FACTORS AFFECTING THE PELLETING OF HAY By James Lee Butler AN ABSTRACT Submitted to the School for Advanced Graduate Studies of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Engineering 195B Approve d 7^ 1 Since the physical forms in which hay is harvested at the present time are not adapted to mechanical handling, some other form must be utilized* Compressing the hay into small, dense units or pellets appears to be a method of changing the form into one which can be mechanically handled and at the same time allow more efficient utilization of transportation and storage facilities. A review of literature revealed a considerable amount of material dealing with the advantages of pelleted hay. Most of this literature dealt with very small pellets made in a ring die type of pelleting mill. Since this pro­ cess requires that the hay be finely ground before pellet­ ing, it does not appear practical for field operation. Very little information was available concerning the effect of the various variables encountered in the pelleting of long or chopped hay in sizes practical for a field operation. In order to determine the effect of some of these variables, a pelleting apparatus, using electrical strain gage transducers with amplifiers and a recorder to measure pressure and displacement, was made. Data were taken to determine the effect of pressure, time of application of pressure, pellet size, diameter of pelleting chamber, treatment of hay before pelleting, bind­ ing materials, and moisture content on the production of hay pellets. Preliminary investigation comparing the angle of repose and pellet density-bulk density relationship for James Lee Butler pellets with various length to diameter ratios was made. The handling durability of pellets made from long hay was compared to that of pellets made from chopped hay. These data, obtained with alfalfa hay, indicate the following conclusions: a) the density of pellet obtained is linear with the logarithm of the pressure applied; b) pellet density may be increased, other variables being held constant, by increasing the sample weight; c) pellet density is linear with the logarithm of time of pressure applica­ tion; d) uneven distribution of hay across the pelleting chamber increases unit power requirements and decreases pellet density; e) the use of binding materials to increase pellet density is not practical; f) pellets made from long hay will withstand more handling than pellets of equal den­ sity which are made from chopped hay; g) as much as 63 per cent of the moisture may be squeezed out of freshly cut hay without the dry matter loss exceeding that normally incurred in field curing; h) firm pellets which will withstand mechanical handling can be produced from long hay when the moisture content is 25 per cent or less; i) pellets having a length to diameter ratio approaching unity have better flow characteristics and greater bulk densities for a given pellet density. James Lee Butler FACTORS AFFECTING THE PELLETING OF HAY By James Lee Butler A THESIS Submitted to the School for Advanced Graduate Studies of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Engineering 195$ ProQuest Number: 10008562 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest ProQuest 10008562 Published by ProQuest LLC (2016). Copyright of the Dissertation is held by the Author. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code Microform Edition © ProQuest LLC. ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106-1346 ACKNOWLEDGEMENTS The author wishes to express his sincere apprecia­ tion for the counsel, guidance, and support of Professor H. F. McColly, who supervised the investigation upon which this thesis is based. He also wishes to thank the other members of the guidance committee, Doctors J. S. Boyd, M. L. Esmay, J. S. Frame, R* T. Hinkle, H. Larcher, G. H. Martin, and C. P. Wells, for their suggestions and guidance. A word of thanks is also due Doctor W. F. Buchele for his assistance and many helpful suggestions. The author is grateful to Professor A. W. Farrall, Head of the Agricultural Engineering Department, for pro­ viding the personnel and facilities which made stimulating working conditions for graduate study. Sincere thanks is extended to the J. I. Case Company of Racine, Wisconsin, for the research grant that made this work possible. Special thanks is due Mr. W. J. Weiss of that company for his interest and encouragement. Special thanks is also extended to the Georgia ' Experiment Station for granting one year of sabbatical leave which helped to make this study possible. The faith and support of the authorfs wife, Jane, is gratefully acknowledged. ii Thanks is also extrended to all other persons who contributed of their time and experience to this investi­ gation. iii James Lee Butler candidate for the degree of Doctor of Philosophy Final examination: Dissertation: August 1, 1953, 3:00 A.M., Room 213 Agricultural Engineering Building Factors Affecting the Pelleting of Hay Outline of Studies Major Subject: Minor Subjects: Agricultural Engineering Mechanical Engineering Mathematics Biographical Items Born: January 8 , 1927, Sevierville, Tennessee Undergraduate Studies: Graduate Studies: University of Tennessee 1947-50, B.S., 1950 University of Tennessee, 1950-51, M.S., 1951 Michigan State University, 1956-53 Experience: Cryptographer, United States Air Force, 1944-46 Assistant Agricultural Engineer, Georgia Experiment Station, 1951-56 Graduate Teaching Assistant, Michigan State University, 1956-57 Graduate Research Assistant, Michigan State University, 1957-53 Honorary Societies: Pi Mu Epsilon Society of Sigma Xi Professional Societies: American Society of Agricultural Engineers iv TABLE OF CONTENTS Page INTRODUCTION ........................................... 1 REVIEW OF L I T E R A T U R E .................................. 5 THE INVESTIGATION . . . .............................. 14 Part I Apparatus Part II Procedure .......... .............14 ................... Procedure for 1957 . . . 23 ............. 