U!'.IIVER2,iTY OF HAWA\' I LIBRARY Hydrogen Production from Glycerin Reforming Thesis submitted to the graduate division of the University of Hawai'i in partial fulfillment for the degree of Master of Science In Bioengineering August 2006 By Aurelien MD DOUETTE Thesis Committee Scott Q. Tum, Chairman Charles Kinoshita Stephen Masutani
96
Embed
Hydrogen Production from Glycerin Reforming · PDF file_~3-1 Steam Reforming ... __ .3-6 Types of Reformers: ... __ 4-1 Reformer
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
U!'.IIVER2,iTY OF HAWA\' I LIBRARY
Hydrogen Production from Glycerin Reforming
Thesis submitted to the graduate division of the University of Hawai'i in partial fulfillment for the degree of
Master of Science
In
Bioengineering
August 2006
By
Aurelien MD DOUETTE
Thesis Committee
Scott Q. Tum, Chairman Charles Kinoshita Stephen Masutani
We certify that we have read this thesis and that, in our opinion, it is satisfactory in scope and quality as a thesis for the degree of Master of Science in Bioengineering.
HAWN Q111 .H3
no. 4081
THESIS COMMITTEE
Chairperson
II
Table of Contents
List of Figures .................................................................................................................... v
List of Tables ................................................................................................................... vii
Appendix A ...................................................................................................................... 73 _~A-l Effect Calculation .......................................................................................... 73 _~A-2 A Model Development for Third Order Design ............................................ 75 _~A-3 Example Calculating effect ofX\: ~\ ............................................................ 75
Figure 1: Picture of unused catalyst piece (left) and used piece (right) ............................ 18
Figure 2: Photograph of catalytic refonning test stand. .................................................... 19
Figure 3: Schematic of glycerin, nitrogen and oxygen injection system .......................... 20
Figure 4: Glycerin feed pump calibration curve for increasing and decreasing flow ....... 21
Figure 5: Thermocouples and catalyst positions inside the reformer. (Distances in cm). 22
Figure 6: Sample conditioning system for reformate analysis by GC ............................. 24
Figure 7: Glycerin reformate gas composition predicted by chemical equilibrium analysis as a function of temperature (O/C = 0.5, SIC = 1) ................................................ 29
Figure 8: Refonning conditions selected for experimental investigation ........................ 30
Figure 9: Run 9 reformate gas concentrations .................................................................. 35
Figure 10: Run 12 pressure and temperature profile for thermocouples 1,2, and 3 ........ 38
Figure 11: H2 and CO molar concentration in reformate gas for conditions 1 through 12 . ............................................................................................................................... 44
Figure 12: Run 18 reformer with WGS reformate gas composition at 380 (left) 1420°C (right). Samples 1, 2, 7 and 8 are reformer gas composition ................................ 49
Figure 13: Gas composition for reformer run at condition 12 with WGS set at 420°C .. 50
Figure 14: H2 yield (mole H21 mole glycerin) for each sample during condition 12 test. 52
Figure 15: Condition 12 methane yield (mole Cl4' mole glycerin) for each sample ..... 53
Figure 16: Methane production as a function of OIC ....................................................... 55
Figure 17: (1 bis) Photograph of use catalyst pieces from run 12. Blackened pieces on the left with carbon deposited on them were located nearest to the reformer inlet. ............................................................................................................................... 56
Figure 18: Used and new in-line sintered metal filters that are used downstream of the reformer ................................................................................................................. 57
v
Figure 19: Crude glycerin refonnate gas composition ..................................................... 59
Figure 20: Reformer pressure during test 20 using crude glycerin as feedstock. (Feeding started at 12:00 and ended at 14:24) ..................................................................... 60
Figure 21: Picture of reformer tube plugged by carbon built up from crude glycerin reforming test. ....................................................................................................... 62
Figure 22: System pressure for crude glycerin feed contaminated with methanol (reformer set at condition 12) ............................................................................... 64
