T

Industrial Crops & Products 145 (2020) 111955

Contents ists avai ab e at ScienceDirect

Industria Crops & Products

journa homepage: www.e sevier.com/ ocate/indcrop

Efects of mu ti-stage mi ing method on the energy consumption of comminuting forest residua s

Ya an Liua, Jinwu Wanga,b,*, John C. Barthc, Ke y R. We schd, Vincent McIntyrea, Michae P. Wo cotta

a Composi e Ma erials & Engineering Cen er, Washing on S a e Universi y, Pullman, WA, 99164 -1806, Uni ed S a es b Fores Produc s Labora ory, U.S. Fores Service, Madison, WI, 53726, Uni ed S a es c Depar men of Economics, Washing on S a e Universi y, Pullman, WA, 99164-1806, Uni ed S a es d Depar men of Chemical Engineeing, Washing on S a e Universi y, Pullman, Uni ed S a es

A R T I C L E I N F O

Keywords: Moisture content Hammer mi Mu ti-stage mi ing Comminution aws Energy consumption Partic e size and distribution Size energy Re atiuonship

A B S T R A C T

This study has eva uated the comminution energy reduction efciency of a mu ti-stage mi ing method and estab ished a mathematica mode for estimating the specifc energy consumption as a function of moisture content and partic e size change. A B iss hammer mi was emp oyed to grind the feedstock targeting at two fna screen sizes of 0.84 and 0.69 mm. For the target screen size of 0.84 mm, the sing e-stage, two-stage, and three-stage mi ing process were emp oyed. Whereas, the sing e-stage, two-stage, and four-stage mi ing process were emp oyed for the 0.69-mm target screen size. Four moisture content eve s were emp oyed, i.e. 6 %, 12 %, 18 %, and 28 %. Whi e the fna partic e sizes and distributions were simi ar independent of the intermediate stages, the resu ts showed that the sing e-stage mi ing consumed 43 % more energy than the three-stage grinding process. A so, mu ti-stage mi ing runs more stab e in grinding high-moisture content materia s. In addition, a mathematica mode based on Bond’s comminution aw was estab ished to predict the specifc energy con-sumption with the moisture content and the partic e sizes before and after grinding.

1. Intr ducti n

Fossi fue s, current y the primary source for iquid transportation fue s, bring about many environmenta issues. Forest havestoing re-sidua s are a potentia resource for producing biofue s to provide an a ternative to fossi e fue s. For any route to convert woody biomass to ready-to-use fue s, size reduction is a necessary preprocessing process. However, the process of breaking down forest residua s is energy in-tensive because woody ce wa s and structures are very tough (Dukes et a ., 2013; Liu et a ., 2016a; Repe in et a ., 2010). A so, the woody biomass, especia y softwood, is more reca citrant to various agents than agricu tura biomass materia s (Chen et a ., 2011; Shuai et a ., 2010). Therefore, diferent pretreatment processes were adopted for woody biomass (Gao et a ., 2013; Liu et a ., 2017a, b; Liu et a ., 2016b; Zhu et a ., 2015, 2010), which required severe pretreatment conditions. In order to reduce the chemica pretreatment severity, a mechanica -chemica pretreatment was emp oyed to achieve adequate tota sugar conversion and a fne mi ing process was dep oyed in this combined process (Liu et a ., 2017b; Miura et a ., 2012; Peng et a ., 2013; Zakaria et a ., 2014). Since the feedstock preparation and the combined

pretreatment both invo ve the mi ing process, manipu ating the energy consumption in the mi ing process is of signifcance.