23 Procedure for 1953 .............. RESULTS AND DISCUSSION ................................ The 1957 Investigation............ 34 34 ................. The 1953 Investigation 30 Density Relationships 36 . . 36 Energy Requirement Relationships * ........... 48 Effect of Binding M a t e r i a l s .......... 39 Effect of Moisture Content Handling Characteristics . . ♦ ............. 59 . . 61 S U M M A R Y ................................................. 63 C O N C L U S I O N S .......................................... 70 SUGGESTIONS FOR FURTHER STUDY 72 ....................... R E F E R E N C E S ............................................... 74 v LIST OF FIGURES Figure 1. Page Hydraulic press and pelleting apparatus used in the 1957 investigation..................... • 15 2. Pelleting apparatus used in the 195# investi­ gation ........................................... 16 3. Schematic diagram of pelleting apparatus 4* Pressure transducer • . . ......... 5* Calibration curve for pressure transducer • • • 22 6* Displacement transducer . . • • . • • . • • • • 24 7. Tilting board arrangement used to determine the angle of repose of pellets * • • « • • • • • • 26 Special slide rule constructed to solve the equation for pellet density ................... 26 6. 9* 10. • * . . . . . . . 17 20 Pellet density vs. length of time of applica­ tion of p r e s s u r e ................................ 37 Pellet density vs. length of time of pressure application ( s e m i - l o g ) • • * • 36 11. Per cent of pellet expansion vs. t i m e ........... 39 12. Pellet density vs. pressure— 2.40” diameter chamber .............. 40 Pellet density vs. pressure— 2.40” diameter chamber (semi-log) ....................... 42 Pellet density vs. pressure— 1.50” diameter chamber (semi-log) . . . ...................... 43 Pellet density vs. pressure— 3.42” diameter chamber (semi-log) ....................... 44 16. Effect of sample weight on pellet density . . . 46 17. Wedge-shaped appearance of 40 gram pellet made in 3.42” diameter c h a m b e r ........................ 46 13* 1415. vi LIST OF FIGURES (Cont.) Figure 18* 19* 20. Page Effect of pelleting chamber diameter on density of 40 gram chopped hay pellet (semi-log) * . . 49 Pellet density vs* pressure for long hay-1*50" diameter chamber (semi-log) • * • • • ......... 50 Pellet density vs* pressure for long hay-2*40" diameter chamber (semi-log) ................... 51 21. Pellet density vs. pressure for long hay-3-42" diameter chamber (semi-log) * ....................52 22. Pressure vs. displacement for different samples pelleted in 2.40” diameter chamber ........... 54 Pressure vs. displacement for 40 gram sample pelleted in different size chambers . • • • • • 56 24* Pellet density vs. moisture content (semi-log). 62 25* Appearance of long and chopped hay pellets be­ fore and after being subjected to handling durability test • • • • • ........ • • • • • • 66 Oscillograph chart record 66 23* 26. vii LIST OF TABLES Page Table I II III IV Nutrient loss in juice removed from 75 per cent moisture alfalfa hay by three different methods 34 Effectiveness of Ceredex No. 265 as a binding a g e n t .............................. ............ 60 Comparison of pellet density-bulk density rela­ tionship and angle of repose ................. 63 Comparison of pellet handling durability 65 viii ... INTRODUCTION AND OBJECTIVES Introduction In terms of number of acres harvested and tonnage, hay is the most important harvested feed for livestock. The average annual production in the United States is about 100 million tons (1). Even though it holds this distinc­ tion, Kleis (13) found, in studying materials handled on Michigan livestock farms, that the greatest labor require­ ment in terms of man hours per ton is with long, loose hay, baled hay and chopped hay in that order; the average being 2.77, 1*57, and 1.37 man hours per ton, respectively. These figures include time for unloading from wagon, stor­ ing, and removal from storage and feeding. The amount of hay harvested in the long, loose form has decreased rapidly since the end of World War II. In 1944 this form accounted for about two-thirds of the crop, in 1951 only 30 per cent of the U.S. crop was harvested as long, loose hay, 63 per cent was baled and 7 per cent was chopped (1). Long loose hay is bulky, allowing relatively lowtonnage loads to be hauled and requiring about 400 cubic feet of storage space per ton. In addition, because of its intermingling characteristics, it is difficult to handle and has high labor requirements per ton. Baling puts hay in a more dense and convenient form 2 for handling, thus changing the character of the work. Although the character of the work is changed, in many cases the hay may be handled more times in getting it from the field into storage when in the baled form than is nec­ essary when it is in the long, loose form. It is common practice to allow the baled hay to drop back onto the ground from the baler and then perform the following oper­ ations by hand; load hay onto trailer or wagon, unload onto conveyor, remove from conveyor and stack into hay mow. Before the hay can be fed, it must be moved from the mow to the feeding area and broken apart, requiring additional hand labor. Chopped hay is usually blown from the chopper into the transporting vehicle and pneumatically or mechanically conveyed from this vehicle into storage. It can then be self-fed from storage or (although not commonly done) pneu­ matically conveyed to the feeding area. Thus, of the three methods of harvesting, only chopping lends itself to com­ plete mechanization. (Complete mechanization implies that no lifting by human hand is necessary.) Although it posses this advantage, chopped hay har­ vesting has not been very widely accepted. In fact, many farmers who have field choppers for making silage prefer to bale their hay even though this process requires an addi­ tional investment in equipment and more labor than chopping. Kleis (13) states some of the reasons given by farmers for not using chopped hay: 3 "I* Chopped hay must be drier for safe storage than baled hay. 2. Chopping hay pulverizes the leaves and creates a severe dust problem in the storage and feed­ ing area. 3. Field losses due to shattering and pulverizing are greater for chopped hay. 4* Chopped stems are harsh and cause sore mouths in livestock. 