Figure 23: System pressure and temperatures for crude glycerin feed contaminated with NaCI (reformer set at condition 12) ...................................................................... 65
Figure 24: System pressure for test conducted with crude glycerin feed contaminated with 0.5% (weight) NaOH (reformer set at condition 12) ............................................ 66
Figure AI: Original design and path of steepest ascent, showing H2 yield for each condition (mole H2 / mole glycerin) ..................................................................... 79
Figure Bl: Temperature profile at the feeding tube centerline in thereformer (reformer entrance at 28 cm) ................................................................................................. 87
Figure B2: Regression of 40% glycerin / 60% water mixture viscosity as a function of temperature ........................................................................................................... 89
vi
List of Tables
Table 1. Results of ultimate analysis of crude glycerin samples obtained from Pacific Biodiesei ................................................................................................................. 6
Table 2: GC retention time for reformate constituents ..................................................... 23
Table 3: Initial runs conditions, in real and coded units ................................................... 31
Table 4: Reformate gas yield, hydrogen concentration and setting values for conditions 1 through 10 ............................................................................................................. 40
Table 5: Effects of experimental variables and their interactions on reformate gas H2 concentration and yield (coded units) ................................................................... 42
Table 6: Experimental conditions determined from response surface analysis and resulting H2 and CO production ............................................................................ 43
Table 7: H2 yield and associated error (mole H2 I mole glycerin) .................................... 45
Table 8: Reformate composition for runs 13 and 14 ........................................................ 47
Table 9: Mole H2 per mole glycerin using the WGS at different temperatures, with refonner operated at Condition 12 (O/C = 0.5, SIC = 2.2 and T= 804°C) ........... 47
Table 10: Gas composition from refonner (O/C=O.5, S/C=2.2, T=804°C) with WGS reactor (set point T WGs=420°C, internal reactor temperature = 369°C) ............... 50
Table 11: Methane production for run 1 through 8 ........................................................... 54
Table AI: Refonner experimental variable values and coded unit values for a 23
experimental design with two center points, coded unit values for two and three variable interactions, and average hydrogen concentration at each condition ...... 74
Table A2: Summary of effects! model coefficients for variables and their interactions .. 76
Table A3 Comparison of equation 21 model of refonning reaction vs. real data. ........... 77
Table A4: Values for future conditions along the path of steepest ascent. Point 4 corresponds to condition II and condition 12 lies between point 4 and 5 ........... 78
Table Bl: Extrapolated data for Il, G, Re, and Oh at elevated temperature .................. 89
vii
I-Abstract
Following a factorial experimental design, a series of tests were performed to
investigate the effects of operating parameters; oxygen to carbon ratio (O/C), steam to
carbon ratio (SIC) and temperature (T), on reforming glycerin to a H2 rich gas. A
mathematical model defining the effect of those three variables was derived., and used for
improving the reaction hydrogen yield. From the range of experimental conditions tested
it was concluded that OIC, as well as the interaction between OIC and temperature had
the most important effects on H2 yield. 4.5 mole of hydrogen were produced per mole of
glycerin at experimental conditions ofO/C=I, S/C=2.2, and T=804°C. This is 65% of the
maximum theoretical yield, and 90% of the yield predicted by thermochemical
equilibrium. 1.4 moles of carbon monoxide per mole of glycerin were also produced.,
presenting a potential for an additional 1.4 mole hydrogen per mole glycerin. A water gas
shift reaction was then used., and its operating temperature optimized, in order to convert
the reformate gas CO into hydrogen by combining it with water. Results were satisfying,
with a fmal yield of 5.3 moles H2 I mole glycerin, which is 75% of the maximum
stoichiometric hydrogen yield. Crude glycerin, obtained from biodiesel production, was
fmally tested (without a water gas shift) as a feed to compare it with pure glycerin used
throughout the tests. The initial results were very encouraging, almost identical to those
of pure glycerin, but carbon formation quickly became a problem. Possible contaminants
causing the coking may include methanol, chloride and sodium cations, and free fatty
acids, all present in crude glycerin as byproducts ofbiodiesel synthesis.
