SO'EVQTE F CTGRS AWEC‘E'EEQG THE HYDRAUUC REMOVAL OF MANURE ROM CCNCMTE flue“: gov “10 Degree of M. S. MICHMN STATE UNIVERSE“ .fames Brian McQuitty 1959 THEM. LIBRARY w '. Sun- .“ . "i SOME FACTORS AFFECTING THE HYDRAULIC REMOVAL OF MANURE FROM CONCRETE by James Brian McQuitty AN ABSTRACT Submitted to the College of Agriculture Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Agricultural Engineering Department 1959 \7 A, A , f Approved -' i 1;”ng * J ii Present trends in agriculture are toward more speciali- zation, with larger livestock production units, and toward confinement raising of hogs. These trends should result in more efficient utilization of labor with increased output per man-hour, since mechanization can be more widely adOpted. The trends also have emphasized the need for more efficient manure-handling systems. Recently, considerable interesttms'been shown by farmers in the liquid manure system, where water is usedfbr'cleaning. The washings are permitted either to drain away or to be collected in a storage tank to await disposal. It has been demonstrated that such a system of cleaning lends itself to automation. Thereis evidence availableto indicate that liquid manure is a cheaper source of plant nutrients than commercial ferti- lizers. The system offers a valuable meanscfi'more efficiently conserving and utilizingthe potential crOp producing and soil conserving value of animal excretion. Scientific data related to the liquid manure problemame, at present, very limited and there is an immediate need for information concerning the cleaning process. This investiga- tion was conducted to determine the effect of nozzle type, pressure and degree of surface roughness on the removal of manure from concrete. The experimental apparatus consisted of a water pump. and a trolley wnhfiqwas capable of moving along a track at a uniform rate. A nozzle was clamped to the trolley to direct iii a stream of water, perpendicular to the directicn of travel, on concrete slabs with three degrees of surface roughness. These three degrees of roughness were produced by a steel trowel, a wood float, and a brush. Each surface had three replications. Difficulty was eXperienced when manure was used as the material to evaluate the work. To assist in obtaining data, sand was used as an inert material. The general behavior of these materials was similar. A significant correlation coefficient of 0.9291 was obtained. Preliminary tests were conducted to assist in arbitrary selections of the numerous variables involved. The solid spray nozzle was found superior to all others. With this nozzle, a highly significant improvement in the effectiveness of manure removalwes obtained when the pressure was increased from 60 to 80 p.s.i. Increasing pressure to 100 p.s.i. produced a further significant improvement. The effect: of the degree of surface roughness was found to be negligible. SOME FACTORS AFFECTING THE HYDRAULIC REMOVAL OF MANURE FROM CONCRETE by James Brian McQuitty A THESIS Submitted to the College of Agriculture Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Agricultural Engineering Department 1959 ACKNOWLEDGEMENTS The writer wishes to express his sincere appreciation to all those who contributed to this investigation. Special thanks are due to: Dr. J. S. Boyd, Agricultural Engineering Department, who as my major professor continually provided counsel, guidance, and support throughout the entire investigation and prepara- tion of this manuscript; Mr. C. M. Hansen, Assistant. Professor in the Agricultural Engineering Department, forlns helpful suggestions in design- ing the apparatus; Dr. W. D. Baten, Professor in Statistics, forhis valuable assistance in connection with statistical procedures; Dr. A. W. Farrell, Head of the Department of Agricultural Engineering, and Dr. M. L. Esmay,. Graduate Student Adviser, for their administration and helpful guidance of my program; The W. K. Kellogg Foundation of Battle Creek, Michigan, for the generous award of a Fellowship that permitted a year's study in the United States of America, and the Northern Ireland Ministry of Agriculture for providing the necessary year’s leave of absence; 11 John Bean Division, Food Machinery and Chemical Corpora- tion of Lansing, Michigan, for loan of the pump, and the Spraying Systems Company of Bellwood, Illinois, for the donation of nozzles used in the investigation; Fellow graduate students for their assistance and helpful suggestions; My wife, Melody, for her patience and encouragement throughout the entire year's study. 