Paper No. 923036


Richard A. Gilbert, Chemical Engineering Dept.
William A. Miller, Dept. of Industrial & Management
Systems Engineering
University of South Florida
Tampa, FL 33620 USA

Steven A.Sargent, Jeffrey K. Brecht
Horticultural Sciences Dept.
Institute of Food and Agricultural Sciences
University of Florida
Gainesville, FL 32611 USA
For presentation at the 1992
ASAE Summer Meeting
Charlotte NC June 21-24, 1992

    SUMMARY: The design, implementation and results of automating a commercial tomato packing operation are described. Installation of a motor control center permitted integrated control of packing line speed from the dump rate into the receiving tank through carton filling and inventory control. Two years after implementation, benefits included productivity gains of 35X, more consistent packout quality, rapid access to packout information, higher wages for workers and potential for reduced mechanical injury to tomatoes during handling.
    KEYWORDS: Automation, packing line, safety, quality, instrumented sphere, mechanical damage, inventory control.

This is an original presentation of the author(s) who alone are responsible for its contents.

The Society is not responsible for statements or opinions advanced in reports or expressed at its meetings. Reports are not subject to the formal peer review process by ASAE editorial committees therefore, are not to be represented as refereed publications.

Reports of presentations made at ASAE meetings are considered to be the property of the Society. Quotation from this work should state that it is from a presentation made by (the authors) at the (listed) ASAE meeting.



    Commercial tomato packing lines are designed so as to permit flexibility in the volume of product which can be packed. This flexibility is essential in order for managers to maintain consistent packout quality to meet U.S. Grade Standards, despite quality differences between incoming lots of tomatoes or shortages of grading personnel. The capacity (or throughput) of a packing line is determined by the rate at which the field bins or gondolas are dumped into the water receiving tank (known as the dump tank) and by the speed of the individual components (elevator rolls, undersize eliminator belt, wash brushes, etc.). For maximum packing efficiency at any particular time, the capacity (or percent coverage) of a packing line should be as near to 100% as possible with the packing line running at a speed which permits adequate sorting and grading. One hundred percent capacity refers to the full carrying capacity of the packing line, with complete coverage of the individual components by the tomatoes. Operation at full capacity also lowers impacts at many transfer points by reducing roll distances down transfer plates.

    In typical commercial packing lines, the only means of changing the capacity is to vary the d dump rate, since individual components are driven by fixed-speed motors. Therefore, when a packing line is operated at low capacity, it is underfilled and can permit excessive fruit-to-fruit impacts and fruit-to-conveyor impacts at the numerous transfer points, causing bruising and other mechanical injuries (Sargent, et al., 1992a). When the fixed-speed packing line is operated at high capacity, overcrowding can occur and result in ineffective removal of undersize fruit, and poor washing, waxing and grading.

    Studies have shown that graders have a tendency to remove a certain amount of fruit each minute, despite the quality (Wardowski, et al., 1987). At low capacity, the amount removed per grader is determined by the number of fruit which can be picked up, while at high capacity at the same speed, the amount removed is determined by the number of fruit which can be observed. In other words, when the packing line is run at low capacity graders try to appear busy, while when it is run at high capacity graders tend to reject fruit based on the quality. Therefore, productivity in packing operations could be increased by varying the packing line speed with the dump rate so as to maintain nearly full capacity while permitting graders to properly grade fruit. This should also result in improved packout quality and savings due to fewer rejected shipments because of poor quality.

    In order to help overcome some operational problems in the tomato packinghouse, it was decided to add computer control capabilities to the existing packing line by integrating the individual packing line components and monitoring the packout. This would permit variable speed control of the integrated components to improve packing efficiency and provide process information for inventory control. To implement this solution, a dual- pronged strategy was developed. First, management confidence in the solution was developed and second, the actual computer control equipment was installed in stages. Matjaz (1987) discussed this type of situation.

    In our opinion, it is necessary to obtain a commitment from the packinghouse management prior to beginning the project to purchase the equipment required to complete the entire project. Typically, the large gains in productivity come near the front end of the implementation process, while later changes tend to have fewer productivity gains and probably are more expensive to implement. If large productivity improvements are shown early in a project, it is often difficult to get managers to commit the money to complete the final stages of the project. This has been referred to as the S-curve (Foster, 1987; Hayes and Wheelwright, 1984).

    After extensive discussions with packinghouse management it was clear that to be successful the installation had to meet the following criteria:

This paper reviews the computer installation strategy selected for this application, summarizing project results and highlighting problems that were generated by the project solution. Data are also presented concerning the effect of varying packing line speed on tomato impacts at transfer points and on subsequent fruit quality.