I
2 - Introduction
Society is increasingly looking for clean and renewable fuels to offset the
negative effects of fossil fuel use including greenhouse gas emissions and consumption of
limited resources. This is even truer for a community like Hawaii that relies on fossil
energy resources (-90%), [I] primarily imported oil, and has limited space for waste
disposal. Two possible renewable energy supplies are biodiesel and hydrogen. Biodiesel
is produced on two islands (Maui and Oahu) by the company Pacific Biodiesel and is
made from used cooking oil and grease trap waste obtained from Hawaii's many
restaurants. At present, this biodiesel is used by state vehicles and buses on the island of
Oahu in a blend with regular diesel. The most common mix is B20 (20% biodiesel and
80% diesel). One of the major difficulties facing wide spread biodiesel use on the
mainland U.S. is its cost (-$2.60/gal), which is significantly higher (depending on the
source of oil, size of the facility, etc.) than fossil fuel based diesel (-$2.00/gal). [2]
However, biodiesel in Hawaii is a good and competitive alternative due to soaring gas
prices as well as fossil fuel shipping cost to the islands. In the week of February 20,2005,
retail diesel prices at the pump were $2.69/gal (Chevron) whereas biodiesel from Pacific
Biodiesel was priced at $2.59/gal.
Annually, 180 million pounds of biodiesel are produced nationwide, generating
about 18 million pounds of glycerin. [3] The U.S. consumption of diesel for transportation
is 340 billion pounds per year and the U.S. demand for glycerin is 453 million pounds per
year. [4] Thus, the U.S. glycerin demand could be met from the production of 4.53 billion
The values of SIC, OIC, and T selected for condition 12 defined a point outside of the
experimental cube shown in Figure 8, where the model (equation 13) derived from the
first 10 tests may no longer be valid. In addition, the response surface around the point
defined by Condition 12 may no longer depend on the SIC, T, and OIC variables in the
same manner. To test this, two runs (13 and 14) were perfonned around condition 12.
46
keeping OIC (0.5) and T (804°C) constant, and setting SIC to 1.91 and 2.51, respectively,
to detennine its effect. Table 8 shows the test results.
Table 8: Reformate composition for runs 13 and 14
Condition 13 14
OIC 0.5 0.5
Variable SIC 2.5 1.9
T,OC 804 804
molelmole glycerin H2 CO C02
4.67 1.38 1.34 4.72 1.41 1.34
The effect of SIC turns out to be 0.05 (in coded units) (see Effect Calculation, Appendix
A3), which is very small and shows that changing SIC by a small amount can hardly
improve the production of hydrogen. This does not tell us however how SIC affects
catalyst deactivation, coking, CO and Cf4 production, and other performance indicators.
6-6 Water Gas Shift Results
After optimizing the reforming process, a water gas shift reactor was added
downstream of the reformer to investigate its effect on hydrogen production. The water
gas shift (WGS) reacts CO with H20 to produce C02 and H2 in a low temperature
(-400°C) reactor. The experiments used a monolithic WGS catalyst from Nextech based
on nano-particle, ceria-based, mixed oxide support. In the tests of the shift reactor
performance, the temperature was the only variable that was changed. The gas stream
exiting the reformer flowed immediately to the shift reactor, the two being in series
(Figure 2). For the investigation of the shift reactor, the reformer was operated with
operating variables set to those used for Conditions 12. Five tests were performed (16-20)
47
to reach an optimized was temperature. For each test, the reformer was initially run by
itself to make sure that reformate composition was similar to that determined earlier for
condition 12. When the reformer was operating stably, the reformate was directed to the
water gas shift reactor. Table 9 shows the operating temperature for the shift reactor and
the amount of gas produced for each of the tests.
