b i o m a s s a n d b i o e n e r g y 7 5 ( 2 0 1 5 ) 1 5 5 e1 6 0

Available online at www.sciencedirect.com

ScienceDirect

http://www.elsevier.com/locate/biombioe

The improvement of hog fuel by removing fines,

using a trommel screen

Carla Nati*, Natascia Magagnotti, Raffaele Spinelli

CNR IVALSA, Via Madonna del Piano 10, Sesto Fiorentino, FI, Italy

article info abstract

Article history: The study tested the use of a trommel screen originally designed for compost materials to
Received 6 August 2012 reject oversize particles from hog fuel, processed from several sources and by two different
Received in revised form comminution devices. The experiment consisted in screening material previously
13 February 2015 comminuted by a convertible crusher, designed to use both hammers and knives. Three
Accepted 17 February 2015 different feedstock types were used, and namely: discarded pallets, logs and branches from
Available online 6 March 2015 park maintenance. Each feedstock type came in two different qualities, depending on the
tool used for comminution, i.e. hammers or knives. Trommel screen productivity varied
Keywords: between 4.2 t h 1, and 5.2 t h 1
of oven dry material. Screening hog fuel derived from
Crusher pallets was 30% and 40% less productive than screening fuel derived from logs and
1
Trommel screen branches, respectively. Screening cost varied from 16.2 V t dry material in the case of
1
Solid biofuel branches, to 19.9 V t oven dry material for pallets. Screening allowed an increase of fuel
Low-quality wood waste quality only when applied to pallet-derived hog fuel.
Particle size distribution © 2015 Elsevier Ltd. All rights reserved.
Screening performance

demanding and are currently targeting low quality sources,

1. Introduction
The demand for wood biomass has consistently increased in
the last years in Italy, as a response to the new and ambitious
EU renewable energy strategy. Italy has received the EU
Directive 2009/28 CE in her National Action Plan approved in
2010. In this new plan, biomass plays a major role, becoming
the main renewable source and accounting for 44% of
renewable energy production [1]. As a consequence, there is
now a strong competition between energy and conventional
industrial users for the available wood resources. Wood-based
board producers currently complain about a shortage of raw
material and hold the new energy users as the main culprits.
However, not all the energy users demand for the same raw
material. While small and medium size biomass plants
require high-quality wood fuel, larger installations are less
due to their substantially lower price. Low cost cellulosic
material could also be targeted by the growing biorefinery
sector, for conversion into a whole new generation of bio-
based products [2].
Low-quality wood waste contaminated with nails and dirt
is usually comminuted with crushers, fitted with blunt tools
(hammers) and capable of producing hog fuel, a coarse feed-
stock within the particle-size specification of large-scale
automated boilers. Bigger plants are less sensitive to the
presence of oversize particles, rather abundant within hog
fuel [3]. Among biomass manufacturers there is a constant
effort to ameliorate low quality wood biomass and increasing
its suitability for energy use, with the aim to achieve a higher
market price.
Thus, several machine makers now offer a new line of
horizontal crushers that can be temporarily converted into