5* Baled hay can be stored in rather open shel­ ters or even stacked outside without excessive spoilage. 6. Baled hay may be more easily transported if the hay is to be sold or fed at a location other than near the storage point." For economic reasons, hay handling must be mechan­ ized. Because of the objections to chopped hay and since neither baled hay nor loose hay is adapted to mechanized handling, some other method of hay harvesting must be util­ ized. If the hay were compressed into small packages or pellets that would flow in conveyors, the objections of chopped hay would be eliminated and the pellets could be handled with present mechanical equipment such as ear c o m elevators. This would allow the handling of hay from the field to the feeding area to be completely mechanized. Kleis (13) found the labor requirements in the stor­ age area (including unloading of wagons, storing and removal 4 from storage) to be 0.&9 man hours per ton for baled hay and 0*79 man hours per ton for ear corn. Assuming pelleted hay to be comparable to ear corn, 0.10 man hours per ton could be saved in the storage area alone if the hay were pelleted instead of being baled. This comparison does not take into account the saving in human energy which would be realized by the elimination of the lifting of the heavy bales. In addition, specialized hay storage structures would no longer be needed since hay pellets could be stored in conventional grain bins. Hay in the pelleted form has great appeal to the commercial hay grower. In addition to lower labor require­ ments for moving hay, pellets due to their increased den­ sity can be shipped at a freight rate which is considerably lower than that for baled hay (7). Objectives The primary objectives of this research were to de­ termine the effects of the following on the production of hay pellets: a) pressure, b) time of pressure application, c) Pellet weight, d) diameter of pellet, e) treatment of hay before pelleting, f) binding materials, g) moisture con­ tent. The secondary objectives were to compare the effect of the pellet length to diameter ratio on: a) the angle of repose, b) pellet density-bulk density relationship, c) handling durability, using both long and chopped hay. REVIEW OF LITERATURE The forming of ground feeds into pellets is an established practice, being introduced into the United States in 1929 (15)* Pelleting of such feeds produces a form which is convenient to handle, creates less dust, is more palatable, reduces waste and provides a balanced ration in each bite. In the manufacture of these pellets, the feed which has been dried to a moisture content of 6 or 7 per cent is finely ground and mixed. This mixture is usually condi­ tioned with steam in a small mixer which is an integral part of the ring die pelleting mill. The conditioned feed is then fed by a regulating feeder into the pellet mill proper where it is forced through dies by rollers operat­ ing under very high pressure. The feed is extruded from the dies as hard, glossy cylinders which are broken into the desired lengths by a set of adjustable knives. The pellets then pass over a shaker which separates out the fine material and returns it to the pelleting mill. Before the pellets go into storage or are bagged, they pass through a cooling tower to remove the heat gained in the pelleting operation. Although many shapes of pellets are made, the most common are round or square in cross-section and range in 6 size from 1/8 of an inch in diameter for chick size pellets to 3/4 or one inch cubes for range feeding of livestock. Due to the low output of the mill when producing the smaller sizes, many producers have discontinued this size and now crack the larger, laying size pellet (1/4 inch diameter) into crumbles for chick feeds (22). The power required to produce pellets is quite high, ranging from 12 to 18 horsepower hours per ton (22). The additional cost of pelleting must be borne by the consumer, who is willing to pay for it as indicated by the fact that 64*5 per cent of Ralston Purinafs feed production was sold in the pelleted form in 1957 compared 1952. 90 per cent of the output of Pellets accounted for to 52 per cent in Payway Mills of Kansas City in 1957 (2). In addition to pelleted concentrate feeds, high quality alfalfa hay which is used as a protein supplement is also normally pelleted. This hay is chopped in the field from the standing plant, hauled to the mill and fed into a rotary drier which dries the hay from a moisture contentof about 75 per cent to about 7 per cent in minutes. approximately 5 From this point the pelleting process is essen­ tially the same as that described for concentrates. The hay is dried rapidly in order to decrease the losses normally incurred in field curing. Le Clerc (16) reports that 17 to 18 per cent of the dry matter and 21 to 22 per cent of the protein is normally lost in the field curing process, whereas with rapid drying these losses become insignificant. 