The energy consumption of a mi ing process main y depends on materia , moisture content, mi type, and partic e size change between the feed and the mi ed product. Many studies showed that the specifc energy consumption increased a ong with an increase of moisture content (Liu et a ., 2016a; Mani et a ., 2004; Tumu uru et a ., 2014). Hammer mi s are more energy-efcient in producing fne partic es than knife mi s (Liu et a ., 2016a; Miao et a ., 2011) and are appropriate for producing fne partic es (Liu et a ., 2016a). A so, the re ationship be-tween partic e size change and energy consumption was described by a number of comminution aws (Char es, 1957). In these theories, the energy consumption during wood mi ing is expressed as the function of the partic e size diference between the feed and the mi ed product. A constant of materia s characteristics is used to identify the diference between various feedstocks under a specifc grinding mechanism. It was found that the constant was not on y re ated to the species of the feedstock but a so infuenced by the moisture content of the feedstock (Temmerman et a ., 2013).

It has been noted that the energy consumption is infuenced by both

⁎ E-mail address: jinwu.wang@usda.gov (J. Wang).

https://doi.org/10.1016/j.indcrop.2019.111955 Received 23 Ju y 2019; Received in revised form 31 October 2019; Accepted 6 November 2019 Available online 27 November 2019 0926-6690/ Published by Elsevier B.V.

Corresponding author at: Forest Products Laboratory, U.S. Forest Service, 1 Giford Pinchot Drive, Madison, WI, 53726, United States.

Y. Liu, e al. Industrial Crops & Products 145 (2020) 111955

N menclature

CB Bond’s comminution aw constant, kWh-mm1/2/kg Ck Kick’s comminution aw constant, kWh/kg CVR Von Rittinger’s comminution aw constant, kWh-mm/

kg E1-2 Energy consumption of mi ing the feedstock size into

the product size, kWh/kg MC Moisture content, % SEC specifc Energy consumption, kWh/kg x1 Feedstock size before mi ing, mm x2 Product size after mi ing, mm

the initia size and fna partic e size (Adapa et a ., 2011; Cadoche and Lopez, 1989; Naimi et a ., 2013). When targeting at a very sma screen size, there are strategies of going through severa intermediate screens and then feed the grindings to the target screen size (Temmerman et a ., 2013), which is a mu ti-stage mi ing. A so, onecan direct y feed raw materia s to the fna screen size, which is a sing e-stage mi ing. The diferences in the overa energy consumption among these diferent mi ing methods have not been disc osed. It is a so unc ear wheteher the fna product size is simi ar or not after subject to diferent mi ing routes. A so, the interaction between moisture content and partic e size change is of interest to investigate. Many grinding energy va ues re-ported in the iterature were generated with aboratory-sca e mi s. These mi s are designed to attain a product size without optimization for energy efciency. Moreover, test samp es are often ess than 1 kg, which may not generate a stab e mi ing regime. For examp e, the specifc energy consumption (SEC) of grinding a sawdust to a median size of 233 μm with a aboratory-sca e rotor impact mi was 1.844 kW h/kg (Karinkanta et a ., 2012), which is 46 % of the tota avai ab e wood energy. These aboratory-based va ues have created a perception that the mechanica grinding of wood to fne powders is too energy intense to be cost-efective. Few studies have been carried out to deve op routes to rea ize a cost-efective industria process of fne grinding woody biomass. Energy consumption and energy size re- ationship on an industria sca e have not been avai ab e. This in-formation barrier has hindered the understanding of a sca ab e process on which economic estimates can be based to eva uate commercia i-zation of fne wood grinding. Therefore, an industry sca e B iss hammer mi was emp oyed to eva uate the energy consumption for diferent degrees of size reduction under various moisture content eve s and grinding schemes.

The objective of this study was to investigate the efects of the mu ti-stage mi ing method on the specifc energy consumption of commi-nuting feedstock with various moisture contents and corre ate the en-ergy consumption with partic e sizes. Specifca y, the wood chips were

subject to a series of mu ti-stage grinding and the wood partic es were characterized; the Von Rittinger’s, Kirk’s, and Bond’s aws were used to describe the re ationship between energy consumption and partic e size change. The best ftted mathematic mode was se ected to predict the energy consumption under various moisture content and targeting partic e sizes.

