Compression and relaxation properties of timothy hay

Transcription

1 September, 211 Vol. 4 No.3 69 Compression and relaxation properties of timothy hay S. Talebi, L. Tabil, A. Opoku, M. Shaw (Department of Chemial and Biologial Engineering, University of Saskathewan 57 Campus Drive, Saskatoon, SK, Canada, S7N 5A9) Abstrat: The ompression and relaxation harateristis of timothy hay were investigated with respet to hay moisture ontent, applied load, and hay quality. Experimental trials were performed by using a Baldwin hydrauli universal testing mahine model 6 HVL The applied loads ranged from 9 kn to 24 kn, in 3 kn inrements. Two qualities of timothy hay were used, high quality with moisture ontents of 7.44%, 1.17%, 12.97%, and 16.42% wet basis (w.b.), and low quality with moisture ontents of 6.38%, 8.67%, 16.24%, and 18.94% w.b. The results indiated that the ompat density of hay samples inreased with inreasing moisture ontent and applied pressure. Less maximum applied pressure was required to ahieve the same ompat density with inreasing moisture ontent. Models were fitted to the applied pressure-ompat density data. Relationships were developed between the model onstants and the experimental variables. The use of Faborode-O Callaghan model for bale densities less than 5 kg/m 3 and the simple power law model for bale densities greater than 5 kg/m 3 are the most appropriate models expressing the relationship between density and pressure during the ompression of timothy hay. The relaxation of the hay samples were affeted by the initial maximum applied load or pressure and the moisture ontent. Samples with higher moisture ontents had higher perentage relaxation than low moisture ontent samples. The perent relaxation values ranged from 27.4% to 53.35% for the high quality hay, and for the low quality hay the values ranged from 28.8% to 53.7%. The asymptoti modulus values (E A ) were influened by the maximum applied pressure or load and moisture ontent. A linear relationship was developed between the asymptoti modulus, maximum applied pressure, and moisture ontent. Keywords: densifiation, ompation, moisture ontent, viso-elasti properties, density, asymptoti modulus, disinfestations, Canada DOI: /j.issn Citation: Talebi S, Tabil L, Opoku A, Shaw M. Compression and relaxation properties of timothy hay. Int J Agri & Biol Eng, 211; 4(3): Introdution Hay has been baled or densified into various sizes, shapes, and weights for many years to minimize storage spae, and ut down on transportation osts. A major issue in the business of exporting hay is the phytosanitary Reeived date: Aepted date: Biographies: Sana Talebi, B.S.Eng., E.I.T. Projet Manager, Atmospheri Environments, Stante, 2781 Lanaster Road Suite 2, Ottawa ON K1B 1A7 Canada. Researh and Work Interests: Physial properties of biologial materials, limate hange and air quality assessments, Greenhouse Gas (GHG) Verifiations, potable water assessments, water audits, wastewater sewer sampling and groundwater monitoring. Phone: ; Fax: ; Sana.Talebi@Stante.om; Anthony Opuku, M.S., Researh Assoiate, Department of Chemial and Biologial Engineering, 57 Campus Drive, College of Engineering, University of Saskathewan, Saskatoon SK S7N 5A9 Canada. Researh Interests: Pelleting of feeds and forage and optimizing the proess involved in feed and forage proessing; physial properties of agriultural materials and postharvest tehnology of agriultural rops; bioproess engineering; value-added engineering and postharvest handling of rops. Phone: ; Fax: ; ano38@mail.usask.a; Mark Shaw, M.S., P.Eng., Engineering Department Leader, Crestline Coah Ltd., 82-57th Street East, Saskatoon, SK S7K 5Z1 Canada. Researh Interests: Biomass pretreatment and densifiation. Phone: x 72; Fax: ; mshaw@ restlineoah.om. Corresponding author: Lope G. Tabil, PhD, P.Eng., Professor, Department of Chemial and Biologial Engineering, 57 Campus Drive, College of Engineering, University of Saskathewan, Saskatoon SK S7N 5A9 Canada. Researh Interests: Pelleting of feeds and forage and optimizing the proess involved in feed and forage proessing; physial properties of agriultural materials and postharvest tehnology of agriultural rops; bioproess engineering; value-added engineering and postharvest handling of rops; storage, drying and ooling of biologial materials; infrared spetrosopy; biomass proessing and utilization. Phone: ; Fax: ; lope.tabil@usask.a.