7 More recently the pelleting of lower grades of alfalfa hay, with and without concentrates, has been done in the conventional ring die pelleting mill* Razee (20) reports a commercial sheep-feeding operation in New Mexico which feeds some 65,000 sheep annually. These sheep re­ ceive their roughage as pellets, made from 90 per cent number 3 alfalfa hay and 10 per cent blackstrap molasses. The cost for pelleting ranges from $6 to $10 per ton, yet the feeder believes that this cost is offset by the reduced handling labor and reduced feeding waste. In addition to these savings, better gains per ton of hay fed were re­ corded when sheep were fed pellets made from number 3 hay than when fed number 1 hay which was not pelleted. Neale (16) found that from 25 to 35 per cent less TDN (total digestible nutrients) were required to fatten wethers which were self-fed cubes made of 60 per cent coarse alfalfa hay, 30 per cent sorghum grain and 10 per cent molasses than were required when the wethers were hand fed alfalfa hay and whole sorghum grain. In additional studies (19) he found that cubes made of 70 per cent hay, 20 per cent sorghum grain and 10 per cent molasses gave even better results. In all cases, coarse stemmy alfalfa hay that was hard to sell in the long form was used. Cate, et al (5), found that lambs fed a pelleted ration containing timothy hay as a roughage outgained and had a better carcass grade than those fed an unpelleted ration containing alfalfa hay as a roughage. In addition a to increased gains, Bay, et al (2), reported that many lamb feeders feed pellets to get away from digestive troubles. Recent sheep feeding trials indicate that pellets made of coarsely ground hay are superior to those made of finely ground hay. Esplin (21) reported that sheep fed pellets made from hay ground through a 1/4 inch screen made larger and cheaper gains than when fed pellets made from hay which was ground through a 1/16 inch screen. Tests at Dixon Springs, Illinois (17) show that a ton of alfalfa hay pellets produced 253 pounds of beef, whereas one ton of loose hay produced only 39 pounds. These tests were conducted with self-fed 425 pound calves. The calves eating pellets ate 15.6 pounds per day, those eating baled hay ate 10.96 pounds per day and those eating chopped hay ate 10.7 pounds per day. Since it took 6 or 7 pounds of hay daily for the calves to maintain their weight, those on pellets had about 9 pounds daily to apply to gain compared to only 4 or 5 pounds for those on loose or chopped hay. In order to assure the same quality of hay, the hay was pelleted from the bale as fed. It appears that the increase in gain experienced when pelleted hay is fed may be due largely to increased consumption rather than to some chemical change occurring during the pelleting operation. Ittner (14) found that a lignin analysis of pellets and comparable long hay showed the TDN of the pellets to be 60 per cent compared to 59 per cent for the long hay. 9 Ring die type of pelleting is done almost exclu­ sively by commercial concerns or farmer cooperatives and while it does have the advantage of reducing handling labor in feeding and allowing more efficient feed utilization by the animals, it in no way alleviates the problem of har­ vesting hay# In fact, when one considers the handling re­ quired to get the hay to the pelleting mill, it probably requires more labor than if the hay were fed directly* In order to reduce the handling time to a minimum, the Stewart Brothers of California mounted a pick-up unit, hammermill and pelleting mill, including a boiler to pro­ duce steam, in tandem* This 10 ton machine takes the field-cured hay, with a moisture content of 15 to 20 per cent from the windrow, grinds and pellets it at the rate of 2 tons per hour# The pellets produced by this machine sell $15 per ton higher than baled hay of the same quality (9)* This machine, though large and expensive ($35,000) does put the hay in a very convenient form for mechanized handling from the windrow to the feed bunk# Since the hay is ground before being pelleted, it probably would not be satisfactory as the only roughage for lactating cows unless the farmer is willing to accept a reduction in butterfat content of the milk. Hodgson (11) reports that ground hay is not palatable and milk production is reduced when cows receive it as their only roughage. When ground hay is pelleted, the palatability is improved but it may cause a lowering of the butterfat content of the milk. 10 Tyznik (4, 24 ) found that the butterfat content of the milk, the fatty acids in the rumen and the amount and kind of forage supplied to the dairy cow are closely re­ lated. He found that limiting the amount of roughage or feeding a finely ground roughage caused the fatty acid ratio in the rumen to change from the normal of 65 per cent acetic, 20 per cent butyric and 15 per cent propionic to an interchange of the butyric and propionic* At the same time the butterfat content of the milk was decreased. Upon re­ ceiving the roughage in the form of long alfalfa hay, the fatty acid ratio quickly returned to normal and the butter­ fat content of the milk increased to its normal value. From this, it appears that if hay is intended to serve as the only roughage for dairy cows, it should not be finely ground even if it is to be pelleted. Bruhn (4) has experimented with pelleting ofchopped and long hay. He reported that short-chopped forages can be fed directly into a conventional pelleting mill at the expense of greatly reduced output. This apparently is due largely to the inefficient grinding action which takes place in the pelleting process. He states that the output of a pellet mill may be reduced to 25 pounds per horsepower hour when pelleting a short-chopped forage, whereas the same machine pelleting a finely ground material may have an output of 100 to 150 pounds per horsepower hour. Eke (10) reported that chopped hay was pelleted in a Prest-O-log stoker fuel machine. Hay with an initial 11 moisture content of 17•5 per cent and a density of 10 pounds per cubic foot was pelleted in a 1/2 inch die. finished pellet had a moisture content of 15per cent The and a density of 30 pounds per cubic foot. Hay wafers, compressed from hay chopped into 1 inch lengths, are reported by Jones (12) to be superior in but­ terfat production and equal in all other respects to pel­ lets made from finely ground hay. These findings confirm earlier reports that pellets made from finely ground alfalfa should not be used as the only roughage for lactating cows because they apparently pass through the cowfs rumen too rapidly for maximum production of organic acids which are synthesised into butterfat. These wafers are 3 inches in diameter and about 1 inch thick. No information concerning the manufacture of these wafers, other than the fact that they were made in essentially the same