houston.
Nathan (Nate) A Hatcher is an Engineering Consultant with
OGT. He holds a BS in chemical engineering from the University LINKS
of Kansas.
Email: [email protected]
More articles from: Optimized Gas Treating
Clayton E Jones joined Optimized Gas Treating, Inc as a
Software Development Engineer in 2012. He holds a BS in More articles from the following categories:
chemical engineering from McNeese State University and a MS Gas Processing and Treatment
in chemical engineering from the University of New Mexico. Process Modelling and Simulation
Simon A Weiland is Technical Applications Engineer with

6 PTQ Q2 2019 www.digitalrefining.com/article/1002299