111 TABLE OF CONTENTS INTRODUCTION. . . . . . . . . OBJECTIVES. . . . . . . . . . REVIEW OF LITERATURE. . . . . EXPERIMENTAL APPARATUS. . . . PROCEDURE . . . . . . . . . . ANALYSIS OF DATA. . . . . . . Evaluation of Nozzle . . Evaluation of Pressure . Evaluation of Roughness. Correlation Between Sand DISCUSSION OF RESULTS . . . . SUMMARY AND CONCLUSIONS . . . SUGGESTIONS FOR FUTURE STUDY. BIBLIWRAPHY. O O O O O O O 0 iv 0 Manure. Page 12 15 24 25 27 27 28 31 34 37 38 LIST OF FIGURES The general layout of the experimental apparatus The three degrees of surface roughness . . . . . Uniform distribution of manure on the concrete . slabs (before flushing). . . . . . . . . . . . . Typical break-up pattern of the manure on the. . concrete slabs (after flushing). . . . . . . . . Effect of nozzles Operating at various pressures on sand retained on concrete surfaces. . . . . . Manure and sand retained on concrete surfaces. . using a solid spray nozzle at various pressures. Page 13 13 19 19 26 29 Table 2. LIST OF TABLES Index numbers of farm production per man-hour for the United States (1947 - 49 = 100). . . . . The average daily amount and composition of . . solid and liquid excreta of mature animals . . . Percentages of fertilizing constituents in urine of various farm animals and value in relation to total excrements . . . . . . . . . . . . . . . . Nitrogen losses in storage of liquid manure . . using different methods of sealing . . . . . . . Relationship between pressure and capacity for . the nozzles under test . . . . . . . . . . . . . Summary of analysis of variance: the effect of nozzle type, pressure, and surface roughness on sand removal 0 O O O O O O O O O O I O O 0 O O 0 vi Page 10 16 25 INTRODUCTION Throughout the United States, there is a steady trend toward greater specialization in crap and livestock production. The size of production units is increasing, and dairy herds of 1,000 cows are now a reality. A similar trend toward larger units is apparent in hog production, a trend which is accompanied by a confinement-raising system. These trends will permit more economic production, since the same labor is being used in these new systems as in the old practices. This has been accomplished largelyas a result of recent advances in mechanized feed handling. In general, however, output per man-hour in livestock production still lags considerably behind thatcfl'crop produc- tion (11*. This is illustrated in the following table. TABLE 1 INDEX NUMBERS OF FARM PRODUCTION PER MAN-HOUR FOR THE UNITED STATES (1947 - 49 = 100) _.-.._..__,__ - - V- ________.. -_._._____ _. _ ~_——p_._-_ .— _ _ .._..__...—.._.—__._—— fl... Year Farm Output Livestock and Products Crops 1910 46 71 46 1937 66 80 65 1947-49 100 100 100 1957 143 123 154 * Numbers in parentheses refer to appended references. 2 Mechanizing the feeding operations undoubtedly will be of great importance, but other items must be considered if livestock production is to become more efficient. Kleis and Wiant (2), in their study of materials-handlingon 320 Michigan livestock farms, indicated that manure handling is an import- ant consideration. In general, most of the time spent on this chore was in cleaning and not in disposal. With large-scale production units, the problem of efficient manure-handling methodskms‘been magnified. Regular cleaning of individual pens in confined hog set-ups, and feeding on paved areas in loose-housing systems for cattle puts a heavy demand on labor. Numerous reports have appeared in the farm press recently on how individual farmers are attempting to overcome the problem. The methods used have centeredyaround handling manure in liquid form. Jedele (3) reports that there is a definite trend to such a systemin.the midwest states.. A similar trend is indicated by Coady and Ratcliffe (4) in areas of Britain. Methods in current use involve hosing down with water the area to be cleaned. The manure and washings then are permitted to drain away or are collected in a storage tank prior to subsequent disposal on crOp or grassland. Complete automation would be ideal and would result in a major contri- ‘bution toward higher efficiency in livestock production. This may not be practicable, especially for cleaning paved cattle lots. However, as yet, the possibilities have not been eXplored to any degree. Before this canbe considered, 3 basic information is needed on the use of water for the cleaning of manure from concrete surfaces. The purpose of this investigation is to obtain such information. OBJECTIVES The specific objectives of this investigation are: 1. To determine to what extent degree of roughness.of the floor influences the efficiency with which manure can be removed from concrete surfaces; 2. To study the effectcfifwater pressure on the effici- ency of removal; 3. To determine the relative efficiency of nozzles with various spray patterns in removing manure from concrete surfaces. REVIEW OF LITERATURE A review of gliterature reveals that, as yet, little scientific investigationlns been conducted, particularly from an