XY-4巖心鉆機升降機的設計 (2)帶CAD圖,XY-4巖心鉆機升降機的設計,(2)帶CAD圖,XY,巖心,鉆機,升降機,設計,CAD Journal of Stored Products Research () Stored Grain Advisor Pro: Decision support system for insectmanagement in commercial grain elevators$P.W. Flinna,?, D.W. Hagstruma,1, C.R. Reedb, T.W. PhillipscaUSDA-ARS Grain Marketing and Production Research Center, Manhattan, KS, USAbDepartment of Grain Science and Industry, Kansas State University, KS, USAcDepartment of Entomology, Oklahoma State University, Stillwater, OK, USAAccepted 20 September 2006AbstractA decision support system, Stored Grain Advisor Pro (SGA Pro) was developed to provide insect pest management information forwheat stored at commercial elevators. The program uses a model to predict future risk based on current insect density, grain temperatureand moisture. A rule-based system was used to provide advice and recommendations to grain managers. The software was tested in aresearch program conducted at commercial grain elevators in Kansas and Oklahoma, USA. A vacuum-probe sampler was used to taketen 3-kg grain samples in the top 12m of each bin that contained wheat. After the insect species and numbers were determined for eachsample, the data were entered into SGA Pro. A risk analysis and treatment recommendation report for all bins was presented to the grainmanagers every 6 weeks. SGA Pro correctly predicted for 7180% of bins whether the grain was safe or at high risk of dense infestationand grain damage. SGA Pro failed to predict unsafe insect densities in only 2 out of 399 Kansas bins (0.5%) and in none of 114 bins inOklahoma. Grain managers who followed SGA Pros recommendations tended to fumigate only the bins with high insect densitiesinstead of fumigating all bins at their facility. This resulted in more efficient insect pest management because fumigating bins only wheninsect densities exceeded economic thresholds and treating only the bins that required fumigation minimized the risk of economic lossesfrom insects, reduced the cost of pest management, and reduced the use of grain fumigant.Published by Elsevier Ltd.Keywords: Rhyzopertha dominica; Cryptolestes ferrugineus; Decision support system; Model; Integrated pest management; Stored grain; Area-wide1. IntroductionMost cereal grain produced in the USA is stored incommercial facilities known locally as grain elevators.Major insect pests of stored wheat in the USA includeRhyzopertha dominica (F.), Sitophilus oryzae (L.), Crypto-lestes ferrugineus (Stephens), Tribolium castaneum (Herbst),and Oryzaephilus surinamensis (L.). The first two speciescause the most grain damage because the immature stagesdevelop inside the grain kernels. These internal feedinginsects are a major cause of insect contamination in wheatflour because the immature stages and pre-emergent adultscannot be completely removed from the wheat before it ismilled. Grain managers and regulators use the number ofinsect-damaged kernels (IDK) in wheat as an indirectmeasure of the density of internally-infested kernels. Ifmore than 32 IDK are found per 100g of wheat, the grainis classified as sample grade and unfit for humanconsumption (Hagstrum and Subramanyam, 2006). Atmost domestic flour mills, the wheat purchasing specifica-tions include a maximum IDK count of either 3 or 5/100g.Cryptolestes ferrugineus is a very common insect pest thatoften reaches high densities near the grain surface. Younglarvae of this species frequently feed on the germ of wholekernels and on fine material in the grain (Rilett, 1949).They leave the germ before becoming adults and do notcause IDK. Nevertheless, grain infested with this species islikely to receive a lower price than uninfested grain.ARTICLE IN PRESS front matter Published by Elsevier Ltd.doi:10.1016/j.jspr.2006.09.004$This paper reports the results of research only. Mention of aproprietary product or trade name does not constitute a recommendationor endorsement by the US Department of Agriculture.?Corresponding author. Tel.: +17857762707; fax: +17855375584.E-mail address: paul.flinngmprc.ksu.edu (P.W. Flinn).1Retired.Please cite this article as: Flinn, P., et al., Stored Grain Advisor Pro: Decision support system for insect management in commercial grain elevators.Journal of Stored Products Research (2007), doi:10.1016/j.jspr.2006.09.004Typically,controlofstored-graininsectsingrainelevatorsintheUSAincludesmonitoringofgraintemperature and calendar-based fumigations using phos-phine fumigant (Hagstrum et al., 1999). This approachoften fails to distinguish between bins with high and lowinsect densities and does not optimize the timing of thefumigation treatment. Therefore, grain may be unnecessa-rily fumigated, or the fumigation may not be timed toprevent high insect populations and grain damage fromoccurring. Although careful monitoring of grain tempera-ture often alerts the manager to potential mold and insectproblems (Reed, 2006), large populations of insects orsevere mold problems can develop before a temperatureincrease is