Professional Documents
Culture Documents
Paper INFUB1 JBM V4
Paper INFUB1 JBM V4
School of Business and Engineering Vaud, University of Applied Sciences Western Switzerland, 1401
Yverdon-les-Bains, Switzerland
2
HEPIA, University of Applied Sciences Western Switzerland, 1201 Geneva, Switzerland
3
Centre of Appropriate Technology and Social Ecology, Laboratories for Sustainable Energy Systems,
Langenbruck, Switzerland
4
Corresponding author Ph: +41265577594, Fax:+41265577579,
e-mail: jean-bernard.michel@heig-vd.ch
ABSTRACT
Torrefied wood pellets are produced from torrefied chips by thermo-chemical pre-treatment of biomass at
200-320C in the absence of oxygen during about 15-30 minutes. Overall, the torrefaction process
efficiency has been reported to be 90-95% % as compared to 84% for pelletisation. Torrefaction improves
the biomass: 30% higher calorific value and 50% higher energy density resulting in much lower handling
and transport costs. The fuel becomes hydrophobic making long term outdoor storage possible. The
purpose of this project was to compare the combustion and emission characteristics of torrefied vs.
normal wood pellets. With no modification to the feeding and the burner parameters, the ignition and
combustion characteristics of torrefied pellets are found very similar to those of normal pellets. Particulate
emissions per energy output were found very close and directly related to the ash content in the
feedstock. Using the Taguchi approach, it was possible to establish a model of the boiler performance as
a function of the input parameters. Further testing confirmed the validity of the model showing optimum
performance with a defined value of primary and secondary air flow rates which minimized particulate
emissions for both the normal pellets and the torrefied pellets.
Keywords: biomass, torrefaction, combustion, pellets, testing, Life-Cycle-Analysis, Life-Cycle-Impact
N.B. In this paper, all figures of thermal energy content of the fuels are given on the Low Calorific Value
(LCV) basis.
1 INTRODUCTION
1.1
Torrefied wood pellets are an attractive fuel for co-combustion in coal-fired power stations (Maciejewska
et al., 2006). Except for start-up, the process is autothermal (it generates its own energy due to mild
pyrolysis reactions) and the energy of the off-gases, which represent about 10% of the input energy, is
recovered. Overall, the process efficiency has been reported to be 90-95% % as compared to 84% for
pelletisation in one given set of operating conditions (Uslu, 2008). The purpose of this R&D project is to
compare the combustion and emission characteristics of torrefied wood pellets with those of normal wood
pellets.
1.2
Approach
About 1 ton of torrefied pellets have been prepared by ECN on their 100 kg/h pilot facility, using poplar as
the feedstock. Combustion tests have been carried-out on a 50 kW pellet boiler of the company Hoval with
normal pellets and with torrefied pellets. Input and output measurements have been made during start-up
and during stabilized operation. Flue gas concentrations of O2, CO and NO were measured continuously.
Total particulate emissions (TPM) were sampled using a disk filter following the proposed ISO/DIN 13336
standards. A scanning mobility particle sizer (SMPS) was used to determine the size distribution and the
total number concentrations of particles. The analytical set-up is shown in the appendix. The design of
experiment method from Taguchi was used to reduce the number of tests to a minimum while exploring
the complete space of variables with a 9*4 test matrix.
Torrefied
Biomass
pellets
In 1985, Pechiney built a 10000 t/y production plant, to use torrefied wood instead of charcoal in electric
furnaces (Peguret, 1986).
This new type of fuel is very promising because it alleviates a lot of the disadvantages of normal biomass
pellets:
The volumetric energy density is 50% higher than with normal pellets resulting in the same reduction
of handling and transport cost per energy output.
Grinding energy is reduced by 90% and overall, the process efficiency has been reported to be 9094% as compared to 84% for pelletisation (Uslu et al., 2008). See Figure 2.
Torrefied biomass is hydrophobic and therefore not subject to swelling and degradation allowing
outdoor storage and in the long term.
Its greater calorific value should be beneficial for combustion.
Energy
Pelletisation
Grinding
Drying
= 84 %
Torrefaction
Grinding
Pelletisation
Energy
12
Figure 2 Process efficiency comparison. Normal pellets (top), torrefied pellets (bottom) after Uslu et al.
(2008)
Several large scale production plants are planned or in construction in Europe and elsewhere, for the cocombustion of torrefied wood in coal-fired power stations:
Energy Center of the Netherlands, BO2 process (Kiel, J et al., 2008). They now work in association
with Vattenfall.