engineering standpoint. There are many aspects involved in this system of manure handling. It }s necessary therefore to.discuss the material available to emphasize and correlate the relationships between the various aspects. Anderson (5) estimates that the annual production of manure from livestock in the United States is in the order of one billion tons. This has a tremendous potential value as a source of plant nutrients and as a means of conserving and improving soil fertility. The main benefits to be derived from. conserving and applying farm manureto crOps and pasture are largely indirect The humus improves physical characteristicsofthe soil through improved aeration and temperature relationships and by increasing the water-holding capacity of the soil. It stimulates soil micro-organisms which~ increase the avail- ability of.soil nutrients. The relative value of excreta from different animals is shown in Table 2. ‘The values will vary with quality and quantity of feed and live weights. 0\ TABLE 2 THE AVERAGE DAILY AMOUNT AND COMPOSITION OF SOLID AND LIQUID EXCRETA OF MATURE ANIMALS (5) Daily Production Nitrogen P205 K20 Animal Per Animal Solid Solid Solid Solid Liquid Liquid Liquid Liquid Lbs. Lbs. % % . % % % % Cattle ‘ 52.0 20.0 0.32 0.95 0.21 0.02 0.15 0.95 Sheep 2.5 1.5 0.65 1.68 0.46 0.05 0.23 2.10 Hogs ‘ 6.0 3.5 0.60 0.30 0.46 0.12 0.44 1.00 Hens 0.1 ---- 1.00 ---- 0.80 ---- 0.40 ---- Old systems of manure handling involved separate conser- vation of the liquid portion, in addition to storage of the solid manure. The practice still is common today on many EurOpean farms. For economic reasons, their American counter- parts have long abandoned the practice. Lack of sufficient equipment to handle this liquid portion without heavy labor requirements is suggested by Hansen (5) asthe primary reason. In addition, distribution of the liquid portion solved only a part of the overall manure-handling problem. Mixing solid and liquid excreta with additional water is by no means a new practice. In Switzerland, Germany, and more recently in England, the liquids and solids have been stored 86Parately, then mixed with water and distributed over fields through an overhead irrigation system. Cleaning, however, still is done by hand labor, and the method is applicable only to Very intensive taming systems. 7 In the United States, the conventional methodis to allow the liquid portion to drain away. On feed lots in rainy weather, the solid material is of such a consistency that mechanical devices will not pick it up, so that it runs to the edge of the paved lot to accumulate as a mud hole. This is unsanitary, causes offensive odors, and is a breeding ground for mosquitoes and flies. Hydraulic removal of manure, with temporary storage in a holding tank, for both solids and liquids would overcome these difficulties. Table 3 showsthe relative value of urine in relation to total excrements. Urine is a very valuable source of nitrogenand potash, although deficientin phosphate. The constituents are in a readily available form. TABLE 3 PERCENTAGES OF FERTILIZING CONSTITUENTS IN URINE OF VARIOUS FARM ANIMALS AND VALUE IN RELATION TO TOTAL EXCREMENTS(51 W Animal Nitrogen P205 K20 Value 33 % 75 3’5 Cattle 53 5 71 65 Sheep 63 4 86 75 Hogs ' 32 13 55 40 Anderson (5) estimates that only one-third to one-half Of‘the potential crOp-producing and soil conserving value of manureeis actually utilized, even though approximately fifty Percezn.cu‘it is drOpped on pasture. In general, losses are 8 a result of ,failure to save the liquid portion, imprOper fermentation and drying out, and leaching from storage piles exposed to rain. Itis the most soluble and readily available constituents which are lost. The losses around the farmstead are great. This is indicated from the work of Bauman, et al. (7), in a survey of 127 farms in central Indiana,ix1which they noted the percent- age of nutrients recovered under various management systems. With cattle housed on concrete, under a roof, the recovery rate was 79 percent as compared to only 9 percent from concrete with no roof covering. Only 3 percent was recovered from dirt lots. I Prompt spreading of stable manure on the field generally has been recognized as the best method of reducing losses inherent in most storage practices. Even when this is done, loss can be great. Drying winds result in loss of ammonia, and rain may remove soluble constituents in surface run off. Iversen (8) compared the effectcnlcrOp yields of plowing down stable manure at different times after application. In general, when yields resulting from immediate plowing were given a base value of 100, plowing under 6 hours after appli- cation gave 80-85 percent yield. After 24 hours, yields were reduced to 70-75 percent, while after 4 days, theywere reduced to 