Originally appeared in HYDROCARBON PROCESSING®

SEPTEMBER 2014 issue, pgs 71–75. Posted with permission.

Special Report Refining Developments

S. RATAN, Technip, Stone and Webster Process Technology,
Claremont, California; S. FARNAND and J. LI, Air Products
and Chemicals, Allentown, Pennsylvania

Hydrogen perspectives for 21st century refineries

The growth market for transportation fuels is shifting from Despite its high operating and capital costs, H2 is no longer
mature economies to developing and emerging nations, partic- just a critical utility for refinery operations; instead, it is gaining
ularly in China and India. These developing regions, however, status as a valuable asset in the refining process.
often suffer from limited domestic oil and gas availability, high Several proven options can be used to reduce capital invest-
energy costs, questionable power grid reliability, and potential ment and to improve the thermal efficiency of both new and
water scarcity. Today, highly complex deep-conversion refineries existing H2 facilities. Such options include advanced H2 man-
face the challenging task of improving hydro-processing intensity agement, along with the integrated utilization of refinery offgas
and residue upgrading while processing increasingly lower-quali- (ROG), enhanced energy efficiency, increased economies of
ty crudes under more stringent fuel specifications. Therefore, it is scale, augmentation of existing H2 capacity, and strategic over-
exceedingly important for 21st century refiners to manage hydro- the-fence (third-party) H2 supply. Furthermore, most of these
gen (H2 ), power, water and carbon effectively and economically. options provide added benefits of improved availability and re-
As refining market needs and trends continue to support in- duced environmental impact.
creased demand for H2 , its availability and cost are increasingly
critical for many refiners, creating a challenging environment ADVANCED HYDROGEN MANAGEMENT
for securing reliable, efficient and environmentally friendly H2 . H2 is typically supplied and balanced in a refinery through
Still, there are some proven solutions for reducing the net unit a network fed by H2 recovered from offgas streams and mostly
cost of (on-purpose) H2 , including advanced H2 management, supplemented by on-purpose H2 generation sources. Creative
capacity revamps and over-the-fence H2 supply, as well as some solutions are being developed to enhance refining profitabil-
value-added options and case studies demonstrating ways to ity not only through optimized process integration, but also
enhance refining economics. through advanced and smart H2 management methodologies
to optimize refinery H2 networks. The traditional H2 pinch
CHANGING REFINERY LANDSCAPE analysis (HPA) technique is not sufficient to model the com-
The unprecedented development of abundant unconven- plexities and optimize a network design, especially in terms of
tional gas reserves in the US and other regions of the world is realistic capital expenditure (CAPEX), operational and health,
expected to have a substantial impact on energy independence safety, and environment (HSE) constraints.
and security, as well as on the refining and petrochemical sec- One such advanced refinery H2 management methodology,
tor, at least in the US, in the coming decades. Projections of which is based on linear programming (LP) using platform
sustained low-priced natural gas (NG) and high-priced oil re- independent models (PIMs) . Such software carries various
sulting in an “oil-gas price gap” support H2 higher usage inten- process and utility models for rigorous simulation and recon-
sity via H2 addition, rather than the carbon removal route. ciliation to identify an optimized overall H2 balance and its
Deeper hydroprocessing and bottom-of-the-barrel strategies network. It uses a real-life cost database, scaling indices and
are being implemented more extensively to increase the yield various objective functions for economic analysis and case as-
of premium clean-fuel slates via higher H2 usage (scf/bbl). The sessment for grassroots refineries as well as upgrading or ex-
economics of this approach are increasingly attractive when H2 pansion of existing refineries (FIG. 1). It conducts sensitivity
is generated from relatively cheaper feed gas and are further analysis for optimizing the level of H2 recovery from ROG and,
improved by applying various optimization options and smart additionally, can recommend the appropriate steam-power sys-
concepts to lower the unit cost of H2 (i.e., the combined capital, tem along with the overall CO2 footprint based on a refinery-
operating and maintenance cost for a unit of H2 , or UCH). wide energy philosophy without compromising on safety, reli-
H2 is, and will continue to be, an essential element for the ability and operational flexibility.
refining industry, particularly in high-complexity deep-conver-
sion refineries. H2 constitutes a significant portion of refinery Case study. Advanced refinery H2 management methodol-
processing and operational costs, especially in high-growth ogy was applied to identify an optimized master plan for a large
economies. Consequently, there will continue to be strong in- grassroots deep-conversion refinery processing high-sulfur
centives to lower the UCH while enhancing reliability to im- crude. It was also applied to establish an optimum level of H2
prove refinery profitability. recovery based on the trade-off between the investment in the
Hydrocarbon Processing | SEPTEMBER 2014 71
 Refining Developments

