determining machine efficiency parameters for a citrus ...

DETERMINING MACHINE EFFICIENCY PARAMETERS FOR A CITRUS CANOPY SHAKER USING YIELD MONITOR DATA R. Shamshiri, R. Ehsani, J. M. Maja, F. M. Roka

ABSTRACT. A yield monitoring system was used to collect yield data from a commercial citrus canopy shaker during the 2008 harvesting season. A computer algorithm was developed to preprocess the yield data before estimating field efficiency (Ef) and field machine index (FMI) measures of the mechanical harvesting equipment. Total time of the harvesting operation was calculated and divided into primary and support functions, which corresponded to the effective harvesting time and machine time losses, respectively. Time losses related to row-end turning were determined using an algorithm based on linear regression and geometrical methods. Each component of the harvesting operation was then expressed as a percentage of total field time. It was observed that FMI varied from 80% to 98% with 4% standard deviation. Turning time varied from 3% to 8% of the total operational time. Further data analysis showed an exponential relationship between FMI and row-end turning time with R2=0.97. It was also observed that the actual travel distance and the effective time of operation have linear relationships with the theoretical distance of operation with R2=0.96 and 0.93, respectively. Keywords. Citrus mechanical harvesting, Field machine index, Field efficiency, Yield data, GPS data, Operational analysis.

F

lorida is the major producer of citrus in the United States, accounting for more than 75% of all sweet oranges grown in the country (FASS, 2009). More than 95% of Florida’s sweet orange crop is processed into juice, making the use of mechanical mass harvesting systems feasible (FASS, 2009). Between 1999 and 2009, the acreage of Florida citrus that was harvested with mechanical systems increased to more than 14,200 ha (Roka et al., 2009). Growers who utilized mechanical harvesting equipment witnessed improved labor productivity and reduced the number of manual harvest workers. More importantly, net harvesting costs decreased 10% to 20% as compared to conventional hand harvesting techniques (FASS, 2009; IFAS, 2010). Field machine index (FMI) and field efficiency (Ef) are two of the important performance parameters that go into machinery management decisions. The field machine index is the “ratio of productive field time to productive plus turn

time” (Braide, 1989). Hunt (1995) describes nonproductive or machine lost time occurs when the operator is driving the machine but not performing the machine’s primary task. “Non-productive” time includes turning times at row-ends, stops for equipment adjustments, driver breaks, and cleaning various parts of the machine. The FMI, as shown in equation 1 (Renoll, 1970a), requires three data points; total field time (Tf), total stop time (TS), and total turning time (Tr). FMI =

T f − Ts

× 100

(1)

Field efficiency is defined as the ratio of the maximum theoretical productivity of a machine (canopy shaker) to its actual field productivity (ASAE Standards, 1999). This parameter can be calculated using equation 2, where Tt is theoretical time of operation. Ef =

Submitted for review in January 2011 as manuscript number PM 9008; approved for publication by the Power & Machinery Division of ASABE in November 2012. The authors are Ramin Shamshiri, ASABE Member, Graduate Student, Reza Ehsani, ASABE Member, Associate Professor, Joe Mari Maja, ASABE Member, Postdoctoral Research Associate, Agricultural and Biological Engineering Department, University of Florida, IFAS, Citrus Research and Education Center, Lake Alfred, Florida; and Fritz M. Roka, Associate Professor, Food and Resource Economics Department, University of Florida, IFAS, Southwest Florida Research & Education Center, Immokalee, Florida. Corresponding author: Reza Ehsani, Agricultural and Biological Engineering Department, University of Florida, IFAS, Citrus Research and Education Center, 700 Experiment Station Road, Lake Alfred, FL 33850; phone: 863-956-8770; e-mail: [email protected].