2. Materials and meth ds

2.1. Ma erials

The Doug as-fr forest harvest residua s were produced from a roadside s ash pi e in southwest Oregon, USA. The s ash pi e was comminuted using an on-site horizonta grinder into a product passing through a 102 × 102 mm2 screen, which was de ivered to a faci ity and was then screened with a B ack-C awson gyratory screen (Mode 580, Seria 4095-76, B ack C awson, Everett, WA) with a 45-mm round-ho e punched p ate top deck to remove “overs”, and a 6 mesh (3.35 mm) woven-wire bottom deck to remove “fnes”. The accepted materia was air-dried for storage and abe ed as FS. The sieving ana ysis of the who e stock showed the geometric mean diameter of the raw materia s was 10.62 mm.

The moisture content of the forest residua s was determined ac-cording to ASAE Standard S358.2 (2006) for fve rep icate samp es and expressed as a fraction of tota mass. The FS moisture content was 13.1 %. Three other moisture content eve s were prepared by p acing the FS inside an environmenta chamber for two weeks with an equi ibrium moisture content setup at the 6, 18 and 28 % of moisture content, re-spective y. The fna moisture content was confrmed to be 6.7 %, 16.1 %, and 27.8 %, which covers a range of moisture content in typica wood grinding.

2.2. Me hods

2.2.1. Experimen al design The routes of the mu ti-stage mi ing are isted in Tab e 1. The two

target screen sizes of these six routes were 0.84 mm and 0.69 mm. The target screen size of 0.84 mm inc uded a sing e-stage mi ing, two-stage mi ing, and three-stage mi ing process. Whereas, the 0.69-mm target screen size had a sing e-stage mi ing, two-stage mi ing, and four-stage mi ing process. These two sizes are standard screen sizes and the sma est and second sma est size in which the hammer mi can be operationa stab y without causing overcharging and orifce b ocking. The se ected moisture content eve s for assessment are 6 %, 12 %, 18 %, and 28 %. Under each moisture content, a the diferent mi ing routes were carried out. Each mi ing samp e weighed 23 kg. Mass, moisture content, and partic e size distribution of the materia after each grinding were determined; the tota e ectrica energy used in each mi ing was recorded.

Table 1 Hammer mi screen size and mu ti-stage mi ing route se ection.

2

Y. Liu, e al. Industrial Crops & Products 145 (2020) 111955

Fig. 1. Partic e size distribution of the raw materia s at a moisture content of 6 % from (a) the three-stage mi ing and (b) the three diferent mi ing routes.

Table 2 Geometric mean diameter of the raw materia s and the grindings from the three-stage mi ing and four-stage mi ing with a moisture content of 6 %.

Target screen size (mm) Step Step Xgw (μm) Sgw (mm)

0.84 1 FS-0.84 232 2.2 2 FS-1.98 354 2.1

1.98-0.84 235 2.1 3 FS-3.18 486 2.1

3.18-1.17 325 2.1 1.17-0.84 255 2.2

0.69 1 FS-0.69 198 2.0 2 FS-1.17 255 2.0

1.17-0.69 171 2.3 4 FS-6.35 852 2.1

6.35-1.98 344 2.2 1.98-1.17 290 1.9 1.17-0.69 171 2.3

2.2.2. Energy consump ion of grinding process A F uke 1735 Power Logger (F uke Corporation, Everett, WA) was

used to measure the e ectrica energy consumption of the mi ing pro-cess of a hammer mi (B iss E iminator Fine Grind Hammer mi , EMF-24115-TFA, 45 kW, B iss Industries Inc. Ponca City, OK, USA)). The tota energy consumption, a sum of id e energy and net energy, was

Fig. 2. SEC under various routes with a moisture content of (a) 6 %, (b) 18 %, and (c) 28 %.

derived from the recorded active power and mi ing time. The id e energy can be ca cu ated through the base ine power of the hammer mi under a run without a charge and mi ing time. The net energy was ca cu ated by subtracting the id e energy from the tota energy (Liu et a ., 2016a).

