Review of efficiencies in comminuting forest fuels

ABSTRACT

Knowledge about the generalized operational efficiency of comminution machines and systems is of great interest when tailoring solid biofuel supply systems. The objectives were therefore to collect and compare data from the literature to those from machinery manufacturers on the performances of various forest biomass comminution systems, with the aim of identifying areas for further research. Our meta-analysis was based on data from 55 scientific publications and specification sheets from 566 machine models collected from manufacturer’s websites. A majority, 56%, of studies were carried out at the roadside and 54% of studies examined comminuted logging residues, which thus reflects the most common materials and environments of the forest fuel supply system studied in the literature. We conclude that: (1) Chipping is more productive and requires less energy than grinding for machinery with nominal power up to 300 kW. (2) Chipping small trees or residues lowers productivities and increases energy demand compared to logs; the comminution productivity is dependent on the type of machine, the nominal power, and the material handled. (3) The energy demand is a function of the variable type of machinery (chipper, grinder), and maximum particle size produced. (4) Productivities in operational studies are clearly lower than the ones reported by the manufacturers in idealized conditions. Further investigations of equipment such as large chippers and grinders operated in terminals under controlled conditions are needed, in order to gain a better understanding of the different factors affecting the efficiencies of large equipment.

KEYWORDS:

• Bioenergy
• biomass
• chipping
• grinding
• productivity
• quality

Introduction

The global demand for forest biomass for energy and refinery purposes is increasing annually, and markets are currently growing on local, regional, national and global scales (FAO 2014). For any of the processes, including the production of solid fuel for heat and power production, the biomass must be comminuted (reduced in size) before being used in industrial processes, in order to achieve the required properties for high energy conversion efficiency.

Comminuting processes can be carried out in stand, at the roadside, at the terminal or at an industrial site. Comminuting is mostly carried out using diesel-powered units while large-scale (often semi-mobile) electrical powered units may be used at large terminals or at industry sites as they require access to power grids. The larger the scale of machinery being used for comminuting, the higher the productivity and the lower the cost per unit of mass achieved (Spinelli and Hartsough 2001; Ghaffariyan et al. 2013b). Comminuting can also increase the transport efficiency: comminuted residual forest biomass such as stumps, tops and branches and small diameter trees, provide greater payloads compared to loose tree parts. Therefore, in fuel production systems where relatively long transportation distances exist, such as in the Nordic countries, biomass is often comminuted using mobile or semi-mobile diesel-powered chippers (cutting the wood) or grinders (crushing the wood) before being processed at industrial plants. The literature provides extensive knowledge about interactions between work tools/machine configurations and biomass properties, where operational properties such as cutting power, number of cutting knifes, cutting angles etc. and biomass properties such as tree species, wood density, wood temperature etc. have all been considered (Eriksson et al. 2013; Kühmaier and Erber 2018). Many of these studies have been carried out in order to achieve high quality chips for the pulping industry. Forest raw materials for solid fuel production are, in general, more heterogeneous compared to the materials comminuted for pulping (i.e. debarked stem wood) which probably produces greater variations in performance. In order to ensure that delivery of solid fuel fractions can be achieved at low cost, independent of the production system, knowledge of the operational efficiency of machines and systems dependent on parameters such as size and type of machine configuration used, and biomass assortment fractions, is of great interest.

The equipment used for comminution of woody biomass can be classified as follows: disc chippers, drum chippers, cone-screw chippers, shredders, vertical (tub) grinders and horizontal grinders. Chippers use sharp tools (like knives), which have to be sharpened at regular intervals to avoid excessive loss of performance (e.g. Nati et al. 2010). By contrast, the hammers used in many types of grinders may be sharp and almost knife-like, but are designed not to need any sharpening (blunt tools) and are eventually replaced when worn out. Thus, hammers are less sensitive than blades to the contamination which is common, especially in the case of handling logging residues (Naimi et al. 2006). Shredders use one or more rotating shafts, with low speeds of rotation, often below 50 rpm (Pottie and Guimier 1985). Rotors with hammers are used in hammer mills, hammer hogs (the first category is used for fine grinding at rotation speeds above 1200 rpm, while hammer hogs rotate below 1200 rpm), and in vertical-feed tub grinders (Naimi et al. 2006). The term “crusher” is also commonly used in the literature to mean any blunt, coarse comminution process (e.g. stump crushing), especially when the focus is on productivity at the system level rather than on the mechanical details (e.g. Asikainen 2010). The term “shredder” is also used for machines with comparatively high rotation speeds although, usually, “shredders” are defined as blunt comminution devices with rotation speeds up to 500 rpm.

