1
41623R04
Conceptual Design of Optimized Fossil Energy Systems with Capture
and Sequestration of Carbon Dioxide
Final Report
Reporting Start Date: September 22, 2002
Reporting End Date: August 31, 2004
Principal Author:
Dr. Joan M. Ogden
jmogden@ ucdavis. edu
Date Report Issued: September 30, 2004
DOE Award Number: DE- FC26- 02NT41623
Address of Submitting Organization:
Princeton Environmental Institute
27 Guyot Hall
Princeton University
Princeton, NJ 08544
2
DISCLAIMER
This report was prepared as an account of work sponsored by an agency of the United
States Government. Neither the United States Government nor any agency thereof, nor
any of their employees, makes any warranty, express or implied, or assumes any legal
liability or responsibility for the accuracy, completeness, or usefulness of any
information, apparatus, product, or process disclosed, or represents that its use would not
infringe rights. Reference herein to any specific commercial product, process, or service
by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or
imply its endorsement, recommendation, or favoring by the United States Government or
any agency thereof. The views and opinions of authors expressed herein do not
necessarily state or reflect those of the United States Government or any agency thereof.
3
Conceptual Design of Optimized Fossil Energy Systems with Capture
and Sequestration of Carbon Dioxide
Dr. Joan M. Ogden, Dr. Christopher Yang,
Nils Johnson, Jason Ni
Institute of Transportation Studies
University of California
Davis, CA 95616
Joshua Johnson
Information Center for the Environment
Department of Environmental Science and Policy
University of California
Davis, CA 95616
ABSTRACT
In this final progress report, we describe research results from Phase I of a
technical/ economic study of fossil hydrogen energy systems with CO2 sequestration. This
work was performed under NETL Award No. DE- FC26- 02NT41623, during the period
September 2002 through August 2004.
The primary objective of the study is to better understand system design issues
and economics for a large- scale fossil energy system co- producing H2 and electricity with
CO2 sequestration. This is accomplished by developing analytic and simulation methods
for studying the entire system in an integrated way. We examine the relationships among
the different parts of a hydrogen energy system, and identify which variables are the most
important in determining both the disposal cost of CO2 and the delivered cost of H2.
A second objective is to examine possible transition strategies from today’s
energy system toward one based on fossil- derived H2 and electricity with CO2
sequestration. We carried out a geographically specific case study of development of a
fossil H2 system with CO2 sequestration, for the Midwestern United States, where there is
presently substantial coal conversion capacity in place, coal resources are plentiful and
potential sequestration sites in deep saline aquifers are widespread.
4
TABLE OF CONTENTS
LIST OF FIGURES ……………………………………………………………………... 6
LIST OF TABLES …………………………………………………………………….... 8
EXECUTIVE SUMMARY ……………………………………………………………... 9
1.0 INTRODUCTION................................................................................. 12
1.1. Background and Motivation............................................................................. 12
1.2. Scope of this Study........................................................................................... 14
1.2.1. Task 1.0 Implement Technical and Economic Models of the
System Components ................................................................................. 15
1.2.2. Task 2.0. Integrated Studies of the Entire System to Find the
Lowest Cost Network............................................................................... 16
1.2.3. Task 3.0 Case Study of Transition to a Fossil Energy System with
CO2 Sequestration .................................................................................... 16
2.0 RESULTS AND DISCUSSION............................................................ 18
2.1. Task 1.0. Implement Technical And Economic Models Of The System
Components ..................................................................................................... 18
2.1.1. Task 1.1. Modeling the Fossil Energy Complex ...................................... 18
2.1.2. Task 1.2. Modeling CO2 Compression and Pipeline Transport ................ 25
2.1.3. Task 1.3. Modeling CO2 Sequestration sites............................................ 29
2.1.4. Task 1.4. Modeling H2 Demand Centers ................................................. 30
2.1.5. Task 1.5. Modeling H2 Delivery Infrastructure........................................ 32
2.1.5.1. Modeling Hydrogen Distribution System Components.................... 33
2.1.5.2. A Comparison “ Point- to- Point” Hydrogen Delivery Costs.............. 35
2.1.5.3. Designing a Local hydrogen distribution network ........................... 37
2.1.5.4. Hydrogen Refueling Stations............................................................ 41
2.1.5.5. Summary of Component Costs and Performance for Fossil
Hydrogen Energy System with CO2 Sequestration.......................... 43
2.2. Task 2.0. Integrated Studies of the Entire System to Find the Lowest
Cost Options .................................................................................................... 46
2.2.1. Task 2.1. Develop Simple Model for Entire System and Perform
Sensitivity Studies .................................................................................... 46
2.2.1.1. An Integrated Hydrogen System Model ........................................... 46
2.2.1.2. Preliminary Results........................................................................... 48
2.2.2. Task 2.2 Explore Use of Mathematical Programming Techniques
to Study More Complex Systems............................................................. 50
2.3. Task 3.0 Case Study of Transition to a Fossil Energy System with CO2
Sequestration.................................................................................................... 53
2.3.1. Task Overview ......................................................................................... 53
2.3.2. Estimating Hydrogen demand .................................................................. 56
2.3.2.1. Methodology..................................................................................... 56
2.3.2.2. Sensitivity Analysis .......................................................................... 63
5
2.3.3. Infrastructure Components ....................................................................... 68
2.3.4. Infrastructure Optimization ...................................................................... 70
2.3.5. Early Results for Infrastructure Design and Delivered Cost.................... 76
3.0 CONCLUSION..................................................................................... 79
3.1. Task 1.0 Implement Technical and Economic Models of the System
Components ..................................................................................................... 79
3.2. Task 2.0. Integrated Studies of the Entire System to Find the Lowest
Cost Network ................................................................................................... 79
3.3. Task 3.0 Case Study of Transition to a Fossil Energy System with CO2
Sequestration.................................................................................................... 79
4.0 FUTURE WORK .................................................................................. 81
4.1. Task 1. Improve and Extend Models of Fossil Hydrogen Energy
Systems with Carbon Capture and Sequestration ............................................ 81
4.2. Task 2. Understand The Implications Of New Carbon Capture And
Sequestration Technologies For Widespread Use Of Fossil Hydrogen
As An Energy Carrier ...................................................................................... 82
4.3. Task 3. Carry out a series of regional case studies of a transition to
fossil hydrogen energy systems with CO2 capture and sequestration............. 82
5.0 REFERENCES...................................................................................... 84
6.0 Compressed H2 gas truck ( 1/ day).......................................................... 94
7.0 Liquid H2 truck ( 1/ day)......................................................................... 94
8.0 Onsite electrolyzer................................................................................. 94
9.0 Onsite steam methane reformer ( SMR)................................................ 94
LIST OF ACRONYMS AND ABBREVIATIONS……………………………………. 88
APPENDICES …………………………………………………………………………. 90
APPENDIX 0. CONVERSION FACTORS AND ECONOMIC ASSUMPTIONS........ 91
APPENDIX 0. CONVERSION FACTORS AND ECONOMIC ASSUMPTIONS........ 91
APPENDIX A. MODELING THE FOSSIL ENERGY COMPLEX............................... 95
APPENDIX B. CO2 COMPRESSION AT THE FOSSIL ENERGY COMPLEX ........ 106
APPENDIX C. CO2 PIPELINE CALCULATIONS...................................................... 112
APPENDIX D. INJECTION RATE INTO UNDERGROUND RESERVOIRS,
CALCULATIONS FOR INJECTION SITE COSTS ......................... 121
APPENDIX E. HYDROGEN FUEL DELIVERY INFRASTRUCTURE:
HYDROGEN COMPRESSION, STORAGE, PIPELINE
TRANSMISSION, LOCAL PIPELINE DISTRIBUTION AND
REFUELING STATIONS................................................................... 129
APPENDIX F. An Integrated Hydrogen System Model............................................... 158
APPENDIX G. LITERATURE REVIEW OF MATHEMATICAL
PROGRAMMING METHODS APPLIED TO PIPELINE
SYSTEM DESIGN.............................................................................. 180
APPENDIX H. GIS DATA SOURCES USED IN THIS STUDY ................................ 183
APPENDIX I GEOGRAPHIC INFORMATION SYSTEMS ( GIS)
DEFINITIONS .................................................................................... 189
6
LIST OF FIGURES
Figure 1.1.1. A Fossil Energy System for Production of Hydrogen and
Electricity with CO2 Sequestration. ( Variables for the Study are
Shown in Italics).................................................................................... 13
Figure 1.1.2. Schematic of More Complex Hydrogen System................................... 14
Figure 2.1.1. Hydrogen Production from Natural Gas with and without CO2
Capture .................................................................................................. 19
Figure 2.1.2. Production of Electricity and H2 from Coal with CO2 Capture ............ 20
Figure 2.1.3. Cost of Hydrogen Production from Coal and Natural Gas with
CO2 Separation and Compression versus Hydrogen Plant Size............ 21
Figure 2.1.4. Installed Capital Cost of CO2 Pipelines ................................................ 26
Figure 2.1.5. Levelized Cost of Pipeline Transmission ($/ tonne CO2) vs.
Pipeline Length and Flow Rate ............................................................. 27
Figure 2.1.6. Levelized Cost of CO2 Pipeline for Coal- Based H2 Plant ($/ GJ
H2 HHV) vs. Pipeline Length and CO2 Flow Rate................................ 28
Figure 2.1.7. Levelized Cost of CO2 Pipeline ($/ GJ H2 HHV) for Natural Gas
to H2 Plant vs. Length and CO2 Flow Rate ........................................... 28
Figure 2.1.8. Mapping Hydrogen Demand Density.................................................... 31
Figure 2.1.9. Hydrogen Demand Density in Ohio...................................................... 32
Figure 2.1.10. Levelized Cost of Hydrogen Pipeline Transmission ( including
compression, storage, and pipeline) vs. Pipeline Length and
Energy Flow Rate ( MWth).................................................................... 35
Figure 2.1.11. Minimum cost delivery mode for a range of operating conditions
( P – pipeline, G – compressed gas trucks, L – liquid trucks)................ 36
Figure 2.1.12. Graph of minimum cost for the three modes of hydrogen delivery
as a function of flowrate and transport distance.................................... 37
Figure 2.1.13. Idealized city model with 25 and 125 hydrogen stations
distributed in rings throughout the city. ................................................ 38
Figure 2.1.14. Tradeoff between convenience and delivery network distance for
pipelines and trucks for different numbers and configurations of
stations. (“ P” denotes pipeline distribution, “ T” truck
distribution.) .......................................................................................... 39
Figure 2.1.15. The relationship between the number of stations within the city
and the total delivery distance for pipelines and trucks. ....................... 40
Figure 2.2.1. Capital Cost $/ LDV for H2 Infrastructure vs. Fraction of H2
Vehicles ................................................................................................. 50
Figure 2.2.2. Delivered Cost of H2 ($/ kg) vs. Fraction of H2 vehicles in Fleet
for City of One Million People.............................................................. 51
Figure 2.3.1. GIS modeling flowchart ........................................................................ 55
Figure 2.3.2. Hydrogen demand density given different density thresholds in
Columbus, Ohio..................................................................................... 57
Figure 2.3.3. Demand clusters under different density thresholds in Columbus,
Ohio ....................................................................................................... 58
7
Figure 2.3.4. Demand clusters and associated aggregate hydrogen demand.............. 60
Figure 2.3.5. Demand centers with 10% market penetration...................................... 61
Figure 2.3.6. Demand centers with 50% market penetration...................................... 62
Figure 2.3.7. Number of hydrogen demand centers ................................................... 64
Figure 2.3.8. Percent of statewide hydrogen demand captured .................................. 65
Figure 2.3.9. Percent of statewide land area captured ................................................ 66
Figure 2.3.10. Spatial distribution of demand centers given the three threshold
scenarios ................................................................................................ 67
Figure 2.3.11. Conceptual Network Structure .............................................................. 71
Figure 2.3.12. Nodes and paths for the hydrogen distribution infrastructure
network including demand clusters, coal plants and potential
hydrogen pipeline locations .................................................................. 73
Figure 2.3.13. Layout of the minimum network length for one hydrogen
production plant at the 10% hydrogen vehicle market penetration
level ....................................................................................................... 75
Figure 2.3.14. Costs comparision for central and distributed hydrogen
production for the 10% and 50% market penetration levels. ................ 78
8
LIST OF TABLES
Table 2.1.1. Cost and Performance of Natural Gas Based Hydrogen
Production Plants w/ and w/ o CO2 Capture ( Foster Wheeler
1996)...................................................................................................... 21
Table 2.1.2. Levelized cost of hydrogen production from natural gas with and
without CO2 separation and compression ............................................. 23
Table 2.1.3. Cost and Performance for Hydrogen and Electricity Production
from Coal ( 70 bar gasifier) ( Kreutz 2002) ........................................... 24
Table 2.1.4. CO2 Pipeline Transmission and Storage System for Base Case
H2 Plants Producing 1000 MW of hydrogen output from Natural
Gas and Coal ......................................................................................... 30
Table 2.1.5. Characteristics Of Hydrogen Refueling Stations .................................. 42
Table 2.1.6. Summary Economic Data for Large Central H2 Production
Systems as a Function of Scale ............................................................. 43
Table 2.1.7. Economic Data for Gaseous Hydrogen Pipeline Transmission
Systems as a Function of Scale ( including hydrogen
compression, large scale gaseous storage and transmission
pipeline)................................................................................................. 45
Table 2.2.1. Characteristics of City and Calculated Infrastructure ........................... 48
Table 2.2.2. Capital Costs for Hydrogen Infrastructure Options ( million $) ............ 49
Table 2.2.3. Objective Function Used in Various Pipeline Studies .......................... 52
Table 2.2.4. Mathematical Programming Methods Used in Various Studies to
Model Pipelines..................................................................................... 53
Table 2.3.1. Threshold values for each scenario ....................................................... 63
Table 2.3.2. Results for each threshold scenario....................................................... 63
Table 2.3.3. Data for utility coal plants over 100MW electricity output and
estimates for H2 capacity given complete coal conversion and
efficiency improvements. ...................................................................... 69
Table 2.3.4. Distance matrix for network optimization indicating distance
between demand clusters to other demand clusters and coal
plants ..................................................................................................... 71
Table 2.3.5. Decision table indication which pipelines are built for the
minimal spanning pipeline network for one coal plant source.............. 74
Table 2.3.6. 10% Demand cluster information including intracity distribution
pipeline network length ......................................................................... 76
Table 2.3.7. Details of final hydrogen infrastructure for 10% and 50% market
penetration levels................................................................................... 77
9
EXECUTIVE SUMMARY
In this final report, we present results from Phase I of a technical and economic
assessment of fossil H2 energy systems with CO2 sequestration. This work was performed
during the period September 2002- August 2004 under NETL Award No. DE- FC26-
02NT41623.