* Corresponding author.
E-mail addresses: nati@ivalsa.cnr.it (C. Nati), magagnotti@ivalsa.cnr.it (N. Magagnotti), spinelli@ivalsa.cnr.it (R. Spinelli).
http://dx.doi.org/10.1016/j.biombioe.2015.02.021
0961-9534/© 2015 Elsevier Ltd. All rights reserved.
156 b i o m a s s a n d b i o e n e r g y 7 5 ( 2 0 1 5 ) 1 5 5 e1 6 0
chippers by replacing the standard blunt hammers with sharp
knives [4], therefore producing a more homogeneous end
product. These machines offer improved particle size distri-
bution, with both hog fuel and standard chips. If used as fuel,
wood quality is defined by such parameters as moisture con-
tent, ash content, and particle-size distribution (EN 14961-
1:2010).
Particle size distribution is one of the most important pa-
rameters for an efficient combustion at heating plants [5].
Particle size distribution is important also during storage, as it
affects caloric value [6] and durability [7]. Conversion effi-
ciency and emission rates are also influenced by particle size
distribution [8]. In particular, gasifiers are most sensitive to
this parameter [9], while circulating fluid bed boilers are most
tolerant [10]. Particle size and shape play a crucial role with
handling: long irregular particles tend to bridge over openings
[11], especially when fuel moisture content is high [12]. In
contrast, fine particles reduce the air flow within chips piles,
accelerating microbial decay, raising internal temperature
and increasing the risk of self-combustion [13]. Particle size
distribution is the result of many factors, and especially
chipper type [14], feedstock type [15], blade wear and screen
size [16]. Unfavourable particle size distribution can be
improved through screening, which is routinely implemented
when using wood biomass in fast pyrolysis reactors [17]. In
contrast, screening before direct combustion is not a gener-
alised practice and is performed occasionally, often at a semi-
experimental level [18]. Hence the difficulty in prospecting its
full potential, once the market and the technology will be
mature. Very little information is available on this topic in the
scientific bibliography.
The goal of this study was to analyse the performance of a
trommel screen specifically designed to ameliorate the char-
acteristics of woody and herbaceous biomass from park and
riverbeds maintenance before composting. The experiment
aimed to determine: a) the productivity and the cost of the
mechanical screening process and b) the quality improvement
obtained when applying a composting screen to hog fuel. For
the latter purpose only material of dimensions included be-
tween 3 and 63 mm (“accepts”) was considered suitable for a
medium size biomass plant. Material smaller than 3 mm
(“fines”) and bigger than 63 mm (“oversize”) was considered
unsuitable.
2. Materials and methods
The Authors tested a trommel screen, used in the composting
sector to separate organic wastes into different size fractions.
The machine was a Caravaggi V3000D mobile drum screen
[19], produced in Italy by Caravaggi mfg. in three models: 2000,
3000 and 5000. Model numbers refer to drum length, in mm. In
the test model, the rotating drum measured 1720 mm in
diameter and 3550 mm in length. The rotating drum was fed
by an auger conveyor, which had the additional task of
refining coarse materials. The drum structure supported
perforated screen plates: particles smaller than the screen
mesh size would fall through it and onto a belt conveyor for
moving to the small particle heap, to the side of the machine.
Larger particles were moved to the end of the drum and
dumped on a second conveyor belt for discharge onto a sec-
ond heap, just at the back of the machine. Screen plates were
interchangeable and came in different mesh sizes, to be
selected according to job type. The machine on test was fitted
with 20 mm-wide square mesh screen plates. It was powered
by a built-in 44 kW diesel engine and installed on a two-axle
trailer for transportation. Drum rotational speed was set to
0.25 Hz. All the screening functions were electronically
controlled and visualized on a digital panel. The screen was
served by a 70 kW farm tractor, equipped with a front-end
loader. The tractor was used for relocating and feeding the
screening machine, so that the trailer-mounted screening
device, the tractor and its driver represented a fully indepen-
dent small scale hog fuel screening operation.
Productivity was estimated with a time-motion study,
conducted during commercial operation by one of Caravaggi's
workers, at the firm location in Palazzolo sull'Oglio (BS), in
Northern Italy.
The study consisted in screening material previously
comminuted by a convertible horizontal crusher [4]. The
crusher was equipped with a high-speed drum structure car-
rying 44 hammers, in four rows. As an alternative, the drum
could be fitted with a series of fixed knife mounts, for con-
version into a chipper. In the latter case, the total number of
tools was halved to 22. Therefore, the convertible crusher was
designed to use both hammers or knives, depending on raw
material type and conditions. Conversion between the two
configuration was simple, and would take ca. 4 h. Three
different feedstock types were used for the experiment, and
namely: discarded pallets, forest logs, residues from park
maintenance. Pallets conformed to EU standards and were
made of pine planks (Pinus nigra Arn.) discarded by a local
construction company. Logs came from the cleaning of a
country road and were an even mix of Austrian pine (P. nigra
Arn.) and Black locust (Robinia pseudoacacia L.), in 2e3 m
lengths. Park maintenance residues e originated from urban
greenery in the town surroundings - consisted of branches
from many ornamental tree and shrub species, and especially
Magnolia grandiflora L. and Cupressus Sp., representing
respectively 40% and 55% of the total bulk volume of the park
residues. Minor amounts of palm branches were also included
within the park maintenance residues. The average moisture
content of the different feedstock types was 25%, 42% and 45%
respectively for discarded pallets, logs and park maintenance
residues. All feedstock types were processed with the
convertible crusher in both configurations. Therefore, the
screening experiment included 6 treatments, derived from the
combination of 3 feedstock types and 2 comminution modes.
Each treatment was replicated 3 to 4 times, depending on
product availability. Each replicate consisted of a big bag of
hog fuel, weighing between 170 kg and 410 kg (mean 270 kg).
Big bags were weighed before screening, using a load cell
connected with a hook to the bag handles, and hoisted with a
forklift until the bag was fully suspended. The scale had a
rated accuracy of 200 g. The weight of the big bag and the hook
were subtracted from the reading. A second big bag was tied to
exit of the large size particle evacuation conveyor, to collect all
the product. This was weighed with the same method
described before. The weight of the small size particles passed
through the screen and into the small particle heap was
 b i o m a s s a n d b i o e n e r g y 7 5 ( 2 0 1 5 ) 1 5 5 e1 6 0 157