Three-phase installation strategy

    A programmmable controller-based control system was selected for the packing line (Telemecanique, Westminster, MD, Model TSX-17). This type of computer technology has several advantages over any other candidate system. An important feature is the fact that it is an industrially hardened, microprocessor-base machine that provides a modular design. This feature allowed the packinghouse to invest in the CPU and I/O modules, required to initiate the control scheme, in stages.

    A phase installation strategy using a programmable controller has several advantages. The initial investment is lower. The backbone equipment of the overall control plan can be installed without management becoming over concerned that the technology is beyond their understanding and therefore out of their control. Finally, the observed success with the first phase of the plan would develop a feeling of confidence in the equipment.

    The first phase of the computer control effort in the packinghouse was to develop an operational interlock and speed control for the various belts that feed tomatoes to the grading tables. This was determined to be the best first phase installation activity because of the enormous amount of wasted time and human energy that occurred when the speeds of individual packing line components compensate for variations in the productivity of the human graders. Formerly, the packing line manager could not adjust each of the components to provide an efficient coupling of packing line speed and worker grading capabilities. As a result, two performance extremes were common. Produce would accumulate on the grading table rolls and large numbers of tomatoes would not be graded or would pass by the graders so slowly that incorrect grading would occur.

    To keep within the criteria of maximum integration with present equipment, variable frequency drives were installed to vary the speed of existing motors on the individual components. A digital-to-analog output module was installed in the programmable controller. This output port provided a 0 to 10 volt signal that controlled the variable frequency drives. A 16-bit digital input module received the speed control setting from a 3-digit BCD thumb wheel located at the grading table. Once installed, the grading table supervisor was able to control the speed of all the packing line components from a single input station. An overview of the automated system components is presented (Figure 1).

    The second phase of installation was to develop an audit system that would monitor the number of each type of tomato box after packing and filling. This accounting system included a material balance scheme so that a cull rate for each packout could be determined. Count information was acquired with a combination of micro-switches, photo sensitive eyes and position sensors. Three 16-bit digital modules were used to detect transitions from these detectors. Two dual channel ASCII modules (Telemecanique, Model 67) were used to provide an interface with printers and monitors. The hard copy was provided to the accounting department and to the ripening room supervisor. Electronic copy was delivered to the packinghouse manager and to the sales staff. The monitors were equipped with a screen dump option so that hard copy was also possible at these locations.

    The third stage focused on the packinghouse's motor control center (MCC; Telemecanique, Model 17). Although the first phase of the project integrated the speed control for the conveyor system, the initial work did not include the computerization of the MCC. This was a deliberate decision. If the computer had been allowed to take over this function early in the project, management personnel would have lost their sense of control before having time to develop confidence in the new computer system. By leaving computerization of the MCC until the last phase, operation personnel could still go through the familiar startup sequence of manually engaging the equipment motor starter contacts.

    Even though the computerization of the MCC was to be done last and management had seen the packinghouse efficiency improvements generated from the first two installation phases, they were still hesitant to start phase three. To sidestep management concerns, master control relays were installed between the programmable controller, digital output modules and the MCC contactors. This was a relatively simple and inexpensive compromise and allowed management to override the computer control on any piece of equipment when necessary. This extra psychological security was all that was needed to push the project to completion.

    With the three-phase installation plan completed, the programmable controller had total control of the automated equipment, provided synchronized conveyor speed between the components, and generated inventory reports for each lot of tomatoes packed. The computer system allowed management to override any command decision but executed all safety interlocks whenever any manual override was instigated. The system included a real time clock so that packout records could be time correlated and down times could be monitored.

Benefits of Computer Control

    The main objective, increased productivity, has been achieved. In fact, that increase produced other benefits as well. One year after completion of phase one, productivity was up approximately 26% (Table 1). After two years, with phase one and two completed, productivity had increased approximately 35% (data not shown). Further data were not obtained concerning productivity after implementation of phase three. Packout information on quality, grade and size for each grower is available immediately after each lot is packed, and a complete audit of the entire day's packing activity is available immediately after the last packout of the day. Prior to these modifications, each lot required from 15 to 45 minutes to obtain data, depending on packout size and complexity, and the daily activity report required from 12 to 15 hours.

    The completion of phase three did not significantly alter productivity figures. This would be expected since the phase three installation was focused on facilitating the technical aspects of equipment operation. Computer control of the MCC also reduced human interaction in start up and shut down procedures, enforced safety interlocks and protected various motor drives from improper operation. Therefore, from an accident intervention point of view, existing productive levels are less likely to be affected by human- generated equipment failure.