2 7 September, 211 Vol. 4 No.3 requirements of the importing ountries, due to the possibility of introduing Hessian fly (Mayetiola destrutor (Say)) from infested hay into importing ountries like Japan. The Japanese authorities have set up quarantine proedures in order to exlude infested produts from reahing their ountry. Buntin and Raymer [1] investigated the effet of Hessian fly damage on the forage prodution of suseptible and resistant ultivars of soft red winter wheat and indiated that the total dry matter yield of the forage was redued by 14% to 46%. The Hessian fly pupae live within hollow, stemmed plants (wheat, rie, barley, rye, and Agropyron speies). In ideal onditions, the fly an have up to two generations in one season. After the larva has reahed full growth, the skin hardens, and turns brown; this overing is alled the puparium. It is at this point that the inset enters a period of aestivation, or dormany [2]. Although Hessian fly prefers wheat plants, it an also infest timothy hay bales. It is possible that volunteer plants, suh as wheat or barley whih may be growing in a timothy hay field, may unintentionally be inorporated one the hay is ut and baled. After field baling, the timothy hay bales are transported to a entral proessing plant for rebaling or reompression, so as to further inrease the bale bulk density and inrease the apaity of shipping ontainers. Disinfestation studies performed by Shaw et al. [3,4] revealed that ompressing the bales is an effetive way of destroying Hessian fly puparia, thus, disinfesting the hay. Yokoyama et al. [5] determined that ompression alone ould be a suessful means of disinfesting ompressed hay, provided that a minimum pressure of 2.6 kpa (.21 kg/m 2 or 3 psi) ould be ahieved at all loations in the bale. Hay proessors in Western Canada already have existing re-baling mahines for mehanial reompression of field-ompressed (single-ompressed) baled hay into double-ompressed baled hay for export. Compression harateristis of baled hay depend on the ompression mahine, initial hay moisture ontent, quality, harvest ut number (whih ditates the oarseness/fineness and fiber ontent of hay) and type of hay and the time required (hold or grunt time) for ompression. Pitt and Gebremedhin [6] experimentally determined pressure-density relationship for hopped alfalfa and grass at different moisture ontents, hop lengths, and harvest ut number. They indiated that for alfalfa, the material stiffness (pressure to attain a given hange in density) was higher for seond-harvest material than for the first-harvest material. However, for grass material, the first-harvest was stiffer than seond-harvest. The effets of moisture ontent and hop length on alfalfa material stiffness were variable. The objetive of this study is to investigate how hay moisture ontent, applied load, and quality affet the ompression and relaxation harateristis of timothy hay. Knowledge of the aforementioned harateristis will allow for the predition of pressure in any loation within the hay bales. 2 Materials and methods 2.1 Materials Two different qualities of timothy hay were used throughout this experiment. The hays were obtained from Elan Forage In. of Broderik, SK, Canada, a hay proessor and exporter. The first ut hay ontained higher amounts of hay stalks and was designated as high premium (HP) or high quality hay, whereas, the seond ut hay onsisted mainly of leafy material with less amount of stalks and was designated as low premium (LP) or low quality hay. ontent of 14%. Both hays had an initial moisture Before testing, eah timothy hay quality was separated into 4 groups and their moisture ontents were adjusted to 6%, 1%, 15%, and 18% wet basis, w.b. In order to ahieve moisture ontents below the initial moisture ontent of 14%, the samples were dried until the desired value was ahieved. For moisture ontents above 14%, a alulated amount of deionised water was sprayed evenly over the fluffy material. All of the samples were sealed in large ontainers and kept at 5 in a walk-in ooler for over 14 days. Prior to the start of the experiment, moisture ontents of the samples were again verified to ensure aurate results. Moisture ontent was determined aording to ASABE Standard ASAE S358.2 FEB3 [7], where 25 g sample was dried in a onvetion oven at 13 for 24 h. Moisture was