calculated by subtracting the weight of the large size particles potential differences between treatments. The software was
to the total initial weight of the sample. Feedstock types were used on the productivity data for performing a typical analysis
screened according to a random sequence. Each screening of variance (Anova), especially suited to the factorial experi-
cycle was stop watched individually, using Husky Hunter ment just described. In contrast, the distribution of particle
hand-held field computers running the dedicated Siwork3 size did not satisfy the normality assumption and the statis-
time study software [20]. Productive time and delay time were tical significance of differences between treatments was
recorded separately [21]. A single 1 kg sample was collected checked using non-parametric tests [24].
from each repetition e one big bag - for determining moisture
content and particle size distribution. The former was ob-
tained with the gravimetric method, according to European 3. Results and discussion
standard EN 14774-2:2009; the latter with the oscillating
screen method, according to European Standard EN 15149- 3.1. Productivity and cost
1:2010.
Operating costs were calculated using the procedures The results of the time study refer to pure screening work,
described by Miyata [22], on an estimated annual utilization of excluding all other work time and delays [21]. In particular,
1000 h for both the screen and the loader (Table 1). The cor- delays can represent a significant proportion of the scheduled
responding investment costs were 70,000 V for the screen, and work time, which can reach up to 50% [25]. Delays were
60,000 V for the loader: in both cases the depreciation period excluded from this study for its short duration, which did not
was assumed to be 5 years. Repair and maintenance was allow formulating a reliable estimate of delay time. Delay
estimated to 50% of depreciation, while labour cost was set at occurrences are typically erratic and can only be captured
16 V hour 1. Fuel cost was assumed to be 1.3 V L 1. The with long-term follow-up studies [26]. Net productivity in
calculated operational cost was increased by 20% in order to tonne (t) per productive machine hour (PMH0) is reported in
include administration costs [23]. Table 2. Mean values for the different feedstocks varied be-
For the purpose of the analysis, chips were considered as tween 5.6 t h 1 and 9.5 t h 1 on a fresh weight basis, or 4.2 t h 1
belonging to three different main groups: undifferentiated pile and 5.2 t h 1 of oven dry material. Screening fresh hog fuel
before screening (“in” pile); fraction that did not fall through obtained from discarded pallets was 30% and 40% less pro-
the screen mesh (“out large” pile); fraction that fell through ductive than screening the fresh product obtained from logs
the screen mesh (“out fine” pile). One sample was collected and branches, respectively. Anova shows that original feed-
from each pile on occasion of every screen run, or repetition. stock type was the only factor with a statistically significant
Each sample was analysed for particle size distribution, using effect on screening productivity (Table 3). The effect of feed-
a certified automatic screening device model FTL0200. Five stock type was exerted through the different moisture con-
sieves were used in order to separate the six following chip tent. When the same analysis was repeated for oven dry mass,
length classes: >100 mm, 100 mm e >63 mm, 63 mm e then all differences were obliterated. Hence, mass productiv-
>45 mm, 45 mm 16 mm, 16 mm e 3 mm, <3 mm. Each frac- ity was lower for discarded pallets just because the material
tion was then weighed with a precision scale. The different was lighter than the others, not because its through flow was
fractions were grouped into three functional classes: oversize any slower.
particles (>63 mm), accepts (63 mme3 mm) and fines (<3 mm). Average screening cost amounted to 10.9 V t 1 on a fresh
Data were analysed with the Statview advanced statistics basis, or to 17.9 V t 1 of oven dry material. These figures
software, in order to check the statistical significance of the included all the screened material, including oversize parti-
cles (length >63 mm), whose % incidence was on average 7.8,
with higher values reported for branches (from 14.6 % to
18.3%). In detail, screening cost was 16.2 V t 1, 17.9 V t 1 and
Table 1 e Costing assumptions and machine cost.
19.9 V t 1 of oven dry material for branches, logs and dis-
Operation Unit Screen Loader carded pallets, respectively.
Purchase price V 70 000 60 000
Service life Year 5 5 3.2. Quality
Salvage value Fraction of 0.2 0.2
initial price
Between 70% and 83% of the total mass went through the
Interest rate 0.04 0.04
Fuel consumption L PMH0 1
3.7 4.0 screen and into the small size particle pile (Table 4). On the
Crew N 0 1
Depreciation V year 1 11,200 9600
Annual usage h PMH0 1 1000 1000 Table 2 e Data for screen productivity (t h¡1) according to
Total fixed cost V PMH0 1 15.01 12.86 different raw materials.
Repair and V PMH0 1 5.6 4.8
Productivity (PMH0) Mean SD Minimum Maximum
maintenance
Personnel cost V PMH0 1 0.0 16.0 Total 7.71 2.16 4.35 11.41
Total variable cost V PMH0 1 12.21 29.71 Branches 9.49 1.68 7.61 11.41
Overhead (20%) V PMH0 1 5.4 8.5 Logs 8.09 1.06 6.41 9.53
Total V PMH0 1 32.7 51.1 Pallets 5.65 1.33 4.35 7.78