manner as pressed logs, has been released to the public. Work by Sanderson (23) and Bruhn (3, 4) indicates that firm pellets may be made from either long or chopped hay without additional binding agents by applying pressures in the range of 4#000 to 10,000 psi. These tests indicate that the moisture content of the hay to be pelleted must be below 30 per cent if firm pellets are to be obtained. As the moisture content is decreased, the density of the pel­ let formed increases for a given pressure. Sanderson found that in most cases 30 seconds was a sufficient length of time to maintain pressure application. He also states that 12 for a given diameter pellet, an increase in volume requires an increase in pressure if the pellet density is to be maintained. This indicates that the power requirements would be lower for a thin, large diameter pellet than for one in which the thickness or length approaches the diam­ eter in size* This point should be investigated because the ultimate goal is to produce a size and shape which, in addition to being acceptable to the animals, will be free flowing. These thin pellets may not have both character­ istics. In a report by Chancellor (6), it is stated that pellets may be formed by pressures in the range of 2,000 to 6,000 psi. The moisture content for producing these pellets and its effect on density of the pellet formed is not given. He does state, however, that when the moisture content was above 50 per cent, the cells became ruptured and the material behaved essentially as a liquid. He fur­ ther states that the retention of compactness seems to occur at lower unit pressures for pellets of larger diam­ eters and seems to be greater if the pressure is maintained for 5 to 10 seconds. From these reports, it appears that the power re­ quirements for pelleting may be reduced by using large diameter, thin pellets. Bruhn (4) states that preliminary tests indicate that a field machine making large size (2 inch to 6 inch diameter) pellets will require about twice as much power as baling. He states that Holstein cows, fed 13 a mixed sample of pellets varying from 2 to 5 inches in diameter, seemed to show little preference to size and readily consumed pellets having a density of about 40 pounds per cubic foot. Jersey cows, however, seem to have more difficulty in eating the large pellets. Sanderson (23) formed the opinion that the density range for a pellet 2 inches in diameter should be from 15 to 30 pounds per cubic foot. Pellets in this density range appeared to be durable enough to withstand handling and were not so hard that cows had difficulty in eating them. THE INVESTIGATION Part I Apparatus The pellets for this investigation were made in cylinders of various diameters. The force was applied hy- draulically to a piston which compressed the sample in the cylinder or chamber. The hydraulic press, chamber and pistons shown in Figure 1 were used in the early investiga­ tions. This arrangement, however, left much to be desired. The pressure was built up by a hand pump, making the pro­ cess rather slow and laborious. Pressure was maintained by continual pumping, making it difficult to maintain a steady pressure- With this arrangement it was also difficult to accurately measure the minimum length of the pellet while under pressure and almost impossible to measure the rela­ tionship between pressure and displacement. The device shown in Figures 2 and 3 was constructed to eliminate many of the undesirable features of the appa­ ratus mentioned above. This device consisted of a pellet­ ing chamber, a working piston to apply pressure to the hay sample, a head piston to restrain the sample in the chamber and a spacer which maintained the head piston in a fixed position while the sample was under compression. Three different sizes of pelleting chambers (1.50, 2.40 and 3.42 15 Figure 1* Hydraulic press and pelleting apparatus used in the 1957 investigation* The pelleting apparatus parts, shown on the base of the press, are from left to right: Working piston* Pelleting chamber* Head piston* 16 Figure 2. Pelleting apparatus used in the 1956 investigation. A. 6. C. D. E. F. G. H. I. J. Pump and control valves. Pressure regulating valve. Selector valve. Pressure regulating valve for "booster11. Displacement transducer. Feed opening in pelleting chamber. Pressure transducer. Brush oscillograph. Brush amplifiers: BL-320, upper; BL-520, lower. Spacer for maintaining position of head piston. Figure 3* Schematic diagram CD CO © CV2 Pi M -P pelleting ® O of apparatus Ui to i < Uf • COC o -p o © ^o CO +3 Oa CO rH c3 O © o ■P © © a nd 03 a 3 03 CJ o o •H © CiD CO c»D <+-t o o piston J=3 ■P © to working attached •H -P a Q> £' © o 03 rH o< 03 •H P • CD ■P CV2 Cable -P -P © in -iKV rH ZS «5D •tH 25 came in contact with the head piston* the The recording pen of oscillograph was then shifted to the outermost chart line* By trial and error the gain of the amplifier was adjusted so that when the working piston was backed off 4 inches, the pen completely traversed the chart width of 40 divisions. Thus, each division of the chart was made to equal one-tenth of an inch. By depressing the calibration switch with this amount of gain, it was found that the calibration signal was spread over approximately 14 lines when the attenuator switch was set at 1. In order to get an indication of the handling dura­ bility of the pellets, two chain-and-flight type elevators were crossed in such a manner that the discharge of each was located 6 feet above the hopper of the other. Thus, a sample pellet was subjected to two drops of 6 feet in each cycle through the elevator system* A pivoted board, equipped with a handwheel and screw for tilting, was used to determine the angle of repose for the pellets {Figure 7)* An analytic gram balance was used for weighing the samples. Length and diameter measurements were made with a combination square which had a scale graduated in onehundredths of an inch. In order to express the density of the pelleted sample in terms of pounds per cubic foot, the following equation was used 26 Figure 7* Tilting board arrangement used to determine the angle of repose of pellets. ?, OEI SITV (« /F IT 5 ) ° 5 W, 40 Vi WEIGHT 50 60 TO a'o 90 100 L , U !h G T H ( IK ) )) I W .'AV111 (» T''|IT111! (G R A M S ) * 'I 2io 'o D, OIAMETER (IN .) Figure 3. Special slide rule constructed to solve the equation for pellet density. where O - pellet density in pounds per cubic foot • W Z weight in grams. D - diameter in inches* L = length in inches. 