3 September, 211 Compression and relaxation properties of timothy hay Vol. 4 No.3 71 determined in tripliates and expressed in perent wet basis (% w.b.). 2.2 Compression and relaxation tests The ompression tests were performed using a Baldwin Hydrauli Universal Testing Mahine Model 6 HVL-1254 (Sate System In., Grove City, PA). The mahine is apable of reahing a maximum load of 25 kn. The loading speed for all the samples was onsistently the same, at a value of (1.5±.25) mm/s. A omputer with a Labview (National Instruments, Austin, Texas) generated program was hooked up to the Baldwin mahine in order to reord time, fore, and distane data throughout the test. The omputer had a 624R DAQ ard installed, whih is a 12 bit multi-purpose input/output ard. A able onneted this to a SCXI-1 hassis, whih ontained the slots for data aquisition modules. Two modules were used in this testing proedure. The first was a SCXI 112D w/135 front plate, whih was used to read the voltage input from the load ell as well as the displaement sensor. The other module was a SCXI 118 w/132 front plate; this module allowed the user to feed an output voltage from the 624E DAQ ard to ontrol the loading and unloading rate. The interfae load ell used had a maximum apaity of kn and a resolution of 4.22 mv/v. Linear displaement during ompression was reorded with a Temposonis G-Series linear position sensor (MTS Sensors, Cary, NC). The ompression ell used to ompress the hay had a ross-setional area of 7 mm 7 mm. The initial height of the sample was determined by measuring the distane between the upper and lower loading plates with a digital alliper. During ompression, the displaement measured by the linear position sensor was subtrated from the initial distane measured to determine the instantaneous height of the sample. Hay samples were hand paked into the ompression ell. A sample size of 3 g was paked uniformly into the ell. The samples were ut to 127 mm lengths and layered into the ell in a rissross fashion. The experiments were designed with two fators: hay quality and hay moisture ontent. Two levels of hay quality and four levels of moisture ontent were used in the tests. The hay samples were ompressed to 6 different loads, starting at 9 kn and inreasing at intervals of 3 kn to 24 kn. These loads translate into maximum applied pressures of 5.58 MPa, 7.44 MPa, 9.3 MPa, MPa, 13.2 MPa and MPa. After eah desired load was reahed, the ompression mahine rosshead was stopped and its position was held onstant until a total elapsed time of 2 s had passed. During this time, the relaxation in the material was observed, where the fore with respet to relaxation time were reorded. After eah test, the sample was removed, weighed and sealed in a bag for post-experiment moisture ontent determination. One test was onduted for eah treatment. The data reorded by the omputer was used to determine the instantaneous veloity of the Baldwin mahine rosshead speed, the density of hay, and pressure with respet to time. 2.3 Data analysis The experimental data was analyzed by using several models whih have previously been used to desribe experimental ompression and relaxation data. The models were fitted to the experimental data using both SAS (Statistial Analysis System, Cary, NC) and Mirosoft Exel (Mirosoft Corp., Redmond, WA) programs. Cooper-Eaton model. The SAS program was used in the analysis of analyzed using the Exel program. The rest of the models were For the Mirosoft Exel analysis, the raw data was analyzed with both the solver funtion and non-linear regression. The onstants that were determined from the models were related to the experimental variables. The mean square error and the oeffiient of determination were used to determine the aeptability of the orrelation between the model onstants and the experimental variables. The ompressibility of powders has been analyzed on numerous oasions by many researhers. The models proposed by these researhers were assessed on their appliability to the ompression of timothy hay. Walker [8] desribed the volume ratio in terms of applied pressure for non-metalli powders, and partiles of sulfur, ammonium and sodium hloride, and trinitrotoluene (Equation (1)).