Most of the results found in literature are, however, only true for specific systems and working conditions. Currently, there is a need for generalization of results to be used for systems analyses at national and international levels (Eriksson et al. 2013). Although data are available from the manufacturers, their productivity figures use idealized conditions, while scientific studies often take into account more real-world conditions (e.g. operational waiting times due to intermittent feeding), making it difficult to carry out direct comparisons of efficiencies.

The objectives of this study were, therefore, to collect and compare data from the literature to those from machinery manufacturers about performance (productivity, energy demand, and fuel quality) for various forest biomass comminution systems, with the aim of identifying areas for further research. The study was limited to data on the comminution of fuel wood from primary forest biomass i.e. biomass directly delivered from the forest to the industry.

Materials and methods

Our research was based on collecting information currently available from two main sources:

• Peer-reviewed scientific articles and scientific reports on comminution operations (“literature review”).

• Machine manufacturers’ websites and catalogues (“manufacturers review”).

The two types of information were classified in two separate datasets; afterwards, the results from the analyses of the two datasets were compared to identify similarities, differences and knowledge gaps.

Definitions of comminution equipment

In both reviews, comminution machines were classified either as chippers (machines with cutting knives, including chunkers), grinders (machines with hammers, including crushers, hammer hogs and hammer mills) or shredders (defined as having blunt rotors with rotation speeds below 500 rpm). The category “grinders” was further subdivided into grinders with horizontal feeding systems, and grinders with vertical feeding systems. Accordingly, four equipment categories were used for classification of equipment: chippers, vertical grinders (tub grinders), horizontal grinders and shredders.

Literature review

The available literature, in the form of scientific articles and reports, was searched for information on chipping, chunking, crushing, grinding and shredding operations.

The literature was searched using the “Google Scholar” search engine, along with the “Web of Science” and “Scopus” databases. The search words used were “chipping” (chip*), “grinding” (grind*), “crushing” (crush*), “comminuting” (commin*), “shredding” (shred*) combined with the secondary search words “forest” (forest*), “trees” (tree*), “wood” (wood*), “biomass” (bio*). Only studies of conventional machinery used for primary solid fuel wood production were considered. For each study included in our analyses, the following data were collected to carry out a meta-analysis:

• Location of the study/operation in terms of geography (country) and location of operations (terrain, roadside, terminal, industry).

• Characteristics of the equipment: type (chipper, horizontal grinder, vertical grinder, shredder), nominal power (kW)/power required (kW), specification of models/brands if available, and settings (e.g. drum/disc/screw chipper, number of knives, sieve size).

• Forest biomass quality: type of material (RES = logging residues, SMALL = small trees, SW = stem wood, STUMP = stump wood), tree species (hardwood, softwood), moisture content, particle size distribution (quantity of chips in the different dimension classes listed in the studies), minimum (mm) and maximum (mm) targeted particle sizes and quantity of fines (percentage of particles <3 mm).

• Efficiency of the chipping process: productivity (m³ solid/PMH), energy demand (kWh/m³), torque (Nm). The productivities collected in different units other than m³ solid (tonnes, dry tonnes, m³ bulk were all converted to solid m³ by means of the specific densities stated in the studies and reference values for each type of assortment). The reference for time consumption was the Productive Machine Hour (PMH), defined in each of the studies as the working time excluding delays. The energy demand, if given as diesel consumption, was converted with an energy content of 9.95 kWh/liter diesel.

Manufacturers’ review

Manufacturers included in the review were first selected on the basis of a review of machine catalogues that listed the most common type of equipment used for chipping and crushing operations in four European countries (Hakkila 2003 (Finland); Spinelli and Hartsough 2001; (Archivio Macchine Forestali 2007) (Italy); Bioenergi (Tidskriften Bioenergi) 2010 (Sweden); Krajnc 2011 (Serbia)). The websites of manufacturers listed in the catalogues were all visited and the data for each machine listed retrieved.