=+

Net SEC Total Energy Idle EnergySample mass MC/(1 ) (1)

3

Y. Liu, e al. Industrial Crops & Products 145 (2020) 111955

Fig. 3. The moisture content efects on the net SEC under (a) various mu ti-stage mi ing routes, individua steps (b) in the three-stage mi ing of 0.84 mm and (c) in four-stage mi ing of 0.69 mm.

A previous study found that the tota specifc energy consumption (SEC) was minimized when the mi was operated at its rated power(Liu et a ., 2016a). Therefore, the norma ization of tota SEC was carried out by converting the rea ized power to the rated power, which can be achieved in practice through adjusting feeding rates. One assumption for this norma ization i.e. the net SEC did not change under various feeding rates. The norma ized tota SEC was obtained by the fo owing

Table 3 Parameters obtained from inear ftting with Von Rittinger’s, Kick’s and Bond’s aw.

Moisture content (%)

SEC Type Rittinger’s aw

R2CVR

Kirk’s aw

Ck R2

Bond’s aw

CB R2

6 Net 0.0218 0.89 0.0218 0.68 0.0512 0.85 Tota 0.0303 0.89 0.0303 0.68 0.0711 0.85

12 Net 0.0347 0.93 0.0287 0.50 0.0743 0.86 Tota 0.0489 0.93 0.0402 0.50 0.1039 0.86

18 Net 0.0398 0.82 0.0359 0.65 0.091 0.91 Tota 0.0556 0.82 0.0502 0.65 0.1271 0.91

28 Net 0.0548 0.58 0.0571 0.75 0.1387 0.92 Tota 0.0761 0.57 0.0795 0.76 0.1928 0.92

equation.

= ×Total SEC Net EnergyRated Power Baseline Power

Rated Power (2)

The base ine id e power and rated power were found to be 12 kW and 42 kW for the B iss hammer mi . Whereas, the net energy can be obtained by subtracting id e energy from tota energy from each actua run.

2.2.3. Par icle size charac eriza ion After mi ing, the resu ting partic es were sieved with a Ro-Tap

Shaker (The W. S. Ty er Company, C eve and, Ohio, USA) for 20 min to ana yze the size distribution, geometric mean diameter (Xgw), and standard deviation (Sgw) fo owing ASAE Standard S319.4 (2008).

2.2.4. Energy-size rela ionsgip Three comminution aws were used to describe the re ationship

between energy consumption and partic e size change. Von Rittinger’s aw, Kick’s aw, and Bond’s aw were used to ana yze the re ationship as shown in Eqs. (3)–(5) (Char es, 1957; Temmerman et a ., 2013).

=E Cx x1 1

VR1 22 1

=E C ln xxk1 2

1

2

(3)

(4)

=E Cx x1 1

B1 22 1 (5)

where E1-2 is the required energy per unit mass for reducing a partic e size from x1 to x2. x is a partic e size characteristic, which refers to the geometric mean diameter here. The numbers 1 and 2 represent the partic e size before and after grinding. C is a constant characteristic of the forest residua materia s.

3. Results and discussi n

3.1. Par icle size dis ribu ion and geome ric mean diame er

Fig. 1(a) shows the size distribution of the raw materia s and the resu ting partic es from each step of the three-stage mi ing of the 0.84-mm screen. Fig. 1(b) shows the size distribution of the fna partic es from three diferent mi ing routes of the 0.84-mm screen for the samp es with a mositure content of 6 %. The fna partic e size cumu- ative distribution curves from three diferent mi ing routes over ap most y as shown in Fig. 1b. The geometric mean diameter and standard deviation of the fna resu ting partic es from diferent mi ing routes are summarized in Tab e 2. The comparison of the geometric mean diameter found that there is no substantia diference among these diferent mi ing routes. It indicates that the screen size so y determines product size distribution.