The primary objective of the study is to better understand system design issues and
economics for a large- scale fossil energy system co- producing hydrogen ( H2) and
electricity with carbon dioxide ( CO2) sequestration. This is accomplished by developing
new analytic and simulation tools for studying the entire system in an integrated way. We
examine the relationships among the various parts of a fossil hydrogen energy system, and
attempt to identify which variables are the most important in determining both the disposal
cost of CO2 and the delivered cost of H2.
A second objective is to examine possible transition strategies from today’s energy
system toward one based on fossil- derived H2 and electricity with CO2 sequestration. We
are carrying out a geographically specific case study of development of a fossil H2 system
with CO2 sequestration, for the Midwestern United States, where there is presently
substantial coal conversion capacity in place, coal resources are plentiful and potential
sequestration sites in deep saline aquifers are widespread.
We consider fossil energy complexes producing both H2 and electricity from either
natural gas or coal, with sequestration of CO2 in geological formations such as deep saline
aquifers. The design and economics of the system depend on a number of parameters that
determine the cost and performance of the system “ components”, as a function of scale and
geography ( components include: the fossil energy complex, H2 pipelines and refueling
stations, CO2 pipelines, CO2 sequestration sites, and H2 energy demand centers). If we
know the cost and performance characteristics of the components, designing the system can
be posed as a problem of cost minimization. The goal is to minimize the delivered H2 cost
with CO2 disposal by co- optimizing the design of the fossil energy conversion facility and
the CO2 disposal and H2 distribution networks. Research to perform this cost minimization
has two parts: 1) implement technical and economic models for each “ component” in the
system, and 2) develop optimization algorithms to size various the system components
and connect them via pipelines into the lowest cost network serving a particular energy
demand. Finally, to study transition issues, we use these system models to carry out a case
study of developing a large- scale fossil energy system in the Midwestern United States.
The research consisted of three tasks.
Task 1.0 Implement Technical and Economic Models of the System Components
We utilize data and component models of fossil energy complexes with H2 production, and
CO2 sequestration developed by the principal investigator as part of the Carbon Mitigation
10
Initiative ( CMI) at Princeton University. 1 Models for H2 distribution systems and refueling
stations were adapted from the principal investigator’s previous studies of H2 infrastructure
for the US Department of Energy Hydrogen R& D Program ( Ogden 1998, Ogden 1999a,
Ogden 1999b), studies at UC Davis under the Hydrogen Pathways program ( Yang and
Ogden 2004), and those of other researchers ( Mintz et al. 2003, Amos 1998, Thomas et al.
1998). During the past year the principal investigator worked with the “ H2A”, a group of
hydrogen analysts convened by the USDOE to develop cost and performance estimates for
hydrogen technologies. The H2A is developing an EXCEL- based spreadsheet database, for
hydrogen production, refueling and delivery systems. 2 In addition the National Academy
of Engineering recently released an assessment of the Hydrogen Economy, including data
on hydrogen technologies ( NAE 2004). In Phase II, we propose to update our models to
reflect the new information contained in these studies.
Task 2.0. Integrated Studies of the Entire System to Find the Lowest Cost Network
As a first step, we developed a simple analytical model linking the components of the
system. We consider single fossil energy complex connected to a single CO2 sequestration
site and a single H2 demand center. We developed “ cost functions” for the CO2 disposal
cost and the delivered H2 cost with explicit dependence on important input parameters ( e. g.
size of demand, fossil energy complex process design, aquifer physical characteristics,
distances, pressures etc.). Analytic sensitivity studies of this “ simple system” are used to
provide us with insights on which parameters are most important in determining costs.
As a next step, we extended this simple model, by designing the supply to meet a specified
level of demand. Results were derived for the cost of fossil hydrogen production with CO2
sequestration as a function of geographic factors ( geographic density of demand, location
of fossil energy complexes and sequestration sites), level of hydrogen use ( e. g. size of the
market, market penetration of hydrogen vehicles), and technology ( type of supply
technology, hydrogen vehicle fuel economy). We developed an idealized model of a city as
a basis for designing and costing hydrogen distribution infrastructure ( e. g. a hydrogen
pipeline network or truck delivery routes in cities).
To study more complex and realistic systems involving multiple energy complexes, H2
demand centers, and sequestration sites, we explored use mathematical programming
methods to find the lowest cost system design. From our system modeling, we seek to
distill “ rules for thumb” for developing H2 and CO2 infrastructures.
Task 3.0 Case Study of Transition to a Fossil Energy System with CO2 Sequestration
1 Begun in 2001, the Carbon Mitigation Initiative is a ten- year $ 15- 20 million dollar joint project of
Princeton University, BP and Ford Motor Company to find solutions to global warming and climate
change.
2 During the period February 2003- August 2004, the principal investigator took part in developing the
H2A database, and led the team looking at hydrogen delivery systems. The H2A spreadsheets should
become available in October of 2004, and we plan to include these results as part of Phase II.
11
In this task, the goal is to explore transition strategies: how H2 and CO2
infrastructures might develop in time, in the context of a geographically specific regional
case study. We focus on the Midwestern United States, a region where coal is widely used
today in coal- fired power plants, and good sites for CO2 sequestration are available. The
goal is to identify attractive transition strategies toward a regional hydrogen/ electricity
energy system in the Midwest with near zero emissions of CO2 and air pollutants to the
atmosphere.
To better visualize our results, we use a geographic information system ( GIS)
format to show the location of H2 demand, fossil energy complexes, coal resources,
existing infrastructure ( including rights of way), CO2 sequestration sites and the optimal
CO2 and H2 pipeline networks. We plan to coordinate with other ongoing GIS based
studies of CO2 sequestration potential such as the NATCARB project. Input from these
projects will be used to estimate the best options for sequestration. Optimization tools
available in the ARCView GIS software are used to identify the lowest cost pipeline
network for supplying hydrogen to users.
12
1.0 INTRODUCTION
In this final report, we present results from Phase I of a technical and economic
assessment of fossil H2 energy systems with CO2 sequestration. This research was
performed under NETL Award No. DE- FC26- 02NT41623, between September 2002
and August 2004.
1.1. Background and Motivation
Production of hydrogen from fossil sources with capture and sequestration of
CO2 offers a route toward near- zero emissions in production and use of fuels.
Implementing such an energy system on a large scale would require building two new
infrastructures: one for producing and delivering H2 to users ( such as vehicles) and one
for transmitting CO2 to disposal sites and securely sequestering it.
In Figure 1.1.1, we show a fossil hydrogen energy system with CO2
sequestration. A fossil feedstock ( natural gas or coal) is input to a fossil energy complex
producing hydrogen and electricity. CO2 is captured, compressed to supercritical
pressures for pipeline transport to a sequestration site, and injected into an aquifer or
other underground geological formation. Hydrogen is delivered to users via a pipeline
distribution system that includes compression and storage at the hydrogen production
plant, pipelines ( possibly with booster compressors) and hydrogen refueling stations.
The design and economics of a fossil H2 energy system with CO2 sequestration depend
on a host of factors, many of which are regionally specific and change over time.
( Variable considered in this study are shown in Figure 1.1.1 in italics.) These include:
􀂃 The size, type, location, time variation and geographic density of the H2 demands.
􀂃 Cost and performance of component technologies making up the system. Key
components are: the fossil energy conversion plant [ design variables include the
scale, feedstock: ( coal vs. natural gas), process design, electricity co- production,
separation technology, pressures and purity of H2 and CO2 products, sulfur removal
options including co- sequestration of sulfur compounds and CO2, location ( distance
from demand centers and sequestration sites)], H2 and CO2 pipelines and H2
refueling stations.
􀂃 The location and characteristics of the CO2 sequestration sites ( storage capacity,
permeability, reservoir thickness),
􀂃 Cost, location and availability of primary resources for H2 production.
􀂃 Location of existing energy infrastructure and rights of way ( that could be used for
siting hydrogen transmission pipelines).
For simplicity, in Figure 1.1.1, we have shown a single fossil energy complex,
serving a single demand, and one CO2 sequestration site. However, a future fossil
hydrogen system could be more complex, linking multiple H2 demand centers ( cities),
fossil energy complexes and sites for CO2 sequestration ( Figure 1.1.2).
13
Figure 1.1.1. A Fossil Energy System for Production of Hydrogen and Electricity
with CO2 Sequestration. ( Variables for the Study are Shown in
Italics)
Several detailed technical and economic studies have been carried out for various
parts of the system, including CO2 capture from electric power plants ( Hendriks 1994;
Foster Wheeler 1998; Simbeck 1999), or H2 plants ( Foster Wheeler 1996; Doctor et al.
1999; Spath and Amos 1999; Kreutz et al. 2002), CO2 transmission ( Skovholt 1993) and
storage ( Holloway 1996), and H2 infrastructure ( Directed Technologies et al. 1997,
Ogden 1999; Thomas et al. 1998, Mintz et al 2002). However, relatively little work has
been done assessing complete fossil hydrogen systems with CO2 sequestration in an
integrated way. An integrated viewpoint is important for understanding the design and
economics of these systems. For example, the scale of the fossil hydrogen plant can have
a large impact on the design and cost of both the hydrogen distribution system, and the
system for transporting and sequestering CO2.
scale, H2
purity, time
variation
Fossil
Complex
Fossil
Feedstock
( NG,
Petroleum
Residuals,
heavy oils,
tar sands,
Coal)
CO ( possibl
H2S or
co-
H2
CO2 Sequestration Site
well depth, reservoir
permeability, layer thickness,
pressure, capacity, CO2 purity
plant design,
scale, P, T, purity
of H2, CO2 distance
distance
Electricit
H2 End-pressure,
H2
14
More Complex System:
Optimization for Low Delivered H 2 Cost
What is the lowest cost system for producing and
delivering H2 to serve a growing demand ?
H2 Plant
Primary
Resource 1
CO2 Sequestration Site
H2 Demand
H2
CO2
• H2 Plants: Size and
Location?
• Resources for H2
production:
Characteristics, distance
from H 2 plant?
• Use existing energy
infrastructure/ rights of
way?
• Optimum paths for H 2
infrastructure over
time?
• Design problem is
different than typical oil
or gas pipeline systems
w. r. t time frame and
complexity
Primary
Resource 2
H2 Plant
Onsite H2
Plants
Figure 1.1.2. Schematic of More Complex Hydrogen System
1.2. Scope of this Study
The primary objective of this study is to better understand total system design
issues and economics for a large- scale fossil energy system co- producing hydrogen ( H2)
and electricity with CO2 sequestration. We consider fossil energy complexes producing
both H2 and electricity from either natural gas or coal, with sequestration of CO2 in
geological formations such as deep saline aquifers. We apply various analytic and
simulation methods to study the entire system in an integrated way. We attempt to
identify which variables are the most important in determining both the disposal cost of
CO2 and the delivered cost of H2. We examine the relationships among the system
components ( e. g. fossil energy complexes, H2 and CO2 pipelines, H2 demand centers,
and CO2 sequestration sites), and apply new simulation tools to studying these systems,
and optimizing their design.
A second objective is to examine possible transition strategies from today’s
energy system toward one based on fossil- derived H2 and electricity with CO2
sequestration. We focus on understanding how H2 and CO2 infrastructures might evolve
to meet a growing H2 demand under different regional conditions. If we know the
location, size, cost and performance characteristics of the system components, designing
15
the system can be posed as a problem of cost minimization. The goal is to minimize the
delivered H2 cost with CO2 disposal by co- optimizing the design of the fossil energy
conversion facility and the CO2 and H2 pipeline networks. Research to perform this cost
minimization has two parts: 1) implement technical and economic models for each
component in the system ( Task 1), and 2) explore use of optimization algorithms to size
various the system components and connect them via pipelines into the lowest cost
network serving a particular energy demand ( Task 2). Techniques for studying regional
H2 and CO2 infrastructure development and transition strategies are described, based on
use of Geographic Information System ( GIS) data and network optimization techniques.
To understand the impact of geographic factors, we carried out a case study of
development of a large scale fossil H2 system with CO2 sequestration, for the
Midwestern United States, where there is presently substantial coal conversion capacity
in place, coal resources are plentiful and potential sequestration sites in deep saline
aquifers are widespread ( Task 3).
Three tasks were completed. 3
1.2.1. Task 1.0 Implement Technical and Economic Models of the System
Components
Before developing a total system model, we developed technical/ economic
models for the various parts ( or “ components”) of the system. Performance and cost of
each “ component” of the system is characterized as a function of scale and other
relevant parameters. We utilize data and models of fossil energy complexes with H2
production, and CO2 sequestration developed as part of the Carbon Mitigation Initiative
( CMI) at Princeton University. 4 Models for H2 distribution systems and refueling
stations were adapted from the principal investigator’s previous studies of H2
infrastructure for the US Department of Energy ( Ogden 1998, Ogden 1999a, Ogden
1999b), work at UC Davis under the Hydrogen Pathways Program ( Yang and Ogden
2004), and those of other researchers ( Mintz et al. 2003, Amos 1999, Thomas et al.
1998, NAE 2004). 5
3 Results are given for each task in the “ Results and Discussion” section below. Earlier results were
described in previous progress reports for this contract ( Ogden 2003a, Ogden 2003b, Ogden 2003c).
4 Begun in 2001, the Carbon Mitigation Initiative is a ten- year $ 15- 20 million dollar joint project of
Princeton University, BP and Ford Motor Company to find solutions to global warming and climate
change.
5 During the past year the author worked with the “ H2A”, a group of hydrogen analysts convened by the
USDOE to develop cost and performance estimates for hydrogen technologies. The H2A data should
become available in October 2004. In addition the National Academy of Engineering recently released an
assessment of the Hydrogen Economy ( NAE 2004). In Phase II of this project, propose to update our
models to reflect the new information contained in these studies.
16
1.2.2. Task 2.0. Integrated Studies of the Entire System to Find the Lowest Cost
Network
As a first step, we developed a simple analytical model linking the components of
the system. We consider a single fossil energy complex connected to a single CO2
sequestration site and a single H2 demand center ( see Figure 1.1.1). We developed “ cost
functions” for the CO2 disposal cost and the delivered H2 cost with explicit dependence on
the many input parameters described above ( e. g. size of demand, fossil energy complex
process design, aquifer physical characteristics, distances, pressures etc.). Sensitivity
studies of this “ simple system” provided insights on which parameters are most important
in determining hydrogen costs.
Later, we expanded this simple model to include better models of hydrogen
demand and hydrogen distribution systems. To study more complex and realistic systems
involving multiple energy complexes, H2 demand centers, and sequestration sites, we
explored use mathematical programming methods to find the lowest cost system design.