Table 9: Mole H2 per mole glycerin using the was at different temperatures, with ondition 12 (O/C = 0.5, SIC = 2.2 and T= 8 reformer operated at C 04 °C)
Temperature mole H2/mole Run (0e) glvcerin IS 320 4.31 16 340 4.99 17 360 5.38 18 380 5.9 18 420 5.96 19 420 5.28
Up to two shift reactor temperature conditions were run during a single test period. Each
temperature was only tested for a little over one hour to limit the effect of any change in
the reformer performance over time. Three analyses were obtained per condition. The
results were also dependent on the reformer operation, which were not always stable
especially at the beginning of a test. This caused the results for the was to have greater
variability than the tests using only the reformer. The shift reactor tests for temperatures
of 380 and 420°C were run during the same test period, with each condition run for about
an hour. Figure 12 shows the measured gas compositions for each sample, which were 25
Equation [15] indicates that lowering OIC increases methane production. the likely result
of catalyst deactivation. This was supported by runs I I (O/C = 0.75) and run 12 (O/C =
0.5) which produced 0.04 and 0.14 mole C~ I mole glycerin respectively. Figure 16
54
summarizes these results. Morc tcsts would be needed to better under tand the effect of
oxygen content on catalyst deactivation.
.. :r u
1.6
1.4
1.2
~ 0.6
0.4
0.2
1.2 0.9
OIC
Figure 16: Methane production as a function of ole
55
0.7 0.5
, (
•
, 5cm
)
Figure 17: Photograph of use cata lyst pieces from run 12. Blackened pieces on the left with carbon deposi ted on them were located nearest to the refomler inlet.
Carbon deposits are often responsible for catalyst deactivation and were readily
apparent when the catalyst pieces were recovered from the reformer after a test was
completed. Figure 17 shows the six catalyst pieces recovered from run 12. The catalyst
pieces on the left were the closest to tbe inlet end of the reactor and tbe pieces on tbe
right were further downstream. The pieces on the left are much darker and covered with
carbon, affecting their abi lity to enhance the reforming reaction. Over the length of each
run, the catalysts farthest downstream kept their original, grayish color (and were sti ll
apparently quite active). A possible reason why hydrogen yield did not drop significantly
over the test duration is that the catalyst may have been in excess of amount req uired for
56
complete reaction or the rate of carbon deposi tion may not have been constant over the
test duration. One possible way to test this hypothesis would be to reduce the number of
catalyst pieces and repeat the experiment.
" I
,
I
, . , •
2cm
,
Figure 18: Post-test and pre-test in-line sintered metal filters that are located downstream of the reformer.
Figure 18 shows tbe in-line, sc intered-metal fi lter elements placed in the process line
after the reformer (before the WGS and the GC). The filter element on the right is new
while the one on the left has been used in a reforming test. The used filter is blackened
from carbon collected from the reformate stTeam exiting the reformer. Although the
amount of carbon recovered fro m the refonner and the filter after each test was not
57
significant (between 0.36g to Ig) and accounted for less than 1 % of the total carbon
contained in the glycerin feed, it impacted the reformer by increasing operating pressure
over the test duration. Pressure started at about 20 kPa for each run, and slowly worked it
way up to ~48 kPa. A higher SIC was expected to reduce coking and slow the catalyst
deactivation. From the amount of carbon collected in the reformer and the pressure
recorded in the reformer, it was not possible to determine a clear effect of SIC on the
carbon formation for the duration of these tests. The pressure inside the reformer varied
from 27 to 48 kPa for each condition (with peaks as high as 69 kPa on some occasions
during the earlier tests, probably due to poor feed injection and sudden vaporization of
water).
6-8 Crude Glycerin Test
Crude glycerin is the eventual final target feed to be used for reforming. Its
characteristics were discussed in the experimental part. Reforming of crude glycerin was
tested at the optimal operating condition, (Le. condition 12), without using the WGS.
The conditions were as follow (same as run 12):
O/C = 0.5
SIC = 2.21
Tset = 804°C
58
5 -- ---~
4.6
4
3.6
t 3
L.5 .. a 1
2
1.6
t= : t • • ---- . ~- • 0.5
I I
H lI' ~E )\; H
0 0 2 3 4 6
sample.