T f − Ts − Tr

Tt × 100 Tt + Ts + Tr

(2)

Previous studies, which evaluated machinery performance, have often used tedious procedures such as manual observations and self-recording clocks to record the duration of various field operations that determine FMI and Ef of farm machines (Renoll, 1972, 1981). Availability of geo-referenced data from the yield monitoring system has provided new ways for producers to track the status of machine performance during its field operations. Renoll (1969, 1970a, 1970b, 1979) studied the interaction between machine use and the physical and geometrical

Applied Engineering in Agriculture Vol. 29(1): 33-41

© 2013 American Society of Agricultural and Biological Engineers ISSN 0883-8542

33

characteristics of a field. His results suggested that the FMI is related to row length, turn condition, and terrace systems. Hence, extracting the factors from geo-referenced data and time-motion studies that influence machine performance rates would allow producers to optimize effective field capacity. This, in turn, would translate into improved machine productivity and lower unit cost of equipment operation. If producers could adjust machine operation to fit specific field shapes, row lengths, and row patterns, nonproductive machine time could be minimized. Grisso et al. (2000, 2002, 2004) used spatial data and traffic patterns to determine field efficiency for a soybean combine harvester and a row planter. These studies indicated that field efficiency was higher in flat fields with straight rows as compared to fields with curved row patterns. In a study by Adamchuk et al. (2011), a methodology was presented to define field areas where field efficiency was significantly reduced. It was concluded that the complexity of the machinery maneuvering increased because of “odd field shapes,” obstacles, or contour farming and thereby caused a reduction in a machine’s field efficiency. The main objective of this study was to analyze the yield monitoring data from a commonly used continuous canopy shaker citrus mechanical harvesting machine in order to extract machine operational time parameters and to measure field efficiency and field machine index for typical citrus orchards in Florida. A discussion on economic implications of field efficiency measures for a canopy shaker follows at the end of this article. The uniform terminology for agricultural machinery management (ASABE Standards, 2008a, 2008b) was followed.

MATERIAL AND METHODS Yield data were collected during November and December of 2008 using a yield monitoring system (Maja and Ehsani, 2010) mounted on a continues canopy shaker citrus mechanical harvester (Oxbo 3220, Byron, NY ) at a commercial citrus orchard located at Lykes Bros. Inc. (Fort Basinger, Fla.). Figure 1 depicts the different sections of the

orchard where the mechanical harvester operated during the study period. Individual fields, or blocks, varied in shape, size and row length. It was expected that blocks with irregular shapes would result in lower field efficiency and field machine index measures because turning time becomes a greater percentage of field time. The blocks of citrus orchards studied in this article are a good representation of typical citrus blocks in Florida. Yield data files contained thousands of global positioning system (GPS) data point in minimum code (GPRMC) string (NMEA Standards, 2005) format. GPRMC string was used because it contained all the needed information such as position, time, and speed. A computer software program was developed in the Delphi programming language (Borland®, Austin, Tex.) to preprocess the yield data and extract the information on harvested mass, machine position, and travel speed. This information was used to accurately determine the duration of separate primary and support functions of the harvesting equipment as well as row length, distance between the rows, field efficiency, and field machine index. A flowchart of the custom algorithm and resulting output is shown in figure 2. Raw yield data stored in ASCII files during the harvesting operation were used as the input to a computer algorithm. Data preprocessing steps involved checking these files for errors (e.g., broken strings) and then sorting and combining the files to create a single ASCII file which contains the entire GPS data of the harvesting operation. Total field time, travel speed, and distance between collected points were calculated for each line of GPS data. Field operations were divided into primary and support functions corresponding to effective harvesting time and machine stop time, respectively. The primary functions included tree canopy shaking and fruit collection. Support functions included machine stop times related to equipment maintenance, routine cleaning, driver breaks, and fruit unloading. The above functions were not related to area. Support functions also included the machine turnings at row-ends which were related to total area. Row-end turning

Figure 1. Plot of latitude vs. longitude for the entire harvesting dataset.

34

APPLIED ENGINEERING IN AGRICULTURE

Figure 2. Schematic depicting the yield data preprocessing steps and output from a custom developed computer algorithm.

times were calculated using geometrical and statistical approaches from the total field time. Machine stop time was extracted from total field time by comparing the travel speed with a threshold value near zero. Finally, the total time loss was extracted from total field time for each harvesting day to calculate field machine index and field efficiency. The entire calculation process was performed by a customized computer algorithm which provided output results in a spreadsheet format. PRELIMINARY DATA ANALYSIS Each yield point collected by the yield monitoring system included five important pieces of information: date and time of the collection, latitude and longitude of the location, and a voltage value corresponding to the mass of the harvest produced by the load cell. A computer application calculated the distance between each consecutive pairs of yield points. The world geodetic system 1984 (WGS-84) spheroid earth model illustrated by Snyder (1987) was used to perform linear conversion of the latitude and longitude pairs. Equations 4-10 were used in this model, where a was the equatorial radius (= 6378137 m), b was the polar radius (= 356752.3142 m), h was altitude, θ1 and θ2 were true angles, r1 and r2 were radius of each point, xy1(= r1cos(θ1π/180)), xy2(= r2cos(θ2π/180)), xy3(= r1sin(θ1π/180)), and xy4(= r2sin(θ2π/180)) were the X-Y coordinates corresponding to the longitude and latitude axes, respectively. D was the distance in meters between two points. Assumptions used by Snyder (1987) in developing these equations resulted in errors of less than 0.1 m in 400 m at 45° latitude.   b2 π    180  θ1 = tan −1  tan  Lat1   2  180    π   a