    The changes made to the packing line increased productivity, not only due to increased speeds, but also from improved quality. The quality of the packing process has improved because of the quicker response to process conditions and the ability to regulate the packing line speeds more accurately. The grade quality can be more closely controlled, allowing higher grade packouts than before computer integration improvements. The growers also receive increased profits as a direct result of better quality control raising the value of the tomato packouts.

    Some of the best benefits from installing this relatively inexpensive computer integrated manufacturing system may be the benefits the workers have received. In this processing environment, the number of workers has remained unchanged, but the wages of these workers have increased significantly.

    The base pay for these workers was between $4 and $5 per hour prior to the modifications. As a direct result of the packinghouse productivity increases, workers can now earn at least $0.50 per hour more than prior to computer integration. Workers are offered $0.25 per hour more when hired and then workers can earn a $0.25 per hour bonus if they remain at the packinghouse for the total packing season. A typical worker can earn at least $20.00 per week more than before the PLC's were installed. Additionally, workers are given an increased hourly rate if they return the
next season.

    The wage increase is usually greater than $20.00 per week because the work week is normally greater than 40 hours per week, (Table 1). When growers' tomatoes arrive at the packinghouse, they must be packed that day. Therefore, the packing continues until all tomatoes received that day are packed and placed in cold storage. Since the computer integration implementation phases, the work day has decreased because of the increased packout rate. The result is that workers are now working fewer hours and earning higher wages.

    Working relations between workers and management have also improved, not just because of increased wages. The ability to alter conveyor speeds from the central location, and because the changes in speed-are small, the workers, often times, do not realize that the speed has increased and do not get the feeling management is trying to make them work harder. Previously, the workers would resent the increase in speed and deliberately allow undesirable tomatoes to pass through the inspection station.

Problems created by computer control

    Because of the ability to control the entire packing line from a central location, the system operates at much higher speeds. This has created new problems, basically mechanical in nature, and beginning at carton filling. The new problems created by the increased operating speed are: accuracy in control of carton weight, spillage of tomatoes from filled cartons prior to ridding, potential for increased bruising, and lack of an economic technology to automatically identify filled cartons. These problems are currently being addressed.


Effects of speed on impacts at key transfer points

    Peak impacts were determined at four transfer points on the packing line using an Instrumented Sphere impact recording device (Techmark, Inc., Lansing, Michigan). The transfer points analyzed were the drop from the elevator rolls to the trash eliminator belt, the drop to the sort rolls, the roll to the dryer brush bed and the roll to the perforated size belts. Average impacts were determined while the packing line was operating at four speeds, 80 and 100 bins/hr (analyzed Dec. 3, 1990), and 130 and 200 bins/hour (analyzed May 17, 1991) (Table 2). Sampling methodology has been previously reported (Sargent, et al., 1992a).

    Overall, the impacts for these transfer points were similar to impacts previously determined for fixed speed tomato packing lines, and in some cases substantially lower (Sargent, et al., 1992a). However, as the dump rate increased to 130 and 200 bins/hour, there was a decrease in impact intensity for two of the transfer points. The impact mean for the roll to the dryer brush bed decreased by 32% at 130 bins/hour and by 54% at 200 bins/hour. Also, fewer impacts were recorded at these dump rates, indicating a gentler transfer. The impact mean for the roll to the size belts was quite high at 80 and 100 bins/hr (125.3 G and 138.3 G. respectively), but decreased by 21% at 130 bins/hour and by 63% at 200 bins/hour. This confirmed visual observations that, at the two faster dump rates, the percent packing line coverage was higher than at the slower dump rates. The higher percent coverage of the packing line caused gentler transfers at these two points by reducing the roll distance. The drop to the trash eliminator remained fairly constant, ranging from 54.5 G to 76.4 G (Table 2); this transfer involved a direct drop of each tomato onto the eliminator belt (no fruit-to-fruit contact). The transfer to the sort rolls was fairly high, possibly due to the rolls moving too fast to permit backup of the fruits onto the previous transfer plate.

    Impacts at inclined transfer plates can be eased by slightly reducing the speed of the downstream component in relation to the speed of the proceeding component. This would cause the tomatoes to back up somewhat on the transfer plate. In these tests we determined that the speeds of the dryer brushes and the sizing belts were better correlated with the dump rates of 130 and 200 bins/hour than the speeds at the lower dump rates. However, we did not have the opportunity to adjust the individual speeds of adjoining components to optimize the transfer points at the various dump rates.