noted.Incontrasttotraditionalinsectcontrolpracticescurrently used for most stored grain, the integrated pestmanagement (IPM) approach uses sampling to determine ifinsects have exceeded an economic threshold (Hagstrumand Flinn, 1992). Adapting IPM principles to insectcontrol in a grain elevator is complicated by the structureand operation of the facility. A large elevator may haveover 100 bins, and the bins may contain different types ofgrain, stored for different durations. The grain temperatureand moisture often vary greatly between bins, which affectsthe rate at which insects and molds grow and damage thegrain.To facilitate the development and implementation ofIPM practices in stored grain in the USA, the USDAsAgricultural Research Service recently funded a 5-yeardemonstration project for area-wide IPM for stored wheatin Kansas and Oklahoma (Flinn et al., 2003). This projectwas undertaken by a collaboration of researchers at theAgricultural Research Service (Manhattan, Kansas), Kan-sas State University (Manhattan, Kansas), and OklahomaState University (Stillwater, Oklahoma). We used twoelevator networks, one in each state, for a total of 28 grainelevators. One of the project goals was the development ofa decision support system for insect pest management forgrain stored in commercial elevators.Avalidatedinsectpopulationgrowthmodelwaspreviously developed for R. dominica in concrete elevatorstorage (Flinn et al., 2004). This model was used in adecision support system to make management recommen-dations based on current insect density, grain temperatureand grain moisture. A decision support system (StoredGrain Advisor) was developed previously for farm-storedgrain in the USA (Flinn and Hagstrum, 1990b). However,that software was not suitable for large commercialelevators because the grain sampling methods and recom-mendations were specific for farm-stored grain.Decision support systems for stored grain have beendeveloped in several countries. In Canada, CanStore, wasdeveloped to assist farmers in stored grain management(www.res2.agr.ca/winnipeg/storage/pages/cnstr_e.htm). InAustralia, Pestman ranks insect pest management recom-mendations by their cost and provides a graphical site planthat allow a manager to quickly find information aboutany bin (Longstaff, 1997). In the UK, Integrated GrainStorage Manager (Knight et al., 1999) is a new version ofGrain Pest Advisor (Wilkin and Mumford, 1994) that wasdeveloped with input from farmers and storekeepers tobetter suit their needs. Grain Management Expert System(Zonglin et al., 1999), was developed from Pestman for usein China. QualiGrain is an expert system for maintainingthe quality of stored malting barley (Ndiaye et al., 2003;Knight and Wilkin, 2004).None of the previously mentioned systems fit therequirements of the USA commercial grain storage system.We needed a management program that was based onintensive grain sampling for insect pests in each elevatorbin (at least ten 3-kg grain samples per bin to a depth of12m). In addition, the system needed to be able to predictinsect population growth for up to 3 months, based oncurrent insect density in the bin, grain temperature, andgrain moisture. In this paper, we describe the validation ofa decision support system that was developed as part of anarea-wide IPM demonstration project. The decision sup-port system uses current and predicted insect densityestimates to provide grain managers with an overall riskanalysis for the grain at their facility and recommendedtreatment options.2. Materials and methods2.1. Grain samplingAn area-wide IPM program for grain elevators wasstarted in 1998. Investigators collected data from twoelevator networks in south-central Kansas and centralOklahoma. Each network consisted of at least 10 ruralelevators and at least one terminal elevator. The ruralelevators typically receive grain from farmers and store itfor a shorter period of time compared to the terminalelevators, where most grain is received from rural elevators.Storage bins at these elevators were either upright concretebins, typically 69m in diameter and 3035m tall, or metalbins that are shorter and wider. Maize and other grainswere stored in the project elevators, but only the wheat wassampled during this project.Various sampling methods to estimate insect density inupright concrete grain bins were tested: probe traps placedat the grain surface, samples taken as the grain was movedon transport belts, and samples taken from grain dis-charged from the bins onto a stationary transport belt.Samples taken with a vacuum probe as the grain was at restin the storage bins provided the best estimate of insectdensity. Data collected with the vacuum probe were highlycorrelated with grain samples taken as the bin wasunloaded (r2 0.79, N 16, P 0.001). In addition,unlike the other sampling methods, the power probeallowed the grain to be sampled at any time, and itprovided a vertical profile of the insect distribution for eachbin. We used a Port-A-Probe (Grain Value