The information collected from the manufacturers’ websites included: manufacturer, model, type of equipment (chipper, horizontal grinder, vertical grinder, shredder), nominal power (kW)/power required (kW), specifications (cutting devices, power sources, mobility, number of knives/hammers, screen opening, cutting lengths, feeding angles, loading and discharge configurations), and achievable chipping productivities (m³ solid/PMH). Productivities stated in tonnes, dry tonnes, and as m³ bulk were converted to m³ solid, assuming average densities for solid stem wood from the web-based calculation tool WeCalc (Larsson and Nylinder 2014; Swedish Forest Research Institute 2018). The reported maximum productivity for each machine was collected.

Analyses

General linear models (GLM) were used for testing the effects of different factors on the productivity and energy demands of equipment included in the literature review. Specifically, as possible predictors, we tested the type of machinery, its power, the type of material, screen sizes and maximum targeted particle sizes. For the manufacturer dataset, in a GLM, we tested the effect of engine power and machine type on the expected productivity. Based on the results of these analyses, predictive models were created for each type of equipment. As a result, the productivity values found in the literature review were compared to the productivity values stated by the manufacturers.

Results

From the literature review, 55 papers published between 1982 and 2017 were identified. These papers contained a total of 179 combinations of equipment and woody raw material studied, with either the productivity (161 cases) or the energy efficiency (142 cases) reported, or both. Most studies took place in Sweden (13), Italy (12), US (10) and Finland (7) but also in Japan (3), Australia (3), France (2), Brazil (2), Bulgaria (1), Germany (1) and South Korea (1). Of the studies, 130 were about chippers, 26 about horizontal grinders, 20 about vertical grinders and three about shredders. The countries of origin of the equipment manufacturers were US (28%), Finland (25%), Italy (23%), Sweden (10%), Germany (7%), Austria (4%) with Asian countries combined <1% (Japan (3 cases), China (1 case) and South Korea (1 case)) (Figure 1 and Table 1).

Table 1. The 55 papers used for the meta-analysis presented in alphabetic order.

References
A-E	G-M	N-Z
Abdallah et al. 2011	Ghaffariyan et al. 2012	Nordén and Eliasson 2009
Aman et al. 2011	Ghaffariyan et al. 2013a	Nurmi 1986
Andersson 2010	Ghaffariyan et al. 2013b	Nuutinen et al. 2014
Andersson 2011	Groover 2011	Pochi et al. 2015
Andersson et al. 2012	Han et al. 2015	Röser et al. 2012
Arola et al. 1982	Hellström et al. 2011a	Spinelli and Magagnotti 2014
Arthur et al. 1982	Hellström et al. 2011b	Spinelli et al. 2011
Asikainen and Pulkkinen 1998	Karlsson 2010	Spinelli et al. 2012a
Assirelli et al. 2013	Kons et al. 2015	Spinelli et al. 2012b
Bertilsson 2011	Kweon et al. 2016	Spinelli et al. 2013
Lorensi do Canto et al. 2011	Labbe’ et al. 2017	Spinelli et al. 2014
da Costa et al. 2010	Laitila and Nuutinen 2015	Spinelli et al. 2015
Di Fulvio et al. 2015	Laitila and Routa 2015	Westbrook et al. 2007
Dukes et al. 2013	Laitila and Vaatainen 2012	Yoshida and Sakai 2014
Edlund 2009	Magagnotti et al. 2016	Yoshioka et al. 2006a
Eliasson and Granlund 2010	Manzone and Balsari 2015	Yoshioka et al. 2006b
Eliasson and Picchi 2010	Marchi et al. 2011	Zamora-Cristales et al. 2015
Eliasson et al. 2015	Mihelic et al. 2015
Eriksson 2008	Mitchell and Gallagher 2007

The materials processed were logging residues (RES) in 96 cases, stem wood (SW) in 51 cases, small trees (SMALL) in 27 cases and stumps (STUMP) in five cases (Figure 1). Grinders were mainly used for RES and STUMP, and shredders only for STUMP. Chippers are mostly used for SW and SMALL and grinders and shredders for RES and STUMP (Figure 1). The nominal power needed for the chippers varied between 18 and 1030 kW, the power of the grinders varied between 120 and 783 kW, and the power of the shredders varied between 315 and 429 kW (Figure 2). Most of the machines were driven by diesel engines. Only two chippers were electrically-driven; these were used in test bench experiments.