Atmosclear (Switzerland) large projects planned from 130 to 270 kt/y (Atmosclear web site)
Integro Earth Fuels, Wyssmont process, USA, 84 kt/y Roxborrow, NC (Integro Earth Fuels web site)
Topell, NL , Polow Torbed reactor technology, planned 60 ktons/y in Arnhem (NL) together with RWE
(Maaskant, E, 2009)
4Energy Invest (B), 38 kt/y in Ambleve (B) and Stramproy (4Energy Invest web site)
Essent trading (RWE) and Stramproy : 90 kt/y in Steenwijk (NL) (Essent trading web site)
A special session was devoted to torrefaction during the 18th biomass conference and exhibition (Lyon,
France) where the status of some of these projects were presented. In Spain, the 500 kg/h pilot plant built
by CENER was presented (Celaya et. al, 2010)
However, there seems to be no project so far directly targeted to domestic heating and cogeneration.
4 COST ANALYSES
Several economic comparisons have shown the benefits of using of torrefied pellets instead of normal
pellets. The table below provides a comparison of the cost of pellets for power generation with biomass
from Canada and from South-Africa shipped to Europe.
Hamelink (2005) reported that feedstock costs contribute around 2065% of the total delivery cost
whereas pre-treatment and transport contribute 2025% and 2540%, respectively, depending on the
location of the biomass resources.
According to Uslu (2008) TOP pellets can be delivered at costs as low as 3.3 /GJ (73.5 /ton) with a
biomass cost of 10 /ton as compared to 3.9 /GJ (66.3 /ton) for normal pellets. This is mainly due to
higher energy density compared to conventional pellets, which lowers both the road and sea transport
costs. This is also in agreement with the work of Peng et al. (2008) for pellets processed in South-Africa
with the ECN process and transported to Europe. The comparison with pellets produced in Vancouver and
processed in Europe after Herold, (2009) is presented in Table 1.
Similarly Kiel (2007) reported delivery costs for sawdust pellets supplied to North-West Europe: 4.7 /GJ
for torrefied and 5.9/GJ for normal pellets which confirms the economic advantage of torrefied pellets.
Table 1 Pellet costs from various sources
Cost item
Production capacity
(ktons/y)
40
80
56
Pellets
Pellets
Torrefied pellets
(ECN)
23.6
11
15
Production
70
41
45
Transport
62.6
54
42
Product
Margin
Total
23.9
(/ton )
180.1
106
102
(/GJ)
11.2
6.61
4.99
Test set-up
Combustion tests were carried-out on a 50 kW pellet boiler of the company Hoval shown in Figure 3.
Figure 3 - Schematic of the 50 kW Hoval Biolyt boiler and photograph of the sampling system
A forced draught burner is used on this boiler (and not a grid or a drum), allowing a rather accurate control
of primary and secondary combustion. A scanning mobility particle sizer (SMPS) was applied to determine
the size distribution and the total number concentrations of particles in the range from 0.01-0.400 m.
Exhaust gas is taken with a probe, which is also fed with particle free air. The resulting dilution factor is
adjusted by the flow rate of the diluting air and the total flow. To prevent condensation of water onto the
particle surface, the dilution factor is chosen high enough, to achieve a dew point below ambient
temperature. (Wieser and Gaegauf, 2000).
The design of experiment method from Taguchi was used to reduce the number of tests to a minimum
while exploring the complete space of variables. A first test campaign was carried out in order to identify
most relevant parameters and their levels. This resulted in a matrix of 4 variables and 3 levels as shown in
Table 2 with a total of 9 test cases.
Level 1
Pellet type
C1 (classical,
swiss forest mix)
Level 2
Level 3
T (torrefied)
Poplar
T (torrefied
poplar)
35%
45%
60%
35%
40%
45%
30%
50%
65%
In the first and second test campaign the so-called classical pellets were commercial Swiss pellets and
unfortunately not poplar for comparison with the ECN pellets. The third and last test campaign was carried
out with poplar pellets made specifically for this purpose, referred to later as C2 pellets.
Consequently, all the results were fitted with a model and the results are discussed in the next section.
5.2
The combustion behavior of the torrefied pellets was found very similar to that of the normal pellets:
Measured
Model
Model
Measured
Model
Figure 4 Raw test results and comparison with the model data
The raw results obtained with the various test campaigns are displayed in Figure 4. The points referred to
as model are the calculated values from a curve fitting model (using a second order polynomial). This
approach is necessary to show the separate influence of the various parameters, which otherwise is not
possible with the raw results.