Systems,Shawnee Mission, Kansas), which consists of a vacuumARTICLE IN PRESSP.W. Flinn et al. / Journal of Stored Products Research () 2Please cite this article as: Flinn, P., et al., Stored Grain Advisor Pro: Decision support system for insect management in commercial grain elevators.Journal of Stored Products Research (2007), doi:10.1016/j.jspr.2006.09.004pump powered by a 5.3KW gasoline engine connected byflexible plastic tubing to sections of rigid aluminum tubes1.2m long by 3.5cm wide. The probe was insertedvertically into the grain and a 3.9l (about 3kg) sample ofwheat was taken during every 1.2m transect of grain to adepth of 12m. In the concrete upright bins, the grain wassampled through the entry port. In metal bins, the probewas inserted at 35 locations across the surface. Grainsamples were extracted from the grain mass by suction andcollected in a cyclone funnel.Samples were processed twice over an inclined sieve(89?43cm, 1.6mm aperture) (Hagstrum, 1989) to sepa-rate insects from the wheat. Material that passed throughthe screen was collected on a pan below the screen, whichthen slid into a funnel at the bottom of the pan. A re-sealable plastic bag was attached to the funnel to collect thematerial that was separated from the grain sample. Ahopper above the screen held the grain sample and a funnelat the base of the screen directed material passing over thescreen into a plastic bucket. The sieve was inclined 241from horizontal and the opening of the hopper wasadjusted such that the sample passed over the screen inabout 25s. Each sample was passed over the sieve twotimes. Validation data for SGA Pro were selected from binsthat were sampled at least twice, starting in autumn, inwhich the wheat was not moved or fumigated. In a typicalbin (69m wide and 3035m tall), the sampling rate forvacuumprobesampleswas0.070.13kg/tofgrainsampled. In most cases, only the grain in the top 12m ofthe bin was sampled.2.2. Decision support softwareThe Stored Grain Advisor Pro (SGA Pro) softwarewas initially developed using Microsoft Access. Theprogramwasthenmodifiedandre-writtenusingVisual Basic 6.0. We designed a graphical user interfacethat provides a bin diagram for each elevator location(Fig.1).Datawereenteredusingthreedata-entryforms: insects, grain quality, and grain temperature. Dataentered in the insect form were: sample type (bottom,movingsample,probetrap,orvacuumsample)and the number of insects found in each sample for fiveprimary stored-grain insects (Cryptolestes spp., R. domin-ica, O. surinamensis, Sitophilus spp., and Tribolium spp.)(Fig. 2). Data entered for grain quality were: grade, %dockage,testweight,moisture,foreignmaterial,%shrunken or broken kernels, insect damaged kernels, %protein(Fig.3).Graintemperaturedatawerealsoentered into the database for each bin (data entry is similarto the grain quality form and is not shown here). Mostelevator bins were equipped with one or more cablescontaining up to 20 thermocouple-type sensors per cable.In bins not equipped with temperature sensors, investiga-tors inserted temporary probes to collect information ongrain temperature.The SGA Pro system will recommend either fumigation,aeration, or waiting untilthe next sampling periodbased on current insect density in the bin, grain tempera-ture, aeration capability, time of year, and predictedinsect density in 1, 2, and 3 months. For example,ARTICLE IN PRESSFig. 1. Elevator bin diagram from Stored Grain Advisor Pro; on the computer screen, bins of grain at high, moderate, and low risk for insect problems areshown in red, blue and green, respectively. In this figure, bin numbers that are light grey are at low risk, bins 615, 620 and 621 are at high risk, and the restare at moderate risk. Bin 620 is currently selected (using the mouse), and the information for this bin is shown in the bottom half of the screen.P.W. Flinn et al. / Journal of Stored Products Research () 3Please cite this article as: Flinn, P., et al., Stored Grain Advisor Pro: Decision support system for insect management in commercial grain elevators.Journal of Stored Products Research (2007), doi:10.1016/j.jspr.2006.09.004for bin 620 (Fig. 4), the program indicated that thecurrent insect density was 2.5kg-1and predicted a densityof 14.6insects/kg in 1 month. Twenty-eight percent of theinsects from the samples were species that caused IDK, soSGA Pro recommended fumigation followed by aeration tocool the grain.ARTICLE IN PRESSFig. 