4 72 September, 211 Vol. 4 No.3 V m ln P b V (1) s Where, V = volume of ompated hay, m 3 ; V s = void-free solid volume, m 3 ; P = applied pressure, MPa; m, b = onstants. Jones [9] indiated that the relationship between pressure and density of industrial metal powders ould be represented by a straight line when plotted on log-log sales (Equation (2)). ln m ln P b (2) Where, ρ=ompat bulk density, kg/m 3 ; m', b'= onstants. Kawakita and Lüdde [1] related pressure to volume redution of metalli powders, Equation (3). P 1 P (3) C a b a C V V V (4) Where, C = volume ratio; V = initial volume at zero pressure, m 3 ; a 1, b 1 = onstants. The last model for ompression of powders that was used to desribe the hay ompression data was by Cooper and Eaton [11]. Equation (5) was reated to numerially depit the ompression of erami powders through two independent probabilisti proesses. The first of whih was the filling of large voids through material sliding past one another and slight fratures. The seond proess inluded the filling of small voids through plasti flow and fragmentation. V V V V s a e k1 k2 P P 2 a3e Where, a 2, a 3, k 1, k 2 = onstants. The ompression harateristis of straw were analyzed by both Pitt and Gebremedhin [6], and Faborode and O Callaghan [12]. Faborode and O Callaghan s [12] model (Equation (6)) was assessed to determine its appliability to timothy hay ompression. Pitt and Gebremedhin [6] provided a model (Equation (7)) that was a slight modifiation of Faborode and O Callaghan s model. Both are exponential models that express pressure in terms of ompressed straw density. The ompression of straw an be modeled by these equations up to a density of 5 kg/m 3, after whih the equations beome inaurate. (5) P A b b2 ( 1) [ e 1] (6) 2 P h e f ( ) [ 1] (7) Where, ρ = initial bulk density, kg/m 3 ; β = dry matter density, kg/m 3 ; β = ompat dry matter density, kg/m 3 ; A, b 2, f, h = onstants. A model to desribe the relaxation harateristis of solid foods was developed by Peleg and Moreyra [13]. Relaxation urves at different deformation levels were normalized and fitted to the model (Equation (8)). The onstants in the model give the shape harateristis of different types of material and an be used to ompare different materials. F t k3 k4 t (8) F F( t) Where, F = initial relaxation fore, kn; F(t)= relaxation fore at time t, kn; t =time, s; k 3, k 4 = onstants. Moreyra and Peleg [14] (198) modified Peleg and Moreyra s [13] model and indiated that the slope should be greater than one (Equation (9)). The slope an be used as an index whih gives an indiation of the solidity of the ompressed material. The slope an also be used in determining the asymptoti modulus of food powders and solid foods [15]. The asymptoti modulus is the residual stress in the Peleg s relaxation model and an be defined as the ability of the ompressed material to sustain un-relaxed stresses. E A F 1 1 (9) Aa k4 Where, E A = asymptoti modulus, MPa; ε= strain; A a = ross-setional area, m 2. The perent average relaxation was alulated by using the initial fore at the beginning of the relaxation phase and the final fore after an elapsed time of 2 s. The equation for alulating the perent average relaxation is given in Equation (1) as follows: 1 ( F F ) Perent average relaxation e (1) F Where, F e = final relaxation fore, kn. 3 Results and disussion 3.1 Moisture ontent

5 September, 211 Compression and relaxation properties of timothy hay Vol. 4 No.3 73 The moisture ontent values of timothy hay measured before and after testing showed that all of the samples exept for those adjusted to 6%, lost moisture during storage and testing (Table 1). The variation in the moisture ontents before and after testing ould be attributed to the variable moisture distribution in the hay during onditioning and storage. Also, relative humidities are approximately 3% or lower in the storage, making samples to lose moisture during storage. Table 1 Moisture ontent of hay samples before and after eah test Moisture ontent /% w.b. Sample ID Before test After test LP-18* (1.17) (1.37) LP (.29) (.39) LP (.6) 8.76 (.62) LP (.14) 6.38 (.26) HP-18** (.3) (.57) HP (.6) (.27) HP (.31) 1.17 (.24) HP (1.98) 7.44 (.88) Note: * LP-## stands for low quality hay, followed by its expeted moisture ontent in % wet basis. ** HP-## stands for high quality hay, followed by its expeted moisture ontent in % wet basis. Values in brakets represent standard deviation, n = 3. an inrease in moisture ontent signifiantly redued the maximum applied pressure required to ahieve the same ompat density. Bilanski [16] The same trend reported by Watts and was also observed in this investigation. During the relaxation portion of the test, the densities remained onstant. Figure 1 shows a typial ompression urve of the timothy hay samples at different levels of preset ompression pressures. Table 2 Density of ompressed hay (kg/m 3 ) at different maximum applied pressures and moisture ontents Moisture Applied pressure /MPa ontent /% w.b High quality timothy hay Low quality timothy hay Hay density Table 2 shows the effets of applied pressures and moisture ontent on the ompat density of the timothy hay samples. The ompat density of the hay samples inreased with inreasing moisture ontent. Comparing the ompat density of the hay samples at almost similar moisture ontents of 16.24% for low quality and 16.42% for high quality, it is observed that there were marginal differenes in their ompat densities. Hay quality did not appear to influene the ompat density of the hay samples. The density of the high