67-70 16/03/2004 11:58 am Page 67

here is a growing focus in the refining industry on hydro- they may conclude that the most cost effective solution over

T gen capacity. Hydrogen is generally required for sulfur

removal and quality improvement of hydrocarbon prod-
ucts. As sulfur restrictions on gasoline and diesel become
increasingly stringent, the refining demand for hydrogen con-
tinues to grow.
By evaluating the hydrogen utilisation in their facilities,
refiners are realising that they need additional hydrogen sup-
ply. There are a number of options available to address this
need. Refiners may be able to meet the increased demand by
improving the operations of their existing hydrogen plants.
They may choose to separate hydrogen from waste or offgas
streams or even purchase hydrogen from third parties. As an
alternative, after careful technical and economic evaluation,
the longer term is to build a new hydrogen plant.
In recent years, substantial advancements in hydrogen
plant technology have significantly improved overall life cycle
costs. Based on experience with hydrogen plant benchmark-
ing, it has become clear that the optimum economic solution in
some cases may be to replace an existing hydrogen plant with
a new modern hydrogen plant.
Hydrogen plant technology
Old style
Many refiners are still operating hydrogen plants that were
designed and built 20 or more years ago. These older plants

Time for a new

hydrogen plant?
Clay A. Boyce, M.Andrew Crews and Robin Ritter, CB&I
Howe-Baker, USA, discuss the available options when making decisions
about hydrogen plant efficiency and debottlenecking.

Figure 1. Typical
modern hydrogen plant.
67-70 16/03/2004 11:58 am Page 68
Figure 2. Old style plant.
Figure 3. Modern plant.
were generally designed with the best available technology of
the time. Typically, these plants consist of a steam methane
reformer (SMR) to convert the hydrocarbon feed to a syngas
mixture followed by a high temperature shift converter (HTSC)
and low temperature shift converter (LTSC) to shift most of the
CO to hydrogen. This hydrogen rich gas is purified by a CO2
removal unit where a hot potassium carbonate or MEA solu-
tion removes the CO2 and then a methanator converts the
remaining CO and CO2 to methane and water. The final prod-
uct gas is normally 95 - 97% hydrogen. Figure 2 shows a typ-
ical layout for an old style hydrogen plant.
Modern
The main process difference between a modern hydrogen
plant and an old style hydrogen plant is the hydrogen purifica-
tion technology. A pressure swing adsorption (PSA) unit
replaces the CO2 removal unit and methanator and allows the
modern plant to produce hydrogen product with a much higher
purity. In addition, many significant improvements in technol-
ogy and design allow the modern plant to operate at much
higher efficiency and with substantially lower operating costs.
A modern hydrogen plant includes an SMR followed by an
HTSC. A PSA unit purifies the syngas effluent from the HTSC.
The final product gas is typically 99.99% hydrogen. Since the
PSA unit removes the impurities from the syngas, an LTSC is
not required to further reduce the CO content.
The PSA unit produces an offgas stream that can be used
by the SMR as the primary fuel source. In addition, the
increase in product purity from the PSA unit has potential ben-
efits in downstream units. For example, the higher purity
hydrogen makeup to a hydrotreater lowers the recycle flow
and reduces compression costs. Figure 3 shows a typical lay-
out for a modern hydrogen plant.
Additional hydrogen capacity
options
There are a number of options available to refiners to meet the
Reprinted from HYDROCARBON ENGINEERING FEBRUARY 2004
increase in hydrogen demand. Before proceeding, refiners
should conduct a comprehensive technical and economic
evaluation of their existing operations and evaluate the bene-
fits of the options available to them. The option that provides
the optimum economic and operational benefits will be differ-
ent for each situation and will depend on such things as the
existing steam balance, the cost and availability of utilities, plot
limitations, and the condition of existing hydrogen plants.
The first option for refiners who are operating old style
hydrogen plants is to consider ways to increase the capacity
of these plants. Following an evaluation of the condition and
efficiency of the existing plant, they may be able to effectively
increase the capacity by either operating with tighter control or
selectively upgrading portions of the old plant.
In many refineries, hydrogen is treated as a utility. There
may not be much of a focus on the details of the actual pro-
duction and the plants are sometimes operated loose. In this
case, simple operational changes could significantly increase
production and efficiency. Refiners could also debottleneck an
existing hydrogen plant by revamping or upgrading portions of
it. Debottlenecking options can also have a positive efficiency
impact. Some of these potential upgrades may include:

Replacing reformer tubes with upgraded metallurgy and
thinner walls will allow for more throughput and a
higher heat flux, which would increase capacity.