Figure 1. The number of studies of the different assortments (STUMP = stump wood, RES = logging residues, SMALL = small trees, SW = stem wood) for the different machine categories studied. The total number of studies was 179.

The operations were, in most cases, carried out at landings/roadsides (101 cases), in industry (38 cases) and at terminals (24 cases). In some cases, the studies took place in laboratory/bench trials (nine cases) or in-field (seven cases). For both chippers (66 cases) and grinders (35 cases), landings were the most common locations. The second most common location for use of chippers was in industry (37 cases) while for grinders, it was at terminals (nine cases).

From the manufacturer reviews, 566 models were identified, with the following geographical distribution: the US (25%), Germany (20%), Finland (14%), Italy (13%), Austria (12%), Sweden (8%), Denmark (5%), Belgium (1%), Netherlands (1%) and the UK (1%).

Figure 2 clearly shows that there is a gap in the range of the nominal power measured in the studies compared to the ones stated by the manufacturers. In the case of chippers, the gap can be seen for large machines (600–1000 kW), while for grinders and shredders there are only a few studies of relatively small machines (<200 kW) and very large ones (>700 kW).

Figure 2. Nominal power for the models studied (shaded bars), compared to figures given by manufacturers (black bars) for chippers, horizontal grinders, vertical grinders and shredders.

The productivity of the chippers in the studies varied between 2 and 137 m³solid PMH⁻¹ (Figure 3). The productivity of the studied horizontal grinders ranged from 9 to 166 m³ solid PMH⁻¹, and for the vertical grinders and the shredders from 3 to 64 m³solid PMH⁻¹ and from 33 to 35 m³solid PMH⁻¹, respectively. When including machine power (kW) as a covariate, grinders were significantly less productive than chippers (Table 2). Specifically, vertical grinders were significantly less productive than chippers, horizontal grinders significantly exceeded the productivity of chippers and shredders had similar productivity to chippers (p > 0.05). In chipping operations, the productivity increased with 0.11 m³ for each kW increase of nominal power (R_Pearson = 0.64, p < 0.001). For grinders and shredders, the productivity increased with 0.18 m³ for each kW increase of nominal power (R _Pearson = 0.65, p < 0.001). Hence, when the machine power exceeded 500 kW, grinders appear more productive than chippers.

Table 2. Results from ANCOVA for machine groups (treatments) horizontal grinders, vertical grinders and shredders with machine power (kW) as a covariate. Results are compared to the reference machine chipper. Multiple R-squared: 0.7383, Adjusted R-squared: 0.7316.

Coefficients	Estimate	Std. Error	t value	Pr (>|t|)
Intercept	1.168541	2.443625	0.478	0.6332
Horizontal grinder	9.119454	4.603238	1.981	0.0493
Shredder	−10.566954	9.550764	−1.106	0.2703
Vertical grinder	−18.291687	3.980903	−4.595	<0.001
Machine Power (kW)	0.121809	0.008172	14.906	<0.001

Material has a significant effect on chipping productivity when tested as a covariate together with machine power (p ≤ 0.01). Chipping SMALL and SW is more efficient than chipping RES. Screen size is also relevant, as the hole size of the screen was varied (for example Figure 3, at 300 kW). The effect of different materials was not detected in the case of grinders (p > 0.1). Also, we could not detect an effect of the different places where the operations were carried out (laboratory, forest, landings, terminals, industry).

Figure 3. Productivity (solid m³ PHM-1) measured in the published studies as a function of nominal power of machines.

For chippers with a nominal power less than 500 kW, the productivity values given by the manufacturers increase by 0.245 m³ kW⁻¹ (n = 92, p < 0.001). For chippers with a nominal power above 500 kW, the corresponding figure is 0.048 m³ kW⁻¹ (n = 27, p = 0.027).