quality hay at 7.44% moisture ontent was slightly higher than the low quality hay at 8.76% for applied pressures of 13.2 MPa and MPa. The density of the samples inreased with the inreasing applied pressures for a partiular moisture ontent. Watts and Bilanski [16] reported that an inrease in moisture ontent inreased the ompat density of alfalfa samples, and that the maximum applied pressure also inreased the ompat density. They indiated that Figure 1 Typial urves of hay density as a funtion of ompression and relaxation time at different preset levels of ompression pressures for high quality hay at a moisture ontent of 1.17% w.b. As the pressure inreased, the rate of density hange dereased. This is a result of the hay displaing and depleting the amount of air gaps and voids ontained within the sample ell. Narayan and Bilanski [17] also observed the same trend in their studies of uniaxial ompression of wheat straw. They indiated that at low pressure, partile reorientation dominated the load

6 74 September, 211 Vol. 4 No.3 response; whereas at high pressure, individual partile deformation affeted the load behavior. The horizontal setions of the urves are when the ompression of the hay samples ends and relaxation begins. During relaxation, the density remains onstant at onstant strain and the pressure dereases with time. 3.3 Compression models The relationship between pressure, volume, and density of the timothy hay during the ompression portion of the tests were fitted to models that have been developed for powders. The first of whih was Walker s model [8]. This model yields values of volume ratio whih derease linearly as the pressure inreases. The graph shows that the volume ratio of ompressed hay was high at low hay moisture ontent. In other words, low moisture ontent hay samples produed ompressed samples of higher volume ratios whih indiate that low moisture samples require more energy to ompress than high moisture samples. Walker s model produed an average oeffiient of determination value of 9% (R 2 =.9). Figure 2 shows the relationship between the volume ratio and the natural logarithm of applied pressure. standard deviation of.4957 (14%). The high quality hay showed a higher slope (absolute value) indiating higher ompressibility than the low quality hay. The differenes ould probably be attributed to the make up of the hays. The low quality hay onsisted of more leafy material ompared to the high quality hay whih had more stalks. The leafy material in the low quality hay resulted in its low ompressibility. The slope m, slightly dereased with inreasing moisture ontent. A relationship was established between the hay moisture ontent and the slope m. The value of b peaked between a moisture ontent of 1%-15%, before and after whih, the value of b dereased. A relationship was observed between the moisture ontent of the samples and the onstant b. These relationships varied slightly between the two hay qualities and are shown in Equations (11) to (14). The R 2 of the model onstants of high quality hay were higher ompared to the low quality hay. High quality hay: m M.2 M [R 2 =.999, Standard error =.36] (11) b M.22 M [R 2 = 1., Standard error =.17] (12) Low quality hay: 3 3 m M.4 M [R 2 =.783, Standard error =.165] (13) 3 Figure 2 Fitting of Walker's model to the experimental ompression data for high quality timothy hay The value of ompressibility onstant, m, the slope, is relatively onsistent, and is seen to stay more onstant in the high quality hay samples. For high quality samples m had an average value of with a standard deviation of.573. While the low quality samples m had an average value of with a b M.4 M [R 2 =.826, Standard error =.1561] (14) Where, M = hay moisture ontent, % w.b. Jones [9] model was not suessful in modeling the ompression harateristis of timothy hay. No onsistent relationship was determined between the model onstants and the experimental variable. R 2 values obtained from Kawakita and Lüdde [1] model ranged between.997 and 1.. However, no orrelation was found between the model onstants in this equation and the experimental variables. The values of a 1 and b 1 inreased with moisture ontent, but no suitable trend ould be established. Cooper and Eaton s [11] model eluidates the two mehanisms that exist during the ompression of powders. 3

7 September, 211 Compression and relaxation properties of timothy hay Vol. 4 No.3 75 The mehanisms inlude the filling of large pores through partile rearrangement and the filling of smaller pores through fragmentation and plasti flow. These two proesses are represented by the onstant a 2 and a 3, respetively, in the Cooper-Eaton model. The model analysis yielded values for a 2 ranging from.63 to.99, while the values of a 3 ranged from values of. to.26 (Table 3). This suggests that majority of the ompression ours by filling the large pores through partile rearrangement. There is a slight indiation of ompression through plasti flow and fragmentation, whih is most likely due to the presene of finer partiles distributed throughout the majority large partiles. The values of k 1 and k 2 showed a lot of variations and no trend was observed between them and the experimental variables. The R 2 between the experimental and alulated values of the volume ratio were all higher than.72 with majority of them near.9. Table 3 Cooper-Eaton model (Equation (5)) onstants for both hay qualities Sample ID a 2 a 3 k 1 k 2 R 2 LP- 6* LP LP LP HP- 6** HP HP HP Note: * LP-## stands for low quality hay, followed by its expeted moisture ontent in % w.b. ** HP-## stands