Adding a pre-reformer unloads the primary reformer so
capacity can be increased.

Adding a secondary reformer increases methane
conversion, which increases capacity.

Adding combustion air preheat lowers the fuel requirement
and potentially unloads the waste heat recovery unit and
fluegas fan, resulting in a capacity increase.

Upgrading the CO2 removal unit can minimise hydrogen
loss in the methanator, resulting in a potential capacity
increase.

Adding a PSA unit decreases hydrogen production, but
would typically produce more cost effective and higher
purity hydrogen.
A second option is to separate hydrogen from a waste
stream or an offgas stream that is currently being sent to fuel.
This typically requires adding separation equipment and pos-
sibly some compression. In addition, separating hydrogen out
of the fuel system will usually result in additional makeup fuel.
This could change the heating value of the refinery fuel system
and possibly have an impact on other fuel burning equipment.
A third option is to buy hydrogen from a third party. Various
industrial gas suppliers are willing to sell hydrogen to refiners
either by pipeline or, depending on location, by a stand-alone
plant. This option requires a minimal capital investment by the
refiner but the delivered hydrogen is generally more expensive
per unit than if self produced.
The last option is to build a new hydrogen plant. Building a
modern hydrogen plant is typically the most capital intensive of
the options available. However, the capital investment could
pay off if there is a significant gain in efficiency. Below the typ-
ical overall production cost of hydrogen between a modern
and an old style hydrogen plant is analysed and compared.
Overall hydrogen production cost
The most significant economic factor in evaluating options to
increase hydrogen capacity is the overall cost of producing
hydrogen. The overall production cost can be estimated over
the life of the hydrogen plant by using the different cost para-
meters of constructing, operating, and maintaining the hydro-
gen plant. This overall production cost reflects a complete pic-
ture of the hydrogen plant economics over the life of the plant.
The efficiency of producing hydrogen and byproducts is
67-70 16/03/2004 11:58 am Page 69