Figure 19: Crude glycerin refonnate gas composition
... --
jI(
6 7
-+-H2 .___ CO
-+-C02 ""*"CH4 -+-C2H4
Figure 19 shows the refonnate composition for the crude glycerin test. The crude glycerin
feed rate was measured in the same way as pure glycerin. The hydrogen yield reached 4.4
moles H2 / mole crude glycerin, when the best yield reached with pure glycerin was 4.5
moles. This shows that the process worked well at first. The amount of CO and C02
produced was significantly lower though, with about 1 mole of each produced in the
second sample. The amount of methane produced, 0.23 moles per mole crude glycerin,
was almost twice as much as in run 12 (0.14 moles per mole glycerin). In addition, C2H.
was produced at levels equal to methane, where previously it was almost negligible in all
runs performed using pure glycerin. The hydrogen yield decreased quickly after the
second sample (-0.41 mole H2 per mole glycerin per hour) and the amount of CH. and
Figure 22: System pressure for glycerin feed contaminated with methanol (reformer set at condition 12)
The experiment performed well with the methanol, with the pressure staying within
normal range «34 kPa). The hydrogen yield averaged 4.7 mole H2 per mole glycerin,
which is close to the yield obtained from condition 12 with no contaminants.
Figure 23 shows the system pressure for a test of glycerin contaminated with
sodium chloride (0.5% by weight) using condition 12 reformer settings.
64
1000 ---- - ---------- 60
~Tl
900 .1'2
13 60
800
700 40
~ 800
~ Ii!
~ 500 aoi! l i E ~ 400 ...
20 300
2DO 10
100
0 0
13:00 14:00 15:00 16:00 17:00
TIme
Figure 23: System pressure and temperatures for glycerin feed contaminated with NaCI (reformer set at condition 12)
Although the reformer temperatures were stable, the pressure inside the system rose to
over 30 kPa in about I hour. The pressure then started to increase more rapidly causing
the experiment to be shut down, the same as happened with crude glycerin.
65
70,-----
60
50
20
10
o~~~~----~----~------------------~ 13:00 14:00
Time
15:00
Figure 24: System pressure for test conducted with glycerin feed contaminated with 0.5% (weight) NaOH (reformer set at condition 12)
Similar system performance occurred in a test conducted with pure glycerin
contaminated with sodium hydroxide [0.5 % by weight], as shown in Figure 24. The
pressure climbed to over 45 kPa very quickly. Both tests using pure glycerin
contaminated with sodium compounds at condition 12 have shown a fast pressure rise,
and in both cases, the reformer was found to contain a lot of carbon built-up on the
catalyst and the reformer surfaces. This may suggest that sodium hydroxide and sodium
chloride are possible components causing coking.
66
7- Conclusion
A statistical approach to study and improve the reforming of glycerin to produce
hydrogen was undertaken with the goal in mind to reform crude glycerin obtained from
the transesterification of vegetable oil in the production of biodiesei. Using a fIXed bed
reactor of nickel based catalyst and a 23 factorial design, the effects of the reformer
temperature (T), oxygen to carbon ratio (O/C), and steam to carbon ratio (SIC) were
studied and quantified, so that they could be used to improve the hydrogen yield. The two
levels ofT, OIC, and SIC were 770 and 8S0·C, 0.9 and 1.2. and 2 and 2.7, respectively,
and the hydrogen yield was successfully modeled in terms of these three variables. Over
the range of conditions tested, oxygen to carbon ratio was found to have the greatest
effect. The hydrogen yield obtained, 4.6 mole H2 I mole of glycerin, was as high as 6S%
of the stoichiometric maximum yield (7 mole H2 I mole glycerin) and 8S% of the
maximum theoretical yield based on chemical equilibrium (S.4 mole H2 I mole glycerin).
A path of steepest ascent method was used to arrive at the reformer operating conditions
that produced the greatest hydrogen yield - an oxygen to carbon ratio of 0.5, a steam to
carbon ratio of2.2 and a reformer internal temperature of 760·C. A water gas shift reactor
was then added after the reformer to convert carbon monoxide to hydrogen. Operating the
water gas shift reactor at 369"C yielded S.3 mole H2 per mole glycerin. Crude glycerin
was then tested in the reformer. The initial results were very close to that of pure
glycerin, but the hydrogen yield quickly decreased due to catalyst deactivation and
coking. Carbon accumulation in the reformer plugged the system and caused the pressure
to rise above IOpsig necessitating system shut down. Tests conducted with pure glycerin
doped with potential contaminants, methanol, NaOH, and NaCl, suggest that Na may be
67
one of the elements responsible for the reduced performance. Methanol did not
negatively affect the reforming process. Free fatty acid not converted during the
transesterification process could also be responsible for poor system performance using
crude glycerin but their effect was not investigated. All stated objectives (page 6, section
2-5) of the this research were achieved except objective 5, which was only partially
achieved since the effect of free fatty acids was not investigated.
