ision of Agricultural Sciences UNIVERSITY OF CALIFORNIA CALIFORNIA AGRICULTURAL EXPERIMENT STATION BULLETIN 762 In the hop-drying studie Photo below shows the pilot dryer used in these studies. reported in this bulletin . . . l.Two electrical conductivity-type moisture meters were calibrated for ground hops. The conductivity-type meter appears promising for rapidly determining the moisture content at the end of the drying period. 2. An equilibrium moisture-content curve was developed. 3. Basic drying rates were established. 4. Evidence indicates that hops may "case harden" at very low relative humidities and consequently have a slow basic drying rate. 5. Quality was found to be an opinion of the individual judge rather than the weighted average of a number of finite factors. 6. Samples dried at nO°-115°F were judged superior to those dried at higher temperatures. 7. Partial drying with 150°F air and finish drying with unheated air at a moderate air-flow rate produced a quality product. This procedure might be used to increase the capacity of the kiln. 8. An air rate of 80-85 cu. ft. per min. per sq. ft. will lift a 34-in.-deep hop mass when dry. Sixty cu. ft. per min. per sq. ft. is believed to be the highest practical air rate for a 34-in. kiln. 9. Evidence indicates that mixing resulting from removing the dried hops from the kiln by pelican, emptying the pelicans into the transfer cart, and then emptying the transfer cart, mingles the wet and dry portions sufficiently to bale directly. Small bales made in this way yielded good-quality hops; full size bales were not made. 10. Reversing the direction of air flow when the mass is approximately two-thirds dried will reduce the moisture range, top to bottom of the mass, and decrease the in-drier time. Air-direction reversal before the mass is two-thirds dry may cause insufficient drying of the central portion of the mass. 11. Closed recirculation of the air at the drying temperature after the final desired average moisture content has been reached will reduce the moisture gradient through the mass but will extend the in-drier time. THE AUTHOR: S. M. Henderson is Associate Professor of Agricultural Engineering and Associate Agricultural Engineer in the Experiment Station, Davis. MAY 1958 SOME HOP-DRYING STUDIES S. M. HENDERSON D, Tying studies of hops were conducted by agricultural engineering personnel of the Experiment Station during the hop harvest seasons of 1953 through 1956. This publication contains a summary of the significant portions of the investiga- tion. The report is directed specially to- ward those doing research and engineer- ing design in hop drying. In addition, we believe operators of hop drying plants will find the report valuable. Hop drying is a deep-bed process. Depths of 32 to 36 in. with initial wet mass densities of 4.5 to 5.0 lbs. per cu. ft. are dried with 110° to 150°F air temperature and air rates of 20 to 50 cu. ft. per min. per sq. ft. of floor. They are dried from a moisture content of 70 to 80 per cent (wet basis) to about 8V2 P er cent. When the average moisture content of the mass is about 8V2 per cent, the top of the mass may have a moisture content as high as 16 or 18 per cent and the bottom of the mass as low as 2 per cent. The moisture content of the over- and underdried portions is equalized by mov- ing the loose dried hops to a floor where they are stored for a few days before baling. This intermediate equalizing op- eration requires extra labor and equip- ment and at least as much floor space as the driers; hence its elimination is desir- able. The low moisture content of the bottom of the mass contributes to cone shatter, which lowers the market value. Submitted for publication May 27, 1957. The studies were directed toward elim- inating the tempering operation and reducing cone shatter. In addition, basic hop-drying data pertinent to improving the conventional drying system or pro- viding a new system were desired. Continuous-flow hop driers were not studied even though their use is recog- nized as a solution to many drier prob- lems. Many of the results are indicative rather than conclusive. The short experi- mental season of three to four weeks per year did not permit results to be avail- able soon enough to modify equipment