very important in minimising the production cost of hydrogen. Table 1. Total utility cost of hydrogen
Utility costs are the major operating cost in hydrogen produc- Utilities per 1000 ft3 of contained hydrogen
Old style Modern
tion. Utilities typically include usage of feed, fuel, boiler feed
Natural gas feed, million Btu LHV 0.275 0.317
water, power, cooling water, and generation of export steam Natural gas fuel, million Btu LHV 0.2 0.126
(steam is typically a by product of the hydrogen production Total feed + fuel, million Btu LHV 0.475 0.443
process). Of these, feed and fuel make up the largest portion HP export steam, lbs 20 90
Boiler feedwater, lbs 45 120
of the utility costs. In addition, the credit for export steam can
Power, kWh 0.65 0.52
have a significant impact on utility costs, especially when refin- Cooling water, gal. 530 8
ery utility costs are favourable for steam production. The Cost (US$) per 1000 ft3 of contained hydrogen
remainder of the utilities combined generally makes up less Old style Modern
Natural gas feed @ US$ 4/million Btu LHV 1.1 1.268
than 10% of the total operating cost.
Natural gas fuel @ US$ 4/million Btu LHV 0.8 0.504
These utility costs, together with other economic parame- Total feed + fuel @ US$ 4/million Btu LHV 1.9 1.772
ters applicable to the plant being evaluated, can be incorpo- HP export steam @ US$ - 5/1000 lbs -0.1 -0.45
rated into a cash flow model, and the overall production costs Boiler feedwater @ US$ 0.5/1000 lbs 0.023 0.06
Power @ US$ 0.05/kWh 0.033 0.026
of hydrogen can then be evaluated. The other economic para-
Cooling water @ US$ 0.1/1000 gal. 0.053 0.001
meters include such things as capital cost, startup cost, other Total utility cost, US$ 1.908 1.409
operating costs (including catalyst replacement and tube
replacement), and maintenance costs. From this model, the Table 2. Overall production cost of hydrogen
internal rate of return (IRR), net present value at various rates Old style Modern
of return (NPV), net cash flow, and a generated income state- Plant capacity, MMSCFD (contained) 90 90
Capital cost, million US$ - 55
ment can also be developed.
Average H2 production cost, US$ per 1000 1.996 1.602
standard ft3
Overall production costs Average H2 production cost, US$ million per year 65.6 52.6
Average annual production cost savings, US$ - 12.9
comparisons million
Building a new hydrogen plant is typically not the most appeal-
ing alternative to refiners. A new hydrogen plant requires a sig- design of a typical new plant. Table 1 shows the utilities and
nificant capital investment, and although hydrogen is required utility cost of hydrogen for each plant.
to support many of the refinery unit operations, it is generally Table 1 clearly illustrates a lower total utility cost for pro-
not viewed as a direct money maker. However, once all fac- ducing hydrogen in the modern plant than in the old style plant.
tors are taken into account and a total production cost of The modern plant produces hydrogen at a rate of US$ 1.409
hydrogen over the life of the plant is determined, the best eco- per 1000 standard ft3 of contained hydrogen, while the old style
nomic solution may be replacing an old style hydrogen plant plant produces at a rate of US$ 1.908 per 1000 standard ft3 of
with a new modern hydrogen plant. The following evaluation contained hydrogen. For a plant producing 90 MMSCFD of
illustrates this potential. contained hydrogen, this difference results in an annual utility
savings for a modern plant of US$ 16.4 million. Of course, the
Evaluation basis
utility cost alone does not complete the picture of the overall
To demonstrate the economics of building a modern hydrogen
production cost of hydrogen. A number of other economic and
plant versus continuing to operate an old style plant, two rep-
operating factors must also be evaluated.
resentative plants, each producing 90 MMSCFD of contained
hydrogen from a natural gas feed, will be compared, standard Capital cost
refers to conditions of 60 F and 14.7 psig. Natural gas will also Capital must be invested to build the new plant in order to cap-
be used for fuel, and both plants will export 600 psig super- ture the utility savings of the modern plant. A total capital
heated steam. The old style plant will consist of the major pro- investment of about US$ 55 million would be expected for a
cessing units described above and will produce hydrogen with new typical 90 MMSCFD hydrogen plant. This approximate
a purity of 95%. Other parameters for the old style plant will be installed cost is based on inside battery limits, natural gas
based on typical observed values. The modern plant will con- feed, no compression, no buildings, no water treatment units,
sist of the major processing units described above and will pro- no SCRs or MCCs, industry engineering standards, and deliv-
duce hydrogen with a purity of 99.99%. The modern plant ery to the US Gulf Coast. For purposes of comparison, no cap-
design will be based on producing maximum export steam. ital investment is considered for the old style plant.
This evaluation could be done based on a variety of other
plant configurations, but for demonstration purposes, this eval- Life of the plant economics
uation is limited to the plant types described. For comparison A cash flow economic model can be generated using the util-
purposes, the cost of utilities will be based on the following: ity costs developed for each plant, the capital cost required for

Natural gas: US$ 4 per million Btu LHV. the modern plant, as well as a number of other economic fac-

HP steam: US$ 5 per 1000 lbs. tors. These other economic factors, along with their associated
values, are listed below:

Boiler feedwater: US$ 0.5 per 1000 lbs.

Power: US$ 0.05 per kWh.
New plant construction length: two years.

Cooling water: US$ 0.1 per 1000 gal.
Escalation rate 1.5% for all feed, product and utilities.

Labour: two operators per shift.
Utilities
Overhead: 50% of labour.
As previously discussed, the utility costs of a hydrogen plant
Maintenance: 2% of plant cost per year.
are among the most important economic factors in determin-

Miscellaneous: 1% of plant cost per year.
ing the overall production cost of hydrogen. Simulation models

Catalyst costs accrued in year of change out.
are used for both the old style and modern plants to calculate
utility costs. The old style plant utilities are based on a simula-
Reformer tube replacement: every 10 years.
tion model built to reflect typical plant performance. The mod-
Onstream time: 98.5%.
ern plant utilities are based on a simulation model for the
Working capital: 45 days.

Reprinted from HYDROCARBON ENGINEERING FEBRUARY 2004

67-70 16/03/2004 11:58 am Page 70
Figure 4. H2 production cost versus natural gas
price.
Figure 5. H2 production cost versus export steam
credit.

Debt level: borrow 75% for new plant.

Cost of capital: 8%.