To facilitate regionally specific case studies, we developed an interface between our cost
models and the Geographic Information System ( GIS) database developed in Task 3. This
allows us to make hydrogen system design and cost calculations based on quantities easily
derived from GIS maps.
Through system modeling, we seek to distill “ rules for thumb” for developing H2
and CO2 infrastructures.
1.2.3. Task 3.0 Case Study of Transition to a Fossil Energy System with CO2
Sequestration
In this task, we explore how H2 and CO2 infrastructures might develop, in the
context of a geographically specific regional case study. We focussed on the Midwestern
United States, a region where coal is widely used today in coal- fired power plants, and
good sites for CO2 sequestration are available. We consider how fossil energy systems
might develop over time to meet an evolving energy demand. The goal is to identify
attractive transition strategies toward a regional hydrogen/ electricity energy system in the
Midwest with near zero emissions of CO2 and air pollutants to the atmosphere.
To better visualize our results, use a geographic information system ( GIS) format to
show the location of H2 demand, fossil energy complexes, coal resources, existing
infrastructure ( including rights of way), CO2 sequestration sites and the optimal CO2 and
H2 pipeline networks. First, a survey of relevant GIS data sets was conducted, and a
database was built including hydrogen supply, demand and existing infrastructure.
Network optimization methods were combined with the “ Network analyst” capabilities of
GIS software ( ARCView) to find low cost hydrogen distribution networks. We used this
17
database to make preliminary design and cost studies of fossil energy systems with CO2
sequestration.
18
2.0 RESULTS AND DISCUSSION
2.1. Task 1.0. Implement Technical And Economic Models Of The System
Components
In this Task we implement technical/ economic models of various parts of a fossil
hydrogen system with CO2 sequestration. These include:
Task 1.1. The fossil energy complex for producing hydrogen and electricity from natural
gas or coal ( Appendix A)
Task 1.2. CO2 compression and pipeline transport ( Appendices B, C)
Task 1.3. CO2 injection into underground geological formations ( Appendix D)
Task 1.4. Hydrogen demand for vehicles
Task 1.5. Hydrogen fuel delivery infrastructure ( including hydrogen compression,
storage, pipeline transmission and refueling stations) ( Appendix E)
Key results from the technical/ economic models for each part of the system are
summarized below. 6
2.1.1. Task 1.1. Modeling the Fossil Energy Complex
In the fossil energy complex, a synthetic gas ( or syngas) is produced via gasification of
coal or steam reforming of methane. The syngas undergoes a water gas shift reaction to
increase the hydrogen content. CO2 is removed from the syngas using a separation
system ( such as an amine scrubber, a physical adsorption system like Selexol or a
pressure swing adsorption system or PSA) and is available at near atmospheric pressure.
CO2 is then compressed from capture pressure to a supercritical state and pumped to
pipeline transmission pressures of 15- 20 MPa ( 150- 200 bar). In some cases, electricity
is co- produced with hydrogen. Simplified diagrams of the processes for producing
hydrogen from natural gas and coal shown in Figure 2.1.1 and Figure 2.1.2.
As a basis for modeling natural gas- based hydrogen plants, we use a recent study by
Foster Wheeler ( 1996) and data from Air Products and Chemicals ( Ogden 1999). As
part of the CMI, researchers at Princeton have developed ASPEN- plus process and cost
models for a variety of coal- based systems co- producing H2 and electricity with CO2
capture ( Kreutz, Williams, Socolow and Chiesa 2002), that include alternative options
for sulfur removal and disposal. We use the results of these detailed process design
6 Base case economic assumptions are given in Appendix 0. Model details are given in Appendices A- E.
19
Figure 2.1.1. Hydrogen Production from Natural Gas with and without CO2
Capture
Synga
H2, CO,
CH4, H20
Steam
Natural Gas
Heat
Pure H2
Compressor
Waste
gases
Steam
Hydroge
Purification
Shift Reactor
CO+ H20-
Steam
Reformer
H2
Storage
to H2 Users
Steam
Natural Gas
Heat
Pure H2
Compressor
CO2
Steam
Hydroge
Purification/
CO2 Separation
Shift Reactor
CO+ H20-
Steam
Reformer
H2
Storage
to H2 Users
CO2 Vented CO2 Captured
Compressor
to CO2 pipeline
and sequestration
Synga
H2, CO,
CH4, H20
20
Figure 2.1.2. Production of Electricity and H2 from Coal with CO2 Capture
Syngas
H2, CO, CO2,
CH4, H20
Steam
Heat
Pure H2
Compressor
Steam
Hydrogen
Purification
Shift Reactor
CO+ H20-> CO2+ H2
Steam
Reformer
to H2 Users
Gasifier
Coal, Wastes or Biomass
Compressor
to CO2 pipeline
and sequestration
CO2
21
studies to produce a simplified model for the cost and performance of fossil H2 plants as
a function of scale, feedstock and process design. Summary costs for natural gas and
coal- based hydrogen production systems are given in Table 2.1.1 through Table 2.1.3.
For each coal- to- hydrogen case in Error! Reference source not found., the sizes,
capital costs and O& M costs of the various fossil energy plant components were
estimated, along with the energy consumption, hydrogen and electricity production, and
carbon emissions ( Kreutz 2002). From these studies, we can examine the impact of
plant design on the economics of H2 production and CO2 capture ( Table 2.1.3). This is
complicated, because the plant design changes in several ways, depending on whether
CO2 is captured, and whether sulfur compounds are separated.
CO2 capture and compression add ~ 10- 25% to the hydrogen production cost depending
on the plant design. In Figure 2.1.3, we plot the levelized cost of hydrogen production
from natural gas and coal as a function of plant size, assuming the CO2 is either vented
or captured. Fossil hydrogen plants exhibit strong scale economies. Because coal plants
are more capital intensive than natural gas plants, the hydrogen cost is slightly more
sensitive to scale for coal.
Figure 2.1.3. Cost of Hydrogen Production from Coal and Natural Gas with CO2
Separation and Compression versus Hydrogen Plant Size
Table 2.1.1. Cost and Performance of Natural Gas Based Hydrogen Production Plants
0
2
4
6
8
10
12
0 500 1000 1500
H2 Plant Size ( MWth)
NG CO2 vented
NG CO2 Seq
Coal CO2 Vented
Coal CO2 Seq
Coal Co- seq of CO2 +
H2S
22
w/ and w/ o CO2 Capture ( Foster Wheeler 1996)
CO2 vented CO2 captured
Hydrogen Production MWth ( at 60 bar
output pressure)
1000 1000
First law efficiency HHV basis 81% 78%
CO2 emission rate ( kgC/ GJ H2) 17.56 2.74
CO2 Sequestration Rate ( tonne/ h) 0 204
Capital Investment ( million $)
Reformer 48.65 67.90
Purification 23.65 58.08
CO2 Compression 0 35.67 ( for an
estimated CO2
compressor power
of 18.6 MWe)
Other 123.95 174.67
Subtotal 196.25 336.32
Subtotal ( excluding CO2 compressor) 196.25 300.65
Added costs
Engineering, construction
management, commissioning, training
9.13 16.94
Catalysts and chemical 8.75 9.00
Clients costs 24.00 28.00
Contingency 23.81 39.03
TOTAL INSTALLED CAPITAL
COST ( million $)
261.94 429.3
Incremental Installed Capital Cost for
CO2 Capture ( million $)
167.36
23
Table 2.1.2. Levelized cost of hydrogen production from natural gas with and without
CO2 separation and compression
Levelized Cost of H2
Production with CO2
separation, excluding CO2
compression
($/ GJ H2) HHV
CO2 vented CO2 captured
Capital ( excluding CO2
compression)
1.56 2.28
Natural Gas Feedstock 4.20 4.36
Non- fuel O& M 0.42 0.61
CO2 Compressor Capital
and O& M
n. a. 0.34
CO2 Compressor Electricity n. a 0.27
Total
6.17 7.86
Incremental cost of CO2
separation and
compression
n. a.
$/ GJ H2 HHV 1.69
$/ tonne CO2 29.8
24
Table 2.1.3. Cost and Performance for Hydrogen and Electricity Production from
Coal ( 70 bar gasifier) ( Kreutz 2002)
CO2
Vented,
sulfur
removal
CO2 Capture,
sulfur removal
CO2 capture,
co- sequestration
of CO2 and H2S
H2 Production MWth 1000 1000 1000
Electricity production ( net power out) MWe 52.2 30.9 30.9
First law efficiency HHV 0.736 0.705 0.705
CO2 emission rate ( kgC/ GJ H2 HHV) 35.6 2.61 2.61
CO2 captured ( tonne/ h) 0 437.4 437.4
Installed Capital Cost of Fossil Energy Complex
( million $) = 1.16 x Bare Capital Equipment Cost
H2 Plant excluding CO2 compressor 658.6 707.2 612.6
CO2 Compressor 0 51.7 ( 36.6 MWe) 51.7 ( 36.6 MWe)
H2 Plant including CO2 Compressor 658.6 758.9 663.4
Incremental plant cost for CO2 capture
including CO2 compression
0 100.3 4.8
Incremental plant cost for CO2 separation
excluding CO2 compression
0 48.7 - 46.0
Levelized Cost of H2 Production ($/ GJ HHV)
Plant capital except CO2 Compressor 3.92 4.20 3.64
Non- fuel O& M 1.04 1.12 0.97
Feedstock cost 1.26 1.32 1.32
CO2 compression capital + O& M 0.39 0.39
CO2 compressor power 0.37 0.37
Electricity credit incl comp pwr - 0.52 - 0.675 - 0.675
Total without CO2 compression 5.70 5.97 5.23
Total with CO2 compression 6.73 6.01
Incremental Cost of CO2 Capture, excluding CO2
compression
$/ GJ H2 ( HHV) 0.27 - 0.44
$/ tonne CO2 2.22 - 3.56
Incremental Cost of CO2 Capture, including CO2
compression
$/ GJ H2 ( HHV) 1.02 0.31
$/ tonne CO2 8.43 2.56
25
2.1.2. Task 1.2. Modeling CO2 Compression and Pipeline Transport
Once CO2 has been captured at the fossil energy complex, it must be compressed
to supercritical pressures and transported by pipeline to a suitable sequestration site.
CO2 Compression
Equations for compressor power requirements and cost models for CO2 compressors are
developed in Appendix B. The electric power required for compression of CO2 to
supercritical pressures ( 15 MPa) is modest, perhaps 6% of the total hydrogen power
output ( in MW thermal, based on the higher heat value of hydrogen). The levelized cost
of compression is found to be about $ 4- 6/ tonne CO2, for compressor electricity costing
3.6 cents/ kwh.
CO2 compression costs show the following sensitivities to varying parameters:
􀂃 The cost of electricity dominates the levelized cost of compression. For our base
case assumptions, about $ 3- 3.5/ tonne CO2 is due to power costs, the remainder to
capital costs.
􀂃 Compressor capital costs are sensitive to scale.
􀂃 Compression costs are somewhat sensitive to the compressor outlet pressure. This
pressure is typically at least 15 MPa, to assure that the CO2 stays above the critical
pressure throughout the pipeline. There is a modest incremental cost of about
$ 1/ tonne CO2 to increase the CO2 outlet pressure from 80 to 150 bar for pipeline
transmission.
CO2 Pipeline Transmission
We use a technical/ economic model for supercritical CO2 pipeline transmission
developed by the principal investigator under the CMI program. Our model is based on
pipeline flow equations developed in ( Farris 1983) and ( Mohitpour 2000). [ Details of
CO2 pipeline flow and cost calculations are given in Appendix C.] This model has been
benchmarked with existing CO2 pipeline models in the literature ( Farris 1983, Skovholt
1993), and with industry practice through conversations with engineers at BP.
One of the issues in estimating CO2 pipeline costs is the wide variation in published
estimates. This is shown in Figure 2.1.4, where installed CO2 pipeline costs ( in $/ m of
pipeline length) according to various studies are plotted versus pipeline diameter
( Doctor 1999; Skovholt 1993; Holloway 1996; Fisher, Sloan and Mortensen 2002). We
have selected a mid- range value for our studies, recognizing that published estimates of
capital costs for CO2 pipelines vary over more than a factor of two above and below the
midrange value. The wide variation is probably due to differences in local terrain,
26
Figure 2.1.4. Installed Capital Cost of CO2 Pipelines
construction costs and rights of way, all of which are important variables in determining
the actual installed pipeline cost.
Using a cost function fit to published pipeline data, and inlet and outlet pressure of 15
MPa and 10 MPa, respectively, we find a pipeline capital cost per unit length ($/ m), in
terms of the flow rate Q and the pipeline length L:
Cost( Q, L) =($ 700/ m) Q
Q0






0.48 L
L0






0.24
[ 1]
Where Qo = 16,000 tonnes CO2 / day and Lo = 100 km.
Figure 2.1.5 and Figure 2.1.6 show the cost of CO2 pipeline transmission as a
function of pipeline flow rate and pipeline length.
The levelized cost of pipeline transmission ($/ t CO2) scales approximately as
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
0 0.5 1 1.5 2
Pipeline diameter ( m)
IEA
Skovholt Low
Skovholt
Skovholt high
Joule II
Various Real CO2 lines
Western US
Christodoulou
Argonne
27
( CO2 flow rate)- 0.52 x ( pipeline length) 1.24
The cost per tonne of CO2 is lower for the coal hydrogen plant than the natural gas
hydrogen plant, because of its larger CO2 flow rate. However, the cost per GJ of hydrogen
produced is higher for the coal plant, because more CO2 is produced per unit of hydrogen
( Figure 2.1.7).
0
5
10
15
20
25
30
35
0 50 100 150 200
Pipeline Length ( km)
($/ tonne CO2)
Flow = 1000 tonnes/ d
Flow = 3000 tonnes/ d
Flow= 10,000 tonnes/ d
Figure 2.1.5. Levelized Cost of Pipeline Transmission ($/ tonne CO2) vs. Pipeline
Length and Flow Rate
28
0
5
10
15
20
25
30
0 200 400 600 800 1000
Pipeline Length ( km)
($/ GJ H2 HHV)
Flow = 1000 tonnes/ d
Flow = 3000 tonnes/ d
Flow= 10,000 tonnes/ d
Figure 2.1.6. Levelized Cost of CO2 Pipeline for Coal- Based H2 Plant ($/ GJ H2
HHV) vs. Pipeline Length and CO2 Flow Rate
0
5
10
15
20
25
30
0 200 400 600 800 1000
Pipeline Length ( km)
($/ GJ H2 HHV)
Flow = 1000 tonnes/ d
Flow = 3000 tonnes/ d
Flow= 10,000 tonnes/ d
Figure 2.1.7. Levelized Cost of CO2 Pipeline ($/ GJ H2 HHV) for Natural Gas to H2
Plant vs. Length and CO2 Flow Rate
29
2.1.3. Task 1.3. Modeling CO2 Sequestration sites
At the CO2 sequestration site, CO2 is injected into an underground geological
formation such as a deep saline aquifer or depleted hydrocarbon reservoir. A CO2 booster
compressor might be needed at the injection well- head depending on the well depth and the
aquifer pressure. Several injection wells might be needed, which would be connected via
above ground piping. Models for injection rate and capacity of underground geological
formations are described based on fundamental reservoir parameters ( see Appendix D for
details). The injection rate of CO2 into an underground reservoir depends on the
permeability and thickness of the reservoir, the injection pressure, the reservoir pressure, the
well depth, and the viscosity of CO2 at the injection pressure. A practical upper limit on the
injection rate per well is taken to be 2500 tonnes per day, limited by pressure drop due to
friction in the well at higher flow rates, assuming practical well diameters ( Hendriks 1994).