4

Y. Liu, e al. Industrial Crops & Products 145 (2020) 111955

Fig. 4. The inear fttings between net SEC and partic e size change under various moisture contents.

3.2. Mul i-s age milling efec on SEC

Fig. 2 summarizes the net SEC and tota SEC under various mi ing routes and moisture contents. Fig. 2(a) shows the SEC under a moisture content of 6 %, whi e Fig. 2(b) and (c) show the corresponding data with a moisture content of 18 % and 28 %, respective y. As shown in Fig. 2(a), the sing e-stage mi ing process of the 0.84-mm screen con-sumed 21 % more than the three-stage mi ing process in both tota SEC and net SEC, whi e 43 % more when compared the sing e-stage mi ing of the 0.69-mm screen to the four-stage mi ing. Fig. 2(b) and (c) show the advantage of mu tip e-stage mi ing with a target size of 0.84 mm decreased a ong with an increase of moisture content from 6 % to 28 %.

Fig. 5. The re ationship between (a) Von Rittinger’s Constant and moisture content, (b) Bond’s Constant and moisture content.

However, the four-stage mi ing sti consumed ess energy than the two-stage mi ing with a target size of 0.69 mm. It can be conc uded that at moisture content eve of 28 %, the advantage of the mu tip e-stage mi ing method decreased when producing 0.84 mm partic e size grindings. In addition, the sing e-stage mi ing at the moisture content eve of 18 and 28 % were not ab e to be carried out due to the overheat of the mi motor resu ting from current spikes. Therefore, the mu ti-stage mi ing method is essentia when grinding materia s with a high moisute content and targeting an extra fne partic e size.

3.3. Mois ure con en efec on SEC

Fig. 3 summarizes the moisture content efects on the net SEC under mu ti-stage mi ing routes. Fig. 3(a) shows that the increasing moisture content resu ted in an increasing net SEC; the net SEC at the moisture content of 28 % was 2–3 times more than the one at the moisture content of 6 %. This fnding was a so in agreement with other studies (Liu et a ., 2016a; Mani et a ., 2004). Fig. 3(b) and (c) show the moisture efects on the energy consumption in each step of the three-stage mi ing for producing 0.84-mm grindings and the four stage mi ing for 0.69-mm grindings. It can observe that the moisture content substantia y infuences on the frst two stages of these two mi ing routes. As these stages are consecutive, the resu ted grindings become the feed materia s for the next step. A so, the moisture content de-creased due to the drying efects of the mi ing process. Therefore, the

5

Y. Liu, e al. Industrial Crops & Products 145 (2020) 111955

moisture content appears to have ess efect on mi ing energy con-sumption.

3.4. Analysis wi h he comminu ion laws

The three comminution aws were emp oyed to describe the re- ationship among the energy consumption, partic e size, and the moisture content of the feedstock. The experimenta data were ftted with the Von Rittinger’s, Kick’s and Bond’s aws as expressed by Eq. (3), (4), and (5), respective y (Temmerman et a ., 2013). The parameters were extracted by the east squaresmethod. Tab e 3 summarizes the resu ts from the inear ftting for a the four moisture content eve s with these comminution aws. Fig. 4 shows the inear ftting between the net SEC and stardardized size diferences. The inear ftting ob-tained by Von Rittinger’s aw and Bond’s aw are better than the ftting from Kick’s aw as indicated by R squares. At the high moisture content eve s, Bond’s aw performs better than Von Rittinger’s aw in the inear ftting. Genera y, the inear ftting with Bond’s aw obtained the best resu ts with the R-square va ues ranging from 0.85 to 0.92. Further-more, the increasing moisture content in feedstock resu ted in an in-creasing comminution constant, which is characteristic of grindabi ity of the feedstock.Fig. 5 demonstrates the changes of the Von Rittinger’s and Bond’s constants a ong with an increase of the moisture content in the feedstock. A inear ftting was emp oyed to obtaine the re ationship between moisture content and the constants. Substituting the Bond’s contant as a function of moisture content into Eq. (5), the fo owing equations are obtained:.