52-57 percent. These results have been confirmed in more recent studies and Danish workers (9) have shown similar results with liquid manure. Applying liquid manure to the soil, followed by 9 harrowing, resulted in a loss of 25 percent of the nitrogen when compared to incorporating the manure in the soil with a sub-surface drill applicator. Application with no attempt to incorporate it in the soil resulted in a loss of at least 50 percent. It is evident then that under present systemscfi’handling and disposal, manure losses are high. Liquifying the manure appears to provide the most practical approach to overcoming these losses, since being in liquid form, it would be compar- atively. easy to apply below the surface of the soil. In addition to saving labor in cleaning, the liquid manure system offers a better means of conserving the valuable by-products from livestock production. With the rising cost of fertilizers, the practice may result in an economic source of plant nutrients. The cost of disposal must be wieghed against current fertilizer prices. Jedele (1) suggests, however, that it costs something even to throw liquid manure away. A.septic tank would be required, with a filter system or disposal beds for instance. At the University of Illinois, preliminary work indicates that application is a sound economic pr0position. Costs of liquid manure systems in Scotland, described by Turner, et a1. (10), show that liquid manure can supply plant food at a cheaper rate than fertilizers when production units are medium to large in size. They stress, however, that when liquid manure is diluted with much water or where conservation and storage practices are poor, the value of the liquid is such that disposal as a source of plant food becomes meeonanic. 10 Iversen (11) stresses the need for adequately sealed storage tanks for liquid manure. The extent of nitrogen loss which occurs during storage is shown in Table 4.. Initial losses are rapid, but the rate of loss soon becomes very low. Since there are periods of the year when distribution of liquid manure would not be practicable, it would appearthat, to reduce losses, an adequately sealed tank is necessary. Hansen (6) suggests that, to reduce dilution, it may be more practical to first scrape the manure into gutters befOre flushing it into a holding tank. Such a method may be most suited to paved lots, but where hogs are raised in confinement, an automatic flushing system would have obvious advantages. TABLE 4 NITROGEN LOSSES IN STORAGE OF LIQUID MANURE USING DIFFERENT METHODS OF SEALING w Pencent of Nitzggeg Loss, Method of Sealing Date of Date of Percent of Filling Emptying Initial August 17 April 27 Nitrogen 76 % % 1. Protected from rain. only . . . . . 0.527 0.269 49 2. Roofed with loose. . planking . . . . . 0.515 0.385 25 3. Roofed-ti ht . . . planking Tunsealed). 0.529 0.407 23 4. Roofed-ti ht . . . . planking sealed). . 0.523 0.490 6 5. As No. 3, covered. . with soil. . . . . . 0.513 0.479 7 6. Sealed with concrete 0.519 0.500 4 11 Puckett, et al. (12) suggest the following system to be practical. Two nozzles, with solid spray patterns, were located a few inches from the floor at the end of a revolving boom. The manure is flushed into a centrally placed drain leading to a holding tank. They found that a pressure of at least 70 p.s.i., delivering 5 g.p.m., from each nozzle was required to dislodge the manure and wash it into the drain. A washing period of 4 minutes, 7 times a day, was adequate to keep the pen clean for 60 pigs. An automatic timing device controlled the cleaning boom and water pump. Ice accumula- tion during very cold weather was reported to be no problem. Although this system was apparently satisfactory, such factors as pressures used, speed of rotation of the boom,and volume of.water required were unfortunately arbitrary selec- tions. In some cases, they were chosen as the result of observations only. From the data given by these workers, 'it is .computed that the volume of water used per pig was in the order of 4.6 gallons daily. This appears to be an excessive dilution rate and emphasizes the need for controlled investigations of factors affecting the hydraulic removal of manure from concrete. EXPERIMENTAL APPARATUS The assembled apparatus is shown in Figure 1. It con- sisted of a John Bean, piston-type, water pump capable of deliveringup to 400 p.s.i., with a capacity rating of10 g.p.m. The 3/4-inch hose from the pump was fitted with nozzles having a spray pattern of a solid stream, a cone, and a flat Spray. Pressures were measured with two gauges. Onevms located at the delivery outlet of the pump. The second, a check, was mounted immediately behind the nozzle. The nozzle was carried on a trolley that traversed a 14-inch gauge track. Adjustable clamps held the nozzle in