68
8 - References
[1] Hawaii State Government Website (www.state.hi.us)
[2] US Government Energy Information Website (www.eiadoe.govD
[3] National Biodiesel Board Website (www.biodiesel.org)
[4] Innovation Group Website (www.the-innovation-group.com)
[5] Ram Ramachandran and Raghu K. Menon, An overview of industrial uses of hydrogen, International Journal of Hydrogen Energy. Volume 23. Issue 7, July 1998, Pages 593-598
[6] Jorge Ancheyta, Mohan S. Rana and Edward Furimsky, Hydroprocessing of heavy oil fractions, Catalysis Today. Volume 109, Issues 1-4.30 November 2005. Pages 1-2
[7] R. Prins, M. Egoro, A. Rothlisberger, Y. Zhao, N. Sivasankar and P. Kukula, Mechanisms ofhydrosulfurization and hydrodenitrogenation, Institutefor Chemical and Bioengineering. Swiss federal Institute of Technologies. November 2005
[8] Fei Xiang Long and BOrje S. Gevert , Modeling initial decay of hydro de metallization catalyst with simultaneous adsorption and reaction mechanism, Journal of Catalysis. Volume 222. Issue 1.15 February 2004. Pages 1-5
[9] Francisco Alcaide, Pere-Lluis Cabot and Enric Brillas, Fuel cells for chemicals and energy cogeneration, Journal of Power Sources, Volume 153, Issue 1, 23 January 2006, Pages 47-60
[10] Matthias Duwe, Hydrogen Technology Overview, Climate Action Network Europe Workshop, September 2003
[II] S.A. Grigoriev, V.1. Porembsky and V.N. Fateev, Pure hydrogen production by PEM electrolysis for hydrogen energy, International Journal of Hydrogen Energy, Volume 31. Issue 2. February 2006, Pages 171-175
[12] Bilge Yildiz and Mujid S. Kazimi, Efficiency of hydrogen production systems using alternative nuclear energy technologies, International Journal of Hydrogen Energy, Volume 31, Issue 1. January 2006. Pages 77-92
[13] Infotech Inc. "Biotechnology Company to Build Bioreactor for Low Cost Hydrogen Production" Biotech Week. February 2005 pp466
69
[14] Biospace News 10118/05, NanoLogix. Inc. UFEC) And Welch Foods Inc. Sign Agreement For Hydrogen Bioreactor (http://biotechxus.biospace.com/news storv.aspx?StorvID=21475020&full= I)
[IS] M. P. Rzayeva and O. M. Salamov, Photoelectrical plant for hydrogen and oxygen productions by water electrolysis under pressure, Renewable Energy. Volume 24. Issue 2. October 2001. Pages 319-326
[16] J. R. Benemann Feasibility analysis of photobiological hydrogen Production, International Journal of Hydrogen Energy. Volume 22, Issues 10-11. October-November 1997. Pages 979-987
[17] Glycerin safety Data Sheet (http://www.safetv.duke.edulmsdslProdPharmacylPhenol 5 Glycerin.pdf)
[18] Hazen research, inc. (www.hazenusa.com)
[19] Wuyin Wang, Scott Q. Tum, Vheissu Keffer, Aurelien Douette, Parametric Study of Authotherma1 Reforming of LPG, Hawaii Natural Energy institute. UniVersity of Hawaii
[20] Finn Joensen and Jens R. Rostrup-Nielsen, Conversion of hydrocarbons and alcohols for fuel cells, Journal of Power Sources. Volume 105. Issue 2. 20 March 2002. Pages 195-201
[21] Xueping Song and Zhancheng Guo Technologies for direct production of flexible Hz/CO synthesis gas, Energy Conversion and Management. Volume 47. Issue 5. March 2006. Pages 560-569
[22] Kaihu Hou and Ronald Hughes, The kinetics of methane steam reforming over a Nila-AIzO catalyst, Chemical Engineering Journal. Volume 82. Issues 1-3. 15 March 2001. Pages 311-328