for high quality hay, followed by its expeted moisture ontent in % w.b. Faborode and O Callaghan s [12] model (Equation (6)) was able to predit aurately the relationship between pressure and ompat density for ompat densities less than 5 kg/m 3. The R 2 were always equal to 1.. Unlike Pitt and Gebremedhin s [6] model (Equation (7)), the values of the model onstants determined by using Faborode-O Callaghan s model ould be predited using the experimental variables. It was determined that the onstant b 2 (Equation (6)) did not orrelate well with the hay moisture ontent. The average and standard deviation of b 2 for all the hay samples were 1.2 and.16. The values of A ould be related to the wet basis moisture ontent, M, by the following relationship: A = -95M + 21 (15) Therefore, the new form of Faborode-O Callaghan model for ompressing timothy hay at densities of less than 5 kg/m 3 an be written as follows. ( M ) 1.2( 1) P [ e 1] (16) 1.2 The R 2 values between the experimental and alulated A values for the high quality and low quality hay were.844 and.91, respetively. The mean square error for these same data points were and for high quality and low quality hay, respetively. Pitt and Gebremedhin [6] model aurately desribed the relationship between pressure and density with R 2 values equal to 1.. However, the values of the model onstants did not have any relationship with the experimental variables. Therefore, this equation was not useful for modelling the orrelation between the model onstants and the experimental variables. It was determined that a simple power law model of the following form: P B (17) ( B1 ) 2 ould aurately relate the applied pressure to the hay sample ompat density above 5 kg/m 3. Where, B 1, B 2 = onstants. The average value of onstant B 1 was determined to be 3.36 with a standard deviation of.17 for both hay qualities. Van Pelt [18] obtained average B 1 values of.29,.24,.24 and.23 for dry orn stalks, soybean straw, wet orn and dry alfafa hay, respetively. The value of B 2 varied with moisture ontent. A relationship between the onstant B 2 and moisture ontent was established for both hay qualities. The relationship between B 2 and moisture ontent was different for the two hay qualities used in this experiment. Equation (18) shows the power law model for both hay qualities at densities above 5 kg/m 3. P B (18) ( 3.36 ) 2 where, for high quality timothy hay: and for low quality hay: ( ) M B2 exp (19)

8 76 September, 211 Vol. 4 No.3 B M (2) The R 2 values for the alulated versus experimental values were.834 and.832 for high and low quality hays, respetively. When omparing the alulated values of pressure versus the experimental values, the oeffiient of determination for all the samples was 1.. This shows that a simple power law model an suessfully predit the relationship between pressure and density during the ompression of timothy hay above densities of 5 kg/m Relaxation harateristis The hay samples experiened a relatively uniform inrease in ompression pressure as the loads were applied. After the desired pressures were reahed, the relaxation harateristis of the ompressed samples were observed. The initial loading of the samples followed the same urve pattern, while the relaxation urves were almost equally spaed and followed the same pressure deaying pattern. The slight flutuations in the urves may be due to the relaxation harateristis of the hay samples. The relaxation urves were affeted by the initial maximum applied load. The higher the applied pressure, the higher is the residual stress that is left in the sample. Typial ompression-relaxation urves for the hay samples are presented in Figure 3. Tabil and Sokhansanj 19] observed similar ompression-relaxation harateristis during the ompation of alfalfa grinds. The ompression-relaxation harateristis of the hay samples show that the relaxation behavior of timothy hay is time-dependent. Hay samples exhibited both elasti and visous harateristis, a behavior exhibited by most agriultural materials. This shows that timothy hay is a visoelasti material. Sine large strain stresses were applied and all the strains were not reovered, timothy hay samples might have exhibited nonlinear visoelasti behavior. Zoerb and Hall [2] determined that agriultural produts suh as pea bean exhibited visoelasti behavior. Negi et al. [21] similarly reported that silages exhibited nonlinear visoelasti behavior. As the moisture ontents of the timothy hay samples were inreased, the perent relaxation also inreased for both hay qualities (Table 4). The moisture ontent of the samples had an evident effet on the perent relaxation. The high quality hay illustrates this trend more than the low quality hay. The effet of applied load on the perent relaxation of hay was random. The hay qualities did not seem to affet the perent relaxation, as no major trend was observed. The perent relaxation values ranged from 27.4% to 53.35% for the high quality hay, and for the low quality hay the values varied from 28.8% to 53.7%. Figure 3 Typial pressure vs. time urves for varying maximum applied pressures during ompression and relaxation for high quality timothy hay at 1.17% Table 4 Applied pressure /MPa Effets of applied pressure and moisture ontent on perent relaxation of high quality and low quality hay High quality hay perent average relaxation /% Low quality hay perent average relaxation /% MC* = 7.44% MC = 1.17% MC = 12.97% MC = 16.42% MC = 6.38% MC = 8.76% MC = 16.24% MC = 18.94% Note: *MC = moisture ontent of the hay before testing or after testing.