Notes: PMH0 ¼ productive machine hours without delays. Notes: PMH0 ¼ productive machine hours without delays.
158 b i o m a s s a n d b i o e n e r g y 7 5 ( 2 0 1 5 ) 1 5 5 e1 6 0

Table 3 e Anova's table for productivity (t h¡1).

DF Sum of squares Mean square F-value P-value Power
Tool 1 5.833 5.833 2.933 0.1061 0.35
Material 2 60.188 30.094 15.132 0.0002 0.99
Tool*material 2 0.004 0.002 0.001 0.9990 0.05
Residual 16 31.820 1.989

about the same for the input pile and the two output piles. The
Table 4 e Mass fraction in % after screening application, only significant differences were found in the chips obtained
i.e. large and fine chips. from discarded pallets through knife action. Only in this case
Raw material Tool Mass fraction Mass fraction p did screening allow a significant increase of accepts in the
(%) large (%) fine large particle pile. This was obtained through the total elimi-
Branches Hammers 18.4 81.6 a nation of all fine particles from this pile. In fact, the main ef-
Knives 22.8 77.2 a fect of screening as applied in this test was that of removing
Logs Hammers 19.9 80.1 a fine particles from the origin material and moving it to the
Knives 16.9 83.1 a
small particle pile. This effect was strong and significant for all
Pallets Hammers 30.3 69.7 b
feedstock and tool types (Table 6). On the contrary, screening
Knives 29.7 70.3 b
had little effect on coarse particles, which were displaced in
Note: different letters on the p column indicate that differences are
significant amount only when handling branches had been
statistically significant at the 5% level, according to the non-
comminuted with blunt tools (hammers). Hence, the balance
parametric KruskaleWallis test.
of accepts could only be changed when it was strongly
affected by the presence of fines. Otherwise, even a massive
other hand, the fact that there were still coarse particles in the removal of fine particles could not shift this balance. This is in
out-fine pile was easily explained by the irregular shape of contrast with the results obtained by Spinelli et al. [18], who
comminuted wood particles. Particle length was generally detected a main effect on coarse particles. However, that
larger than particle width and thickness, which allowed some study was conducted with a completely different screening
coarse particles to pass through the screen even if their length device (oscillating pan), specifically designed for removing
was larger than the screen mesh size. That did not happen coarse particles.
with laboratory analysis, where coarse particles were sepa- Very few studies are available on the mechanical screening
rated manually before mechanical screening, leading to the of wood chips for energy use, but it seems clear that the re-
detection of coarse particles which should have not been sults are closely related to screening method and type. We
there. Concerning % allocation, the KruskaleWallis non- suspect that the same screening unit used for this test could
parametric test found a statistically significant difference have offered completely different results if we had used a
only for the material obtained from discarded pallets, different mesh size. More tests should be conducted in order
regardless of comminution tool (i.e. knives or hammers). A to find the best match between mesh size and target product.
higher fraction of this material ended up into the large particle
pile, even if the original batch contained less coarse particles
than the original batches obtained from the two other feed- 4. Conclusions
stock types (Table 5). This may indicate a secondary indirect
effect of moisture content, with drier and lighter particles Screening is not overly expensive, but the screening units
tending to “float” over the screen mesh rather than falling developed for the composting sector need some adaptation
through. before they can be effectively used with wood chips. Under
Concerning the actual quality upgrade, the test did not their standard configuration, these units cannot offer an
result into any significant shift of the fraction of accepts from equally good job with coarse particles. The increasing interest
one pile to the other. The % incidence of accept particles was in using waste wood for energy production in a wide range of