Using a standard equation for flow into an injection well ( Hendriks 1994), this upper limit
implies that for a layer thickness above 50 m and permeabilities above 40 milliDarcy , the
flow rate is limited not by the reservoir characteristics, but by the pipe friction flow
constraints. For the base case 1000 MW natural gas ( coal) to H2 plant, producing about 5,000
( 10,000) tonnes CO2 per day, 2 ( 4) wells are needed. The installed capital cost of each well is
( Hendriks 1994):
Capital ($/ well) = $ 1.56 million x well depth ( km) + $ 1.25 million.
In our base case, we assume a well depth of 2 km. CO2 is distributed by surface piping at
the injection site from well to well. We require each reservoir to store 20 years of CO2
production from the H2 plant. For our base case ( reservoir thickness of 50 m), the length
of surface piping required at the injection site is found to be 12 ( 37) km for the natural
gas ( coal) based H2 plant. This implies a cost of $ 3.2 ( 9.2) million, based on a piping
cost from Equation [ 1], but assuming that the minimum cost is $ 155,000/ km
($ 250,000/ mile) ( Ogden 1999). As long as the aquifer characteristics allow such a
relatively high injection rate, the cost of injection wells and associated piping is quite
small, less than $ 2/ tonne CO2.
The total levelized cost of CO2 pipeline transmission and storage is shown in Table
2.1.4, for hydrogen plants producing 1000 MW of hydrogen per day from natural gas
and coal. Per tonne of CO2, the cost of CO2 disposal is higher for natural gas, but
because the coal plant produces about twice as much CO2 as the natural gas H2 plant, the
contribution to the levelized cost of H2 ($/ GJ) is higher for coal.
30
Table 2.1.4. CO2 Pipeline Transmission and Storage System for Base Case H2 Plants
Producing 1000 MW of hydrogen output from Natural Gas and Coal
H2 from natural gas H2 from coal
CO2 captured ( tonne/ h) at full capacity 204 406
CO2 Disposal System ( 100 km pipeline, 2 km well depth, injection rate = 2500 t CO2/ day/ well)
CO2 100 km Pipeline Diameter ( m) 0.25 0.34
Number of CO2 Injection Wells 2 4
Injection Site Piping length ( km) 12.2 37
System Capital Cost ( million $)
CO2 100 km Pipeline 40.5 55.7
CO2 Injection Wells 8.8 17.5
CO2 Injection Site Piping 3.2 9.2
Total CO2 Pipeline Transmission and Storage System 52.5 82.4
Levelized Cost of CO2 Disposal ($/ tCO2)
CO2 100 km Pipeline 5.26 3.45
CO2 Injection Wells 1.16 1.17
CO2 Injection Site Piping 0.44 0.61
Total CO2 Pipeline Transmission and Storage System 6.87 5.23
Total CO2 Pipeline Transmission and Storage System ($/ GJ H2) 0.39 0.59
2.1.4. Task 1.4. Modeling H2 Demand Centers
Designing a hydrogen fuel delivery infrastructure depends on the characteristics of the
hydrogen demand. We model the magnitude, spatial distribution, and time dependence
of hydrogen demand, based on Geographic information system ( GIS) data on
populations, estimates of vehicles per person, and projections for energy use in
hydrogen vehicles, and market penetration rates. Our method for calculating a hydrogen
demand map is described below ( see Figure 2.1.8).
􀂃 First, population density is mapped as a function of location. This information is
available in GIS format from US Census data.
􀂃 On average in the US there are about three light duty vehicles for every four people
( Davis 2000). From this, we can approximate the numbers of light duty vehicles as a
function of location ( vehicles/ km2). This obviously a simplification, as numbers of
vehicles will not exactly track population. If more detailed information is known
about the locations of vehicles, this could be shown as well. In addition, early
markets for hydrogen might be found in heavy duty applications, such as fleets. If
information is known about these vehicles, this could be added as well.
􀂃 Next, a market penetration rate for hydrogen is estimated ( fraction of new vehicles
using hydrogen). This could be done in various ways. For example, one could
assume that a “ ZEV mandate” is put in place, so that a fixed fraction of new vehicles
sold must use hydrogen. Alternatively, one could devise other criteria for estimating
31
how many new hydrogen vehicles are sold each year, based on projections of when
they become competitive with competing technologies like gasoline internal
combustion engine technologies. From the market penetration rate, the number of
hydrogen vehicles can be found as a function of location and time ( H2 vehicles/ km2
versus time).
Vehicle Population
Density ( veh/ km 2 )
=
H2 Demand
Density
( kg H2 / d/ km 2 )
Number,
Size and
Location of
H2 refueling
stations
CREATING A H2 DEMAND MAP
H2 veh
H2 Vehicle
characteristics,
drive cycle and
mileage
X
Energy Use per
Vehicle
( kg H2 / veh/ day)
H2 Vehicle
Population Density
( veh/ km 2)
X =
Fraction
H2
vehicles
( time)
Technical
progress,
Economic
competitiveness,
Policy
Market
Penetration
rate
Customer
convenience
Refueling
pattern
End- user
req. H2
pressure
purity
Census
Data on
vehicles
by type
and
location
Figure 2.1.8. Mapping Hydrogen Demand Density
􀂃 The hydrogen use per vehicle ( kg H2/ d/ vehicle) is estimated from assumptions about
hydrogen vehicle fuel economy and miles traveled. A map of hydrogen demand
density versus location and time can be calculated ( kg/ d/ km2). This is shown in
Figure 2.1.9, for the state of Ohio. The lighter colors are low demand density, the
darker colors higher density. The cities of Cleveland, Columbus and Cincinnati are
obvious areas of high demand.
􀂃 Once the hydrogen demand density is known, one has to decide how many refueling
stations are required and where they should be sited. The number, location and size
of refueling stations have a major effect of the cost of infrastructure. In the United
States, on average, there is one gasoline refueling station for every 2000 light duty
vehicles ( Davis 2000). For several cities we examined, stations tend to cluster along
major roads in “ spoke” or “ ring” like patterns. Often, more than one station is found
at major intersections or at freeway exits. Recent analyses suggest that today’s
convenience level could be preserved, if perhaps 10- 30% of current gasoline stations
32
offered hydrogen ( Nicholas 2003). Methods for siting and sizing stations are
discussed further in section 1.5 and in Appendix H.
Figure 2.1.9. Hydrogen Demand Density in Ohio
An application of this hydrogen demand model is described for a case study of Ohio
( Task 3).
2.1.5. Task 1.5. Modeling H2 Delivery Infrastructure
There are many options for producing and delivering hydrogen to users. These include
centralized production options ( e. g. fossil energy complexes with CO2 capture), and
decentralized options ( such as small reformers or electrolyzers located at refueling
stations). We have developed cost and performance estimates for a variety of possible
hydrogen supply and delivery options, which are likely to be important in future
hydrogen energy systems:
Centralized, large- scale production of hydrogen from:
􀂃 Steam reforming of natural gas with and without CO2 sequestration
􀂃 Coal gasification with and without CO2 sequestration
􀂃 Large scale electrolysis
Distributed production of hydrogen at refueling sites from:
􀂃 Natural gas reforming
􀂃 Electrolysis using off- peak power
For centralized production, we consider hydrogen delivery via truck ( compressed gas
tube trailer or liquid tank truck), or via gas pipeline.
33
At refueling stations, we assume that hydrogen is dispensed to vehicles as a compressed
gas for onboard storage at 5000 psi.
2.1.5.1. Modeling Hydrogen Distribution System Components
Models for hydrogen delivery infrastructure components are described in detail in
Appendix E. These include:
􀂃 Hydrogen compressors
􀂃 Gaseous hydrogen bulk storage
o Above ground pressure vessels
o Underground storage
􀂃 Compressed gas tube trailer trucks
􀂃 Hydrogen gas pipelines
o Long distance transmission lines
o Local pipeline distribution networks
􀂃 Liquefiers
􀂃 Liquid hydrogen bulk storage
􀂃 Liquid hydrogen trucks
􀂃 Hydrogen refueling stations
o LH2 truck delivery
o Gas pipeline delivery
o Onsite small steam methane reformers at station
o Onsite small electrolyzers at station
Hydrogen compression
Electricity needed for compression is about 5- 10% of the energy content of the hydrogen
( on a higher heat value basis), depending on the inlet and outlet pressures. 7 Compression
typically adds less than $ 1/ GJ ($ 0.14/ kg) to the cost of hydrogen. Most of this cost is
due to the electricity cost. ( See Appendix E for details.)
Gaseous Hydrogen Storage
In the case of large centralized fossil hydrogen production, it is desirable to run the
hydrogen production plant continuously. However, the system- wide demand profile for
transportation fuel will vary over the day, weekly and even seasonally, so that some
storage capacity ( ranging from ½ day to several days plant output) will be needed in the
system.
Hydrogen can be compressed and storage as a high- pressure gas. For a gaseous
hydrogen pipeline distribution system, several options are available. Hydrogen could be
7 Compression energy requirements are higher for hydrogen as compared to natural gas, by roughly a
factor of three.
34
stored: 1) in the pipeline, 2) at the refueling station, 3) at the production site. We
assume the last option is used, although some storage is also located at the refueling
station.
Bulk gaseous storage at the central plant can be accomplished in several ways ( Taylor
et. al 1986). First hydrogen is compressed from production pressure ( typically 200 psi for
steam reforming or gasification systems) to storage pressure of perhaps 1000 psi
( assuming that the pipeline will be fed from storage). For very large quantities ( on the
order of 100 million scf or more), underground gas storage might be used.
Hydrogen Liquefaction and Liquid Hydrogen Storage
Alternatively, hydrogen can be liquefied ( at 20 K), stored in a dewar and delivered to
refueling stations via cryogenic tank trucks. Liquefaction is more energy intensive than
compression: electricity needed for liquefaction is about 33- 40% of the energy content
of the hydrogen ( on a higher heat value basis). Liquefiers have strong scale economies,
making them most suitable for use with large central plants. Liquid hydrogen
distribution is preferred when small quantities of hydrogen are shipped long distances.
Hydrogen Transmission Pipelines
The cost of a hydrogen pipeline depends on the pipeline diameter and length. If the flow
rate, pipeline length and inlet and outlet pressures, temperatures and gas properties are
known, we can use steady- state fluid flow equations to estimate the pipeline diameter
and the cost. In some cases, it may be desirable to add “ booster” compressors along the
pipeline to recompress the gas.
In Appendix E, we develop equations for hydrogen pipeline transmission costs as a
function of pipeline flow rate and length. The levelized cost of the hydrogen pipeline
( not including compression or storage) is given approximately by:
Cpipe[$ / GJ] = 0.15 Q[ MW]
1000MW






− 0.5 L[ km]
100km






1.25
Pipeline capital costs scale inversely with hydrogen flow rate and almost linearly with
distance.
Levelized costs are shown for hydrogen pipeline transmission including compression,
storage at the central plant, and the pipeline are shown Figure 2.1.10, as a function of
pipeline length and flow rate. We see that long distance transmission can add up to a few
dollars per GJ to the cost of hydrogen. Hydrogen pipelines are well- suited for delivery
of large quantities of energy.
35
Figure 2.1.10. Levelized Cost of Hydrogen Pipeline Transmission ( including
compression, storage, and pipeline) vs. Pipeline Length and Energy
Flow Rate ( MWth)
Hydrogen Truck Delivery
Hydrogen can be delivered by truck as well as by pipeline. For truck delivery, hydrogen
is compressed to high pressure and carried in a tube trailer or liquefied and carried in a
cryogenic tank truck.
Recent studies by NREL ( Amos 1998) and SFA Pacific ( Simbeck and Chang 2002)
have given estimates for the cost and performance of tube trailers and LH2 trucks. The
precise cost of truck delivery depends on the delivery route and the amount of hydrogen
delivered.
2.1.5.2. A Comparison “ Point- to- Point” Hydrogen Delivery Costs
The detailed cost models described above are used to determine the cost of “ point- to-point”
hydrogen delivery for different transport modes as a function of hydrogen flow
rate and transportation distance. 8 Figure 2.1.11 and Figure 2.1.12 show the least cost
8 Hydrogen delivery includes compression or liquefaction and hydrogen storage at the central plant, and
hydrogen transport via pipeline or truck. By “ point to point”, we mean delivery from the central H2
production plant to the edge of the city. Transport within the city via a local pipeline network is NOT
included. Local distribution costs are estimated in the next section.
0
2
4
6
8
10
12
14
16
0 500 1000 1500 2000
Energy Flow Rate ( MWth)
transmission ($/ GJ)
L= 100 km
L= 300 km
L= 1000 km
36
mode for any given flow and distance for point- to- point hydrogen transport. We see that
at large flow rates, pipeline transport is the lowest cost option. For small quantities of
hydrogen, compressed gas trucks are best at short distance and liquid hydrogen trucks at
longer distance. The overall cost of point- to- point transmission ranges from several $/ kg
to less than $ 0.5/ kg ( for pipelines with large flow rates). Figure 2.1.11
Flowrate Transport Distance [ km]
kg H2/ day] 25 50 75 100 125 150 175 200 225 250 275 300 325 350 375 400 425 450 475 5
2000 G G G G G G G G G G G G G G G L L L L
4000 G G G G G G G G G G G L L L L L L L L
6000 G G G G G G G G G L L L L L L L L L L
8000 P G G G G G G G L L L L L L L L L L L
10000 P G G G G G G L L L L L L L L L L L L
14000 P P P G G G L L L L L L L L L L L L L
18000 P P P P P P L L L L L L L L L L L L L
22000 P P P P P P P L L L L L L L L L L L L
24000 P P P P P P P L L L L L L L L L L L L
28000 P P P P P P P P L L L L L L L L L L L
32000 P P P P P P P P P L L L L L L L L L L
36000 P P P P P P P P P P P L L L L L L L L
40000 P P P P P P P P P P P P L L L L L L L
44000 P P P P P P P P P P P P P L L L L L L
48000 P P P P P P P P P P P P P P L L L L L
52000 P P P P P P P P P P P P P P P L L L L
56000 P P P P P P P P P P P P P P P P L L L
60000 P P P P P P P P P P P P P P P P P L L
64000 P P P P P P P P P P P P P P P P P P L
68000 P P P P P P P P P P P P P P P P P P P
72000 P P P P P P P P P P P P P P P P P P P
76000 P P P P P P P P P P P P P P P P P P P
80000 P P P P P P P P P P P P P P P P P P P
84000 P P P P P P P P P P P P P P P P P P P
88000 P P P P P P P P P P P P P P P P P P P
92000 P P P P P P P P P P P P P P P P P P P
96000 P P P P P P P P P P P P P P P P P P P
100000 P P P P P P P P P P P P P P P P P P P
Figure 2.1.11. Minimum cost delivery mode for a range of operating conditions ( P –
pipeline, G – compressed gas trucks, L – liquid trucks).