position. This arrangement made it possibleto vary the angle at which the water struck the surface of the floor and the angle normal to the direction of travel. A 1/2 H. P. electric motor, Operating through a variable Speed, reversible, hydraulic transmission,vdth.a V-belt final drive, pr0pelled the trolley to provide constant velocity. Clamping the V-belt to the underside of the trolley reduced the possibility of overturning. Nine concrete slabs, each 2 feet x 1 foot x 2 inches, represented three replications of 'three degrees of surface finish. A steel trowel, a wood float, and a brush finish were used to obtain a smooth, medium, and rough degree of finish. (Figure 2) -The slabs were laid in randomized position on a wooden frame parallel to the track. They were placed with the 2-foot FIGURE 1 - THE GENERAL LAY-OUT OF THE EXPERIMENTAL APPARATUS 1 a 1 ' \ 0 ~ ‘ 1‘. THE THREE DEGREES OF SURFACE ROUGHNESS. LEFT T0 A RIGHT: woos FLOAT, STEEL TR0wEL, AND BRUSH FINISHES FIGUREZ- 14 dimension at right angles to the track and 20 inches on centers. Partition panels were placed between the Slabs to prevent Splashing to adjacent Slabs. Each Slab was placed at a gradient of 1/4 inch to 1 foot, the directionof 810pe being perpendicular to the track. PROCEDURE Preliminary consideration of the problem emphasized the large number of variable factors involved. Since pressure nozzle type and surface roughness were being investigated,zfll other sources of variation had to be held .constant. The difficulty was in selecting suitable arbitrary conditions as no data were available on which to base the selection. The variables should be such that future investigations could be added. To assist in the selection,’ a series of preliminary tests and observations were conducted. It was observed that, when the stream of water hit the floor at a narrow angle to the horizontal, ,a more effective cleaning Job was accomplished. This angle was fixed at 20 degrees. Care was taken each time a nozzle was changed. to ensure that this angle remained the same, since any deviation would have introduced an error. - The angle normal to the direction of travel of the trolley was fixed at 90 degrees. This was considered necessary, from a practical standpoint, to flush all the area ina square pen; 0r‘ where travel was in a circular direction, to flush the manureto a central drain. As nozzle types available vary considerably, the number U3 be tested had to be reduced. Preliminary work using a nfizzle that was adjustable to give different conical patterns 1nClicated that the wider the distribution of water, the poorer "353 the Job of manure removal. Nozzles with conical spray 16 patterns were observed to be less effective than those with either a solid stream or a flat or fan-shaped pattern. It was decided then to test four nozzles that had the same capacity of water per minutefbr any particular pressure- The relationship between capacity and pressure is shownin Table 5. The nozzles had spray angles at 40 p.s.i., of 0 degrees, 15 degrees, 25 degrees, and 40 degrees. In each case, the equivilent orifice diameter was 5/32 inches, the pipe connec- . tion 1/4 Inch, and the nozzle length 1 Inch. It was observed that the distance from the end of the nozzle to the surface to be. flushed affected the efficiency of removal. This distance was fixed at 30 inches, which is approximately. the distance at which a man would hold the nozzle when flushing the floor down. TABLE 5 RELATIONSHIP BETWEEN PRESSURE AND CAPACITY FOR THE NOZZLES UNDER TEST Pressure ,. Capacit (P.S.I.) (G.P.M. 40 4.0 60 4.9 80 5.7 100 6.3 120 5.9 The Speed at which the trolley moved affected the removal efficiency. A speed of 30 feet per minute was used and the 17 transmission was set accordingly. It was timed frequently throughout the investigationto be sure no variation occurred. It was decided that the range of pressures used should be kept within reasonable limits and within the.capabilities of relatively inexpensive pumps. Below 40 p.s.i., the efficiency of removal was poor. Since the volume of water used had to be kept in mind, to prevent excessive dilution, an upper limit of 120 p.s.i. was selected. Consequently, the range of pressures tested varied from 40 to 120 p.s.i. at intervals of 20 p.s.i. ' As manure may vary widely in both physical and chemical composition, a considerable error could be introduced. This variation was thought to be a minimum with manure from fattening hog pens, as the feed fed to the hogs would be constant. To reduce variation further,it was very thoroughly mixed in bulk. In the preliminary work, it became evident that the distribution pattern of the manure placed on the concrete slabs also was of considerable importance. To reduce these errors, various methods of placing the manure were tried. As would be expected, large lumps were less readily removed than were small lumps. When the manure was placed in one-pile on each slab, considerable variation was observed in the cleaning process. This was attributed to the fact that the leading edges of the Piles were uneven, allowing the stream of water to work more e1’1‘1cient1y in some instances than in others. 