4*$5 z constant obtained by combining the unit conversion factors and TT/4. To expedite the numerous calculations required in this investigation, a slide rule (Figure 3) was constructed to solve Equation (1). This was done by taking logarithms of Equation (1) and setting it equal to zero, giving log Q - log W + log L -f 2 log D - log 4*35 = 0 (2) A logarithmic scale modulus of twelve inches was used for the ^ , W, and L scales, making it necessary to use a modulus of twenty-four inches for the D scale in order that the functional modulus of all scales be the same. In order for all the scales on the slide rule to run from left to right, the signs of Equation (2) must alternate. Since the sign of 2 log D did not follow this pattern, it was neces­ sary to run this scale from right to left. The log of 4.35, being a constant, was not used as a scale but was allowed for in the positioning of the scales. Density values de­ termined by the use of this slide rule were found to be within one per cent of those determined by the use of a calculator. Part II Procedure Since alfalfa is the primary hay crop, it was used for this investigation. The apparatus shown in Figure 1 was used to make the pellets for the 1957 investigation and the procedure was slightly different than that followed for the 1953 investigation. Procedure for 1957 In 1957 the hay was cut with a rotary-type lawn mower if chopped hay were desired or with a sickle if long or crushed hay were desired. The hay was crushed by pass­ ing it between a pair of smooth, 6 inch diameter crushing rolls which had an adjustable clearance. The minimum clear­ ance which would allow the hay to pass between the rollers was used. After receiving the above treatment, the hay was then allowed to sun-cure until the desired moisture content was reached. Late in the season when unfavorable drying conditions existed, the hay was dried in a solar drying unit. Due to the difference in drying rates for the differ­ ent treatments, it was difficult to obtain samples from the three treatments (crushed, long, and chopped) which had the same moisture content. In order to obtain a uniform mois­ ture content for each of the treatments, the samples were 29 stored in a constant temperature-humidity room and allowed to come to equilibrium moisture content* The maximum equi­ librium moisture content available with this temperature and humidity (5#° F . and 70 per cent relative humidity) was about 12 per cent* In the pelleting operation, the head piston was inserted in one end of the pelleting chamber and a weighed sample of hay and the working piston inserted from the other end* The pelleting chamber containing these compon­ ents was then placed in the hydraulic press so that the working piston was directly beneath the ram of the 5 inch hydraulic cylinder* Pressure in the hydraulic cylinder was built up by a hand pump until a predetermined pressure (which gave the desired pressure on the hay sample) was indicated by a gage* This pressure was maintained by a con­ tinual pumping action for 15 seconds, except in the tests for nutrient losses in hay juices, and released. After the pressure was released, the pelleting chamber was set on a block so that the pelleted sample could be forced out of the chamber by the working piston. The force necessary to push the sample out of the chamber was applied by rotating a hand-wheel which acted through a pinion and rack arrange­ ment on the ram of the hydraulic cylinder* It was noticed that a very rapid rate of longitu­ dinal expansion of the pellet occurred immediately after ejection from the pelleting chamber; hence, the pellet was weighed and set aside, and then measured after the rate of 30 expansion became negligible. For the pellets having a binding agent added, the procedure above was followed after the desired amount of binding agent had been thoroughly mixed with the hay sample. Hay for the nutrient loss studies was cut with a rotary mower and, to prevent loss of moisture, put in mois­ ture sample cans. These samples were then brought into the laboratory where the different treatments were applied. The juice and hay residue from these treatments were col­ lected separately in moisture sample cans and taken to the Agricultural Chemistry Department where the chemical anal­ yses were made. Moisture content determinations for all above sam­ ples, except nutrient loss samples, were made by drying the pelleted sample in an electric oven. The pellet was then reweighed after drying and the moisture content was calcu­ lated from the difference in weights before and after drying. Procedure for 1953 Since larger quantities of hay were used in the 1953 investigation, a field chopper was used whenever chopped hay was desired. This hay was chopped green, to a theoretical length of 1/2 inch, and dried to the desired moisture con­ tent in a solar drying unit. Depending upon weather condi­ tions, long hay was either sun-dried or dried in the solar unit. Eight pellets were made for each of the variables 31 investigated* Because of the time required to complete a test, it was impossible to study the effects of the various variables on pellets made from hay with any moisture con­ tent other than at or near the equilibrium moisture content which was approximately 12 per cent. By drying hay to this moisture level, it was possible, by thoroughly mixing chopped hay to obtain a sufficient number of samples having a relatively uniform moisture content. Since long hay could not be easily mixed, more difficulty was experienced in obtaining samples of this hay having uniform moisture, even