= × + ×Total SEC MCx x

(0.0055 0.0364) 1 12 1 (6)

= × + ×Net SEC MCx x

(0.0039 0.0259) 1 12 1 (7)

With these equations, one can rough y estimate the net and tota energy consumption in the range of investigated moisture content and partic e size change. These predictors of Eqs (6) and (7)wi faci iate the techno-economic ana ysis of hammer mi grinding.

4. C nclusi n

The mu tip e-stage mi ing process substantia y decreased the spe-cifc energy consumption of hammer mi ing process for producing fne partic es at ow moisture content. It a so provides advantages in grinding the materia s with high moisture content compared to sing e-stage mi ing. Bond’s aw was found to better corre ate the specifc energy consumption with partic e size change. A so, the constant in the Bond’s aw has a good inear re ationship with moisture content. Therefore, the deve oped Bond’s aw equations are recommended to be used to estimate the tota specifc energy consumption under various moisture content and partic e size change.

Declarati n f C mpeting Interest

The authors dec are that they have no known competing fnancia interests or persona re ationships that cou d have appeared to infu-ence the work reported in this paper.

Ackn wledgement

The study was supported by the Northwest Advanced Renewab es A iance (NARA), sponsored by the Agricu ture and Food Research

Initiative Competitive Grant (no. 2011-68005-30416) from the USDA Nationa Institute of Food and Agricu ture.

References

ASAE Standard S319.4, 2008. FEB. Method of determining and expressing fneness offeed. ASABE Standards. American Society of Agricu tura and Bio ogica Engineers,St. Joseph, Michigan, USA, pp. 663–665.

ASAE Standard S358. 2, 2006. FEB03. Moisture measurement-forages. ASABE StandardsVo . 608. American Society of Agricu tura and Bio ogica Engineers, St. Joseph,Michigan, USA, pp. 1–2.

Adapa, P., Tabi , L., Schoenau, G., 2011. Grinding performance and physica properties ofnon-treated and steam exp oded bar ey, cano a, oat and wheat straw. Biomass Bioenerg 35, 549–561.

Cadoche, L., Lopez, G.D., 1989. Assessment of size-reduction as a pre iminary step in the production of ethano from ignoce u osic wastes. Bio Waste 30, 153–157.

Char es, R., 1957. Energy-size reduction re ationships in comminution. Trans. AIME 208,80–88.

Chen, W.H., Pen, B.L., Yu, C.T., Hwang, W.S., 2011. Pretreatment efciency and struc-tura characterization of rice straw by an integrated process of di ute-acid and steamexp osion for bioethano production. Bioresour. Techno . Rep. 102, 2916–2924.

Dukes, C.C., Baker, S.A., Greene, W.D., 2013. In-wood grinding and screening of forest forbiomass feedstock app ications. Biomass Bioenerg 54, 18–26.

Gao, J., Anderson, D., Levie, B., 2013. Saccharifcation of reca citrant biomass and in-tegration options for ignoce u osic sugars from Catch ight Energy’s sugar process(CLE Sugar). Biotechno . Biofue s 6, 10.

Karinkanta, P., I ikainen, M., Niinimäki, J., 2012. Pu verisation of dried and screened Norway spruce (Picea abies) sawdust in an air c assifer mi . Biomass Bioenergy 44, 96–106.

Liu, Y., Wang, J., Wo cott, M., 2017a. Mode ing the production of sugar and byproductsfrom acid bisu fte pretreatment and enzymatic hydro ysis of Doug as-fr. Bioresour.Techno . Rep. 224, 389–396.