18 Finally, a mould was made into which a weighed amount of manure was placed. This provided for a uniform pattern for all samples (Figure 3). Figure 4 shows the typical break-up pattern of the manure on the slabs after flushing. It was originally intended to use difference in weight as the criterion for measuring efficiency of removal. A weighed amount of manure was placed, using the mould, on the dry surface of each slab. The trolley, carrying the nozzle, made one run past the slabs. The slabs previously weighed with wet surface, were again weighed with the manure residue. The weight of the residue was found by difference. It was soon noted, however, that the slabs themselves were a source of variation, due to the amount of water absorbed by the concrete. This was overcome by keeping the slabs soaked in water. The trial was continued, weighing the wet blocks, and then placing the manure on the slabs once the surface was dry. The purpose of applying the manure to the dry surface was to Simulate as far as possible what would be most likely to occur in practice. This method was adOpted throughout the investigation. The reason for weighing the slabs when wet was to obtain more accurate dataibr the residues, since these were weighed on the slabs immediately after flushing. I Evaluating efficiency by the procedure proved unsatis- factory, due tO‘a combination of factors: a. the weight of the slab in relation to ,the amount of manure flushed off was too great to give accurate results; ’ Zed! , FIGURE 3 - muons DISTRIBUTION OF MANURE ON THE SLABS (BEFORE FLUSHING) f} 6! ~ - | . . , .. w L I c. ‘ . Q , f‘ ,. . 0 lg .‘- ' i n 0‘ K ‘n_ ', . u ' . s - _.n . .' ‘ _‘. . u A a .' r a Ring ‘1‘ _ féJi- T ' A"; o ‘8' 4““ ' I . .3 9—. [#3: t3 . O ¢' ,‘ :- b ‘ ‘ ' ' , 'b'; V'. o o g -' " I _- ' o ' O‘ .' o . l 9...! . . . ‘ 4 . . l." . .. . .‘ w - .. ‘ "a“ _. Wm .- . «5“. .. ‘Q. A. P a o ' .L (5 _‘-":‘Jo "at. . D t $5 “II-flawe- ’ - ' FIGUIE 4 - TYPICAL BREAK-UP PATTERN OF MANURE ON THE SLABS (AFTER FLUSHING) 20 b. the manure absorbed-part of the water; 0. the break-up of the manure resulted in water being trappedxnlthe slab, the quantity varying considerably. Results were meaningless, since a combination of factors ”b" and ”c" resulted in some instances in an actual increase in weight in ~the residue compared to the weight of manure applied. . To increase accuracy, it was thought that washing the residue from the slab and then filtering off the surplus liquid would be a possibility. However, this proved imprac- tical, due to the presence of large amounts of colloidal material in the manure. These colloids rapidly plugged the filter. The problem then was to determine some criterion for evaluating the efficiency of manure removal. Dry matter appeared to offer the best possibilities. However,. in view of the large number of samples .involved and the problem of drying the residues, the use of an inert material to replace the manure seemed desirable. Consequently, sand was selected as a material to evaluate cleaning efficiency. Clean, air-dry sand, which passed through a. BOemesh screen was used. The finer particles were removed and discarded by passing the sand over a IOO-mesh screen. Five hundred grams of sand were placed on the dry surface of each slab. The distribution of the sand was kept uniform by a mould, and the position on the slab was kept constant. 21 The first nozzle was fitted and the required pressure built up before the trolley was set in motion. The trolley had approximately 2 feet to travel before the stream of water hit the first slab. This was designed to permit the trolley to gain the required speed before the first slab was reached, and so reduce possible error by having the stream of water travelling more slowly. The stream of water flushed each slab only once. The trolley was stopped at the end of the track. The water pump then was stapped and the trolley returned to its original position by reversing the transmission drive. The residue of sand on each block was washed off into a collecting basin and the washings filtered. A milk strainer was used for this purpose. The water rapidly filtered off and within a few minutes had ceased to drip. After a further 3 minutes, the strainer with the wet sand was weighed. Since the weight of the wet strainer and filter previously had been determined, the weight of the wet sand was obtained by difference. This method was usedtx>collect data for each of the four nozzles at the five pressures. The accuracy of this technique was tested by placing a known quantity of sand on a slab, washing it all off into the basin, filtering and weighing as before. This was repeated several times, using the same quantitycfl'dry sand. Variation in the results showed a range of only 1.0 percent, which was considered satisfactory. 