(4)

  b2 π    180  θ2 = tan −1  tan  Lat 2    a2 180    π   

(5)

 2 π   2 π  cos  180 θ1  sin  180 θ1    +   r1 =    a2 b2    

29(1): 33-41

 2 π   2 π  cos  180 θ2  sin  180 θ2    +   r2 =  2 2   a b    

X=

−0.5

( xy1 − xy2 )2 + ( xy3 − xy4 )2

 xy + xy2  Y = 2π  1  ( Long1 − Long 2 )  720 

D=

X 2 +Y2

+h

(7)

(8) (9) (10)

This spheroid model was tested on 110 GPS data points that were randomly selected from the actual yield data. The distances estimated by the model between each pair of these selected data points were compared to actual field measurements. The comparison showed a maximum error of 1.6% between the model and the ground truth measurements. The distance between two consecutive tracked points and their corresponding time difference was used to calculate the travel speed and to determine machine field time. The most prevalent machine ground speed was less than 1.0 m/s with an average travel speed of 0.53 m/s and a standard deviation of 0.48 m/s (fig. 3). Yield data were collected from a separate experimental harvesting site (Latitude: 29.6404, Longitude: -82.3618) and used during development and testing of the algorithm that extracted machine time losses at row-end turnings.

−0.5

+h

(6) Figure 3. Travel speed histogram for the canopy shaker for the entire harvesting operation.

35

This algorithm calculated the amount of changes in both latitude and longitude values and compared the results with a threshold that was adjusted for each dataset by trial and error. To improve the results, linear regression analysis was performed on latitude and longitude pairs using samples of 5 to 100 data points. Based on the R-square value, the points with largest residuals were considered as the turning points. Due to the large size of the block, the maps of the harvester’s stop and turning points were provided for only a small portion of the trial operation as shown in figures 4 and 5, respectively. Row lengths for three blocks labeled 1, 2, and 3, varied considerably from the field shown in figure 5. This would increase the time for the row-end turnings, and consequently increase overall turning times, while harvesting the shorter rows of block 3 as compared to other two blocks.

FIELD MACHINE INDEX AND FIELD EFFICIENCY Three parameters determine field machine index (FMI): total field time, total stop time, and total turning time (eq. 1: Tf, TS, and Tr, respectively). These parameters were estimated from yield data and used to calculate FMI (eq. 1) for each harvesting day. In order to calculate field efficiency (eq. 2), theoretical field time (Tt) had to be estimated. First, average machine speed was set equal to 0.5 m/s throughout the study period. As shown in figure 2, the speed of canopy shaker during harvesting did not exceed 0.89 m/s. Second, theoretical travel distance (dt) for

Figure 4. Sample map of stop points plotted from raw yield data collected during the harvest operation.

27.343

3 2

Latitude

27.342

1 27.341

27.34

the operation was determined by calculating distance between each row-end and then summing the results. Having the theoretical travel distance and average travel speed, the theoretical time of operation (Tt) was calculated and used in equation 2. Field efficiency accounts for the time losses due to field area and support function activities whereas the FMI accounts for the “turn time” of the canopy shaker during field harvesting operations.