or experimental procedures during a cur- rent season. Thus, the experience and results of one set of studies could be used only for designing a set of studies to be made the following harvest season. Frequently, the results indicated the need for continued detailed investigations which, unfortunately, would have re- quired more funds and personnel than were available. Even though many of the results are inconclusive, the findings are believed generally significant enough to contribute toward improving of the deep- bed hop-drying process. MOISTURE-CONTENT DETERMINATIONS Commercial moisture content is based on the wet weight of the material. The percentage moisture (wet basis) equals 100 times the pounds of water divided by the sum of the pounds of dry matter [4] Fig. 1. Relation of the three moisture- content indexes. 50 45 35 a 1* •4- a 15 10 5 0.50 0^5 0.40 035 0.30 O 025 ^ •I- 2 91 ^> a U I 3.8 a 19 2 1.0 12 16 T\n7e — Hours Fig. 6. Deep-bed drying pattern for 110° F. 20 2U < [12] mathematical procedure for relating air rate, drying time, and moisture differen- tial in deep-bed drying processes has not been found. After elevating the relative humidity, the lower portions of the mass increased in moisture content. The increase took place in 1 to 2 hours. The moisture con- tent in the lower portion remained con- stant after this period, whereas the upper portions continued to dry, but at a slower rate. The reduction in moisture differential was satisfactory but the extended time in process would increase the cost of drying and reduce the drier capacity. Additional studies of equalization procedures are reported later. The steady-state moisture content of the lower portions of the masses during equalization was lower than the equilib- rium moisture content (fig. 5). For ex- ample, the steady-state moisture content (dry basis) for the 130°F study was 5% per cent, whereas the value for 55 per cent relative humidity and 130° calcu- lated from the equilibrium equation is 9.2 per cent. Perhaps overdried hops are altered physically so that the equilibrium curve is altered. Or perhaps the tempera- k & ~ \z Time — Hours k & Time — Wours Fig. 7, Deep-bed drying pattern for 130° F. Fig. 8. Deep-bed drying pattern for 150° F. [13] ture feature of the equation does not apply as shown for hops, which are phys- ically different from most other agricul- tural materials. BASIC DRYING INDEX The drying rate of agricultural mate- rials in a thin layer or layer one unit deep can be represented by: -ftEft-*-'" M = moisture ratio at the start of a drying period M e = moisture ratio after a period of time, 0, in hours M E - equilibrium moisture ratio at the drying temperature and relative humidity p. 5 = saturated water vapor pressure of the atmosphere at the dry- ing temperature, in lbs. per sq. in. V n = air rate through the thin layer, in cubic feet per minute per square foot, raised to a power n k = the drying index 6 — time in hours The equilibrium moisture ratio (data from fig. 5 converted to moisture ratio by fig. 1) for each of the four laboratory runs was subtracted from the observed moisture ratios of the 1%-hi- layers at the bottom of the drying stack and the differences plotted against time in figure 9 (solid lines). The kV n term in the drying equation was found to have the following values for the respective curves: 110°F, 50 cu. ft. per min.— 0.706 130°F, 50 cu. ft. per min.— 0.760 150°F, 50 cu. ft. per min.— 0.750 150°F, 30 cu. ft. per min.— 0.584 If the l^-in. layer dried as a true thin layer, the index difference between the 150°F curves, assuming no experimental error, must be a function of the drying- air velocity. A measure of the air-velocity factor can be provided by evaluating n thus : 0.750 /50\ n 0.584 -( 30/ from which n is found to be 0.484. The theoretical range for n is to 0.8. Wheat, corn, and other small grains have drying indexes nearly independent of air rate; open water surfaces have an n approach- ing 0.8. Since the drying layers are not true thin layers, the difference between the 150°F indexes may be affected by depth. The true exponent, n, would be slightly less than 0.484 and the true k slightly larger than noted. The complete drying equation is: logN^# = - 0-112 p.F"«" M -M E or M -M K = e 0.112 p s "P"0.4S4 M -M E The drying index, 0.112, in this equation is the average of the &'s at 50 cu. ft. per min. per sq. ft. divided by 50 to the 0.484 power. The dotted 135 °F curve of figure 9 is from the run of sample 6, table 1. The kV n value is 0.587. The performance of this run at 106 cu. ft. per min. per sq. ft. through an 8-in. depth would be com- parable to 20 cu. ft. per min. per sq. ft. through 1%-in. depth. The index equals that found in the laboratory tests for 30 cu. ft. per min. per sq. ft. at 150 °F, whereas we would expect the index to be less than at 30 cu. ft. per min. This test was made a year later than the laboratory studies with hops grown under different conditions. Natural biological variation could easily account for this deviation from the laboratory-determined indexes. Many drier operators report that hops "case harden" when dried with high- temperature, low-relative-humidity air and that drying with high-humidity air at the same temperature may actually permit faster drying. The drying curves for samples 6 and 7, table 1, shown in figure 10, tend to support this report: the [14] 1 2 3 4. G 5 6 f 2 I 9 For llO°Oir\>€ 10 12 14 Fig. 9. Basic drying curves for IVfc-in. depths at temperatures noted. Air rates were 50 cu. ft. per min. per sq. ft. except for one test, marked (30). [15] 80 70 ^60 4- VJ50 J) 30 3 <20 10 12 3 Tirr2^^^ 1 Fig. 10. Drying curves for two samples at dif- ferent relative humidi- ties. enough for use during the entire season. The few runs that were made were ex- ploratory and yielded no significant data. The over-all dimensions were 38 in. by 50 in. by 88 in. high. The space for hops was a 34-in. cube. A portable gas burner was used for heat. The air direction through the hops was reversed by revers- ing the rotational direction of the fan, and moving the burner to the top or out- let port. The air inlet and outlet ports were provided with gasketed covers. The door (D in fig. 11) could be opened to permit recirculation of the air with the system closed. [16] The pilot drier (see frontispiece) was modified for the 1954 and 1955 studies. The heat source and fan were rearranged so the heated air could be delivered either up or down through the hop mass. A per- forated-bottom steel basket was fabri- cated to fit the drying chamber. The basket, 33 in. by 33 in. by 34 in. deep, was supported so that it and the con- tained hops could be weighed frequently during the drying period. Thus, the aver- age moisture content at any time could be estimated since freshly harvested hops have a moisture content of 70 to 80 per cent (wet basis). A hand-operated bale press was constructed to make a cubical bale 7 in. on each side. It was designed for a bale ram pressure of approximately 15 lbs. per sq. in., which is representa- tive of commercial bale pressures. 1954 Test Procedures The quantity of hops for test was ob- tained from the mechanical picker of the E. C. Horst Co. at approximately the same time each day in order that initial quality, test to test, would be as com- parable as possible. The experimenters were unable to control or determine the time elapsed between harvest and pickup for test. This may or may not have affected the quality of the dried product. Table 1. 1954 Hop Drying Tests Sample no. Air temp., deg. F Bed depth, ins. Air rate, cu. ft. per min. per sq. ft. Drying time, hrs. Final* moist cont., per cent wet basis Run details 1 115 34 47 MX 4.1 2 115 34 47 nx 3.3 Tempered with convection air, 16 days 3 115 34 47 17M 3.3 Tempered with forced air, 16 days 4 115 34 47 iiy 2 12.8 Air direction reversed when % dried 5 120 8 96 GVs 10.7 6 135 8 106 5 6.7 Unhumidified air at 9 per cent relative humidity 7 135 8 101 m 9.0 Air humidified to 24 per cent relative humidity 8 150 8 95 3 6.1 9 150 8 96 IX 30 Finished with convection air, 8 days 10 150 8 96 IX 30 Finished with forced air, 2 days 11 165 8 86 2 3.7 12 8 95 3% 17.0 Dried at a constant discharge air tem- perature of 117° F 13 8 105 5% 11.3 Dried at a constant discharge air tem- perature of 117° F 14 140 24 8.4 Dried in a laboratory vacuum oven 15 115 34 18 25 6.2 Random sample as hops were dumped on cooling floor 16 115 34 18 25 6.2 Mixture from top and bottom of kiln floor 17 115 34 18 25 6.3 From top and bottom of kiln floor, lay- ered