9 September, 211 Compression and relaxation properties of timothy hay Vol. 4 No.3 77 The internal pressure of the ompressed hay samples delined towards an asymptoti value during the relaxation period. It was observed that the slope (k 4 ) of Equation (8) dereased with inreasing moisture ontent and was relatively unaffeted by the maximum load or pressure. This suggests that timothy hay is less rigid with inreasing moisture ontents. Moreyra and Peleg [14] and Peleg and Moreyra [13] reported that the slopes (k 4 values) of wet materials were lower than dry materials. Peleg and Moreyra [13] suggested that this phenomenon ould be attributed to the internal flow of liquid in the sample and rearrangement of liquid bridges. The values of E A an be thought of as a measure of the sample rigidity, the less rigid a sample is, the smaller its asymptoti modulus. Similar to the values of k 4, the values of E A dereased with inreasing moisture ontent. Furthermore, the values of E A inreased with inreasing maximum load or pressure. The applied load affeted E A in this manner beause a higher applied load ould ause a higher density or solid-like properties. Table 5 shows the asymptoti modulus of all samples tested. The inrease in E A values with respet to inreasing applied pressures have been reported by other researhers [19,23] as well. Table 5 Asymptoti modulus (E A ) of timothy hay as a funtion of hay quality, maximum pressure, and moisture Hay quality Moisture ontent/% ontent Maximum applied pressure /MPa Asymptoti modulus /MPa Low Low Low Low High High High High The manner in whih the E A values varied with respet to the moisture ontent and the maximum applied pressure were analyzed. It was determined that a linear relationship existed as indiated by Equation (21). The relationship an be applied to both hay qualities. The R 2 values between experimental and alulated E A was determined to be.9, and the mean square error was.63. Tabil and Sokhansanj [19] reported a power law relationship between the maximum applied pressures and the E A values. Mani et al. [23] presented a linear relationship between the E A values and the applied pressures. In this study, E A was a linear funtion of both the maximum applied pressure and moisture ontent of the hay sample. E.787P.93M (21) A M Where, P M = maximum applied pressure, MPa. 4 Conlusions The ompression and relaxation harateristis of timothy hay samples were investigated and the following onlusions an be drawn: 1) The ompat density of the hay samples inreased with inreasing moisture ontent and maximum applied pressure. Less maximum applied pressure was required to ahieve the same ompat density with inreasing moisture ontent. 2) Models were fitted to the applied pressure-ompat density data. Relationships were developed between the model onstants and the experimental variables. 3) The use of Faborode-O Callaghan model for bale densities less than 5 kg/m 3 and a simple power law model for bale densities greater than 5 kg/m 3 appropriately expressed the relationship between pressure and hay density during the ompression of timothy hay. 4) The perent relaxation of the hay samples were affeted by the maximum applied pressure and the moisture ontent of hay. Hay samples with higher moisture ontents had higher perent relaxation than low moisture ontent samples. The perent relaxation values ranged from 27.4% to 53.35% for the high quality hay, and from 28.8% to 53.7% for the low quality hay. 5) The asymptoti values (E A ) were influened by the maximum applied pressure and hay moisture ontent. A linear relationship was developed between the asymptoti modulus, maximum applied pressure and moisture ontent. Aknowledgements Aknowledgments for the finanial support are the