22 From the data, as will be discussed later, the nozzle producing the solid spray pattern was outstanding in its ability to flush the sand from the concrete. The problem now was to find out if the manure would behavein.a similar manner in its reaction to pressure and surface roughness. One thousand grams of manure were placed on each slab, using the mould, with each lot placed in exactly the same position. The surfaces then were flushed as before, using the nozzle with the solid stream. The trial was repeated for the five pressures. In this test, the dry matter of the residue was used as the basis for evaluation. Because of the practical difficulties involved in drying the residues after washing from the slabs, it instead was dried on the slabs by a battery of infra-red heat lamps. These were suspended directly over the row of slabs, with the utmost care taken to ensure that the height of the lamps were so adjusted thatzx>charring of the.material occurred. During the drying process, a hard crust formed on the surface of the manure which retarded drying. This had to be broken regularly. After a drying period of eighteen hours, the dried residue was carefully removed from each slab by scraping and by use of a steel brush, and then weighed. The eighteen-hour drying period was rigidly adheredto throughout Numerous manure samples were taken during the course of this phase of the investigation. They were dried in the same manneras the residues under the heat lamps for eighteen hours. 23 The average dry matter was 30.15 percent, while the range was only 1.7 percent. The dry matter was higher than expected, but the manure-~as collected from the hog lot--had been eXposed to normal drying effects. The narrow range indicates that mixing was effective in producing a uniform sample. ANALYSIS OF DATA The primary objectives were to determine which nozzle did the most effective cleaning job, at what pressure this was most efficient, and to what extent surface roughness influenced cleaning efficiency. The experimental set-up represented a four classification problem. The data were treated statistically, using the analysis of variance (13). ' To simplify computations in the analysis of sand residues, data obtained at 120 p.s.i. for all four nozzles were analysed separately from that obtained at the four lower pressures, since these residue values were small in relation Moths rest. In Table 6, a summary of the analysis of variance of four nozzles, four pressures, and three degrees of surface rough- ness with three replications are given. Throughout the analysis, the significance of differences between each of the nozzles, the pressures, and the surface roughnesses were detected by using the Studentized Ranges for the new multiple Range Test (14). This uses special protec- tion levels based on degrees of freedom. The table shows that both nozzle and pressure were highly significant, while surface roughness was not significant. 25 TABLE 6 SUMMARY OF ANALYSIS OF VARIANCE: THE EFFECT OF NOZZLE TYPE, PRESSURE, AND SURF ACE ROUGHNESS ON SAND REMOVAL :— Source of Variation Degrees Sum Mean of of Square Freedom Squares Tatalo o o o o o o o o o o o 143 2,909,836 Replications . . . . . . . . . 2 58,825 29,413 Roughness. . . . . . . . . . . 2 95,633 47,817 Nozzle . . . . . . . . . . . . 3 2,106,410 702,137** Pressure . . . . . . . . . 3 387,107 129,036** Nozzle x Pressure (a). . . . . 9 98,948 10,994 Nozzle x Roughness . . . . . . 6 31,792 5,299 Roughness x Pressure . . . . . 6 5.930 997 Roughness x Replications (b) . 4 46,462 11,616 Nozzle x Roughness x Pressure. 18 18,615 1,034 Residual . . . . . . . . . . . 90 60,064 667 ** Significant at 1 percent level of probability. (a) Error term used to test significance ofnozzleand pressure. (b) Error term used to test significance of roughness. Evaluation of Nozzles The nozzle with the 0 degree spray angle was much superior to the other three in its ability to remove sand from the surfaces. This superiority was significant at the 1 percent level of probability. Further analysis showed that this highly significant superiority held throughout all the ' pressures tested. Very careful observation confirmed this superiority when manure replaced the sand. The outstanding ability of the solid stream nozzle compared with the other three tested is expressed graphically Jh1 Figure 5, where the mean residue values for a particular Pressure are plotted against the pressure. 