when near the equilibrium moisture content. This was primarily due to the difference in stem size of the hay, the larger stems having a higher moisture content. With the small sample sizes used in the smaller diameter pellet­ ing chambers, a small range in stem size of the hay was sufficient to cause considerable variation in the sample moisture content. Because of the difficulty in obtaining uniform samples of long hay, the effect of sample weight and size of pelleting chamber diameter was studied with chopped hay. Since the 2.40 inch diameter pelleting chamber was easier to load than the 1.50 inch chamber and did not re­ quire the use of the booster to obtain the maximum desired pressure, it was used as the pilot size in this investiga­ tion. Before any pellets were made, the amplifiers for the transducers were allowed to warm up until the circuits 32 were stabilized. calibrated. The amplifiers were then balanced and Next, the head piston was moved to the proper position and the spacer inserted. The working piston was brought into contact with the head piston and the relief valve adjusted until the desired pressure was indicated on the recorder chart. The pellet was made by inserting a weighed amount of hay in the pelleting chamber and compres­ sing it for the desired length of time. Time was measured by counting the number of chart lines as they moved by a fixed point. The pressure was then released, the spacer removed and the head piston moved back to allow the pellet to be ejected through the discharge opening of the chamber when the working piston was moved forward. The pelleted sample was then weighed and set aside to be measured when the e ^ a n s i o n rate became negligible. Moisture content was determined by the same method described in the 1957 pro­ cedure • Bulk density values for the pellets were determined by loosely filling a cubic foot measure and weighing this volume. The angle of repose was measured by piling the pel­ lets on the tilting board (Figure 7) and turning the hand­ wheel until enough of the pellets slid off to expose a con­ tinuous surface. The angle was measured by aligning the blade of a combination square along this surface, the pro­ tractor head of the square being held against a horizontal beam which was located above the pellets. Eight measurements 33 were made for each test* Handling durability was measured by weighing the test pellets at the end of a specified number of cycles through the crossed elevator arrangement# RESULTS AND DISCUSSION The 1957 Investigation The 1957 investigation showed that up to 68 per cent of the juice could be squeezed out of freshly cut alfalfa hay with a dry matter loss of less than that normally incur­ red in field curing* Table 1 shows the analysis of the losses for different treatments* TABLE I NUTRIENT LOSS IN JUICE REMOVED FROM 75 PER CENT MOISTURE ALFALFA HAT BY THREE DIFFERENT METHODS Per Cent Lost * rr- Dry Matter 1 Ether Extract N.F • Extract Ash Fiber 14.6 2 6.5 0.33 0*31 58*5 21*4 16.5 2 17.4 27.6 0*44 0 *44 66.5 25.6 19.8 3 13.3 26 *4 4.02 4.02 36.3 23.6 14.8 Water Protein ^Treatment 1 = 1100 psi applied for 1 minute. Treatment 2 = 4400 psi applied for 1 minute. Treatment 3 = Centrifuged for 5 minutes in a 9 inch diameter basket at 6000 rpm. **Nitrogen free. The moisture content (wet basis) of the hay after the above treatments was 56.5 per cent, 43*5 per cent and 65 per cent for treatments 1, 2, and 3 respectively. 35 Although none of the treatments reduced the moisture content of the hay to a safe storage level, the amount of material required to make one ton of 20 per cent moisture hay was reduced f r o m 62*00 pounds (for 75 per cent moisture content hay) to about 3100 pounds by treatment number 2* Since the hay which received this treatment would require the removal of only about 1100 pounds of water to produce one ton of 20 per cent moisture hay, some process such as this might be economically feasible. The pellets which were formed from this high moisture hay were not firm enough to withstand handling until after they had been dried. Preliminary investigation of the effect of two bind­ ing agents, bentonite (a volcanic clay which was recommended by the W. E. Thompson Co., Detroit, Michigan) and blackstrap molasses, on the density of both long and chopped hay pel­ lets indicated that their use was not practical. No in­ crease in density was noted when bentonite was used. Black­ strap molasses, in amounts of 5 and 10 per cent by weight, did show some binding effect. This material, however, formed a coating on the piston and chamber walls, making it very difficult to operate the apparatus. With a different type of pelleting arrangement, the use of blackstrap molas­ ses as a binding agent might be practical. The 1957 investigation provided the information which was used as the basis for the design of the 1953 apparatus. 36 The 1956 Investigation Density Relationships The effect of length of time of pressure applica­ tion on pellet density is shown in Figure 9 for 12 per cent moisture content hay. These data form straight lines when plotted on semi-logarithmic graph paper (Figure 10). Thus, the increase in density of pellet formed is linear with the logarithm of time of pressure application. Because of the relatively flat slope of the pellet density vs. time of pressure application curve, it would seem desirable from a production standpoint to use short time of pressure appli­ cation. For this investigation, however, a 5 second pres­ sure application was used so that the difference in pellet density due to small errors in timing would be negligible. The rate of linear expansion was measured to deter­ mine the necessary minimum time which should elapse after pelleting so that the value of this measurement would not change an appreciable amount. As is shown in Figure 11, most of the expansion took place in the first 15 minutes and after about 30 minutes the rate of expansion became neg­ ligible. Hence, all pellets made in this investigation were allowed to set for at least 30 minutes after pelleting before measurement. The density vs. pressure relationship for different sample weights of chopped hay pelleted in the 2.40 inch diameter chamber is shown in Figure 12. This relationship Figure 9. Pellet density vs. length of time of psi 3200 application psi 4000 of 37 u 00 CO t) c C l. f-i I) ^ x s u s p 3©XI©d 38 39 2400 psi 80 60 4000 psi (percent 40 Expansion of minimum pellet length) 100 T 2.0 Chopped hay, 12$ m.c. 80 gram s a m p l e , 2.40" pelleting chamber 0 15 30 T im e Figure 11. 