The productivity values of grinders and shredders from the manufacturers were combined due to the relatively low number of entries in each category (i.e. including shredders, horizontal grinders and vertical grinders). The productivity for grinders and shredders with a nominal power less than 500 kW increased by 0.210 m³ kW⁻¹ (n = 15, p > 0.1), the corresponding increase for machines with a nominal power exceeding 500 kW was 0.048 m³ kW⁻¹ (n = 15, p > 0.1).

The literature values for productivity and efficiency of the machines studied were compared to values declared by the manufacturers (Figure 4).

Figure 4. Productivity for chippers (top figure), and shredder and grinders (bottom figure), as given by manufacturers (“manufacturer”) and as observed in the published literature studies (“observed”). Triangles and squares represent the averages for each series, the errors bars represent confidence intervals (95% level).

For chippers, the productivity observed in the literature was generally 36% lower than the maximum ones quoted by manufacturers. The observed difference became significant for machines with powers of 100–200 kW and for 300–400 kW.

For grinders and shredders, the productivity values stated in the literature were generally 9% lower than the maximum ones quoted by manufacturers. The difference became significant for machines with powers of 350–600 kW.

The energy demands of chippers varied between 2.3 and 64.6 kWh/m³ solid (n = 19). For horizontal-feed grinders, the energy demand varied between 5.1 and 61.5 kWh/m³ solid (n = 19). For vertical-feed grinders, the range was 10.6–102.0 kWh/m³ solid (n = 103). For shredders, only one figure was available (23.5 kWh/m³). There was a weak correlation between the energy demand and the nominal power (Figure 5). For horizontal grinders, it was observed that there was a reduction in energy demand as the power increased (R² = 0.34, p = 0.005), however this relationship was not confirmed for chippers. Only weak differences were observed when considering the different materials, however further studies could clarify these relationships.

Figure 5. Chippers’ and grinders’ energy demands as a function of machine nominal power, based on the literature review data.

There was a significant reduction in energy demand both for chippers and grinders as the maximum particle size of chipped material increased (p = 0.029 and p = 0.003), with a stronger correlation in the case of grinders (R _Pearson = −0.550). The studied chippers generally had a lower power demand than grinders/shredders for the production of particles of the same given size (p < 0.001, Figure 6).

Figure 6. Energy demand of chippers and grinders as a function of the maximum produced particle size (dotted lines indicate observed trends), based on the literature review data.

Therefore, another possible interpretation for the higher efficiency of chippers could be that the grinders are generally used for more difficult materials than chippers (such as logging residues and stumps).

The moisture content of materials varied widely between 11% and 63%, but appeared not to have significant effects on the energy demands and did not differ significantly for the different types of materials fed to chippers or grinders.

Discussion

Our analyses considered the findings of different scientific studies of comminution equipment, comparing them to the manufacturers’ information. The majority of the literature studies were based in Europe (55% of the studies) and the US (19%). Similarly, the equipment considered in the studies and the ones from the manufacturers were mostly sourced from Europe and the US.

The observed productivity values from the literature were generally 10–30% lower than the ones declared by the manufacturers. The differences between the declared productivity values and findings in the studies can be due to different reasons. One reason could be the use of feeding systems in the studies which were dependent on the heterogeneity of comminuted materials. Limitations in the feeding systems, such as e.g. ineffective intermittent feeding with loaders, clearly influence the operations, as shown by the differences in productivity between experiments where stem wood was used and where stumps, small trees or logging residues were comminuted. On average, stem wood may be smaller than in an idealized situation. Smaller stems will reduce the maximum engine power needed, and the diesel consumption per m³ is likely to increase as there will be more time when the diesel engine is running without full use of its chipping capacity. This explanation is supported by Spinelli and Hartsough (2001), who noticed that larger stems resulted in higher productivity (e.g. an increase in stem weight from 10 to 100 kg resulted in an increase in productivity of about 2–3 times), a result which also suggested that frequent interruptions in the feeding reduced the productivity. The log diameter varied between 5–25 cm in the reviewed studies, but, many of the studies did however not report any stem wood dimensions which thus made it irrelevant to include stem diameter as a covariate in our analysis. The rate at which raw material is fed into the machine may also be part of the explanation. Liss (1987) found that the productivity measured in practical tests was about two-thirds that of the theoretical productivity (i.e. the one calculated from feed rate and maximum tree size). Additionally, in our analysis we considered only the highest productivity values reported by the manufacturers, as it is likely that this value is the one used in promotion of the technology. However, the differences would have been much lower if considering the lower range of productivity values, or calculating average values based on the productivity range given.