Table 5 e Mass fraction in % of particle size distribution in the three piles (in, out large and out fine).
Material Tool Accepts (mass fraction %) Coarse (mass fraction %) Fine (mass fraction %)
In Out large Out fine In Out large Out fine In Out large Out fine
Branches Hammers 56.8 42.9 66.4 29.5 54.2 11.0 13.7 2.9 22.5
Knives 70.2 62.0 75.3 16.4 35.3 0.2 13.5 2.6 24.5
Logs Hammers 82.2 76.0 85.3 10.2 22.9 0.1 7.6 1.0 14.6
Knives 85.5 84.3 88.5 8.6 15.6 0.0 5.9 0.1 11.5
Pallets Hammers 90.2 94.3 87.0 3.0 5.6 0.2 6.7 0.0 12.7
Knives 93.5 96.2 90.2 1.3 3.8 0.0 5.2 0.0 9.8
 b i o m a s s a n d b i o e n e r g y 7 5 ( 2 0 1 5 ) 1 5 5 e1 6 0 159

Table 6 e Significance (p-values) matrix for the particle size distribution figures presented in Table 5.
Accepts Coarse Fine
In vs. out In vs. out Out large vs. In vs. out In vs. out Out large vs. In vs. out In vs. out Out large vs.
large small out small large small out small large small out small
Branches Hammers 0.1288 0.1841 0.1496 0.0429 0.0526 0.0458 0.0007 0.0014 0.0007
Knives 0.3456 0.5040 0.4122 0.1005 0.0687 0.0848 0.0010 0.0019 <0.0001
Logs Hammers 0.4080 0.4231 0.4130 0.1290 0.0579 0.0923 0.0220 0.0311 0.0152
Knives 0.7128 0.2834 0.3415 0.0720 0.1122 0.0345 0.0174 0.0718 0.0400
Pallets Hammers 0.1096 0.1213 0.1093 0.0656 0.1243 0.0946 0.0057 0.0012 0.0013
Knives 0.0194 0.0096 0.0018 0.0914 0.1884 0.0590 0.0019 0.0024 0.0015

Bold p-values indicate a difference significant at the 5% level.