RESULTS AND DISCUSSION Harvesting operations were performed over 22 days from 17 November 2008 through 16 December 2008. A summary of the time parameters estimated from the yield data for each harvesting day are reported in table 1. The relative percentage difference (RPD) between the effective and theoretical time of operation is calculated based on their absolute difference divided by their arithmetic mean (table 1). This parameter is a quantitative indicator that shows by how much the effective and theoretical field times differ from each other. The performance of a machine depends on operator skill, weather and soil conditions, machine conditions and adjustments (Edwards, 2009). These factors along with variability in harvesting load can affect machine travel speed. In this study, effective field time (Te), as reported in table 1, is the total field time (Tf) minus total time loses (Tr+Ts). Because total field time is directly affected by machine performance, RPD values provide field managers with an index for quality insurance of the effective field time. The RPD can be used to determine the possibility of improving the effectiveness of operation with respect to the targeted time (Tt). While, greater total time losses (Tr+Ts) produce larger differences between actual and theoretical field time as represented by field efficiency, the quality of effective field time (Te) due to machine performance is represented only by the RPD values. For example, if a particular section of field is to be harvested with different drivers with different weather/soil conditions and harvesting load, and if same time loses (Tr+Ts) are produced in different field times, then field efficiencies would be the same in each repetition. This is because the parameters that affect machine performance (i.e., how close to the theoretical harvesting speed the machine has operated) are not considered in field efficiency calculation. In this study, the average RPD between the effective and theoretical time over the entire harvesting period was 22.3% (std. dev. = 12.3%) with minimum and maximum values of 1.21% and 42.8% corresponding to days 6 and 16, respectively. Table 1 also reports time percentage of operation (PO) for each day with respect to the entire time of operation. At the end of operation, these data were used to observe daily progress of the operation. For this harvesting operation, the average PO was 4.5% for each day.

27.339 -81.131

-81.13

-81.1285

-81.1275

Longitude

Figure 5. Sample map of row-end turnings plotted from raw yield data collected during the harvesting operation.

36

APPLIED ENGINEERING IN AGRICULTURE

Days 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

Date 11/17 11/18 11/19 11/20 11/21 11/22 11/24 11/26 11/28 11/29 12/01 12/02 12/03 12/04 12/05 12/06 12/08 12/09 12/10 12/11 12/12 12/16

Tf 3.97 4.53 2.09 4.30 3.57 3.89 4.04 6.19 8.92 6.06 9.33 1.22 10.03 5.84 5.62 7.61 6.54 7.17 6.55 9.86 8.77 3.32

Table 1. Different time segments of operation in each day.[a] Operation Time (h) Tr + Ts Ts Tt Δt Tr 2.24 1.74 0.11 0.88 0.99 2.84 1.69 0.12 0.96 1.08 1.08 1.01 0.04 0.48 0.52 2.34 1.97 0.09 1.05 1.14 1.81 1.76 0.07 0.87 0.94 1.92 1.97 0.07 0.85 0.92 2.20 1.84 0.07 1.06 1.13 3.25 2.94 0.10 1.84 1.94 5.17 3.75 0.21 2.54 2.75 3.16 2.90 0.36 1.66 2.02 5.23 4.10 0.37 2.98 3.35 0.42 0.80 0.01 0.60 0.61 7.28 2.75 0.33 1.86 2.19 4.44 1.40 0.16 1.09 1.25 4.08 1.54 0.44 1.83 2.27 5.30 2.31 0.18 2.19 2.37 4.21 2.32 0.16 1.89 2.05 4.70 2.48 0.16 2.00 2.16 3.93 2.62 0.13 1.95 2.08 3.82 6.04 0.16 4.80 4.96 3.90 4.88 0.40 3.01 3.41 1.38 1.94 0.49 0.87 1.36

Total 129.42 74.69 54.74 4.22 37.26 Min 1.22 0.42 0.80 0.01 0.48 Max 10.03 7.28 6.04 0.49 4.80 Avg. 5.88 3.40 2.49 0.19 1.69 [a] Tf: total field time, Tt : theoretical time of operation, Δt: difference between total field time and theoretical time of operation, Tr: total turning time, TS: total support function time, Te: effective time of operation, RPD: relative percentage difference between the actual and theoretical time of operation, PO: Time percentage of operation for each day with respect to the entire time of operation.