Hydrogen balance Syngas distribution Default efficiency while satisfying captive power
Users Mass rate Nm3/h Sweet syngas distribution t/day Default split needs and minimized unit cost of H2 .
H2 for Naphtha HDT 6.39 t/d 2,985 Nm3/h Feeds % % %wt
H2 for Kero HDS 2.25 t/d 1,052 Nm3/h
H2 for Diesel HDS 17.12 t/d 7,994 Nm3/h
Sweet syngas to power
Sweet syngas to hydrogen
2,943.0 t/d
2,646.3 t/d
52.65
47.35
52.65
47.35
52.65%
47.35% REFINERY OFFGAS
237,297 Nm3/h
H2 for Hydrocracker
H2 for ARO Complex
508.22 t/d
17.00 t/d 7,936 Nm3/h
Sweet syngas to fuels
Total
0.0 t/d
5,589.3 t/d
0.00 0.00
100.00
0.00%
100.00% INTEGRATION
Total 551.0 t/d 257,264 Nm3/h To capture their full potential, there
Producers Mass rate Nm3/h
CCR-H2 (high purity) –151.93 t/d –70,937 Nm3/h
Refinery fuels balance has been an increasing trend to use ROG
H2 from gasification –241.96 t/d –112,977 Nm3/h Refinery energy balance for on-purpose H2 generation. The un-
–38.86 t/d –18,146 Nm3/h Refinery fuels availability MMKcal/kg 503.24
H2 from gasification
Hydrogen generation unit –118.23 t/d –55,204 Nm3/h Refinery fuels demand MMKcal/kg 755.84 derlying driver is that H2 burned as fuel
Total 551.0 t/d 257,264 Nm3/h
Delta (+import/-excess) MMKcal/kg 252.60 is a loss of its “asset” value above its heat-
Fuels import LHV
Hydrogen unbalance 0.0 t/d 3
0,0 Nm /h Feeds Kcal/Kg T/d MMKcal/hr
ing value. Accordingly, by integrating
(+/–; shortage/overproduction) Stop blink Natural gas 11,700.00 518.16 t/d 252.60 ROG in a H2 generation plant, not only
Fuel oil M 100 10,680.00 – –
Naphtha warning Fuels export LHV Others
does most of its H2 content get recov-
No shortage Feeds Kcal/kg t/d MMKcal/hr ered, but the price attached to the ROG
Refinery fuel gas (flare)
Fuel oil M 100
12,464.55 0.00 t/d
10,680.00 0.00 t/d
0.00
0.00 is often lower than that of hydrocarbon
feedstocks for H2 production.
FIG. 1. Select optimized H2 network case output display. There are enough financial incentives
to identify potential ROG streams in the
Makeup fuel Purge gas fuel refinery that can be integrated cost-effectively with the H2
plant to enhance its economics. Utilization of ROG through
HC feedstock Reforming Hydrogen integration with the H2 plant can be mostly accomplished by
PSA
+ shift three routes (FIG. 2):
1 •  Low contribution—H2 recovery by mixing with the
2 3 process gas upstream of the H2 generation PSA
Pretreatment Recovery •  Medium contribution—Direct use as (part) feedstock
PSA
for H2 generation
To RFG net •  High contribution—Dedicated recovery PSA with
Optional optional extended integration of its purge gas as (part)
feed for H2 generation.
ROG The typical reduction in net H2 costs can be between 2%
1 Low and 10% depending upon the relative pricing of ROG vs. the
Potential H2 contribution 2 Medium base feed, available quantity of ROG stream(s), H2 (or hydro-
(quantity  H2 fraction)
3 High
carbon) fraction, available pressure and level of impurities.
To illustrate this concept, a case study was undertaken using
FIG. 2. Routes for ROG integration with H2 plant. ROG (with 40 vol% H2 ) as the primary feed for H2 generation
based on medium level contribution. Comparative econom-
ics are presented in TABLE 1. Though ROG pricing was similar
TABLE 1. ROG integration as SMR feed (45 MMscfd H2 ) to NG in terms of its heating value basis, the operating cost
Case 1 Case 2 benefit from the potential H2 contribution was substantial
ROG as partial feed NG feed only (approximately 7%). The payout of the additional investment
Operating cost savings , US$/hr 1
188 Base was less than three years, without downsizing the reformer and
downstream section, which were still sized for NG feed as the
Additional investment , MM, $ 2
4.1 Base
controlling design case.
Payout, years 2.5 Base If NG compression was required, the payout for ROG inte-
1
Pricing data: ROG: 260 $/t (avg. 40 vol% H ); natural gas: 180 $/t; steam: 8 $/t;
and power: 60 $/MW
2 gration would have been even shorter. Other inherent benefits
2
For compression and pretreatment of operating with a ROG feed mix included relaxation on re-
cycle H2 , easier startup and longer reformer tube life. Various
pressure swing adsorption (PSA) recovery system and the po- ROG integration schemes have been implemented and proven,
tential savings from recovered H2 . ultimately providing better refinery margins.
Based on the optimized H2 network configuration and the
objective function, the program identified the most cost-ef-
fective H2 recovery level of about 68%, beyond which it was ENHANCED ENERGY EFFICIENCY
not economical in terms of incremental capital payoff against AND DESIGN OPTIMIZATION
extra H2 recovery credit. The integrated H2 network, with On-purpose H2 generation plants are capital intensive due
dual purity headers and minimized losses, resulted in lowering to high-temperature catalytic processing and necessary gas-
the on-purpose H2 generation capacity by 30% compared to phase purification. The total investment can vary considerably
the base case without the advanced refinery H2 management depending upon site-specific factors such as location, feed-
methodology. The program also conducted the overall fuel- stock, export steam conditions, degree of utility integration
steam-power balance together with CO2 loads and related C- and reliability needs.
72 SEPTEMBER 2014 | HydrocarbonProcessing.com
 Refining Developments