in bale Moisture content when removed from drier. [17] The hops were placed on test within an hour after obtaining them. At the end of each test a sample of the dried hops was baled in moistureproof paper and placed in a constant temperature of about 68°F. The bales were submitted to a panel of eight experts for quality evaluation (table 2) after being in storage for two and one half months. Seventeen tests were run; test conditions, in addition to those given in table 1, were as follows: Samples 1, 2, 3. These are from a single run made August 16. The heated air was pulled down through the mass. Sample 1 was taken as an average of the dried hops at the end of the drying run. Two samples were taken from the sur- face, the driest part of the mass, at the end of the run. These, nos. 2 and 3, were Air Out Ueated Air \n 6»oo Vb 6 ,0b C-.fe Wuops<«?; Fig. 11. Pilot drier for deep-bed studies; D, door permitting recirculation of air with the system closed. exposed to room air conditions in con- tainers 16 in. square and 5 in. deep. In one, the bottom was tight and the top open; the outside to inside air exchange was by diffusion. The other was force- ventilated at 150 cu. ft. per min. per sq. ft. The moisture response of these samples was determined by periodic weighings of the samples and containers (fig. 12). Sample 4. August 17. The direction of air movement was reversed when the average moisture content was about 35 per cent. An average bale was made at the conclusion of the test. Samples 5, 6, 8, 1 1 . August 19, 20, 23, 25, respectively. Bales of an average sample were made at the end of the run. Sample 7. September 3. Steam was added to the heated incoming air to in- crease the relative humidity to 24 per cent. This run was otherwise comparable to no. 6, in which the relative humidity of the incoming air was 9 per cent. The drying curves for these two tests are noted graphically in figure 10. Samples 9, 10. August 31. The hops were removed from the drier when they had reached an average moisture content of 30 per cent. The containers of samples 2 and 3 were filled and the hops dried with the same procedure. The relations of moisture content and time are reported in figure 13. Samples 12, 13. August 18. These are duplicate runs. Because of difficulty in adjusting the incoming air tempera- ture and since no. 12 was removed from test before the accepted final moisture content was reached, a second run, no. 13, was made. The average moisture con- tent and incoming air temperature are plotted in figure 14. Sample 14. August 16. A small quan- tity of the green hops secured for the sample 1 run were placed in a vacuum oven at 140 °F and a minimum vacuum of 29 in. of mercury. The heat transferred poorly to the center of the mass and a small portion of the mass was still wet when the hops were removed and baled. [18] The bale was spoiled when opened for evaluation. Samples 15, 16, 17. August 30. These were from a commercial drier at the time of hop removal. The moisture content of the top of the mass was 9.3 per cent; the bottom, 4.6 per cent; and the center, 5.9 per cent. Sample 15 was taken within 20 minutes after the hops were removed from the drying floor. Sample 17 was made so that bottom and top thirds had a moisture content of 4.6 per cent and the center third 9.3 per cent when baled. Results and Conclusions The results are summarized in tables 1, 2, and 3, and figures 12, 13, and 14. Table 2. Quality Indices* for 1954 Tests Based on Individual Ratings by Eight Quality Judges Sample Color Aroma Shatter Lupulin Over-all no.f Av. Std. dev. Av. Std. dev. Av. Std. dev. Av. Std. dev. av. 1 3.38 0.74 3.25 1.03 3.67 0.52 3.25 0.96 3.29 2 2.75 0.71 3.12 0.84 3.57 0.54 3.75 0.50 3.21 3 3.38 0.74 3.12 1.13 3.72 0.49 3.00 1.00 3.13 4 3.00 0.76 3.12 0.64 3.28 0.76 3.00 0.82 3.04 5 1.63 0.74 2.25 0.64 2.86 1.07 2.67 0.57 2.18 6 3.00 0.93 2.62 0.52 2.14 0.69 3.00 1.00 2.87 7 2.75 1.17 2.62 1.06 2.28 0.49 2.50 1.00 2.62 8 2.62 1.05 2.38 1.06 2.00 0.00 2.75 0.96 2.58 9 2.12 0.84 2.37 1.19 2.28 1.03 2.75 0.96 2.41 10 3.38 0.52 3.00 1.07 3.43 0.79 3.00 1.15 3.13 11 2.75 1.03 2.38 1.14 1.86 0.37 3.00 0.82 2.71 12 2.50 1.20 2.62 1.06 3.00 0.56 2.50 0.57 2.54 13 1.75 0.71 1.88 0.94 2.83 0.98 1.67 0.57 1.77 15 3.12 0.64 3.38 0.72 2.72 