Table 2 reports the percentage of turning time, stop time, and effective time, as well as the Ef and FMI parameters. Actual field distance (da) corresponding to the total distance that canopy shaker travelled each day and the theoretical distance (dt), which is the minimum travel distance required to finish the harvesting operation, are given in table 3. The difference between actual and theoretical distances for each day are reported as Δd. Travel distance at row-end turning for each day is given by dr. The average distance between rows was calculated and multiplied by the field distance in order to determine machine field area for each day. The Ef and FMI (table 2) over the 22 days of the study period ranged between 41% to 78% and 80% to 98%, respectively. The FMI was above 91% every day except on 15th and 22nd day. Low FMI values on 15th and 22nd day are caused by extra distance harvester travelled (table 3). The differences between the actual covered area (AA) and the actual harvesting area (AE) on these two days are also significant (14600 and 16270 m2, respectively). Particularly of 22nd day of harvest, when harvester covered maximum additional area than required, it can be seen from table 1 that the time percent of operation was only 2.57% (average daily time percent of operation = 4.5%). An exponential model was fitted (eq. 11) to establish the relationship between FMI and turning time. The model was a good fit (fig. 6) with a coefficient of determination (R2) of 0.975. Parameters of the model fit are: A = -0.0012, , B =

29(1): 33-41

41.49 0.52 4.96 1.90

Te 2.98 3.44 1.57 3.16 2.64 2.97 2.92 4.24 6.17 4.03 5.97 0.61 7.84 4.59 3.35 5.25 4.48 5.02 4.47 4.90 5.37 1.97

RPD (%) 28.6 19.4 37.3 30.0 36.8 42.9 27.9 26.6 17.6 24.4 13.4 36.6 7.4 3.4 19.7 1.2 6.3 6.5 12.9 24.8 31.7 34.9

PO (%) 3.1 3.5 1.6 3.3 2.8 3.0 3.1 4.8 6.9 4.7 7.2 0.9 7.7 4.5 4.3 5.9 5.0 5.5 5.1 7.6 6.8 2.6

87.95 0.61 7.84 3.99

1.2 42.8 22.3

0.9 7.7 4.5

19.25, C = 98, and D = -0.086. Overall, increased turning time resulted in an exponential decrease in field machine index. It is important to note that this relationship (fig. 6) is valid only for Oxbo 3220 citrus mechanical harvester under typical citrus orchards in Florida. The relationship between FMI and turning time might vary for different types of citrus harvesting machines or different orchard design. The relationship between effective time and the theoretical distance of operation (fig. 7) followed a linear trend (R2 = 0.930) as given by equation 12. Actual and theoretical distance of operation, plotted in figure 8, also demonstrated a linear relationship as given by equation 13 (R2 = 0.964). Given the size of the field, the linear relationships established above can be used to estimate the effective time and actual travel distance required to complete a harvesting operation. These two estimates could assist in determining harvest costs, which include machine and operator time. For example, knowing what size of field area is to be harvested, the minimum required travel distance (dt) to finish the harvesting operation for this area can then be used in equations 12 and 13 to have an approximate idea on how much effective time should be expected on the field and what distance the machine is required to travel. Predicting effective time would be important for estimating equipment leasing rates and costs of hiring equipment operators. Approximating the actual distance to be traveled would allow fuel and lubrication costs to be estimated.

37

Days 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Min Max Avg.

Table 2. Field efficiency (Ef), field machine index (FMI) and the percentage of different time of operations with respect to the total field time. Percentage of Operation (%) Tr/Tf TS/Tf Te/Tf (Tr + Ts)/ Tf 2.9 22.0 75.1 24.9 2.7 21.2 76.1 23.9 1.8 22.9 75.3 24.7 2.1 24.5 73.4 26.6 2.0 24.2 73.8 26.2 1.7 21.9 76.4 23.6 1.7 26.2 72.1 27.9 1.7 29.8 68.6 31.5 2.3 28.5 69.2 30.8 6.0 27.5 66.5 33.5 4.0 32.0 64.0 36.0 1.0 49.4 49.7 50.4 3.3 18.5 78.2 21.8 2.7 18.7 78.6 21.4 7.8 32.6 59.7 40.4 2.3 28.7 69.0 31.0 2.5 29.0 68.5 31.5 2.2 27.9 69.9 30.1 2.0 29.7 68.3 31.7 1.6 48.7 49.7 50.3 4.5 34.3 61.2 38.8 14.7 26.1 59.2 40.8 1.0 14.7 3.3

18.5 49.4 28.4

49.7 78.6 68.2

FMI = Ae BTr + Ce DTr

(11)

Te = 1.93dt + 2600

(12)

d a = 0.807dt + 6.629

(13)