TABLE 2. H2 flowsheet initial optimization guidelines based on relative price ratios

Higher reformer outlet Lower steam/carbon ratio, Level of CO conversion, Enhanced PSA H2 recovery,
temperature, > 1,600° F < 2.8 > high temperature shift > 86%–89%
OPEX
Feed/fuel price > 1.1 <= 1.0 > 1.2* > 1.1
Steam/fuel price > 1.0 > 0.9 > 1.2 N/A
Steam/feed price N/A >1.2 N/A N/A
CAPEX >>  <  > Medium temperature shift > 
>> Low temperature shift
* Relevant for steam/carbon ratios > 2.7

The energy-related costs become more important for

Hot utility
optimizing larger H2 plant designs and for lowering the hy-
drogen unit cost (UCH). Accordingly, it allows higher incre- Hot composite curve QH min
mental capital investment payoffs for applying enhanced and
Pinch
advanced heat recovery below the heat pinch by extending the
so-called “cold composite,” as shown in FIG. 3.
A typical H2 plant energy balance is illustrated in FIG. 4.
The main thrust for improving the thermal efficiency lies in Temperature
reducing eventual heat loss through the flue gas to stack and
Process to process
through process cooling to cold utility (air or cooling water). heat recovery potential
Energy efficiency optimization based on operating ex- DT min
penses (OPEX)/CAPEX evaluations depends upon the H2
QC min
generation unit (HGU) capacity and related steam reformer Cold composite curve
(SMR) size. It is further sensitive to whether application of Cold utility
pre-reforming is required (for feed flexibility) or is optional.
Based on case-specific data, a sensitivity analysis was con- Enthalpy
ducted for key variables such as feed/fuel price ratio and fuel/
FIG. 3. Enhanced heat recovery below pinch.
steam price ratio to optimize OPEX against incremental CA-
PEX investment while keeping overall economics within the
given evaluation criteria. The results, presented in TABLE 2, can Flue gas to stack
Export steam
be summarized as follows:
1.  When feed is more expensive than fuel, which is often 0.1 – 0.15 0.2 – 1
the case with liquid feeds, it calls for higher feed con-
version and H2 yields by increasing SMR outlet tem- Feed 3 – 3.2
H2 plant
perature and/or increasing the steam/carbon (S/C) 2.26 H2 product
ratio, extended shift conversion level and higher PSA Fuel 0.5 – 1.2
recovery in order to lower the OPEX. For CAPEX,
however, it increases appreciably when raising the Extending the Cooling
~0 0.5 – 0.75
SMR outlet temperature beyond a certain level (cho- cold composite
DM water
sen as 1,600°F for the study). It may be more economic
if steam credit is high relative to fuel value in view of
reduced radiant efficiency. FIG. 4. Energy balance of a H2 plant, where 1 Gcal/kNm3 = 104.31 Btu/scf.
2.  When feed and fuel are similarly priced, increasing re-
forming severity in terms of lower S/C ratios and high- HIGHER ECONOMIES OF SCALE
er SMR outlet temperatures can improve OPEX, but For larger H2 plants, variable costs linked to specific energy
CAPEX must be carefully considered. To lower S/C consumption costs tend to govern the unit cost of H2 . TABLE 3
ratios below 2.7, a medium temperature (MT) shift is provides an overview of the sensitivity of variable and fixed
necessary, since HT shift is restricted due to concerns costs to plant capacity. The unit cost of H2 (UCH) is the sum
of over-reduction. CAPEX can be slightly reduced with of variable and fixed costs.
lower S/C ratios based on reduced heat recovery load A 1% reduction in energy costs in a 100-MMscfd H2 plant
downstream of the reformer. Having higher credit for can result in approximately $750,000 savings per year based
export steam further supports lowering of S/C ratios. on $5/MMBtu natural gas. For larger H2 plants, the focus is
3.  Additional CAPEX is necessary to increase PSA recov- more on efficiency optimization and there is a bigger incen-
ery beyond 86%–89%, and therefore should be evaluat- tive for incremental investment in terms of extended heat re-
ed on a case-to-case basis: The economics get favorable covery and flowsheet optimization based on a typical payback
when feed is more expensive than fuel. period of two to four years.
Hydrocarbon Processing | SEPTEMBER 2014 73
 Refining Developments