[23] Mohammad Nurunnabi, Yuya Mukainakano, Shigeru Kado, Baitao Li, Kimio Kunimori, Kimibito Suzuki, Ken-ichiro Fujimoto and Keiichi Tomishige, Additive effect of noble metals on NiO-MgO solid solution in oxidative steam reforming of methane under atmospheric and pressurized conditions, Applied Catalysis A: General. Volume 299. 17 January 2006. Pages 145-156
[24] Jens Sehested, Four challenges for nickel steam-reforming catalysts, Catalysis Today. Volume 111. Issues 1-2, 15 January 2006. Pages 103-110
[25] Kelfin M. Hardiman, Cyrus G. Cooper, Adesoji A Adesina and Ruediger Lange, Post-mortem characterization of coke-induced deactivated alumina-supported Co-Ni catalysts, Chemical Engineering SCience. In Press. Corrected Proo/. Available online 4 January 2005
70
[26] S. Rakass, H. Oudghiri-Hassani, P. Rowntree and N. Abatzoglou, Steam reforming of methane over unsupported nickel catalysts, Journal of Power Sources, In Press, Corrected Proof, Available online 8 November 2005
[27] Joelmir A. C. Dias and Jose M. Assaf, The advantages of air addition on the methane steam reforming over Nily-Ah03, Journal of Power Sources, Volume 137, Issue 2, 29 October 2004, Pages 264-268
[28] Abayomi J. Akande, Raphael O. Idem and Ajay K. Dalai, Synthesis, characterization and performance evaluation of Nil Ah03 catalysts for reforming of crude ethanol for hydrogen production, Applied Catalysis A: General, Volume 287, Issue 2,22 June 2005, Pages 159-175
[29] Thomas Sperle, De Chen, Rune Lsdeng and Anders Holmen, Pre-reforming of natural gas on a Ni catalyst: Criteria for carbon free operation, Applied Catalysis A: General, Volume 282, Issues 1-2, 30 March 2005, Pages 195-204
[30] N. Laosiripojana and S. Assabumrungrat, Hydrogen production from steam and autothermal reforming of LPG over high surface area ceria, Journal of Power Sources, In Press, Corrected Proof, Available online 28 November 2005
[31] F. Melo and N. Morlanes Naphtha, steam reforming for hydrogen production, Catalysis Today, Volumes 107-108,30 October 2005, Pages 458-466
[32] Andrew E. Lutz, Robert W. Bradshaw, Leslie Bromberg and Alex Rabinovich, Thermodynamic analysis of hydrogen production by partial oxidation reforming, International Journal of Hydrogen Energy, Volume 29, Issue 8, July 2004, Pages 809-816
[33] Yu Yang, Jianxin Ma and Fei Wu, Production of hydrogen by steam reforming of ethanol over a NilZnO catalyst, International Journal of Hydrogen Energy, In Press, Corrected Proof, Available online 12 September 2005
[34] Praveen K.. Cheekatamarla and Alan M. Lane, Catalytic autothermal reforming of diesel fuel for hydrogen generation in fuel cells: I. Activity tests and sulfur poisoning, Journal of Power Sources, Volume 152, 1 December 2005, Pages 256-263
[35] J.R. Salge, G.A. Deluga and 1.D. Schmidt, Catalytic partial oxidation of ethanol over noble metal catalysts. Journal of Catalysis, Volume 235, Issue 1, 1 October 2005, Pages 69-78
[36] S. Tum, C. Kinoshita, Z. Zhang, D. Ishimura and J. Zhou, An experimental investigation of hydrogen production from biomass gasification, International Journal of Hydrogen Energy, Volume 23, Issue 8, August 1998, Pages 641-648
71
[37] Jose Comas, Fernando Marifto , Miguel Laborde and Norma Amadeo, Bio-ethanol steam reforming on NilAhOJ catalyst, Chemical Engineering Journal, Volume 98, Issues 1-2,15 March 2004, Pages 61-68
[38] Unni Olsbye, Oddrun Moen, Ase Slagtern and Ivar M. Dahl, An investigation of the coking properties of fixed and fluid bed reactors during methane-to-synthesis gas reactions, Applied Catalysis A: General, Volume 228, Issues 1-2, 28 March 2002, Pages 289-303