0.49 3.50 0.57 3.33 16 3.12 0.64 3.25 0.71 2.57 0.54 3.00 0.00 3.12 17 3.50 0.54 3.12 0.72 2.57 0.54 3.67 0.57 3.43 Judge average Judge no. Color Aroma Shatter Lupulin 1 2.38 2.56 2.34 2.44 2 3.32 2.82 2.56 2.94 3 3.06 3.12 2.67 4 2.50 2.32 3.00 5 3.56 2.31 2.62 2.62 3.82 2.50 3.12 2.08 3.25 3.07 2.81 2.34 6 7 3.54 8 * Excellent, 4; good, 3; fair, 2; poor, 1. t Sample 14 was completely spoiled and was not rated. [19] Table 3. Summary of Statistical Analysis of Data Reported in Table 2 Attribute and source of deviation Sum of deviations squared Degrees of freedom Mean square F for significance At 5% level Atl% level COLOR Samples. . . Judges. . . . Error Total items AROMA Samples . . . Judges Error Total items SHATTER Samples . . . Judges Errors Total items LUPULIN Samples . . . Judges Error Total items 38.85 23.72 36.15 98.72 24.67 34.61 60.02 119.30 35.40 9.36 31.57 76.33 11.98 9.19 17.81 39.98 15 7 105 127 15 7 105 127 15 6 87 108 15 3 38 56 2.59 3.39 0.34 1.64 4.94 0.57 2.36 1.56 0.36 0.80 3.06 0.47 7.62 9.97 2.88 8.67 6.56 4.33 1.70 6.51 1.77 2.10 1.77 2.10 1.78 2.20 1.90 2.82 2.22 2.82 2.22 2.82 2.25 3.02 2.47 4.25 Details of each run needed to clarify table 1 are given above. The results of the quality evaluation by the panel of experts (table 2) were treated statistically to determine the sig- nificance of the data. The "judge aver- age" is the average of all ratings by a single judge for a single factor. The re- sults of the statistical study 4 are summar- ized in table 3. The factors to be used by the judges in evaluating the samples were not specified. Each judge used color 4 The procedures were based on Snedecor (Snedecor, George W. Statistical methods. 3d ed. 422 pp. Iowa State College Press, Ames, Iowa. 1940) and on recommendations of G. A. Baker, Mathematics Department, University of California. Relative Humidity 3 \i 15 l~\me — Days Fig. 12. Response to unheated air; depth, 5 in. [20] 16 70 ± 50 1 30 J 35 30, JD Us a i 4- c Sis <3 10 \ Diffusion ^S. 150 fp.m. 2 3 L 5 Tirr2(? — Days Fig. 13. Finish drying with unhealed air; depth, 5 in. and aroma as factors. One judge did not consider shatter. Three judges did not consider shatter on all samples. Four (half) of the judges did not consider lupulin in their evaluations, and two of the others considered lupulin in only part of the samples. The mean square of the statistical analysis of table 3 is a measure of the variation between the averages weighted on the basis of the number of items and number of averages considered. Conse- quently, any variance can be compared with any other variance. Considering the color data, note that the variance for the samples is smaller than for the judges. This indicates that there was a greater variation between the opinions of the judges than between the samples. The error mean square indicates the variation between items which is due to factors other than sample treatment and judges' opinions. The F index is the mean square under study divided by the error mean square. If F is larger than the F for sig- nificance at the 5 per cent level, the differences are considered significant, if larger than at the 1 per cent level, highly significant. Therefore the test runs are signifi- cantly different, one from another, when evaluated on the basis of color, aroma, [21] 2 5 4. ~T\me — Hours Fig. 14. Average moisture content and required incoming air temperature when discharge air temperature is fixed at 1 17° F. Numbers in parentheses are sample numbers, table 1. and shatter. The judges who considered lupulin as a quality factor were unable to rate the samples on this basis. The judges' opinions are significantly differ- ent, one from another, when considered on the basis of these quality indexes, and vary more than the samples. This means that in all cases the judges could not agree as to the rating of a specific sample; however, the average of their opinions permitted the samples to be rated significantly except in the case of lupulin. The samples dried at 115°F (samples 1-4 and 15-17) were superior in quality to those dried at higher temperatures. Sample 10, which was partially dried at a high temperature and finished with [22] unheated air, was rated as "good." This indicates that plant capacity might be increased by using a drying