37
Figure 2.1.12. Graph of minimum cost for the three modes of hydrogen delivery as a
function of flowrate and transport distance.
2.1.5.3. Designing a Local hydrogen distribution network
Idealized City Model
Once hydrogen from a central production plant is delivered to the city gate, it must be
distributed to refueling stations located throughout the city. ( These stations are sited for
adequate customer convenience.) Distribution could be accomplished via trucks
traveling to stations or a network of small- scale pipelines.
2000
8000
14000
20000
26000
32000
38000
44000
50000
56000
62000
68000
74000
80000
86000
92000
9802050
125
225
325
425
$ 0.00
$ 0.50
$ 1.00
$ 1.50
$ 2.00
$ 2.50
$ 3.00
$ 3.50
$ 4.00
cost
[$/ kg]
flowrate [ kg/ day
transport
distance [ km]
L
P
G
38
To estimate the cost of local distribution, it is important to know the location and
size of refueling stations. Several researchers have looked at possible configurations for
a network of refueling stations ( Ogden 1999, Mintz 2002, Nicholas 2003). We have
modeled the distribution network serving hydrogen stations using an “ idealized city
model”. 9 We develop general expressions for a “ generic” city in terms of its size,
hydrogen demand and the resulting hydrogen infrastructure required to support this
demand. This design is used to determine costs for hydrogen distribution. Using
generalized, idealized city models speeds up the analysis and provides information about
these distribution system characteristics for a wide range of cities. 10
Figure 2.1.13. Idealized city model with 25 and 125 hydrogen stations distributed in
rings throughout the city.
As shown in Figure 2.1.13 we assume the city is circular, with a radially distributed
population. The city size is not specified as a fixed number of kilometers, but rather
distribution system lengths are characterized as a function of the city radius. Distances
are calculated in this city by following a grid ( i. e. rectilinear) road network. The
refueling stations are configured into rings that are concentric around the city center.
Each city configuration consists of one or more rings of stations with varying numbers
of stations in each ring. For a given station configuration, the radii of the rings of
stations were varied in order to minimize the overall weighted average distance traveled
9 Each demand area is treated as an ideal circular city. The layout of the distribution network ( including
the number of refueling stations and the length or distance of distributing hydrogen to those stations) is
estimated as a function of the city’s physical size ( area) and hydrogen demand within the demand region.
10 Where data is available, more detailed models can be used to determine station numbers, locations,
convenience and distribution system layout using a detailed geographic study of the distribution system of
a specific city/ region using GIS tools ( such as in Nicholas 2003).
39
for users. This analysis does not find an optimal configuration of stations, because the
average distance between users and stations is only one criteria among many that will be
used to optimally site refueling stations. Reducing the length and cost of the pipeline
network to supply these stations is another important criteria. As a result, a comparison
is made as to how convenience trades off against the distribution network length ( i. e. the
length of pipe required to connect each of the stations together and to the edge of the
city ( city gate)). In Figure 2.1.14, we show the total distribution length ( in city radii) and
the average distance between stations for cases with 5 to 75 refueling stations. 11
0
20
40
60
80
100
120
0.05 0.10 0.15 0.20 0.25 0.30 0.35
average distance from users to station [ radius]
delivery network distance [ radius]
75 Stations- T 75 Stations - P
40 Stations - T 40 Stations - P
25 Stations - T 25 Stations - P
10 Stations - T 10 Stations - P
5 Stations - T 5 Stations - P
Figure 2.1.14. Tradeoff between convenience and delivery network distance for
pipelines and trucks for different numbers and configurations of
stations. (“ P” denotes pipeline distribution, “ T” truck distribution.)
In Figure 2.1.15, the pipeline length ( Lpipeline) is shown to be a power law function of the
number of stations, while the truck route distance scales linearly with the number of
stations. Thus as the number of stations grows, the pipeline distribution modes become
more efficient than trucks. The model results are plotted to compare length of the
pipeline network or truck driving distance as a function of the number of stations. The
11 Other studies ( Nicholas 2003) have indicated that if 10- 25% of current gasoline stations offered
hydrogen, this might be sufficient for customer convenience. In a typical US city ( where there are about
3000- 4000 people per gasoline station), 10% coverage corresponds to 1 hydrogen station for every
30,000- 40,000 people. ( For cities ranging from 100,000- 3,000,000 people, the number of hydrogen
stations needed varies from about 3- 100 stations)
40
data for pipeline length vs station number is fitted to a power function and for the
homogeneous population density, the equation that describes this relationship is:
L pipeline = β ⋅ N stations
γ
where Lpipeline is the length of the pipeline ( as a multiple of the city radius), Nstations is the
number of stations, β is 3.524 and γ is 0.4115.
For the truck delivery scenario, it is assumed that trucks do not travel to multiple
stations on a given trip so that a linear equation describes this distance:
D truck = 1.44 ⋅ N stations
As demand increases along the demand profile, additional stations are added to the
network of stations. Although this model is not designed to calculate the marginal
increase in pipeline length resulting from adding new refueling stations, the curve fit can
be used to estimate, on average, the length of pipeline needed to supply additional
refueling stations.
Given the hydrogen demand in a city of a certain physical size, an estimate can be made
of the required number of refueling stations and using the equations above, the total
length of pipeline or truck travel distance required to supply the network of refueling
stations. The cost for the network can be calculated using cost models for truck or
pipeline hydrogen delivery.
Lpipeline = 3.5238Nstations
0.4115
Dtrucks = 1.4374Nstations
0
50
100
150
200
250
300
350
0 20 40 60 80 100 120 140 160 180 200
# of stations
Total delivery distance [ radius]
Total pipeline distance - cartesian
Total truck distance - cartesian
- 10
0
10
- 10 0 10
- 10
0
10
- 10 0 10
Figure 2.1.15. The relationship between the number of stations within the city and
the total delivery distance for pipelines and trucks.
41
2.1.5.4. Hydrogen Refueling Stations
Costs for hydrogen refueling stations have been discussed by a number of authors ( DTI
et a. 1997, Ogden et al. 1998, Thomas et al 2000, Simbeck and Chang 2002, TIAX
2003, DTI 2003). 12
In Table 2.1.5, we list the capital and operating costs for four types of refueling stations,
including pipeline- delivered hydrogen, LH2 truck- delivered hydrogen, onsite steam
methane reformers and onsite electrolyzers. A range of sizes is shown for stations
dispensing 100,000 to 2 million scf H2 per day ( 240 – 4800 kg H2/ day). H2 is dispensed
to vehicles at refueling stations as a high- pressure gas for storage in onboard cylinders
( at 34 MPa). Each station could serve a fleet of several hundred to several thousand cars.
There is a wide range of estimates. The cost of hydrogen refueling stations scales
approximately linearly with size. This suggests that the capital cost for refueling station
equipment would be about the same for a few large stations or many small ones. Of
course, other costs such as land or permitting, that don’t scale with size, might be higher
if many small stations were built.
12 Currently, the H2A group is analyzing the costs of refueling station designs. We will update these
estimates as newer data become available. Analysis is also ongoing at UC Davis on today’s hydrogen
refueling station costs ( Weinert 2004) and on hydrogen energy stations that reform natural gas to produce
power and heat for a nearby building as well as hydrogen ( Lipman 2004).
42
Table 2.1.5. Characteristics Of Hydrogen Refueling Stations
Type Reference
Size ( kg/ d)
Capital Cost
as a function
of size
Conversion
Efficiency
Feedstock ->
H2
Electricity
Use
( kWhe/ kgH2)
Total O& M
cost $/ y
Assumptions
ONSITE SMR
Princeton –
100 units
240- 4800 $ 951.07 x
( kg/ d) +
300,352
NG-> H2
η = 0.707
HHV
2.26 kWhe/ kg
H2
425.96 x kg/ d
+ 53747
NG =
$ 3/ MBTU,
Elec =
$ 0.072/ kWh
DTI – first
unit
37- 7500 $ 1155.6 x
( kg/ d) +
199,770
NG -> H2
DTI – 100
units
37- 7500 $ 435.11 x
( kg/ d) +
54266
DTI – 1000
units
37- 7500 $ 273.04 x
( kg/ d) +
34,054
Simbeck 2002 470 1,480,000 η = 70% LHV
$ 119,000 NG
$ 5.5/ MBTU
2 kWhe/ kg
H2
$ 19,000/ yr @
7 cent/ kwh
$ 235,000 NG=$ 5.5/ MB
TU; elec=
$ 0.07/ kWh
TIAX mature
tech. 2003
690 1,175,000
PIPELINE DELIVERED H2
Princeton 240- 4800 $ 602.64 x kg
H2/ d + 34667
2.48 kWhe/ kg
H2
$ 195.92 x
( kg H2/ d) +
43100
Elec =
$ 0.072/ kWh
Simbeck 470 520,000 elec=
$ 0.07/ kWh
TIAX 690 352,500
LH2 TRUCK DELIVERED H2
Princeton 240- 4800 $ 225.51 x kg
H2/ d + 94664
0.27 kWhe/ kg
H2
$ 93.334 x kg
H2/ d + 45082
Elec =
$ 0.072/ kWh
Simbeck 470 680,000 Elec
=$ 0.07/ kWh
TIAX 690 423,000
ONSITE ELECTROLYSIS
Princeton 240- 4800 $ 2528.7 x kg
H2/ d + 20433
Electricity
η = 80%
HHV
49 kWhe/ kg
electrolysis +
4.16 kWhe/ kg
H2
compression
$ 736.63 x ( kg
H2/ d) +
45990
Off- pk power
Elec = 3
cent/ kWh
DTI – first
1000 stations
37- 75 $ 2258.9 x kg
H2/ d + 69760
Electricity
η = 80%
Simbeck 470 4,150,000
$ 2157/ kW
Electricity
η = 63.5%
LHV
55 kWhe/ kg
H2
Electrolysis +
2.3 kWh/ kg
H2
Compression
700,000 elec=
$ 0.07/ kWh
TIAX 690 1,128,000
43
2.1.5.5. Summary of Component Costs and Performance for Fossil Hydrogen
Energy System with CO2 Sequestration
In Table 2.1.6 and Table 2.1.7, we summarize the costs and performance for various
components of a hydrogen energy system. These simplified formulas allow us to
estimate component capital and O& M costs as a function of size, feedstock, and
electricity costs.
Table 2.1.6. Summary Economic Data for Large Central H2 Production Systems as a
Function of Scale
So = Reference
H2 plant size
Cost( So) =
Capital
Investment for
Ref. H2 Plant
( million $)
α= Plant capital
Scale factor
( scale range)
η = Feedstock
Conv. Eff to H2
on HHV basis
Co- products Source
SMR, CO2
vented
613 tonne H2 / d 262 0.7
( 153- 613 t/ d)
0.81 Foster Wheeler
( 1996, 1998)
SMR, CO2
captured
613 tonne H2 / d
( 5000 tCO2/ d)
384 for plant
+
45 ( CO2
compressor)
= 429 total
0.7
( 153- 613 t/ d)
0.7
( CO2 comp)
0.78 Foster Wheeler
( 1996, 1998)
Coal Gasifier,
CO2 vented
613 tonne H2 / d 659 0.828
( 153- 613 t/ d)
0.736 Electricity
( 2.04 kWh/ kg
H2)
Kreutz 2002
Coal Gasifier,
CO2 captured
613 tonne H2 / d
( 10,000 tCO2/ d)
613 for plant +
50 ( CO2
compressor)
= 663 total
0.828
( 153- 613 t/ d)
0.7
( CO2 comp)
0.705 Electricity
( 1.21 kWh/ kg
H2)
Kreutz 2002
CO2
Sequestration
( CO2
compressor is
included in fossil
H2 plant cost
estimates above)
16000 tonne
CO2 / d
100 km pipeline
2500 tonne
CO2 / d/ well
$ 70 million x
( Q/ 16000) 0.48 x
( L/ 100) 1.24
+ Q/ 2500 x $ 4.4
million/ well
+ ( Q/ 2500- 1) x
$ 3.2 million
Pipeline
+ injection well
+ injection site
piping
Ogden ( 2002)
Biomass
Gasifier, CO2
vented
165 tonne/ d 172 0.7
( 150- 750 t/ d)
0.636 Larson 1993;
Simbeck and
Chang 2002
Electrolysis 150 tonne/ d
250 MW H2
$ 75- 150 million
($ 300- 600/ kW)
0.9
( 20- 613 t/ d)
0.8 Oxygen
( 8 kg/ kg H2)
Ogden ( 1998)
CRF = 15%; non- fuel O& M = 4% of capital investment/ y
Capital Cost at plant size S ($) = Cost ( S) = Cost ( So) x ( S/ So) α
S = H2 plant capacity ( tonne/ d)
44
O& M Cost at plant size S ($/ y) = O& M( S) = 4% x Cost ( So) x ( S/ So) α
Feedstock Cost ( S) ($/ y)
= S x 365 d/ y x capacity factor x HHV H2 ( GJ/ kg)/ η x feedstock Cost ($/ GJ)
Byproduct credit ( S) ($/ y)
= S x 365 d/ y x capacity factor x Byprod ( unit/ kg H2) x Byprod price ($/ unit)
Levelized cost of H2( S) $/ kg
= [ CRF x Cost( S) + O& M( S) + Feedstck Cost( S) + Byproduct credit( S)]/( capacity factor x S x 365 d/ y)
45
Table 2.1.7. Economic Data for Gaseous Hydrogen Pipeline Transmission Systems as
a Function of Scale ( including hydrogen compression, large scale
gaseous storage and transmission pipeline)
Reference
equipment size
Capital Investment
($/ kWe)
Εquations with scaling factors
H2 compressor
( note: in some
studies H2
compression is
included as
part of the
central H2
plant cost)
20 MWe $ 1600/ kWe
( multi stage)
$ 900/ kWe
( single stage)
Scale factor of 0.9 for large H2 compressors
( Simbeck and Chang 2002). Costs match well
with Kreutz et al. 2002)
H2 compressor electricity input = 2- 10% of
higher heating value of hydrogen compressed
depending on compressor inlet and outlet
pressures ( see Appendix E). Assuming inlet
pressure of 1.4 MPa, and outlet pressure of 6.8
MPa, and compressor efficiency of 70%, the
electricity use is about 2% of the H2 energy.