In the literature studies, efficient chipping time did not include time for setup, repositioning and other similar interruptions. On the other hand, the time when use of the crane limited the chipping speed was included. Waiting times due to the chipper/grinder and crane interactions could become relevant and dependent on the grapple size (Röser et al. 2012). These facts should be considered when modelling the systems work-place time and thus the overall productivity.

Most of the studies took place at roadside/landings, according to the most common practices in Europe (Asikainen et al. 2008). We could not observe a significant effect of location on machine performance. It is well documented in the literature (Kühmaier and Erber 2018) that the logistical setting of centralized operations in terminals/industries could increase utilization and reduce costs of comminution, compared to landings/roadsides. However, when only considering efficient chipping times, the difference is expected to be small.

There were clear gaps in studies of large chippers and very large grinders compared to the models reported by the manufacturers. Current trends in the Nordic Countries show an increase of comminution operations at terminals and in industries (Kärhä 2011), thus, it is expected that more knowledge should be produced for large machinery in the future, given the rising demand for centralized comminution operations.

Chippers were more productive when chipping small trees and stem wood compared to logging residues while for grinders, we could not find a clear difference caused by materials. Eliasson and Granlund (2010) found that crushing logging residues and stumps with a large horizontal grinder gave, respectively, a 29% and 51% lower productivity than stem wood. Similar results were also found by Eriksson (2008), who noticed that productivity values for logging residues were 11% lower compared to stem wood. In most of the collected grinding studies, logging residues and stumps were ground, and only in five cases were stem wood and small trees ground. The different size of samples could have made it difficult to detect the expected differences due to materials.

Vertical grinders and shredders generally showed a lower level of productivity when compared to chippers or horizontal grinders. Small grinders (<300 kW) were less efficient than chippers having similar nominal power. As the nominal power increased, the difference between chippers and horizontal grinders disappeared. Chippers are usually considered more efficient than grinders. As shown in a recent study by Spinelli et al. (2012b), changing the configuration of a horizontal 120 kW grinder from hammers to knives increased productivity by 30–80% with a reduction in the energy demand of 15–30%; this study supports our findings for relatively small size machines.

Theoretically, chipping is preferable to grinding in terms of efficiency. However, for materials such as logging residues and stumps, grinding and shredding may have practical advantages. For logging residues, small branches (twigs) may be randomly oriented, and could fit in the space between the drum pocket and the anvil, and be moved to the drum case outlet without being comminuted, thus remain uncut. Oversized particles can cause severe practical problems in feeding systems at heat- and power-plants. To solve this problem, Firus and Belter (1998) designed an experimental drum chipper that used radial knives in addition to the tangential knives where the knives were fitted into slots in a flat plate close to the drum and were used to shear off branches lying parallel to the drum’s long axis (see also the summary of another solution in Eriksson et al. 2013). Thus, it seems worthwhile to explore innovative chipping solutions. Another solution is to subsequently screen off oversized particles and feed them through the chipping process again.

The studied chippers generally had a lower power demand than grinders/shredders for the production of particles of the same given size (p < 0.001, Figure 6). In particular, the grinders/shredders were less efficient than chippers when particle sizes to be produced were below 100 mm. For chippers, we noticed a higher energy demand for the case of chipping small trees compared to stem wood (p = 0.008) while for grinders, there were generally no significant differences caused by different materials (i.e. mostly logging residues and stumps).