ECONOMIC IMPLICATIONS OF FIELD EFFICIENCY MEASURES Field efficiency measures have important implications on the overall cost of any mechanical harvesting system. For illustrative purposes, assume that orange trees are spaced at an average distance of 0.4 m (12 ft) down the row. Furthermore, the average yield of each tree is 3 boxes (41 kg/box) and the harvesting system recovers 2.5 boxes per tree, 83%. Also, assume that the mechanical harvesting system analyzed in this research, a canopy shake and catch system, operates for 800 h during a season and runs at an average speed of about 0.5 m/s during harvesting. Finally, the annual cost to own and operate this harvesting system is $750,000 over one season, or $938/h. Considering the above assumptions on tree spacing and harvest speed, the theoretical capacity of the harvesting system would be 484 trees/h. If field efficiency is 60%, then 290 trees (484 × 0.6) and 725 boxes (290 × 2.5) are harvested per hour. The unit cost of harvesting is $1.29 ($938/725) per box. If Ef falls to 40%, the unit cost of harvesting increases to $1.94 per box [$938 / (484 trees × 40% × 2.5 boxes)]. Costs to employ a hand-harvest crew during the 2009-2010 season ranged from $1.60 to $1.90 per box (Roka et al., 2009). Thus, a reduction in Ef to 40% makes the mechanical system more costly than a handharvest crew. A decrease in field efficiency could impose additional and more significant costs if harvest timeliness is a factor.

38

21.4 50.4 31.7

Ef (%) 69.3 72.4 67.5 67.2 65.8 67.6 66.0 62.6 65.3 61.0 61.0 40.9 76.9 78.0 64.2 69.1 67.2 68.5 65.4 43.5 53.3 50.4

FMI (%) 96.3 96.5 97.6 97.2 97.3 97.8 97.7 97.6 96.7 91.7 94.1 98.0 95.9 96.6 88.4 96.7 96.4 96.9 97.1 96.8 93.1 80.1

40.9 78.0 63.7

80.1 98.0 95.3

Some crops, such as tart cherries, have a limited number of days over which they can be harvested without substantial loss of yield or quality. In the case of citrus, timeliness of harvest is not as important of an issue. The harvesting window for early season (Hamlin) and late season (Valencia) orange varieties is more than two months, during which there are few significant changes in overall fruit quality. Typically, juice processing plants impose limits on the daily volume of fruit that can be harvested at any particular citrus orchard. The daily harvesting quotas are a function more of maintaining cost efficiencies at the processing plant rather than maximizing harvesting capacity of a particular harvest method, or attempting to time harvest at optimal fruit quality. Field efficiency measures are influenced by both field conditions and operator management. Field conditions, such as tree spacing, row length, overall block size, and tree yields are beyond the control of a harvesting manager. If block characteristics diminish field efficiency of a particular harvesting machine, the analysis described in this article could assist a harvesting manager to predict the extent to which a Ef value for a specific block affect harvesting costs, and ultimately determine whether a mechanical system could be a lower cost option then a hand-harvest crew. Predicting Ef and FMI values could motivate harvesting managers to adopting new or adjusting existing equipment handling procedures and in so doing, may increase field efficiency by reducing downtime due to equipment repairs. For instance, placing a mechanic as part of the harvesting crew could eliminate hours of downtime from waiting for a shop mechanic to arrive or from having to transport broken equipment back to a centralized shop facility. As new citrus blocks are planted, field characteristics such as row length and block shape become

APPLIED ENGINEERING IN AGRICULTURE

Days 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

[a]

Table 3. The actual and theoretical field distance and field area.[a] Field Distance (m) Field Area (m2) dt dr AA AT Δd 4023.3 1459.8 246 84857 62265 5103.6 777.6 268.2 91018 78984 1936.1 413 80.4 36355 29963 4204.1 831.6 196.8 77933 65063 3265.6 672.1 153.6 60940 50539 3456.2 691.4 141 64189 53488 3959.7 501.6 146.4 69044 61281 5848.6 933.5 222.6 104960 90514 9307.2 1752.8 446.4 171170 144040 5685.2 2396.6 784.2 125070 87985 9407.9 2818.1 798.6 189210 145600 760.75 261.55 25.8 15821 11773 13100.5 2503.9 708 241500 202740 7993.7 1397.5 346.8 145340 123710 7350.2 3144.4 943.2 162420 113750 9546.8 1402.4 378.6 169450 147750 7585.2 1192.7 351.6 135850 117390 8451.1 1324.7 341.4 151290 130790 7077.3 1331.8 281.4 130140 109530 6879.1 1718.7 345 133060 106460 7012.5 3028.5 855.6 155400 108530 5480.9 2137.6 1051.8 117900 84823

dA 5483.1 5881.2 2349.1 5035.7 3937.7 4147.6 4461.3 6782.1 11060 8081.8 12226 1022.3 15604.4 9391.2 10494.6 10949.2 8777.9 9775.8 8409.1 8597.8 10041 7618.5