While CAPEX is relatively less critical for larger plants, its equipment and the level of modifications required. Normally, a
reduction becomes vital for lowering the UCH. The UCH pro- capacity revamp not only provides additional H2 at lower cost,
duced by a single-train large-capacity H2 plant can be apprecia- but also offers the benefits of shorter time schedules because
bly lowered by economies of scale. FIG. 5 illustrates the UCH of already existing interfacing-facilities/offsites, and can also
reduction from a 50 MMscfd to a 200 MMscfd H2 plant based provide feedstock change or flexibility, improved efficiency and
on US Gulf Coast economics with $5/MMBtu NG price. For environmental performance.
higher energy pricing, as observed in Europe and Asia, the ben- There are various proven options to cost-effectively augment
efits of economy of scale continue to hold true. H2 capacity up to 30%; in each case, however, bottlenecks must
Though economies of scale favor larger plants, there is a be identified and a proper assessment conducted to select the
capacity limit above which a single train plant starts becom- most appropriate option.
ing cumbersome and requires detailed evaluation to establish If the target capacity increase is substantial (> 15%) with
the breakpoint for two or more trains. Physical size, weight and major limitations on the reforming section, an effective solu-
transportable limits on the equipment, valves and piping, and tion without overloading the reformer is regenerative reform-
construction facilities must be considered. Such limits have pro- ing. The underlying concept is to use the reformed gas’ high-
gressively increased from 100 MMscfd up to a recent project level heat to reform additional feed through convective heat
size of 220 MMscfd H2 based upon compacting equipment, ad- exchange, also known as post-reforming. This option also pro-
vanced equipment design, piping modeling and modular con- vides a 10%–15% reduction in CO2 , NOx and SOx emission
struction concepts. levels per unit of H2 .
With additional H2 capacity range between 2 MMscfd and
CAPACITY REVAMP OF EXISTING H2 PLANTS 25 MMscfd, a regenerative reforming retrofit investment largely
Refiners are often faced with a H2 shortfall when addressing depends upon the percent increase in H2 capacity, design con-
changes in crude mix against the market-based clean fuels prod- ditions, available design margins and the level of modifications
uct slate. Such variations can be large enough to impact overall required in the existing plant. Generally, when compared with a
operation, H2 balance, and refinery profitability, but may not be new H2 plant for the same additional capacity, together with the
large enough to justify a new dedicated H2 plant. necessary offsite/utilities and auxiliaries, such investments can
Achieving additional H2 by revamping existing plants can be be economically attractive, as shown in TABLE 4.
an attractive alternative with lower UCH through cost-effective
retrofitting. Actual economics will depend upon the degree RELIABLE H2 OUTSOURCING
of uprate, available design margins, the condition of existing H2 is the lifeblood of modern refineries and is essential to the
production of cleaner-burning transportation fuels. When refin-
ers need H2 , they typically have two choices: buy a H2 plant de-
TABLE 3. H2 generation cost split sign license; pay other parties to build the plant; and then own,
Capacity Small Medium Large operate, and maintain the plant themselves (known as the make