[39] Qiangshan Jing, Hui Lou, Liuye Mo and Xiaoming Zheng, Comparative study between fluidized bed and fixed bed reactors in methane reforming with C(h and 02 to produce syngas, Energy Conversion and Management, Volume 47, Issue 4, March 2006, Pages 459-469
[40] S. Czernik, R. French, C. Feik, and E. Chornet, Production of Hydrogen from Biomass-Derived Liquids, Proceedings of 2000 DOE Hydrogen Program Review
[41] J.P. Holman, Experimental Methods for Engineers, MaGraw-Hill Series in Mechanical Engineering, 2001
[42] G. E. P. Box, W. Hunter, J. S. Hunter, Statistic for Experimenters: An Introduction to Design, Data Analysis, and Model Building, Wiley Series in Probability and Mathematical Statistics, 1978
[43] A. H. Lefebvre, Atomization and Sprays, Combustion: An International Series, 1989
[44] N.E. Dorsey, Properties of Ordinary Water-Substances, New York, p184, 1940
72
Appendix A
A - Experimental design part 2: Calculation.
A-I EtTect Calculation
The initial experimental design initially shown in Figure 8 is reproduced below.
s,<c T
oond'8 / oolld'1
/ .----c . .:-OOM'5 /~
27 aeJJ /
- Replacing variables real units by coded units:
I oolld'g
:f) I .olld'10
Jt----;' colid'S
;'
.olld'l
lOS ,.:
OOM"
--_."
.o...r2
OlC
The cube can be seen as having a width of 2 units, the center of it (cond 9and 10) having
coordinates of (0, 0, 0).
The following names will be used for the coded variables:
XI =O/C
73
Replacing the real units by coded units would be done as foIIow:
XI = (O/C - 1.05) I 0.15 [17]
where 1.05 is the middle value for OIC and O. ISis the deviation on each side. Similarly
X2 = (SIC - 2.35) I 0.35
X3 = (T - 810) I 40
[18]
[19]
The values for OIC, SIC and T and their values in coded units are shown in Table AI.
Also included in the table are the interactions between the variables X .. X2 and X3, and
the hydrogen concentration obtained for each condition. From this table, we can calculate
the effects of each variable and their interaction on the hydrogen yield. Note that the
center points (conditions 9 and 10) are not used in determining the effects of the
variables.
Table AI: Refonner experimental variable values and coded unit values for a 23
experimental design with two center points, coded unit values for two and three variable interactions, and average hydrogen concentration at each condition.
T ~i Condition OIC SIC (Cl XI X2 X3 X1X2 X1X3 X2X3 X1X2X3
The model is always within I % of the real data, therefore, it was deemed accurate and
could be used for improving the hydrogen yield.
A4 Path of Steepest Ascent
After determining a model equation that defmes yield (or concentration), it can be
used to define the path of steepest ascent toward better yield. This path is defined by the
first order effects shown in equation 21 above. For each -6.73 units moved in the Xl
direction, X2 moves by -0.67 units and X3 moves -0.26 units or for every I unit moved in
X}, move 0.1 units in X2 and 0.04 units in X3. Following this path is called the path of
steepest ascent. Changing X}, X2, and X3 according to these results should produce a
higher hydrogen yield from the reformer. Table A4 shows the improved values of the
variables in coded units as well as their equivalent values in real values.
77
Table A4: Values for future conditions along the path of steepest ascent. Point 4 corresponds to condition 11 and condition 12 lies between point 4 and 5.