There was a clear relationship between maximum particle size produced and energy demand, with an increase of energy demand as the particle size decreased. This was expected and also documented in other studies of knife lengths and screen sizes for chippers (Liss 1987; Nati et al. 2010) and grinders (Hoque et al. 2007). Trends in energy demand as a function of machine power and materials were not completely identified, and some aspects need further investigation, such as the effect of moisture content. The studies reported only indicative moisture content and more accurate measurements would help to identify its specific effects. On the other hand, Spinelli et al. (2011) determined that species and moisture content have a secondary effect on chipper productivity and fuel consumption, which are primarily controlled by piece size for a given power.

For any type of machine, the waiting time (due to feeding limitations) and the power of its engine during that time, both contribute to the energy requirement per m³. The influence of the feeding rate on energy efficiency suggests that the energy used when the machine is waiting is important, which is also confirmed by measurements (e.g. Liss 1987; Spinelli and Hartsough 2001). Hence, a strategy to increase the energy efficiency could be to reduce the maximum required power by providing some storage of energy, such as a hybrid diesel-electric drive system (c.f. Di Fulvio et al. 2015), or a flywheel to store rotational energy for the peak demands. Some storage capacity for the raw material should also increase the energy efficiency of the process. Indeed, some machines already have some storage capacity (such as a large feeding board or the tub of a vertical-feed grinder). Our study does however not report any productivity benefits for vertical/tub grinders compared to horizontally fed ones.

The method used in this study has important limitations. There is not much information on the number of different machines in use, and their frequency of use. The surveys are mostly European, so the study is not global and manufacturers with most of their production for markets outside Europe are likely to be missing. Some countries that are known to use considerable amounts of woody fuels are missing from the list of manufacturers, notably Poland and the Baltic states. This may be partly due to language barriers, however, it may also indicate some potential for increased local use of the raw material, rather than export. Additional limitations of the study are the methodology differences in how the data were produced both from literature reviews and manufacturer specifications, and secondly, the severe dataset unbalance. Some productivity figures were given in tonnes or loose cubic meters, and assumptions had to be made for the conversion, which may have introduced errors. Shredders are often used more for solid waste than for wood, which means that figures for productivity could also be misleading. Factors known to be of importance, such as the influence of moisture content, wood density and cutting speed, were not fully investigated at this stage. Data from the manufacturers should be interpreted with caution, as not all operational parameters are specified, notably stem sizes.

Conclusions

This meta-analysis evaluated empirical data based on 55 scientific publications and the data for 566 machine models collected from manufacturers’ websites. A majority, 56%, of studies were carried out at roadsides with 54% comminuting logging residues, which thus reflects the most common materials and environments of the forest fuel supply system studied in the literature. Even though the coverage of studies are limited to those mostly carried out in Europe, we can still generalize the following results:

• Chipping is more efficient (more productive and requires less energy) than grinding for machinery with nominal power up to 300 kW.

• Chipping small trees or residues reduces productivity and increases energy demand compared to sawlogs.

• The comminution productivity is a function of the variables type of machine (chipper, grinder), the nominal power, and the material handled.

• The energy demand is a function of the variables type of machinery (chipper, grinder), and maximum particle size produced.

• Productivity values reported in operational studies are clearly lower than the ones reported by the manufacturers in idealized conditions.

We also can also conclude that: (1) crushing should only be used for contaminated biomass and chipping should only be used for clean material, (2) chipping requiresless fuel consumption than crushing (with a few exceptions), and (3) the cost of comminuting as a function of its location in order, from lowest to highest, is: industry < terminal < at roadside < in stand.

Further studies to investigate the factors constraining the performance in operational studies are important to perform. In particular, it would be relevant to study improved methods for feeding difficult materials (such as bunching of residues) and methods to reduce the energy demand (such as hybrid engines) or improved feeding/storage capacities. At the same time, further investigation of large chippers and grinders operated in terminals under controlled conditions are needed, in order to gain a better understanding of the different factors affecting the efficiencies of large equipment.

Acknowledgments

The authors thanks PhD Gunnar Eriksson for helping out with data collection.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This study was part of the Forest Refine and BioHub projects, which were financed by the Botnia-Atlantica program, part of the European Regional Development Fund and the Mobile Flip project, which is financed by the European Union’s Horizon 2020 research and innovation program under grant agreement No. 637020−MOBILE FLIP.