Debt length: seven years.

Depreciation life (of capital cost): 10 years.

Tax rate: 34%.

Project life: 25 years (includes two years of
construction).

Internal rate of return: 0% to obtain actual cost of
hydrogen production.
This resulting average overall production cost of
hydrogen, calculated for the life of the plant, is shown in
Table 2.
The evaluation shows that taking all relevant eco-
nomic factors into consideration, the overall production
cost of hydrogen is lower for the modern plant. The mod-
ern plant produces hydrogen at a rate of US$ 1.602 per
1000 standard ft3 of contained hydrogen while the old
style plant produces hydrogen at US$ 1.996 per 1000
standard ft3 of contained hydrogen. This lower overall pro-
duction cost results in an annual savings for the modern
plant of approximately US$ 12.9 million. This evaluation
demonstrates that for the cases analysed, it is economi-
cally feasible and potentially advantageous to build a new,
more efficient hydrogen plant.
Sensitivity to economic variables
Economic parameters for each refinery are different. The
major parameters that can significantly alter these results
are the efficiency of the existing plant, the feed and fuel
price, and the export steam credit. The efficiency of the
existing plant can span a wide range and should be
assessed for each plant independently. For the remaining
evaluations, the old style and modern plant utilities will be
held constant. The overall production cost of hydrogen will
be analysed as a function of the other two major factors,
feed and fuel price and export steam credit.
Feed and fuel price
The cost of feed and fuel is typically the largest compo-
Reprinted from HYDROCARBON ENGINEERING FEBRUARY 2004
nent of the overall production cost of hydrogen. Feed and
fuel usually account for more than 80% of the total before
the steam credit is taken. The overall operating cost
changes significantly as the natural gas price varies.
Figure 4 shows the effect of varying the natural gas price.
For this evaluation, the export steam credit to natural gas
price ratio was held constant.
As the price of natural gas increases, the modern
plant becomes more favourable. This is due to the over-
all feed and fuel efficiency advantage of the modern
plant. For example, if the natural gas price were
changed from US$ 4 to US$ 8 per million Btu (double the
base case credit), the overall production cost would
increase by US$ 1.326 per 1000 standard ft3 of hydro-
gen for the modern plant and US$ 1.804 for the old style
plant. The higher natural gas price would increase the
average annual savings for the modern plant from US$
12.9 million to US$ 28.6 million.
Export steam credit
The export steam credit also has a significant effect on
the overall production cost. The value a refinery places on
steam depends on utility factors and the existing steam
balance in the refinery. For example, during the winter,
steam tracing is generally used more heavily and the
value of steam could be higher than average. Conversely,
in the summer, when less steam tracing may be utilised,
steam may have a lower than average value. Modern
hydrogen plants typically export much more steam due to
the fact that they are more efficient and do not have a CO2
removal regeneration requirement. Figure 5 shows the
effect of varying the export steam credit.
As the export steam credit increases, the economics
for the modern plant become more favourable. For exam-
ple, if the export steam credit were changed from US$ 5
to US$ 10 per 1000 lbs (double the base case credit), the
overall production cost would drop by US$ 0.45 per
1000 ft3 of hydrogen for the modern plant and US$ 0.1 for
the old style plant. The higher steam credit would
increase the average annual savings for the modern plant
from US$ 12.9 million to US$ 24.4 million. For this evalu-
ation, the price of natural gas stays the same.
Conclusion
Refiners have a number of different options to consider in
addressing the demand for additional hydrogen. With the
cost of utilities and particularly natural gas on the rise,
overall plant efficiency has become a key factor in defin-
ing the economics of what option to pursue. A compre-
hensive technical and economic evaluation of existing
operations is required to determine the optimum solution
for each specific application.
The optimum solution for each refinery will be different
and will depend on such things as the refinery steam bal-
ance, the cost and availability of utilities, plot limitations,
and the condition of any existing hydrogen plants. Todays
modern hydrogen plants take advantage of numerous
technological improvements and offer a much lower over-
all life cycle cost. Once technical and economic evalua-
tions are done, it should not be a surprise to anyone that
the most economically attractive and feasible approach
may be to build a new hydrogen plant.
References
1. TINDALL, B.M., and CREWS, M.A., Economics of export steam for
hydrogen plants, Hydrocarbon Engineering, February 2003, pg 39.