Compressor power ( MWe)
= [ S ( tonne/ d) x ( 1000 MWH2/ 613 tonne/ d)
x ( 2- 10% MWe/ MWH2)]
Capital cost of H2 compressor($) =
( Compressor Power/ 20 MWe) 0.9 x $ 1600/ kWe x
20 MWe
S= H2 plant size ( tonne H2/ d)
H2 Storage High pressure
cylinders
Bulk aboveground
compressed gas
storage
Advanced
automotive
pressure cylinders
Underground
storage
$ 700/ kg ( kg of H2
storage capacity)
“
$ 200- 250/ kg
$ 280- 420/ kg
Compressed gas storage is modular with little
scale economy.
For a H2 central plant, we assume storage
equivalent to 1/ 2 day’s production is needed.
If S = plant output in tonne H2/ d,
Cost = $ 700,000 x 0.5 x S,
for aboveground gas storage
Cost = $ 280,000- 420,000 x 0.5 x S,
for underground storage
H2 Pipeline
H2 Flow
Length
100 km length;
( Pin= 6.8 MPa
Pout= 1.4 MPa)
H2 Flow=
60 t/ d
150 t/ d
300 t/ d
600 t/ d
Pipe
Diam. Cost ( inch)
( million$)
D= 4.8”;$ 16- 62
D= 6.7”,$ 16- 62
D= 8.7”$ 16- 62
D= 11.4”$ 17- 62
Pipeline capital cost ($/ m)
= max 0.3354 x D2+ 11.25 x D + 2.31;
155- 620 ( for rural- urban sites)
D = pipeline diameter in inches
( D is found from hydrogen flow rate, pipeline
inlet and outlet pressures, pipeline length, and
flow regime ( see Appendix E)
46
2.2. Task 2.0. Integrated Studies of the Entire System to Find the Lowest Cost
Options
In Task 2, we combine our “ component” models of hydrogen production, CO2 capture,
transmission and sequestration, hydrogen compression, storage, distribution and refueling
to describe an integrated fossil hydrogen system with CO2 capture and sequestration.
2.2.1. Task 2.1. Develop Simple Model for Entire System and Perform
Sensitivity Studies
In Task 2.1, we studied total system design and economics, for the special case of a single
large fossil energy complex connected to a single geological CO2 sequestration site and a
single H2 demand center ( such as a city with a large concentration of H2 vehicles).
Results for this task were described in the first progress report for this contract. The
system is shown in Figure 1.1.1. Using the component models from Task 1, we developed
a simple analytical model linking the components into a total system. We then estimated
the total delivered cost of H2 with CO2 sequestration for hydrogen produced from coal and
natural gas ( Figure 2.1.3). We conducted sensitivity studies to examine which parameters
are most important in determining delivered hydrogen costs. For our base case
assumptions ( large CO2 and H2 flows; a relatively nearby reservoir for CO2 sequestration
with good injection characteristics; a large, geographically dense H2 demand), H2
production, distribution and refueling were found to be the major costs contributing to the
delivered H2 cost. CO2 capture and sequestration added only ~ 10%. Better methods of H2
storage would reduce both refueling station and distribution system costs, as well as costs
on- board vehicles.
As a second step, we expanded this simple model to include better models of
hydrogen demand and hydrogen distribution systems. Further, this improved model
provides a potential interface with GIS database being developed in Task 3, allowing
hydrogen system design and cost calculations based on quantities easily derived from GIS
maps ( see Figure F. 1). In the next sections we present results for the cost of fossil
hydrogen production with CO2 sequestration including distribution of hydrogen to
vehicles, as a function of geographic factors ( size of demand, geographic density of
demand, location of fossil energy complexes and sequestration sites), level of hydrogen
use ( e. g. market penetration of hydrogen vehicles), and technology.
2.2.1.1. An Integrated Hydrogen System Model
We consider a variety of possible hydrogen supply and delivery options, which are
likely to be important in future hydrogen energy systems:
Centralized, large- scale production of hydrogen from:
􀂃 Steam reforming of natural gas with and without CO2 sequestration
􀂃 Coal gasification with and without CO2 sequestration
􀂃 Large scale electrolysis
47
Distributed production of hydrogen at refueling sites from:
􀂃 Natural gas reforming
􀂃 Electrolysis using off- peak power
For centralized production, we consider hydrogen delivery via truck ( compressed gas or
liquid), or via gas pipeline. For fossil hydrogen with CO2 sequestration, we consider a
disposal system for CO2.
For each supply pathway, we estimate infrastructure costs as a function of a relatively
small number of input variables embodying averaged and/ or simplified information
about geography, markets and technology.
INPUT variables:
Geographic factors:
Total number of vehicles in a region
Region size ( km2)
Market Factors:
fraction of hydrogen vehicles in fleet
refueling station coverage factor ( fraction of all refueling stations that must offer
H2 to assure adequate customer convenience)
Number of vehicles per gasoline refueling station today
Vehicle use miles/ year
Technical Factors:
Vehicle Fuel Economy
Cost and performance of infrastructure components
Layout of distribution system ( from idealized city model in Task 1.5)
From these inputs. we estimate for different production and delivery pathways:
OUTPUT OF MODEL:
H2 production capacity needed
Number of H2 refueling stations
H2 dispensed per station
Layout of hydrogen stations
Delivery system layout ( pipeline length; truck route length)
Cost of entire system from production through delivery for different production
and delivery options
Levelized cost of hydrogen
Details of this model are given in Appendix F.
48
2.2.1.2. Preliminary Results
We have just begun to work with this model to estimate the lowest cost alternatives as a
function of market and geographic factors. As an example, we consider a city of 1
million people, where 10% of vehicles run on hydrogen ( see Table 2.2.1 and Table
2.2.2).
Table 2.2.1. Characteristics of City and Calculated Infrastructure
Geographic Factors
People 1 million people
Light Duty Vehicles ( LDVs) 750,000 LDVs
LDVs/ km2 1500
Area of city 500 km2
City radius ( for circular city) km 12.6 km
Market factors
Fraction H2 vehicles = fH2 10%
Gasoline Vehicles/ gasoline station 3000
Coverage factor 20%
Vehicle performance
H2 Vehicle Fuel Economy
= 2.8 x Today’s Gasoline LDV
57 mpgge
Miles travelled/ y 15,000
H2 energy use/ LDV/ d 0.7 kg H2/ d/ LDV
H2 Vehicles and Refueling Stations
# H2 vehicles in city 75,000
Total H2 production required kg/ d 52.5 tonne H2/ d
# H2 refueling stations 50
H2 refueling station size 1050 kg/ d/ sta
H2 cars/ H2 sta 1500
Central Production Model
Central production capacity tonne H2/ d 65.6 tonne/ d
Central plant storage capacity tonnes 26.25 compressed gas
52.5 Liquid H2
Pipeline Distribution Model
Local distrib. pipeline length/ city radius
( Figure 2.1.15)
20
Local distrib pipeline length 252 km
Truck Distribution Model ( assumes each
truck makes 2 deliveries per day)
Compressed Gas Trucks required 55
LH2 Trucks Required 7
49
Table 2.2.2. Capital Costs for Hydrogen Infrastructure Options ( million $)
Central
production
SMR +
pipeline
delivery, CO2
vented
Central
production
SMR + LH2
truck
delivery, CO2
vented
Central
production
SMR + comp
gas truck
delivery, CO2
vented
Onsite SMR Onsite
Electrolyzer
Capital costs Million $
Central SMR 55 50.5 55
Liquefier - 54 -
Comp Gas
storage
18.3
1/ 2 day
2.54
1/ 2 day
18.3
1/ day
Local
Pipeline
($ 155- 620/ m)
38- 150 - -
Trucks - 4.4 29.5
Refueling
stations
33.3 16.6 33.3 64.9 122
TOTAL
Capital cost
($ million)
145- 257 127 136 65 122
TOTAL
Capital cost
$/ LDV
1933- 3427 1699 1814 866 1628
Operating Costs ( million $/ yr)
Natural Gas 12.60 12.60 12.60 20.06
Electricity 2.85 8.91 2.85 2.60 30.56
Other O& M 6.23 5.75 10.58 2.60 4.88
Total O& M 21.67 27.26 26.03 25.26 35.44
LEVELIZED COST OF H2 $/ kg
Capital 1.42- 2.51 1.25 1.33 0.64 1.19
NG 0.82 0.82 0.82 1.31 0.00
Electricity 0.19 0.58 0.19 0.17 1.99
Other O& M 0.41 0.38 0.69 0.17 0.32
Total 2.84- 3.93 3.03 3.03 2.28 3.51
For this level of hydrogen vehicle use, in this size city, onsite SMR gives the lowest
capital costs and delivered hydrogen costs. In Figure 2.2.1, we plot the capital cost of H2
infrastructure per car as a function of hydrogen market penetration rate. For this set of
assumptions, onsite SMRs are the lowest capital cost option for all values of fH2 > 1%
of the fleet ( at these very low H2 penetration rates, electrolyzers are less costly).
50
Figure 2.2.1. Capital Cost $/ LDV for H2 Infrastructure vs. Fraction of H2 Vehicles
The delivered hydrogen cost ($/ kg) is plotted versus fH2 in Figure 2.2.2. At very low
hydrogen use, compressed gas trucks or electrolyzers give the lowest delivered costs. At
very large fractions of H2 use, pipeline hydrogen gives the lowest delivered cost. This
result is consistent with Figure 2.1.11.
2.2.2. Task 2.2 Explore Use of Mathematical Programming Techniques to
Study More Complex Systems.
Although studies of the simple system in Task 2.1 are useful, a mature fossil hydrogen
system would potentially involve a number of hydrogen production sites, hydrogen
demand centers, and CO2 sequestration sites. To study more complex and realistic
systems involving multiple energy complexes, H2 demand centers, and sequestration sites,
we are exploring use of mathematical programming methods to find the lowest cost
system design. Thus far, we examined the suitability of several mathematical
programming methods that could be used to optimize the design of a hydrogen energy
system with CO2 sequestration.
The basic design problem is shown in Figure 1.1.2. We have several hydrogen demand
centers ( shown in yellow) and primary resources. The question is how to connect these
using the lowest cost system ( including hydrogen production plants, hydrogen
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
0.01 0.1 1
Fraction of H2 vehicles in fleet
Central production SMR +
pipeline delivery
Central production SMR + LH2
truck delivery
Central production SMR + Comp
H2 truck delivery
Onsite SMR
Onsite Electrolyzer
51
distribution and for fossil hydrogen options, a CO2 disposal system.) The longer- term
goal is to compare various possible transition pathways to find the lowest overall cost.
Figure 2.2.2. Delivered Cost of H2 ($/ kg) vs. Fraction of H2 vehicles in Fleet for
City of One Million People
This is a complex nonlinear optimization problem. As a first step, we reviewed the
literature to understand how mathematical programming techniques had been applied to
modeling pipeline systems ( see Appendix H). ( This is a subset of the overall design
problem, as hydrogen production systems are not specifically included in this analysis.)
Several general classes of problems have been studied, relating to optimizing pipeline
systems.
Design Optimization: In this category, we consider the design of a new pipeline
network. Since the network doesn’t exist yet, we must decide how many compressor
stations ( if any) are needed, where they should be located, where the interconnection of
two ( or more) pipes should happen, and what size ( diameter and length) each pipe
segment should be. Constraints may include mass balance at each node, gas flow
equation in every pipe segment, the work equation of compressors and limits on the
pressure or flow rate. Infinitely many designs can meet the constraints. The design and
0
1
2
3
4
5
6
7
8
9
10
0 0.2 0.4 0.6 0.8 1
fraction H2 vehicles in fleet
Central production SMR + pipeline
delivery
Central production SMR + LH2
truck delivery
Central production SMR + Comp
H2 truck delivery
Onsite SMR
Onsite Electrolyzer ( off- pk elec)
52
building cost is used as the objective function to select one design out of the design
space.
Steady- State Operation Optimization: In this case, the network already exists, so
pipe size, number and location of compressor stations are already known. The objective
is to minimize the fuel consumption by compressors, which is determined by the suction
and discharge pressure at each compressor station and the flow- rate of gas going through
these compressors.
Table 2.2.3 summarizes the objective function ( e. g. the cost function to be minimized),
the constraints, and the optimization variables. Table 2.2.4 shows some of the
approaches that have been applied to these two classes of pipeline design problems.
Dynamic operation has also been treated, but we do not consider this here, because of its
complexity.
Table 2.2.3. Objective Function Used in Various Pipeline Studies
Design Optimization Steady- state Operation
optimization
Objective
function
# , , )
( , ,
cos
of compressor terrain K
f pipe diameter pipe length
building t
=
( , , )
cos ( )
f pressure flowrate K
operation t fuel comsumed
=
Constraints
1. mass flow balance equation at
each node
2. gas flow equation at each pipe
segment, i. e. pressure drop
equation
3. working equation of each
compressor
4. limits imposed on pressure or
flow- rate
Same as the left
Optimization
variables
pipe diameter and length, location
of compressor stations and other
interconnection points, etc.
flow- rate, suction and
discharge pressure at each
compressor station, etc.
In our studies so far we have concentrated on minimizing pipeline distances as a
surrogate for minimizing costs. As described below ( Task 3) this was accomplished by
finding the minimum spanning tree connecting hydrogen supply ( a central hydrogen
plant) with demand centers ( cities).
53
Table 2.2.4. Mathematical Programming Methods Used in Various Studies to Model
Pipelines
Traditional
optimization
techniques
Pure linear programming
Nonlinear programming
Sequential linear programming ( SLP)
General reduced gradient method ( GRG)
Inter- point method
Newton- Raphson method
Sequential unconstrained minimization technique ( SUMT)
Dynamic programming
Nontraditiona
l optimization
techniques
Genetic algorithms
Simulated annealing
Neural network
Artificial ants
2.3. Task 3.0 Case Study of Transition to a Fossil Energy System with CO2
Sequestration
2.3.1. Task Overview
In this task, we explore how H2 and CO2 infrastructures might develop in the context of
a geographically specific regional case study. We focus on the Midwestern United
States, a region where coal is widely used today in coal- fired power plants, and good
sites for CO2 sequestration are available. The goal is to identify attractive transition
strategies toward a regional hydrogen/ electricity energy system in the Midwest with
near zero emissions of CO2 and air pollutants to the atmosphere.