A second measurement campaign was carried out, with the objective of finding the best air settings in
terms of CO and particulate emissions. Surprisingly, the optimum settings were the same for C and T
pellets with the following values:
NO=f(P)
C2
mg/Nmat13%O2
mg/Nmat13%O2
120
CO=f(P)
118
116
114
112
110
20.0
30.0
40.0
Pin kW
50.0
C2
50.0
60.0
1500
1000
500
0
20.0
60.0
30.0
40.0
Pin kW
NO=f(P)
C2
mg/Nmat13%O2
mg/Nmat13%O2
120
CO=f(P)
118
116
114
112
110
20.0
30.0
40.0
Pin kW
50.0
C2
50.0
60.0
1500
1000
500
0
20.0
60.0
30.0
40.0
Pin kW
Figure 5 with these settings. It shows that torrefied pellets can potentially produce less CO than classical
pellets and at the same time make it possible to reduce the excess air, thereby increasing the thermal
efficiency. Particulate emissions were found to depend strongly on the fuel ash content. In this case, the
so-called C2 pellets are from poplar with a higher ash content than the poplar used for torrefied pellets
which explains their higher particulate emissions. The particulate size distributions were also very similar.
NOx emissions are found to be similar in this case, but one could expect lower NOx emissions depending
on the amount of fuel nitrogen that has been released during torrefaction.
5.3
The comparison of impacts of the two biomass fuels was performed using the Impact 2002+ life cycle
impact assessment (LCIA) method. The functional unit was the MJ of heat produced by the boiler. Results
are summarized in the following table showing an overall gain of 50% mainly due to the improvement of
the overall process efficiency.
Indicator
Unit
1MJ-Heat-T
1MJ-Heat-C2
Gain T
Human Health
2.54316E-08
4.66562E-08
45%
Ecosystem
Quality
[PDF] (Potentially
species)
0.006238811
0.006907042
10%
Damage
Fraction
of
Climate Change
[kg CO2]
0.004685857
0.013145215
64%
Resources
0.083996764
0.254237773
67%
Single Score
[Points normaliss]
5.06726E-06
1.00833E-05
50%
efficiency compared to the existing plant by 45%. In the same time, the output of valuable products, such
as gas, power, and storable, dense pellet-fuel increase by over 50% from 50 to 76.6 /t of wet plant input.
This could be achieved using four innovative technologies.
Autothermal torrefaction of the biowastes by firing the reactor with the off-gas of the torrefaction
itself.
The externally fired gas turbine, additionally increasing in its cycle efficiency:
by after-firing it with biogas from the anaerobic digestion process (Kompogas-Plant) and
Low temperature Combined Cycle by heat recovery to generate power using ORC-Technology.
A full RENEC (20000 t/a) represents polygeneration of heat, power, SNG and pellet fuel from one site
from regional green wastes, collected from a region of approximately 80000 persons (38000 households)
supplying power for 1500 households (if they are not equipped with an electric water boiler), pellet fuel
for 1000 households and SNG to run 1500 cars. Meaning, RENEC could contribute to up to 3.5% of the
energy supply of the region built into.
Fermenter
Heating
65 kWth
Offgas
recirculation
8%
1280 kW th x 8000 h/a
Green-Matter
8000 t/a
30% TS
Screening
3
BG 160 m /t, 50% CH 4 = 400kW
Gas engine
e = 37%
400 kW Hu
320000 CHF/a
@10 Rp/kWh
=105 CHF/t
132 kW e Strom
Kompost
3100 t/a
40% TS
Wood
Wood
PSA
Methanisation
tot = 92%
3
BG 160 m /t, 50% CH 4 = 400 kW
Kompogas
Fermenter
400
kW
730
kW
50 kWth
Biogas
EFGT
Wood-CHP
e = 22%
tot = 86%
Foging
82% Efficiency
Sketch for Kompogas AG from
CATSE (kozentrum Langenbruck),
9.9.2008 /ms
320 kWth
@ 290C
265 kWe
Power
530000 CHF/a
@25 Rp./kWh
=105 CHF/t
Waste Heat
(radiation to heat
the fac.) 60 kWth
Rekuperated
Band Dryer
Particle
Exhaust
Filter
KOMPOGAS
Thermo-Oil
Boiler
tot = 62%
200kWth
@ 170C
275 kWth
@ 70C
ORC
heat CHP
e = 16%
tot = 95%
LCV
Burner
110 kWe
Torefaction
tot = 90%
33 kWe Strom
Waste Heat
(radiation) 10 kWth
KOMPOGAS
Pelletisation
tot = 100%
Figure 6 - RENEC I as it should look after the project. All installations mentioned above the dashed line
are the existing 8000 t/a Kompogas-plant (SOTA) near Zurich.
10
REFERENCE LIST
Wieser U. and Gaegauf C.K., 2000 - Nanoparticle emissions of wood combustion processes.1st
World Conference and Exhibition on Biomass for Energy and Industry, June 2000 Available at
www.oekozentrum.ch/files/nanoparticles.pdf (accessed 22.03.2010)
12