procedure comparable to that of sample 10. The data suggest that air reversal (sample 4) reduces the drying time; but since the final moisture content of this sample was high, the data are not con- clusive in this respect. All of the samples except nos. 2, 3, 9, and 10, were baled directly from the drier. Since the quality did not appear to have been reduced by this procedure, elimination of the cooling floor seems possible. Note, however, that the air rate was higher and bed depth less in some samples than in field practice. Also note that in many cases the final moisture content was lower than desired. The results (table 2) from samples 2, 3, 9, and 10 (see figs. 12 and 13) indi- cate that exposure to static air may be detrimental to quality. Forced air ap- pears superior. Note that sample 10 dried from 30 per cent moisture content to 10 per cent in one day and that the quality in all respects was high. This sample test indicates that partially dried hops might be finish-dried satisfactorily with un- heated air if a high air rate were used. The results with samples 12 and 13 (fig. 14) indicate that the constant-air- discharge-temperature procedure does not yield a quality product. Drying at a constant incoming-air temperature of 135°F (sample 6) produced better- quality hops in slightly less time. Exposed dried hops respond in mois- ture content to the relative humidity of the air surrounding them. Thus from figure 5 the moisture content of dried hops might be expected to approach 8 per cent at the normal time of harvest. The data of figures 12 and 13 show that the moisture content of approach for the 1954 season was under the normal. The data also show that the response may be relatively fast (note the change in mois- ture content in the first 2 days in fig. 13) . Thus, since the relative humidity may vary from 30 per cent during the day to 35 per cent or more at night, consider- able variation in moisture content of ex- posed hops can be expected within any 24-hour period. 1955 Tests of Closed Recirculation The studies conducted during the 1955 harvest season, described in table 4, were designed to investigate further the possi- bility of reducing the final moisture dif- ferential, top to bottom of the mass, by using high-relative-humidity air during the entire drying period or by equalizing with high-relative-humidity air after dry- ing is completed. High-humidity or "closed recircula- tion" conditions were produced by opening door D (fig. 11), partially or completely covering the air-discharge port, and partially covering the air- intake port. The conditions of the different runs are summarized in table 4. The hop depth given there is that when drying starts; the depth reduces about one third during drying. The air temperature and relative humidity (from wet- and dry-bulb ob- servations) are of the drying air entering the mass. The moisture contents were determined by the calibrated Delmhorst small-grain moisture meter. Equalizing was started or the run was terminated when the average moisture content reached approximately 8.0 per cent. Wind extinguished the heating flame during the night of run 4; run 5 is a rerun. Run 11 was intended to be oper- ated with an air rate just under that which would lift the hop mass. It was started with an air rate of 100 cu. ft. per min. per sq. ft. Motor trouble developed X /4 hour later, and the rate was reduced to 60 cu. ft. per min. per sq. ft. for the remainder of the run. The floating velocity — that is, the air velocity required to just lift a single hop — of 14 random single hop cones was determined by a piece of equipment de- signed for this type of observation by [23] Q C *>* o a o z k. o 3 IO lO : ■* oo" t- IO • O © rl* . cm IO IO o • CM* ■"** CO IO o O CM O • • CO CO CM CO CM • efc :§S o o o • a t>o iH G> CO • fl) *> iH i-i tH H" IO © • IO IO IO CM IO IO w 1* «D -^ Tf t- co to CO CD i-l CM i-H CM CM i-4 i-i y-i i-H r-\ iH r-i IO CO Ci 00 CO IO q oq t> cm CO 00 00 00 00 ^ 00 l>" 00 t^ j a 0) S3*. 1 o g w t> o O O IO IO CO IO CO <1) « S id a" *a* Tji xtf* id " V H-O O •*» CO 0) , O 00 IO O l> t- t- lO CM CM - o o 5 d i tj< cm cm CM CM CM tH a-S • lO IO to in to to IO IO o "S M 00 00 00 00 00 00 rj< 00 00 ^C rlWH iH tH r-i M H H CO la i-H CM CO IO CD C- oocnoH i-H tH _ ®