In this task, the goal is to derive insights about.
􀂃 Time constants and costs. How fast can we implement hydrogen fuel infrastructure?
How much will it cost? What are the best strategies? What level of demand is
needed for widespread implementation of H2 energy system?
􀂃 Sensitivities to: technology performance and costs, size and density of demand, local
availability of primary sources, characteristics of CO2 sequestration sites, market
growth, policies.
􀂃 Rules for thumb for optimizing H2 and CO2 infrastructure development.
54
To better visualize our results, we use a geographic information system ( GIS)
format to show the location of H2 demand, fossil energy complexes, coal resources,
existing infrastructure ( including rights of way), CO2 sequestration sites and the optimal
CO2 and H2 pipeline networks.
We developed a GIS database for the state of Ohio, an area where coal- fired power
plants are widely used. A survey of relevant GIS data sets was conducted ( see Appendix
I), and a database was built, including:
􀂃 Population density data, which is used to estimate hydrogen demands
􀂃 Data on the existing natural gas system
􀂃 Information on the electricity system and power plants
􀂃 Information on roads, railroads
􀂃 Data on the existing gasoline refueling infrastructure
􀂃 Information on sites for CO2 sequestration
We combined this data into a single data base showing features such as hydrogen
demand density, location of power plants, etc. 13 This is used a basis for analyzing
alternative configurations for hydrogen supply and CO2 disposal.
The overall flowchart for the GIS- based modeling is shown in Figure 2.3.1.
13 Data sources used in building this database are given in Appendix H.
55
Figure 2.3.1. GIS modeling flowchart
Hydrogen Demand
Part I: Calculating H2 Demand Density
Population
( 2000 Census
Blocks)
Popul ation
Density
( people/ km2)
Vehicle Density
( vehicles/ km2)
0.7 vehicles/ person
H2 Ve hicle
Density ( H2
vehicles/ km2)
Market Penetration
0.6 kg/ vehicle/ day
Idealized city
network model
Blocks of
High Demand
Density
Apply Threshold ( 50,
100, 150 kg/ km2/ day)
High D emand
Density
Clusters
5- km Buffer
Aggregate H2
Demand per
Block ( kg/ day)
Aggreg ate H2
Demand per
Cluster
Dissolve Blocks
into Clusters
H2 De mand
Centers
Apply Threshold
( 1,000/ 3,000/ 5,000
kg/ day)
Infrastructure Optimization
Identifying the Best Production Facilities and Shortest Pipeline Distances that Serve
Demand
Potential
Rights- of- Way
H2 De mand
Centers
Shortest P ath Matrix
Between All Nodes
( Power Plants and
Demand Centers)
Coal P ower
Plants over 100
MW
Optimized Design
for Shortest
Pipeline Network
Network Analysis
Minimal spanning
tree optimization
algorithm
H2 Deman d Centers
- Area
- Population
- Demand
Intracity Distribution and Refueling Station Optimization
Estimating an Idealized Intracity Infrastructure
# of Stations a nd Length/
Distance for Distribution
Systems Within Cities
Levelized Cost of Hydrogen ($/ kg)
Quantifying the Levelized Cost of Hydrogen for the System Design
Refueling
Station
Levelized Cost of
Hydrogen ($/ kg)
for the System
Coal Plant and CO2
Sequestration Model
Distribution Cost Model
– Pipelines and Trucks
Part II: Identifying H2 Demand Centers
H2 Demand
Density
( kg/ day/ km2)
56
2.3.2. Estimating Hydrogen demand
2.3.2.1. Methodology
In developing an optimized hydrogen infrastructure for the state of Ohio, it is first
necessary to identify the quantity and location of hydrogen demand under different
market penetration scenarios. In this study, hydrogen demand was calculated for two
hypothetical steady- state scenarios in which: 1) hydrogen fuel cell vehicles ( FCV’s)
make up 10% of the light duty vehicle ( LDV) fleet and 2) FCV’s make up 50% of the
LDV fleet. The objective is to identify “ demand centers” in which there is sufficient
hydrogen demand to warrant investment in infrastructure.
To complete this analysis, a Geographic Information System ( GIS) was used to derive
hydrogen demand from block- level Census 2000 population data ( US Census Bureau,
2000). The following steps were followed to identify hydrogen demand density in Ohio
under the two FCV market penetration scenarios.
1. As FCV demand will occur in the future, a base year of 2030 was used for the
analysis. Projected population change statistics (%) from 2000- 2030 by county
( Ohio Department of Development, 2004) were used to calculate population in the
year 2030.
2. Population density was calculated by dividing the population of each census block
by its area ( km2) to arrive at persons/ km2.
3. An estimate of total LDV’s per km2 was calculated by multiplying the population
density by an estimate of auto ownership per person. A factor of 0.7
vehicles/ person was derived from Ohio Department of Public Safety data, which
indicates that 8.3 million vehicles are registered among approximately 11,353,140
people.
4. Hydrogen vehicle density ( H2 vehicles/ km2) was calculated for the two market
penetration scenarios by multiplying the total LDV’s per km2 by 10% and 50%.
5. Hydrogen demand density ( kg H2/ day/ km2) was derived by multiplying the number
of H2 vehicles with an estimate of average vehicle fuel use ( 0.6 kg H2/ day/ vehicle).
This estimate is based on the assumption that the average vehicle travels 15,000
miles/ year and has a fuel economy of 65 miles/ kg.
In summary,
kgH2 / day
km2 = persons
km2






total vehicles
person






H2vehicles
total vehicles






kgH2 / day
vehicle






( 0.7) ( 10% and 50%) 0.6
57
Given hydrogen demand density throughout the state, the next step was to identify
census blocks with sufficient demand to warrant consideration for infrastructure. Three
density thresholds ( 50, 100, and 150 kg/ day/ km2) were analyzed to examine their ability
to capture hydrogen demand. The results of this sensitivity analysis will be presented in
the next section. A GIS was used to select census blocks that met each threshold. Upon
examining the results, it was apparent that the selections did not result in uniform areas
of high density, but rather concentrations of high density census blocks with holes
caused by low density blocks. Figure 2.3.2 illustrates this phenomenon within the city
of Columbus for the three thresholds.
Figure 2.3.2. Hydrogen demand density given different density thresholds in
Columbus, Ohio
58
In designing an optimized infrastructure, it was decided to identify uniform demand
centers rather than islands of small disjointed clusters. Consequently, a 5- kilometer
buffer was used to aggregate these clusters into uniform, consolidated shapes. The
buffer was generated from the high demand density blocks and then all census blocks
that were completely contained within the buffer were aggregated to form the demand
clusters. Figure 2.3.3 illustrates the results from this analysis for the city of Columbus.
Using the 100 kg/ day/ km2 threshold, a total of 67 and 98 demand clusters were
identified statewide for the 10% and 50% scenarios, respectively.
Figure 2.3.3. Demand clusters under different density thresholds in Columbus,
Ohio
59
Given these demand clusters, the next step was to identify a subset consisting of clusters
that have sufficient aggregate demand to support a single fueling station. To calculate
aggregate demand, total hydrogen demand was identified for each census block by
multiplying the demand per km2 with the area ( km2) of each block. Aggregate demand
for each demand cluster was then calculated by summing the demand for all component
blocks. A threshold was then used to eliminate clusters that do not have sufficient
demand to support a fueling station. Three thresholds were tested, including 1,000,
3,000, and 5,000 kg/ day. The results of the sensitivity analyses are discussed in the next
section. For the “ base” case that used a density threshold of 100 kg/ day/ km2, it was
discovered that the hydrogen demand varied from 85 to 63,235 kg/ day under the 10%
scenario and from 115 to 754,836 kg/ day under the 50% scenario. Figure 2.3.4 shows
the results for the 10% scenario.
60
Figure 2.3.4. Demand clusters and associated aggregate hydrogen demand
Using the “ base” case aggregate threshold of 3,000 kg/ day, all demand clusters with a
demand below this threshold were erased, leaving twelve demand centers under the 10%
scenario and thirty- nine under the 50% scenario. The final demand centers using the
“ base” thresholds are illustrated for the 10% and 50% scenarios in Figure 2.3.5 and
Figure 2.3.6, respectively.
61
Figure 2.3.5. Demand centers with 10% market penetration
62
Figure 2.3.6. Demand centers with 50% market penetration
63
2.3.2.2. Sensitivity Analysis
In order to understand the spatial distribution and quantity of hydrogen demand under
the 10% and 50% market penetration scenarios, we used two thresholds to identify areas
of high demand. The first threshold ( density threshold) was used to identify high
demand density and develop demand clusters. The second threshold ( aggregate
threshold) was used to highlight areas with sufficient aggregate demand to warrant
investment in infrastructure. As a result, it served to identify the optimized demand
centers as a subset of the initial demand clusters. In order to determine appropriate
thresholds, we conducted a sensitivity analysis using three threshold scenarios to
analyze their impact on the extent and quantity of hydrogen demand. The three
scenarios considered are shown in Table 2.3.1.
Table 2.3.1. Threshold values for each scenario
Scenario 1
[ low threshold]
Scenario 2
[ base]
Scenario 3
[ high threshold]
Density Threshold 50
( kg H2/ day/ km2)
100
( kg H2/ day/ km2)
150
( kg H2/ day/ km2)
Aggregate Threshold 1000
( kg H2/ day)
3000
( kg H2/ day)
5000
( kg H2/ day)
To compare these scenarios, we calculated the percent of statewide hydrogen demand
captured within the demand centers ( kg H2/ day), the percent of statewide land area
captured ( km2), and the number of demand centers. Table 2.3.2 summarizes the results.
Table 2.3.2. Results for each threshold scenario
Scenario 1
[ low threshold]
Scenario 2
[ base]
Scenario 3
[ high threshold]
H2 Demand
(% of Ohio total)
63.65% 47.21% 32.32%
Area
(% of Ohio total)
8.83% 4.84%
2.59%
Number of Demand
Centers
25 12 8
As expected, a greater percentage of hydrogen demand is captured over a larger land
area and in more demand centers as the threshold is lowered. The following figures
illustrate the results for demand centers with varying levels of hydrogen demand. We
64
categorized the demand centers into five groups based on the quantity of aggregate
hydrogen demand: 0 - 5,000 kg/ day, 5,000 – 10,000 kg/ day, 10,000 – 20,000 kg/ day,
20,000 – 40,000 kg/ day, and greater than 40,000 kg/ day. Figure 2.3.7 identifies the
number of demand centers in each group.
Figure 2.3.7. Number of hydrogen demand centers
This figure indicates that the “ low” threshold results in a large number of centers with
low hydrogen demand ( 1,000 to 5,000 kg/ day). Depending on the location of these
small centers, it may be cost prohibitive to supply them with hydrogen given their low
demand. Consequently, it may be preferable to use a higher threshold to eliminate some
of these smaller demand centers. The “ low” threshold scenario not only results in more
demand centers, but also cause demand centers to increase in size, resulting in more
large demand centers (> 40,000 kg/ day). Figure 2.3.8 illustrates the percent of total
hydrogen demand for each group.
18
1
2
1
2
4
2
1
0
2 2 2
3 3
2
0
2
4
6
8
10
12
14
16
18
20
~ 5000 5000 ~ 10000 10000 ~ 20000 20000 ~ 40000 40000 ~
Category: Size of H2 Demand Center ( kg H2/ day)
Number of H2 Demand Centers Captured
Scenario 1
( DT = 50,
AT = 1,000)
Scenario 2
( DT = 100,
AT = 3,000)
Scenario 3
( DT = 150,
AT = 5,000)
Cumulative # of Demand Centers
Scenario 1 ( Low) = 25
Scenario 2 ( Base) = 12
Scenario 3 ( High) = 8
65
Figure 2.3.8. Percent of statewide hydrogen demand captured
This figure indicates that the “ low” threshold captures more of the hydrogen demand
within Ohio. In particular, it captures more demand in small and large demand centers
because less small demand centers are eliminated and larger centers expand in size.
Although this scenario does result in a 36% increase in the capture of demand over the
“ base” scenario, it requires infrastructure to be installed to over twice as many demand
centers. The “ high” scenario captures only 50% of the demand met by the “ low”
threshold. Figure 2.3.9 indicates the percent of total Ohio land area captured within
each group.
6.21%
1.72%
5.66% 5.57%
1.17%
4.78%
5.97%
4.97%
9.39%
44.49%
30.32%
0.00%
2.09%
4.45%
16.39%
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
~ 5000 5000 ~ 10000 10000 ~ 20000 20000 ~ 40000 40000 ~
Category: Size of H2 Demand Center ( kg H2/ day)
% of Total Ohio H2 Demand
Scenario 1
( DT = 50,
AT = 1,000)
Scenario 2
( DT = 100,
AT = 3,000)
Scenario 3
( DT = 150,
AT = 5,000)
Cumulative % of Total Demand
Scenario 1 ( Low) = 64%
Scenario 2 ( Base) = 47%
Scenario 3 ( High) = 32%
66
Figure 2.3.9. Percent of statewide land area captured
The “ low” threshold scenario captures significantly more land area, especially in small
and large demand centers. This result suggests that the “ low” threshold would require
more extensive intracity infrastructures, resulting in higher costs. The “ low” scenario
occupies 83% more land than the “ base” scenario and 238% more land than the “ high”
threshold. Figure 2.3.10 illustrates the spatial distribution of the three scenarios given
10% market penetration.
1.17%
0.27%
0.89%
0.62% 0.75% 0.52%
0.76%
5.75%
0.16%
0.72%
2.81%
0.41%
0.24%
0.00%
1.18%
0%
1%
2%
3%
4%
5%
6%
7%
~ 5000 5000 ~ 10000 10000 ~ 20000 20000 ~ 40000 40000 ~
Category: Size of H2 Demand Center ( kg H2/ day)
% of Total Ohio Land Area
Scenario 1
( DT = 50,
AT = 1,000)
Scenario 2
( DT = 100,
AT = 3,000)
Scenario 3
( DT = 150,
AT = 5,000)
Cumulative % of Total Land Area
Scenario 1 ( Low) = 8.8%
Scenario 2 ( Base) = 4.8%
Scenario 3 ( High) = 2.6%
67
Figure 2.3.10. Spatial distribution of demand centers given the three threshold
scenarios
This figure illustrates how the demand centers expand in size and number as the
thresholds are lowered. In comparison with the “ base” scenario, the “ low” scenario
captures 36% more of the state hydrogen demand, but requires service to twice as many
demand centers and 83% more land area. However, it does capture 64% of hydrogen
demand in less than 10% of the land area. In contrast, the “ high” scenario captures 32%
less hydrogen demand than the “ base” and addresses 46% less land area and 33% fewer
demand centers. It captures 32% of the hydrogen demand in 2.6% of the land area. The
68
“ base” scenario captures 47% of hydrogen demand in less than 5% of the land area. In
the future, it will be interesting to calculate the levelized cost of hydrogen under each
scenario in order to determine which thresholds are the most cost- effective. For
example, although the “ low” threshold scenario requires extensive expansion of
infrastructure for a relatively small gain in the capture of hydrogen demand, it may
allow for the capture of economies of scale, resulting in favorable economics. An
analysis of cost will be conducted in the near future.