Liu, Y., Wang, J., Wo cott, M.P., 2016a. Assessing the specifc energy consumption andphysica properties of comminuted Doug as-fr chips for bioconversion. Ind Crop Prod94, 394–400.

Liu, Y., Wang, J., Wo cott, M.P., 2017b. Eva uating the efect of wood u trastructura changes from mechanica treatment on kinetics of monomeric sugars and chemica s production in acid bisu fte treatment. Bioresour. Techno . Rep. 226, 24–30.

Liu, Y., Wang, J., Wo cott, M.P., 2016b. Mu tistep process to produce fermentab e sugarsand ignosu fonates from softwood enzymo ysis residues. ACS Sustainab e Chem. Eng.4, 7225–7230.

Mani, S., Tabi , L.G., Sokhansanj, S., 2004. Grinding performance and physica propertiesof wheat and bar ey straws, corn stover and switchgrass. Biomass Bioenerg 27, 339–352.

Miao, Z., Grift, T.E., Hansen, A.C., Ting, K.C., 2011. Energy requirement for comminutionof biomass in re ation to partic e physica properties. Ind Crop Prod 33, 504–513.

Miura, T., Lee, S.-H., Inoue, S., Endo, T., 2012. Combined pretreatment using ozono ysisand wet-disk mi ing to improve enzymatic saccharifcation of Japanese cedar.Bioresour. Techno . Rep. 126, 182–186.

Naimi, L.J., Sokhansanj, S., Bi, X., Lim, C.J., Womac, A.R., Lau, A.K., Me in, S., 2013.Deve opment of size reduction equations for ca cu ating energy input for grinding ignoce u osic partic es. App . Eng. Agric. 29, 93–100.

Peng, H., Li, H., Luo, H., Xu, J., 2013. A nove combined pretreatment of ba mi ing andmicrowave irradiation for enhancing enzymatic hydro ysis of microcrysta ine ce - u ose. Bioresour. Techno . Rep. 130, 81–87.

Repe in, V., Govin, A., Ro and, M., Guyonnet, R., 2010. Energy requirement for fnegrinding of torrefed wood. Biomass Bioenerg 34, 923–930.

Shuai, L., Yang, Q., Zhu, J.Y., Lu, F.C., Weimer, P.J., Ra ph, J., Pan, X.J., 2010.Comparative study of SPORL and di ute-acid pretreatments of spruce for ce u osicethano production. Bioresour. Techno . Rep. 101, 3106–3114.

Temmerman, M., Jensen, P.D., Hébert, J., 2013. Von Rittinger theory adapted to wood chip and pe et mi ing, in a aboratory sca e hammermi . Biomass Bioenerg 56, 70–81.

Tumu uru, J.S., Tabi , L.G., Song, Y., Iroba, K.L., Meda, V., 2014. Grinding energy andphysica properties of chopped and hammer-mi ed bar ey, wheat, oat, and cano a straws. Biomass Bioenerg 60, 58–67.

Zakaria, M.R., Hirata, S., Hassan, M.A., 2014. Combined pretreatment using a ka inehydrotherma and ba mi ing to enhance enzymatic hydro ysis of oi pa m mesocarpfber. Bioresour. Techno . Rep. 169, 236–243.

Zhu, J.Y., Chandra, M.S., Gu, F., G eisner, R., Reiner, R., Sessions, J., Marrs, G., Gao, J., Anderson, D., 2015. Using su fte chemistry for robust bioconversion of Doug as-frforest residue to bioethano at high titer and ignosu fonate: a pi ot-sca e eva uation. Bioresour. Techno . Rep. 179, 390–397.

Zhu, J.Y., Zhu, W.Y., OBryan, P., Dien, B.S., Tian, S., G eisner, R., Pan, X.J., 2010. Ethano production from SPORL-pretreated odgepo e pine: pre iminary eva uation of mass ba ance and process energy efciency. App Microbio Biot 86, 1355–1365.

6