26 NOZZLE SPRAY ANGLE AT 40 PSI 0———-O — 40" H _ 25° 600 H "' '3: * L ‘ ' 500—- ! __.__._.,__. 400 -——~—— ~————-- ——— 1 fl . m I. 2 K < . (r 0 300 | z \. w [I 5 7 3 l 9. U) 200 ‘ Lu . a ‘1 IOO ‘ l i 0 k 40 60 80 I00 I20 PRESSURE- PSI FIGURE 5. EFFECTOF NOZZLES CPERATING AT 'VARIOUS PRESSURES ON SAND RETAINED ON CONCRETE SURFACES° 27 In general, no significant differences existed between the other nozzles, although at any particular pressure, some differences did exist. Evaluation of Pressure Pressure generally was a highly significant variable when sand was used as the material for evaluation. With the solid stream nozzle, each increase in pressure resulted in increased cleaning effectiveness, all of which were significant in the case of sand. The tendency was for-each increase in pressure to result in a lower rate of improvement. The improvement was highly significant with increases in pressure from 40 to 60 p.s.i. and from 60 to 80 p.s.i. Increasing pressure from 80 to 100 p.s.i. was significant at the 5 percent level, as was the increase from 100 to 120 p.s.i. In the latter instance, the improvement was just significant. With manure, pressure also was highly significant when the solid stream nozzle was used. Increasing pressure*from 40 to 60 p.s.i. resulted in little or no improvement in cleaning effectiveness. A highly significant improvement resulted when pressure was increased from 60 to 80 p.s.i. Raising the pressure to 100 p.s.i. gave a further improvement wkdch.was significant at the 5 percent level. A pressure of 120 p.s.i. was not significantly better than 100 p.s.i. Evaluation of Roughness .As shown in Table 6, roughness was.not, in general, a Significant variable. When a separate analysis was made of 28 the data for each of the nozzles, roughness was significant only in the case of the nozzle with the 40 degree spray angle. Both the steel trowel and wood float finishes were signifi- cantly superiorto the brush finish in their ability to permit cleaning. With manure, roughness of surface was shown to have only a negligible effect. No significant differences were detect- able between the degrees of surface roughness'at any pressure. Correlation between Sand and Manure To assist in future investigations, it was considered desirable to determine to what extent the behavior of the manure and sand were similar. When the mean residue values for manure and sand, using the data obtained with the solid spray nozzle, were plotted for the range of pressures, the curves were seen to follow a similar pattern (Figure 6). Consequently, correlation coefficients were computed using the formula, EExy - ELZJEI {’xy = L \H:zx2 - 1.532513] Eye: 9:113] where x = each value in the sand residue sample y = each value in the manure residue sample n = number of values in a sample of manure or sand When the correlation coefficient was computed at each Pressure throughout the range, significant ”r" values were 29 600 ll [L A I o——® MANUREIDRYBASISI H SANDIWETBASISI 500 .0400 2 < I: (9 . I 300 - LIJ D o \ a ?\ Alt-.200 ’\ ‘L‘<}S— I I00 \3 o If 40 so so l00 Izo PRESSURE-PSI FIGURE 6- MANURE AND SAND RETAINED ON CONCRETE SURFACES USING A SOLID SPRAY NOZZLE AT VARIOUS PRESSURE‘S- 30 obtained at 40 and 100 p.s.i. only. The reason for the lack of correlation at the other three pressures may be attributed to the fact that in some instances the water stream would succeed in dislodging and washing off a large fragment of manure from the pile on a slab. This was undoubtedly due to the difficulty of obtaining homogeneity within the mould from one slab to another. This resulted in wider ranges in values for the manure residue thanwas the case with sand. Reference to Figure 4 will illustrate the unevennesscm'manure breakeup. Mean values for the residues for both the sand and manure were obtained at each pressure. The overall correlation coefficient computed for these means was 0.9291. This was significant. DISCUSSION OF RESULTS It was mentioned in the Review of Literature that data related to this work is very limited. Puckett, et al. (6),. with their automatic hog pen cleaning system, recommended the use of a solid spray nozzle for effective cleaning. Their recommendation, based on observation, is confirmed by experi- mental procedure and observation in this investigation. If the curve for manure, in Figure 6, were drawn to approximate the points, then the point of inflection would occur somewhere between 60 and 80 p.s.i. It is suggested that this point reflects a ”breaking” point where the adhesion qualities of manure are overcome. This would mean that the . water pressure should always be above the point of inflection. A movie film was used to demonstrate that pressures over this point would be capablecfl'overcoming the adhesive forces. The camera was mounted