2.3.3. Infrastructure Components
Determination of Hydrogen production capacity - Central plant
Information about coal electricity plants in Ohio is obtained from the EPA’s eGrid
database, including plant data such as electricity output, coal input, CO2 emissions, and
plant efficiency. This information is used to predict the hydrogen production capacity
for each of these locations ( see Table 2.3.3). This H2 capacity can be calculated a
number of different ways. One key assumption is that each coal plant site is currently
limited with respect to its coal supply and handling capacity. This assumption will limit
the ability of these coal plants to increase their coal inputs significantly. The first is to
constrain only the existing coal input and re- direct that feedstock from electricity
production to hydrogen production. Given a 65% conversion coal- to- H2 efficiency, the
hydrogen production capacity can be easily calculated.
A second strategy is to maintain both the coal input and electricity output, while co-producing
hydrogen. Advanced integrated coal gasifier combined cycle ( IGCC)
technology can dramatically increase coal- to- electricity efficiency, requiring less coal
input for the same electricity output. This allows the excess coal input to be converted
to hydrogen by oversizing the gasifier and diverting a stream of hydrogen to be utilized
as a transportation fuel. This can significantly reduce the capital costs associated with
hydrogen production as compared to a standalone plant of the same H2 capacity.
Other strategies can be used to determine the potential hydrogen capacity of existing
coal plant sites as well as other non- existing sites. For the initial analysis, the potential
production plant locations was limited to existing utility coal plants over 100 MW
electricity output and conversion to a dedicated H2 production facility was considered.
In this analysis, coal plants producing mostly hydrogen with some co- production of
electricity were considered with associated capture and compression of CO2 for
sequestration based upon Kreutz et al 2002. The majority of the energy output is in the
form of hydrogen (~ 97%) with the remaining energy output as electricity (~ 3%). The
gross electricity production is about 14% of the total output, but electricity demands
within the plant for compression lead to lower net electricity output. The plant has a
coal input to H2 output efficiency of 66% and an overall net efficiency of 69% ( coal to
H2 + electricity). Hydrogen is compressed to approximately 1000 psi for transport.
69
Table 2.3.3. Data for utility coal plants over 100MW electricity output and estimates
for H2 capacity given complete coal conversion and efficiency
improvements.
ID Plant Name
H2 Capacity - Full
Conversion
( kg/ day)
Plant
Efficiency
H2Capacity - IGCC
Conversion ( 42%)
( kg/ day)
CO2
emissions
( kg/ kWh)
1 ASHTABULA 138,530 40.14% 2,578 0.87
2 AVON LAKE 429,126 34.81% 30,840 1.01
3 BAY SHORE 445,308 32.33% 43,042 1.08
4 CARDINAL 1,449,802 36.62% 77,933 0.96
5 CONESVILLE 1,556,646 33.82% 127,330 1.04
6 EASTLAKE 791,977 32.27% 77,023 1.08
7 GEN J M GAVIN 2,505,969 32.11% 247,850 1.09
8 HAMILTON 51,944 23.22% 9,756 1.46
9 KAMMER 568,833 36.75% 29,841 0.95
10 KYGER CREEK 1,088,682 35.74% 68,132 0.98
11 LAKE SHORE 70,761 23.73% 12,928 1.47
12 MIAMI FORT 1,294,250 31.38% 137,507 1.12
13 MITCHELL 1,213,062 35.39% 80,196 0.99
14 MOUNTAINEER ( 1301) 1,004,843 35.53% 65,047 0.99
15 MUSKINGUM RIVER 1,147,087 35.62% 73,158 0.98
16 NILES 188,992 30.64% 21,474 1.14
17 O H HUTCHINGS 150,727 28.51% 20,334 1.21
18 PHIL SPORN 909,740 36.42% 50,726 0.96
19 PICWAY 67,121 30.36% 7,814 1.15
20 PLEASANTS 1,053,605 34.50% 79,068 1.01
21 R E BURGER 292,972 32.31% 28,389 1.08
22 RICHARD GORSUCH 247,459 26.98% 37,164 1.30
23 W H SAMMIS 1,861,267 33.06% 166,415 1.06
24 WILLOW ISLAND 249,229 28.98% 32,457 1.18
TOTAL 18,777,930 33.87% 1,527,001 1.04
Carbon capture
In the plant configuration chosen, 92% of the CO2 is captured and sequestered while
approximately 8% is emitted to the atmosphere. This system uses “ conventional”
technologies for gas separation: glycol absorption for CO2 capture and pressure swing
adsorption for hydrogen purification. Kreutz et al. also describe advanced technologies
for separations including an inorganic membrane for coupled separation and water gas
shift reaction. CO2 is separated from the syngas stream after the WGS reactors using an
absorption tower with ( Selexol). The CO2 stream is dehydrated and compressed to 2200
psi creating a supercritical stream for transport.
70
On- site production
Natural gas steam methane reformers ( NG- SMRs) are used for producing hydrogen at
the refueling station. Small natural gas steam reformers are curently being developed by
a number of companies including H2Gen, Plug Power, Air Products, and Ztek for
stationary and transportation fuel cell applications. These small reformers, which are
combined with compressors, hydrogen storage tanks and hydrogen fuel dispensers form
the basis of a stand- alone hydrogen vehicle refueling station.
2.3.4. Infrastructure Optimization
Network Analysis
One of the main components of this analysis is the determination of the lowest cost
network for supplying hydrogen to the demand clusters from a hydrogen production
plant that is located at one of the existing coal plant sites. The network components, as
described in Tasks 1.4 and 1.5, include the identified demand clusters, existing energy
rights- of- way ( i. e. natural gas pipelines), coal plants, and CO2 sequestration sites ( some
of which are shown on Figure 2.3.12). The optimization of this network in order to
minimize the cost of hydrogen production, distribution and refueling is a critical
component of this model.
Optimization modeling
The spatial design of the infrastructure, i. e. the location( s) of hydrogen production plant
locations, sequestration sites and the network ( pipelines) for hydrogen distribution to the
demand clusters is carried out by a network optimization algorithm. This optimization
routine minimizes the total pipeline length to connect all demand clusters to one or more
H2 production plants. It is based upon the minimal spanning tree algorithm, which
minimizes pipeline length from a number of potential hydrogen production plant
locations ( sources) to a series of demand clusters ( sinks). The main constraint is that
each of the sinks must be connect to a source, either directly or through another sink.
The input of this optimization routine is a matrix that specifies the shortest network
length for every given pair of nodes ( i. e. sources or sinks). To generate this matrix
( shown in Table 2.3.4), an algorithm was developed using the GIS network analyst.
Network methods used in this study
A network is an interconnected or interrelated chain, group, or system. It can be
represented conceptually and digitally by nodes and links ( Figure 2.3.11). Nodes
represent intersections, interchanges, or confluence points on the network, and links
represent transportation or transmission paths between nodes. Types of nodes can be:
stops, which are locations visited along a path, or; centers, which are locations where
there is a supply or attraction.
71
Figure 2.3.11. Conceptual Network Structure
ArcView 3.2 Desktop GIS was selected for the GIS- based network analysis. This
software package was chosen for being readily available, customizable, expandable, and
familiar to GIS users in the research group. ArcView, with the Network Analyst
extension, allows the user to solve many network- based problems, such as: finding the
most efficient travel route from one location to the next; generating travel directions;
finding the closest service facility to a market; defining service areas based on travel
time; finding the best location for a service center; and determining the number of trips
that will be generated from one location to another.
This project used Network Analyst to find the shortest route between two locations
along the network. In this case, we calculated the shortest routes between all of the coal
power plants over 100MW capacity ( sources) and the centers of the demand clusters
( sinks) along the natural gas pipeline right- of- way ( network). The output is a table of
the shortest distances between all sources and sinks. Given more time, the GIS software
could be re- programmed to optimize routes and locations for a more seamless modeling
effort. However, pre- existing optimization routines were used for calculating the best
routes between locations and minimizing network costs.
Table 2.3.4. Distance matrix for network optimization indicating distance between
demand clusters to other demand clusters and coal plants
72
Demand Clusters
3 7 13 16 23 43 48 49 52 57 62 65
101 283 72.18 171.2 139.9 163.8 288 406.4 308 428 446.7 458.8 481
102 158.2 104.9 152 81.06 106.8 181.3 299.7 201.3 321.3 340 352.1 374.3
103 21.36 209.9 257 186 211.8 226.5 344.9 246.6 366.5 385.2 397.3 419.5
104 363.3 242.3 186.5 181.2 150.1 218.2 311.9 237 333.5 337.4 349.4 371.6
105 266.2 195.8 161.8 131.4 101 69.58 205.3 107 227 245.7 257.8 280
106 229.5 17.19 128.6 90.43 114.3 234.4 352.8 254.5 374.4 393.1 405.2 427.4
107 406.5 338.8 304.2 273.8 243.3 194.2 258.2 189.3 279.8 283.7 295.7 317.9
108 405.2 393.6 411.9 350 337.3 204.4 89.04 156.3 53.8 25.39 40.27 21.93
109 338 218.9 163 156 124.8 192.9 279.6 204.7 301.2 305.1 317.2 339.4
Coal 110 408.7 341.1 306.5 276.1 245.6 196.5 260.5 191.5 282.1 285.9 298 320.2
Plants 111 191.6 67.37 117 57.83 83.61 198.1 316.5 218.2 338.1 356.8 368.9 391.1
112 445.1 433.4 451.8 389.9 377.1 244.2 134 196.2 98.72 70.52 73.13 54.79
113 336.7 217.5 161.7 154.6 123.5 191.6 278.3 203.4 299.9 303.7 315.8 338
114 396.9 329.3 294.7 264.3 233.8 184.7 248.7 179.7 270.3 274.1 286.2 308.4
115 358.3 250.9 216.3 185.9 155.4 146 210 141 231.6 235.4 247.5 269.7
116 262.3 114.3 49.47 81.53 83.2 229.3 357.2 258.8 378.8 397.5 409.6 431.8
117 385 373.4 391.7 329.8 317.1 184.2 57.48 136.1 22.23 6.17 31.08 53.28
118 396.9 329.3 294.7 264.3 233.8 184.7 248.7 179.7 270.3 274.1 286.2 308.4
119 285.2 273.6 291.9 230 217.2 72.96 98.97 51.77 120.6 125.8 137.9 160.1
120 390 291 242.1 226 195.5 177.8 241.7 172.8 263.4 267.2 279.3 301.5
121 336.1 215.2 159.3 154.1 122.9 191 284.7 209.9 306.4 310.2 322.3 344.5
122 366.2 272 235.3 206.9 176.5 154 218 149 239.6 243.4 255.5 277.7
123 309.8 161.8 97.94 126.1 122.1 207.3 343 244.7 364.6 379.5 391.6 413.8
124 389.6 290.6 241.7 225.6 195.1 177.4 241.3 172.4 263 266.8 278.9 301.1
3 0 212.3 259.3 188.4 214.2 228.9 347.3 248.9 368.9 387.6 399.7 421.9
7 212.3 0 111.4 73.24 97.15 217.2 335.6 237.3 357.2 375.9 388 410.2
13 259.3 111.4 0 78.56 80.23 226.1 354 255.6 375.6 394.3 406.4 428.6
16 188.4 73.24 78.56 0 32.13 173.6 292.1 193.7 313.7 332.4 344.5 366.7
23 214.2 97.15 80.23 32.13 0 160.9 279.3 181 300.9 319.6 331.7 353.9
Demand 43 228.9 217.2 226.1 173.6 160.9 0 159.9 69.65 181.5 186.7 198.8 221
Clusters 48 347.3 335.6 354 292.1 279.3 159.9 0 98.34 35.25 63.66 88.56 110.8
49 248.9 237.3 255.6 193.7 181 69.65 98.34 0 120 138.7 150.7 172.9
52 368.9 357.2 375.6 313.7 300.9 181.5 35.25 120 0 28.41 53.31 75.51
57 387.6 375.9 394.3 332.4 319.6 186.7 63.66 138.7 28.41 0 33.64 47.31
62 399.7 388 406.4 344.5 331.7 198.8 88.56 150.7 53.31 33.64 0 49.93
65 421.9 410.2 428.6 366.7 353.9 221 110.8 172.9 75.51 47.31 49.93 0
73
Figure 2.3.12. Nodes and paths for the hydrogen distribution infrastructure network
including demand clusters, coal plants and potential hydrogen
pipeline locations
Network Cost Minimization
The goal of minimizing network costs is approximated by minimizing the total length of
pipeline passing through a specified number of sources and connecting all the sinks.
The minimal spanning tree is by definition the shortest connection path to meet the
specified criteria and by definition reaches links all of the specified nodes without any
loops ( or duplicative pipelines).
74
The optimization routine chooses the shortest path from the sources to all sinks. The
sink that is chosen in this shortest path now becomes a “ source” and again the shortest
path between the new set of sources and remaining sinks is chosen. This process is
repeated until all sinks are connected together via pipeline to other sources or sinks.
Once the locations of the hydrogen production plants and pipelines are determined
( shown in Figure 2.3.13 and Table 2.3.5), the capacity of the hydrogen production plant
and flow through the pipelines is determined and costs can be calculated.
Table 2.3.5. Decision table indication which pipelines are built for the minimal
spanning pipeline network for one coal plant source.
Demand Clusters
3 7 13 16 23 43 48 49 52 57 62 65
Coal Plant 105 0 0 0 0 1 1 0 0 0 0 0 0
3 0 0 0 0 0 0 0 0 0 0 0 0
7 0 0 0 0 0 0 0 0 0 0 0 0
13 0 0 0 0 0 0 0 0 0 0 0 0
16 1 1 1 0 0 0 0 0 0 0 0 0
23 0 0 0 1 0 0 0 0 0 0 0 0
Demand 43 0 0 0 0 0 0 0 1 0 0 0 0
Clusters 48 0 0 0 0 0 0 0 0 1 0 0 0
49 0 0 0 0 0 0 1 0 0 0 0 0
52 0 0 0