Total 170127.4 137435.6 32691.8 9113.4 2632917 2126968 Avg. dA: actual travel distance, dt: minimum required travel distance to finish the harvesting operation, Δd: difference between actual and theoretical distance, dr: travel distance at row-ends turning, AA: actual covered field area, AT: minimum required field area, AE: actual harvesting field area, PD percentage of the difference between the actual and theoretical travel distance of the operation.

less of a constraint as new blocks could be designed with explicit objectives to increase Ef of a mechanical harvesting system. The economics of any mechanical system are enhanced whenever the harvesting capacity of the unit can be increased. The higher a system’s Ef, the higher the hourly harvesting capacity.

CONCLUSIONS A computer algorithm was developed and used to process yield data collected from a citrus mechanical harvesting machine during the 2008 season to determine field efficiency and field machine index. Three separate blocks (labeled as 1, 2, and 3 in fig. 5) demonstrated that the number of row-end turnings increases as field size

AE 81050 86867 35111 74887 58563 62007 66778 101520 164260 112940 176850 15422 230540 139970 147820 163590 130410 146010 125790 127720 142150 101630

PD (%) 26.6 13.2 17.5 16.5 17.0 16.6 11.2 13.7 15.8 29.6 23.0 25.5 16.0 14.8 29.9 12.8 13.5 13.5 15.8 19.9 30.1 28.0

2491885 20.9

becomes smaller. Turning time was greatly influenced by field size, shape, and orchard pattern. Irregular field shapes with rows not intersecting the field boundary at a right angle presented additional turning problems. Row length also had an impact on turning time. Results showed that as row length increases, turning time decreases, and machine capacity increases. Field efficiency can be increased by reducing time lost during the field operations (e.g., row end turning time). Results showed that a turning time of 3% to 8% of the total time for the citrus canopy shaker can be obtained on fields where there was reasonable row length and good turn conditions. Turning time of more than 10% of the total time, however, was considered excessive for operations and was observed only in some blocks. Overall, the results of this study indicated a good adaptability of existing citrus orchards in Florida for the

100

FMI (%)

95 90 85

FMI vs. Turning time FMI=-0.0012e

80 0

0.05

0.1

19.25t

0.15

+98e

-0.086t

2

, (R =0.97)

0.2 0.25 0.3 Turning time (hr)

0.35

0.4

0.45

0.5

Figure 6. Relationship of field machine index with turning time of citrus canopy shaker (Oxbo 3220).

29(1): 33-41

39

4

x 10

Effective time (sec)

3

2

Te=1.93dt + 2600, (R =0.93)

2.5

Te vs. dt

2 1.5 1 0.5 0

0

2000

4000 6000 8000 10000 Theoretical travel distance (m)

12000

Actual travel distance (m)

Figure 7. Relationship between effective time of operation and theoretical travel distance of the canopy shaker.

15000

2

da=0.807dt +6.629, (R =0.96) da vs. dt

10000 5000 0

2000

4000 6000 8000 10000 Theoretical travel distance (m)

12000

14000

Figure 8. Relationship between actual and theoretical travel distance of canopy shaker during harvesting.

use of mechanical harvesters. Field machine index indicates the influence of row-end turning conditions and row length on actual field time and total time losses. The higher the field machine index, the better the field adapted to machine use. The field machine index in this study was observed in the ranges of 80% to 98% with standard deviation of 4%. Time-and-motion studies could help improve the management strategies to decrease turning time when FMI is less than 90%. Accurate field efficiency measurements can help guide growers and/or owners of harvesting equipment to effectively manage their harvesting options, to reduce overall net harvesting costs and increase returns. ACKNOWLEDGEMENTS This work was supported by a grant from the University of Florida IFAS Citrus Initiative program. We would like thank Mr. Bill Barber, Mr. David Bohannon, and Mr. Esa Ontermaa at the Lykes Bros. Inc., Lake Placid, FL for their assistance and support with this study.