2. Insect data entry form for Stored Grain Advisor Pro. The number of insects for each 3-kg sample were entered into the form. For simplicity,common names are used for the insect species (flat Cryptolestes spp., lesser R. dominica, sawtooth O. surinamensis, weevil Sitophilus spp., flourbeetle Tribolium spp.).Fig. 3. Grain quality data entry form for Stored Grain Advisor Pro (grade grain grade, DKG(%) % dockage, TW test weight, moist(%) %grain moisture, DK(%) % damaged kernels, FM(%) % foreign material, SHBK(%) % shrunken or broken kernels, IDK number of insectdamaged kernels per 100g, protein(%) % protein).P.W. Flinn et al. / Journal of Stored Products Research () 4Please cite this article as: Flinn, P., et al., Stored Grain Advisor Pro: Decision support system for insect management in commercial grain elevators.Journal of Stored Products Research (2007), doi:10.1016/j.jspr.2006.09.004An equation was developed to predict insect populationgrowth based on current insect density, grain temperatureand moisture by running simulations for a model of R.dominica (Flinn and Hagstrum, 1990a), over temperaturesfrom 21.5 to 33.51C and moistures from 9.5 to 13.5%.Tablecurve 3D version 3 (SPSS, 1997) was used to fit anequation to the model-generated data, where Z is the rateof increase over 30 days, X is temperature, and Y ismoisture:Z 9:2004 ? 1:6898X 0:07872? 0:0011X2 0:1841Y=1 ? 0:0197X ? 0:0161Y.1:0This equation fitted the data well (R2 0.98, N 25).Because we needed to predict only 13 months ahead, thisequation was adequate for quickly estimating futurepopulations for many bins present in the database (oftenmore than 100). Although C. ferrugineus was often themost numerous species during the first month of storage,we based Eq. (1.0) on R. dominica because it is the moredamaging species, it was more common than C. ferrugineuslater in the season, and the predicted rates of increase forboth species were fairly similar (Hagstrum and Flinn,1990). We did not use a model for S. oryzae because thisspecies was found in about 1% of the wheat samples,whereas, R. dominica was found in approximately 60% ofthe samples.SGA Pro used a rule-based algorithm to determinewhether bins were safe, moderate, or at high risk of havinginsect densities that exceed certain thresholds, based on thecurrent and predicted insect density, grain temperature,and grain moisture. Insect economic thresholds can beadjusted by the user (Fig. 5). In addition, alerts can also beset for: high grain moisture, high thermocouple readings,and high numbers of internally feeding insects in anindividual sample.SGA Pro was tested during the final 2 years of the area-wide IPM study. Bins at each elevator were sampled atapproximately 6-week intervals, data were entered intoSGA Pro, and the report recommendations were shown totheelevatormanagers.SGAProwasvalidatedbycomparing predicted insect densities and control recom-mendations with actual insect densities in the same bins 6weeks later. Validation data came from bins in which thegrain had not been turned or fumigated for at least twosampling periods.3. ResultsIn the Kansas data set from 2002, SGA Pro correctlypredicted that bins were safe or at high risk in 285out of 399 cases (Table 1). Forty-seven of the 399 binsrequired fumigation. SGA Pro failed to predict unsafeinsect densities in only two bins (0.5%), and the insects inthese isolated instances were mostly near the surface,suggestingrecentimmigration.Thesimplegrowthmodel used by SGA Pro tended to overestimate insectdensities in bins that were at medium risk (112 out of 399bins). All of the bins that the software predicted to be athigh risk contained insect densities greater than thethreshold at the next sampling period. In Oklahoma,SGA Pro correctly predicted bins that were safe or athigh risk in 107 out of 133 total bins. Forty-five of the133 bins needed to be fumigated. All of the bins that theprogram determined as being safe turned out to haveinsect densities below the threshold of 2insects/kg 6 weekslater. As in Kansas, SGA Pro tended to overestimateinsect densities in bins that were at medium risk (26 out of131 bins).ARTICLE IN PRESSFig. 4. Stored Grain Advisor recommendation report. The report shows alerts for five elevator bins (Current average insects per kg grain, 1Mon predicted insect density in 1 month, 2 Mon predicted insect density in 2 months, IDK insect % of the insects in the sample that cause IDK,Max SS internal highest number of internal insects in any single sample, and management option recommended actions for the elevator manager).P.W. Flinn et al. / Journal of Stored Products Research () 5Please cite this article as: Flinn, P., et al., Stored Grain Advisor Pro: Decision support system for insect management in commercial grain elevators.Journal of Stored Products Research (2007), doi:10.1016/j.jspr.2006.09.0044. DiscussionCompared to other decision support systems for largecommercial grain stores, SGA Pro is the first that usesintensive sampling of the grain to determine if insectdensities exceed economic thresholds. In countries wherethis type of software has been developed, for example UK,Australia, and Canada, the grain trade operates on a zerotolerance for insects in stored grain. This makes it difficultto implement economic thresholds for insect pests. SGAPro is also the first decision support system for commercialgrain storage that has been field validated, certainly to theextent presented here. Integrated Grain Storage Manager(Knight et al., 1999) was revised with input from storekeepers(atypeofv