MMscfd < 15 15-60 > 60 case) or purchase the H2 requirements from a third party (known
% Variable costs 40–60 50–70 60–80 as the buy case, sale of gas model or over-the-fence supply).
In the buy case, an industrial gas company designs and builds
% Fixed costs 40–60 30–50 20–40
the H2 plant with its capital and supplies H2 directly to the cus-
tomer over the course of a long-term contract. The buy case has
TABLE 4. H2 plant capacity revamp advantages that the refiner can benefit from: the gas company
Option Typical incremental H2, %1 Level of investment2
uses its H2 experience for the refiner, enabling the refiner to
focus on its core refining business; assumes responsibility for
Reformer upgrade 5–15 Medium
operational and maintenance activities; and can provide guar-
Regenerative reformer 15–30 Medium–High anteed on-stream reliability, availability, and efficiency levels.
integration Another way to secure high reliability of H2 supply is to ob-
1
Typical additional H2 between 2 MMscfd and 25 MMscfd tain H2 from a H2 pipeline. Pipeline supply can provide high re-
2
Based on the typical range of $0.5–$2 MM per MMscfd H2 depending upon
% increase and design-specific factors liability and, often, the lowest UCH due to economies of scale.
The various solutions described previously are all able to
105 satisfy the focal objective of lowering the net UCH. These op-
CAPEX + OPEX tions enhance reliability and HSE compliance while providing
100 refiners with improved economics and margins. It is imperative
to the success of 21st century refiners that they manage hydro-
Relative UCH, %

95
Basis: Natural gas @ $ 5/MMBtu gen efficiently and diligently.
7%
90
REFERENCES
85
1 Ratan, S., W. Baade and D. Wolfson, “The Large H2 Plant Challenge,” Hydrocarbon
80 Engineering, July 2005.
50 100 150 200 250 2 Ratan, S., N. Patel and W. Baade, “Driving H2 Plant Efficiencies with an Eye on
H2 plant capacity, MMscfd Environment,” Hydrocarbon Engineering, February 2010.
3 Ratan, S. and M. Pagano, “Refinery H2 management is more than a balancing act,”
FIG. 5. Economy of scale for H2 generation plant. AIChE Spring Meeting, 2011.

74 SEPTEMBER 2014 | HydrocarbonProcessing.com
 Refining Developments

SANJIV RATAN is group deputy vice president for the H2 product

line of Technip, Stone & Webster Process technology, based in
California. His responsibilities include techno-commercial
direction, technology development, competitiveness, strategy,
and business promotion. He has a chemical engineering degree
from the Indian Institute of Technology in Delhi, India.

SARAH FARNAND is the global segment manager for H2 and LNG in Air Products’
Tonnage Gases, Equipment and Energy Division. Her current responsibilities include
forecasting, coordinating competitive analysis, providing strategic analysis and
supporting the global business. She has an economics degree from the College of
William and Mary and an MBA degree from the University of Maryland.

JIMMY LI is the HyCO business technology manager for H2 and syngas in Air
Products’ Tonnage Gases, Equipment and Energy Division. His current responsibilities
include research and development to improve new product competitiveness. He has
a PhD degree in mechanical engineering from Georgia Tech and is a graduate of the
Wharton Management Program at the University of Pennsylvania.

Article copyright © 2014 by Gulf Publishing Company. All rights reserved. Printed in the US.
Not to be distributed in electronic or printed form, or posted on a website, without express written permission of copyright holder.

Hydrocarbon Processing | SEPTEMBER 2014 75