REFERENCES Adamchuk, V. I., R. Grisso, and M. F. Kocher. 2011. Spatial variability of field machinery use and efficiency: Vol. 2: Nutrient Management for Energy Efficiency. In GIS Applications in Agriculture, 135-146. D. E. Clay and J. F. Shanahan, eds. Boca Raton, Fla.: CRC Press.

40

ASAE Standards. 1999. EP496.2: Agricultural machinery management. St. Joseph, Mich.: ASABE. ASABE Standards. 2008a. S495.1: Uniform terminology for agricultural machinery management. St. Joseph, Mich.: ASABE. ASABE Standards. 2008b. D497.5: Agricultural machinery management data. St. Joseph, Mich.: ASABE. Braide, F. G. 1989. Field capacity performance of track type tractors in plough operations under tropical conditions. In Proc. of the 11th Intl. Congress on Agricultural Engineering, 3: 1737-1741. V. A. Dodd, ed. Brookfield, Vt.: Balkema Publishers. Edwards, W. 2009. Farm Machinery Selection. PM-952, Iowa State University. Ames, Iowa: Ag Decision Maker. FASS. 2009. Citrus Summary 2007-2008. Florida Agricultural Statistics Service Publication by USDA, NAAS, Florida Field Office. Washington, D.C.: USDA National Agricultural Statistics Service. Available at: www.nass.usda.gov/Statistics_by_State/Florida/Publications/C itrus/cs/2007-08/cs0708all.pdf. Accessed 4 September 2012. Hunt, D. 1995. Farm Power and Machinery Management. 9th ed. Ames, Iowa: Iowa State University Press. IFAS. 2010. Citrus Mechanical Harvesting. Gainesville, Fla.: University of Florida/ Institute of Food and Agricultural Services. Available at: citrusmh.ifas.ufl.edu. Accessed 4 September 2012. Grisso, R. D., P. J. Jasa, and D. Rolofson. 2000. Field efficiency determination from spatial data. ASAE Paper No. 001013. St. Joseph, Mich.: ASAE.

APPLIED ENGINEERING IN AGRICULTURE

Grisso, R. D., P. J. Jasa, and D. Rolofson. 2002. Analysis of traffic patterns and yield monitor data for field efficiency determination. Appl. Eng. in Agric. 18(2): 171-178. Grisso, R. D., M. F. Kocher, V. I. Adamchuk, P. J. Jasa, and M. A. Schroeder. 2004. Field efficiency determination using traffic pattern indices. Trans. ASAE 20(5): 563-572. Maja, J. M., and R. Ehsani. 2010. Development of a yield monitoring system for citrus mechanical harvesting machines. Prec. Agric. 11(5): 475-287. NMEA Standards. 2005. NMEA Reference Manual. Rev. 1.3. Available at: www.echelleinconnue.net/outils/mobio/ composants/SIRF_NMEA_Reference_Manual.pdf. Accessed 4 September 2012. Renoll, E. S. 1969. Row-crop machine capacity in terraced fields. Bulletin No. 395. Auburn , Ala. : Auburn University, Agricultural Experiment Station. Renoll, E. S. 1970a. Some effects of management on capacity and efficiency of farm machines. Circular No. 177. Auburn, Ala.: Auburn University, Agricultural Experiment Station.

29(1): 33-41

Renoll, E. S. 1970b. Using operation analysis to improve rowcrop machinery efficiency. Circular No. 180. Auburn , Ala.: Auburn University, Agricultural Experiment Station. Renoll, E. S. 1972. Concept of predicting capacity of row-crop machines. Trans. ASAE 15(5): 1025-1030. Renoll, E. S. 1979. Using farm machinery effectively. Bulletin No. 510. Auburn, Ala. : Auburn University, Agricultural Experiment Station. Renoll, E. S. 1981. Predicting machine field capacity for specific field and operating conditions. Trans. ASAE 24(1): 45-47. Roka, F. M., J. K. Burns., J. Syvertsen., T. Spann, and B. Hyman. 2009. Improving the economic viability of Florida citrus by enhancing mechanical harvesting with the abscission agent CMNP. Available at: citrusmh.ifas.ufl.edu. Accessed 4 September 2012. Snyder, J. P. 1987. Map projections:A working manual (Geological survey professional paper 1395). Washington, D.C.: U.S. Government Printing Office.

41