45 60 (minutes) Percent pellet expansion vs. time chamber - 2.40” diameter © a* o o CO I CO o I —I X co Di © Jh 3 02 \ TO TO Figure 12. 8 pressure & vs. t>> C O density in 03 Pellet o e — kA1 o 02 P 1 ©TT 9d o o 41 was in the form of a straight line when these data are plotted on semi-logarithmic graph paper as shown in Figure 13* Since this was true with this diameter chamber, the pres sure-pel let density data for pellets made in the 1*50 and 3*42 inch diameter chambers were plotted directly on semi-logarithmic paper• Figures 14 and 15 show this rela­ tionship for pellets made in the 1*50 and 3«42 inch diam­ eter chambers respectively* Since straight lines on semi- logarithmic paper resulted in each case, the pellet density is linear with the logarithm of the pressure applied and this relationship may be represented by equations of the form In P = In K t or P = Kem ? where P = pressure in psi. K = a constant. e = base of the natural logarithm. P = pellet density in pounds per cubic foot. The equations for these lines were determined by the method of averages. ber the equations are For the 1.50 inch diameter cham­ Figure 13* Pellet 02 in in O O 02 ^fenep qatxad o o chamber o in to CO 02 - 2.40” diameter x 1CT3 ) o pressure {psi (semi-log) LO density* vs. Pressure 42 Figure 14* o o -- CO -- O to o 02 o o 02 in rH c be mbe r(s em i~ log) 03 o to I o X ft tn CD 3 CQ W ft vs. x 10"®) o 731 C P = 367 e°*0675 9 P = 465e°-0555 9 for the 20, 40 and $0 gram samples respectively* For the 3*42 inch diameter chamber, the equations are P = 417e°*°793 9 P = 510e0 *062? 9 P z 479e0 *°^76 9 for the 40 , $0 and 160 gram samples respectively* It may be noted for all three sizes of pelleting chambers the pellet density, obtained with a given pressure, increased when the sample weight was increased* The in­ crease in density obtained by doubling the sample weight was greatest for hay pelleted in the largest diameter. This is shown in Figure 16, where the same weight of hay is represented by each stack* The stack on the left is com­ posed of four 40 gram pellets, the center stack has two SO gram pellets and the stack on the right has one 160 gram pellet. 46 Figure 16* Effect of sample weight on pellet density. All pellets made in 3*42" diameter chamber with 4000 psi pressure* From left to right: Four 40 gram pellets* Two 80 gram pellets* One 160 gram pellet* Figure 17* Wedge-shaped appearance of 40 gram pellet made in 3*42” diameter chamber. 47 The increase in density was least noticeable when the sample weight was increased from 20 to 40 grams in the 1*50 inch diameter chamber# The density vs# pressure curves for these two samples are quite close together, in­ dicating that little or no increase in density can be ex­ pected when the sample weight in the 1.50 inch diameter chamber is increased from 20 grams to 40 grams# One of the reasons for the increase in density which occurred when a larger size sample was pelleted in a given diameter chamber could be that there was less interlocking of the fibers at the pellet surface than within the pellet. This lack of interlocking would allow more expansion to take place near the ends of the pellet. Since a greater portion of the pellet length was near the surface for the thin pel­ lets than for the thick ones, a greater expansion (in per­ centage of length) would take place in the thin pellets. This greater percentage of expansion in the thin pellets would cause them to be less dense. Since the length to diameter ratio was largest for the 40 gram sample pelleted in the 1.50 inch diameter cham­ ber, it is possible that the increase in friction of the material on itself and on the chamber walls absorbed a con­ siderable amount of the pressure which more than compensated for the smaller percentage of fibers near the end of the pellet. This would cause the lower increase in density be­ tween the 20 and 40 gram samples in this size chamber than was found when the sample weight was doubled in all other tests. 4& Figure 18 shows the density vs. pressure relation­ ship for 40 gram samples pelleted in the different chambers. Here it may be noted that, for a given pressure, the pel­ let density was greater for the smaller diameters. This may be explained by the same reasoning as above; because when a given weight was pelleted in a larger diameter cham­ ber, a shorter pellet was obtained than when the same weight was pelleted in a smaller diameter chamber. The same tests were conducted with long hay. As shown in Figures 19, 20 and 21 for the 1.50, 2.40, and 3*42 inch diameter chambers respectively, these data approximate a straight line when plotted on semi-logarithmic graph paper. Since the moisture content for the individual sam­ ples varied from 12 to 20 per cent, a wide range in pellet density occurred within each test. Because of this, the curves showing the density vs. pressure relationship in these figures were approximated. Since the points lie with­ in a reasonable distance from the approximated lines, it appears that the density of long hay pellets is also linear with the logarithm of the pressure applied. Energy Requirement Relationships Pressure and displacement data for 20, 40 and 80 gram samples of chopped hay pelleted in the 2.40 inch diam­ eter chamber were taken from the oscillograph chart record (see Figure 26 for typical record) and plotted on rectan­ gular coordinate graph paper. This relationship for these 1*9 tn 0 X> a CD Cl 0 XI £ CD 5h 0 XI a 0 ~r io a o OD o 0

> -p •i—1 -- ^ O O Cv2 to N i—I CQ a 0 — to CO 1 o — C, 0 0 o 04 0 04 Vl o <* •H Pm ft)43 Q » ctf 33