Effects of speed on subsequent tomato quality

    Since approximately 90% of the tomatoes are harvested at the green stage, green tomatoes were sampled from two locations during packing, the receiving tank and the grading rolls (after final grading but prior to transfer to the size belts). Samples of 96 tomatoes were taken at each location (on the dates noted previously) for each of the four dump rates. The fruit were transported to the laboratory without mechanical injury, stored at 20C and treated with ethylene. Upon reaching the firm, red-ripe stage, each tomato was rated for internal bruising (IB), external bruising and cuts/punctures. IB is a physiological disorder caused by handling impacts (MacLeod and Kader, 1976). For tomatoes sampled at 80 bins/hour, 22% had IB after final grading (cultivar unknown), while those sampled at 100 bins/hour had 14% IB (cv. Heatwave) (data not shown). Tomatoes (cv. Sunny) sampled at 130 and 200 bins/hour had negligible incidence of IB, however this cultivar has been shown to be more resistant to IB than some other cultivars (Sargent, et al., 1992b).

    External bruises and cuts/punctures were rated for the 'Sunny' tomatoes sampled on May 17, 1991. There was a slight increase in external bruises and cuts/punctures from 130 to 200 bins/hour. For tomatoes sampled after grading at 130 bins/hour, 2.1% had at least one bruise greater than 12.7 mm (0.5 in) in diameter; 4.2% of those sampled at 200 bins/hour were bruised (Table 3). Tomatoes with cuts and/or punctures greater than 12.7 mm (0.5 in) accounted for 5.2% and 11.5% of all tomatoes for 130 and 200 bins/hour, respectively. These values were not sufficiently high to affect grade and were significantly lower than those for tomatoes sampled in the float tank, indicating the importance of the grading operation in removing out-of grade tomatoes. The differences in mechanical injuries between the two dump rates may be attributed to impacts on the packing line which were not measured with the Instrumented Sphere, such as in brush rolls.

    The modifications made so far to this packing line have improved productivity, and resulted in packouts with higher tomato quality. Packout quality has improved because the packing line speed can be quickly adjusted based on the quality of the tomatoes being packed. Continued refinements will be made to the system this fall, most importantly, optimizing the speeds of the individual packing line components across the range of packing line speeds so as to maximize percent coverage. By correlating percent coverage of the packing line with efficient grading, impacts at transfer points should also be lower.  


    There are many produce packing environments that could take advantage of computer automation and control techniques. Local packinghouse management may need assistance to plan and implement computer technology for three reasons: lack of computer knowledge from a controls prospective, concern about trying an unfamiliar approach, and other management characteristics required for this type of operation. Making the commitment to new technology and obtaining the correct system will generally be difficult for these companies without outside assistance. The initial assistance may have to come from a third- party, as was the approach described in this paper.


The authors express appreciation to Mr. Frank Diehl, Tomatoes of Ruskin, Inc., Mr. Chuck Fouts, F&L Electric, and the State of Florida High Technology and Industrial Council for support in this project. The assistance of Ms. Judith Zoellner in preparation of data and graphic design is gratefully acknowledged.


Foster, R.N., 1982. A Call for Vision in Managing Technology. Business Week, May 24.

Hayes, R.H., and Wheelwright, S.C. 1984. Restoring our Competitive Edge. Wiley & Sons, New York,

Matjaz, M.1987. Conditions for the Technological Catching Up, Innovation: Adaptation And Growth. Rothwell, R. and J. Bessant (Eds.). Elsevier, Amsterdam.

MacLeod, RF., A.A. Kader and L.L. Morris. 1976. Stimulation of ethylene and CO2 production of mature-green tomatoes by impact bruising. HortScience 11(6):604-606.

Sargent, S.A. J.K Brecht and J.J. Zoellner. 1992a. Instrumented sphere impact analyses of tomato and bell pepper packing lines. Appl. Eng. Agric. 8(1):76-83.

Sargent, S.A., J.K Brecht and J.J. Zoellner. 1992b. Sensitivity of tomatoes at mature green and breaker ripeness stages to internal bruising. J. American Society for Horticultural Science. 117(1):119-123.

Wardowski, W.F., W.M. Miller and W. Grierson. 1978. Packingline machinery for Florida citrus packinghouses. Bull 239. Cooperative Extension Service, Institute of Food and Agricultural Sciences. University of Florida, Gainesville. 33 pp.


Gilbert, R.A. and Llewellyn, J.A.., "Programmable Controllers Practices and Concepts", ITC, Herndon Va, 1985

Hauck, W.C. and Dingus, V.K, Achieving High Commitment Work Systems, Industrial Engineering and Management Press, Atlanta, 1990.

Mantel, S.J. and Rosegger, G., "The Role of Third-Parties in the Diffusion of Innovations: a Survey," Innovation: Adaptation And Growth, Edited by Rothwell, R. and Bessant J., Elsevier, Amsterdam, 1987.