boilers has made screening attractive, in an effort to upgrade
fuel quality and to divide product lots into different streams
directed to different boiler types. Domestic users offer better
prices than industrial ones, but they have stricter particle size
and moisture content specifications. However, upgrading
must consider all fuel quality aspects, and not just particle
size distribution. Even if a screened product met the quality
specifications for particle distribution, it could still miss the
moisture content target for residential use.
These results are based on a short-term study, and should
be taken with caution. Finally, this study tested only one of the
many machines that can be applied to the screening of wood
chips: further studies should test different machines types
and screen size specifications.
Acknowledgements
The research leading to these results has received funding
from the European Union Seventh Framework Programme
(FP7/2012-2015) under grant agreement n 311881 (INFRES
Project). Special thanks are offered to Caravaggi mfg. for
making available their drum screen and their most competent
test team.
references
[1] Pettenella D, Andrighetto N. Le biomasse legnose a fini
energetici in Italia: uno “sleeping giant”? [woody energy
biomass in Italy: a sleeping giant?] AgriRegioniEuropa
2011;24(7):18e22.
[2] Fitzpatrick M, Champagne P, Cunningham M, Whitney R. A
biorefinery processing perspective: treatment of
lignocellulosic materials for the production of value-added
products. Bioresour Technol 2010;101:8915e22.
[3] Strehler A. Technologies of wood combustion. Ecol Eng
2000;16:25e40.
[4] Spinelli R, Cavallo E, Facello A, Magagnotti N, Nati C,
Paletto G. Performance and energy efficiency of alternative
comminution principles: chipping versus grinding. Scand J
For R 2012;27(4):393e400.
[5] Hartmann H, Bo € hm T, Jensen P, Temmerman M, Rabier F,
Golser M. Methods for size classification of wood chips.
Biomass Bioenergy 2006;30:944e53.
[6] Kofman P, Spinelli R. Storage and handling of willow from
short rotation coppice. Report No NEIDK- 3165. Hoersholm,
Denmark: Forskiningscentret for Skovog Landskab; 1997.
ISBN 87-986376-2-2.
[7] Nellist M. The effect of particle size on the storage and drying
of wood fuels. In: Hudson B, Kofman P, editors. Harvesting
storage and road transportation of logging residues.
Proceedings of a workshop of IEA-BA-Task XII activity 1.2
held in October 1995 Glasgow, Scotland. Vejle (Denmark):
FSL; 1995. p. 59e70.
[8] Paulrud S, Nilsson C. The effects of particle characteristics on
emissions from burning wood fuel powder. Fuel
2004;83:813e21.
[9] Asikainen A, Pulkkine P. Comminution of logging residues
with evolution 910R chipper, MOHA chipper truck, and
morbark 1200 tub grinder. J For Eng 1998;9:47e53.
[10] Bridgwater AV. The technical and economic feasibility of
biomass gasification for power generation. Fuel
1995;74:631e53.
[11] Jensen P, Klausner A. Tendency of wood fuels from whole
trees, logging residues and roundwood to bridge over
openings. Biomass Bioenergy 2004;26:107e13.
[12] Mattsson J, Kofman P. Influence of cutting and storage
method on tendency to bridge in biofuels made from willow
shoots. Biomass Bioenergy 2003;24:429e35.
[13] Jirjis R. Effects of particle size and pile height on storage and
fuel quality of comminuted Salix viminalis. Biomass
Bioenergy 2005;28(2):193e201.
[14] Spinelli R, Hartsough B, Magagnotti N. Testing mobile
chippers for chip size distribution. Int J For Eng
2005;16:29e35.
[15] Spinelli R, Nati C, Sozzi L, Magagnotti N, Picchi G. Physical
characterization of commercial woodchips on the Italian
energy market. Fuel 2011;90:2198e202.
[16] Nati C, Spinelli R, Fabbri P. Wood chips size distribution in
relation to blade wear and screen use. Biomass Bioenergy
2010;34:583e7.
[17] Bridgwater AV, Toft AJ, Brammer JG. A techno-economic
comparison of power production by biomass fast pyrolysis
with gasification and combustion. Renew Sust Energy Rev
2002;3(6):181e246.
[18] Spinelli R, Ivorra L, Magagnotti N, Picchi G. Performance of a
mobile mechanical screen to improve the commercial
quality of wood chips for energy. Bioresour Technol
2011;102:7366e70.
[19] Caravaggi. Vaglio/screener 3000 [monograph on the internet]
Pontoglio (BS) Italy. Caravaggi S.r.l. Updated. 2015. Available
from:, http://www.caravaggi.com/vaglio-3000-2/?lang¼en
(accessed 19.01.15).
[20] Kofman P. Siwork 3: user guide. Vejle, Denmark: Danish
Forest and Landscape Research Institute; 1995. p. 37.
[21] Bjo€ rheden R, Apel K, Shiba M, Thompson MA. IUFRO forest work
study nomenclature. Garpenberg: Swedish University of
Agricultural Science, Dept. of Operational Efficiency; 1995. p. 16.
160 b i o m a s s a n d b i o e n e r g y 7 5 ( 2 0 1 5 ) 1 5 5 e1 6 0

[22] Miyata ES. Determining fixed and operating costs of logging
equipment. General Technical Report NC-55. St. Paul, MN.:
Forest Service North Central Forest Experiment Station; 1980.
p. 14.
[23] Hartsough B. Economics of harvesting to maintain high
structural diversity and resulting damage to residual trees.
West J Appl For 2003;18:133e42.
[24] McDonald JH. Handbook of biological statistics. 2nd ed.
Baltimore, Maryland: Sparky House Publishing; 2009.
p. 319.
[25] Spinelli R, Visser R. Analyzing and estimating delays in wood
chipping operations. Biomass Bioenergy 2009;33:429e33.
[26] Spinelli R, Visser R. Analyzing and estimating delays in
harvester operations. Int J For Eng 2008;9:35e40.