10 78 September, 211 Vol. 4 No.3 Natural Siene and Engineering Researh Counil (NSERC) and Saskathewan Agriulture and Food through Agriulture Development Fund. The authors also thank Hunterwood Tehnologies Ltd. (Cohrane, AB) for their tehnial support and Elan Forage In. (Broderik, SK) for providing the experimental samples. [Referenes] [1] Buntin G D, Raymer P L. Hessian fly (Diptera: Ceidomyiidae) damage and forage prodution of winter wheat. Journal of Eonomi Entomology, 1989; 82(1): [2] Webster F M. The Hessian fly. Farmers Bulletin, United States Department of Agriulture; no. 64. Washington, D.C p. [3] Shaw M D, Crerar W J, Opoku A, Tabil L G, Whistleraft J W. Test disinfestations via ompression of Hessian fly in Canadian hay for export to Japan. Progress report, Saskatoon, SK, Canada: Department of Agriultural and Bioresoure Engineering, University of Saskathewan, 24. [4] Shaw M D, Crerar W J, Opoku A, Tabil L G, Ghazanfari-Moghaddam A, Whistleraft J. Evaluation of timothy hay ompression as a disinfestation protool for the Hessian fly (Mayetiola destrutor (Say)). ASABE-CSBE Paper SD5-14, St. Joseph, MI: ASAE, 25. [5] Yokoyama V Y, Hathett J H, Miller G T. Hessian fly (Diptera: Ceidomyiidae) ontrol by ompression of hay for export to Japan. Journal of Eonomi Entomology, 1993; 86(3): [6] Pitt R E, Gebremedhin K G. Effets of forage speies, hop length, moisture ontent, and harvest number on tower silo apaity and wall loads. Journal of Agriultural Engineering Researh, 1989; 44(3): [7] ASAE Standard S358.2 FEB3. Moisture Measurement - Forages. In: ASABE Standards, 68. St. Joseph, Mih.: Amerian Soiety of Agriultural and Biologial Engineers, 26. [8] Walker E E. The properties of powders Part VI: The ompressibility of powders. Transations of the Faraday Soiety, 1923; 19: [9] Jones W D. Fundamental priniples of powder metallurgy. London, UK: Edward Arnold Publishers Ltd., 196; [1] Kawakita K, Lüdde K H. Some onsiderations on powder ompression equations. Powder Tehnology, 1971; 4: [11] Cooper A R, Eaton L E. Compation behavior of several erami powders. Journal of the Amerian Cerami Soiety, 1962; 45(3): [12] Faborode M O, O Callaghan J R. Theoretial analysis of the ompression of fibrous agriultural materials. Journal of Agriultural Engineering Researh, 1986; 35(3): [13] Peleg M, Moreyra R. Effet of moisture on the stress relaxation of ompated powders. Powder Tehnology, 1979; 23(2): [14] Moreyra R, Peleg M. Compressive deformation patterns of seleted powders. Journal of Food Siene, 198; 45(4): [15] Soville E, Peleg M. Evaluation of the effets of liquid bridges on the bulk properties of model powders. Journal of Food Siene, 1981; 46(1): [16] Watts K C, Bilanski W K. Stress relaxation of alfalfa under onstant displaement. Transations of the ASAE, 1991; 34(6): [17] Narayan C V, Bilanski W K. Behavior under high pressure of wheat grains in bulk. ASAE Paper No , St. Joseph, Mih: ASAE, [18] Van Pelt T. Maize, soybean, and alfalfa biomass densifiation. Agriultural Engineering International: the CIGR Ejournal, 23; Vol. V. Manusript EE 3 2. [19] Tabil L G, Sokhansanj S. Bulk properties of alfalfa grind in relation to its ompation harateristis. Applied Engineering in Agriulture, 1997; 13(4): [2] Zoerb G C, Hall C W. Some mehanial and rheologial properties of grains. Journal of Agriultural Engineering Researh, 196; 5: [21] Negi S C, Ogilvie J R, Jofriet J C. Some mehanial and rheologial properties of silages. Canadian Agriultural Engineering, 1987; 29(1): [22] Peleg M, Moreyra R. Effet of moisture on the stress relaxation pattern of ompated powders. Powder Tehnology, 1979; 23(2): [23] Mani S, Tabil L G, Sokhansanj. S. Effets of ompressive fore, partile size and moisture ontent on mehanial properties of biomass pellets from grasses. Biomass and Bioenergy, 26; 3: