CONTROLLING AIRBLAST SPRAYER AIR FOR VARIABLE RATE APPLICATION IN ORCHARDS

Transcription

1 CONTROLLING AIRBLAST SPRAYER AIR FOR VARIABLE RATE APPLICATION IN ORCHARDS By NARESH PAI A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2007

2 2007 Nresh Pi

3 To Archn nd Annt Pi

4 ACKNOWLEDGMENTS I would like to extend my thnks to the mny individuls who hve contributed to mke this project success nd my eductionl experience so enjoyble. I would like to thnk Dr. Msoud Slyni, my dvisory committee chirmn, for his fith in me, nd for his continued support nd inspirtion. His constnt encourgement, timely criticl evlution, nd enthusism for my work hve resulted in the successful completion of my reserch. I m indebted to Mr. Roy Sweeb, Senior Engineering Technicin, for his insightful ides, hnds-on support, nd trining in the workshop nd field, throughout the work. I would lso like to thnk Dr. Thoms Burks nd Dr. John Schueller for giving me vluble knowledge through their courses nd serving on my supervisory committee. I wnt to cknowledge the Agriculturl nd Biologicl Engineering (ABE) deprtment for providing me the opportunity, nd the Citrus Reserch nd Eduction Centre (CREC) for the ssistntship nd technicl resources to conduct my reserch. I m lso grteful to Dr. Rez Ehsni, Mr. Troy Giney, nd the stff of the CREC mintennce deprtment for letting me use equipment needed for the project. On personl note, I would like to thnk my prents nd brother whose support ws of inestimble vlue. A finl word of thnks goes out to ll my friends who hve directly or indirectly contributed to the successful completion of my work. iv

5 TABLE OF CONTENTS pge ACKNOWLEDGMENTS... iv LIST OF TABLES... vii LIST OF FIGURES... viii ABSTRACT...x CHAPTER 1 INTRODUCTION Justifiction Thesis Orgniztion BACKGROUND Pesticide Usge in Florid Pesticide Appliction Technology for Tree Crops Spryers Air-Crrier Spryers Liquid delivery system Air delivery system Testing Methodologies for Spryer Air nd Liquid Output Control Systems in Pesticide Appliction Technology Objectives MATERIALS AND METHODS Airblst Spryer Description Lser Sensor Preliminry Experiments Airblst Spry Distribution Pttern Restricted Air-Input Test Deflecting Air t Output Automtion of Deflector Plte Movement Rel-time Collection of Tree Prmeters Using Lser Sensor Experiment 1: Air Penetrtion through Tree Cnopy...29 v

6 3.7 Experiment 2: Effect of Deflector Plte on Sptil Distribution of Spry Deposition RESULTS AND DISCUSSION Evlution of the Electromechnicl Control System Experiment I: Air Velocity Mesurements Experiment II: Spry Deposition Discussion CONCLUSIONS...48 APPENDIX A B STEP MOTOR SIZING CALCULATION...49 COMPONENT SPECIFICATION...53 LIST OF REFERENCES...54 BIOGRAPHICAL SKETCH...58 vi

7 LIST OF TABLES Tble pge 3.1 Chrcteristic of the horizontlly deflected ir due to vrious plte positions t spryer outlet Input output reltions of the control system...26 A.1 Specifictions of the ctution mechnism...49 B.1 Bll screw nd nut ssembly specifictions...53 B.2 Step motor technicl specifictions...53 B.3 Step motor controller fetures...53 vii

8 LIST OF FIGURES Figure pge 3.1 Schemtic of the PowerBlst irblst spryer Schemtic view of sprying ppliction Airblst spryer with different ir intke re Effect of fn inlet dimeter on the ir output of irblst spryer Schemtic view of the deflector plte motion Effect of deflector plte loction on ir output Components of the control system The step motor controller, mstep Schemtic of ctution mechnism for the deflector plte Reltionship between indexing vlue to the controller bord required for rnge of lser sensor density reding Tree cnopies of different densities Experimentl setup for mesuring ir velocity Schemtic view of spry ppliction experiment nd smpling lyout Reltion between ctul plte position nd tree density Men ir velocity due to different deflector plte loction t 2.15 nd 4.73 km/h Mximum ir velocity due to different deflector plte loction t 2.15 nd 4.73 km/h Mximum ir velocity t two ground speeds of the irblst spryer Effect of two ppliction volume rte on totl deposition t two sptil sections...45 viii

9 4.6 Low ppliction rte: sptil distribution of deposition t different deflector plte position High ppliction rte: sptil distribution of deposition t different deflector plte positions...47 A.1 Block digrm used for stepper motor clcultions...49 ix

10 Abstrct of Thesis Presented to the Grdute School of the University of Florid in Prtil Fulfillment of the Requirements for the Degree of Mster of Science CONTROLLING AIRBLAST SPRAYER AIR FOR VARIABLE RATE APPLICATION IN ORCHARDS By Nresh Pi My 2007 Chir: Msoud Slyni Mjor: Agriculturl nd Biologicl Engineering Spry requirements vry considerbly throughout the grove on ccount of vribility in citrus cnopy size nd folige density. Configuring spryers to suit this tree vribility is vitl for efficient sprying. Currently, crops re spryed uniformly throughout the field bsed on experience. Uniform ppliction of grochemicls not only wstes chemicls but lso hs environmentl implictions. At present, irblst spryers ccount for mjority of spryers for tree crop ppliction in Florid. While moving cross the grove, these spryers rely on strem of ir supply generted by fn(s) to crry the mteril from the nozzles to the cnopy. The ir volume generted in these spryers rnge between m 3 /s. A fully grown tree hs different spry requirement s compred to smll or medium sized tree. It hs lso been observed tht typicl irblst spryer my deposit 2-3 rows beyond the immedite row for which it ws intended. Since the tree size nd density distribution on the field re x

11 usully non-uniform, uniform sprying could result in substntil mteril losses s ground fllout nd drift. My work describes design modifiction of n existing irblst spryer to test the ide of vrible rte spry ppliction by djusting the ir output. My project involved the design, implementtion nd testing of n electromechnicl system to chnge the volume of ir going to the trees bsed on the tree density informtion from lser sensor. This process gives rel time chnge in ir output chrcteristics s the spryer moves cross row of trees. By using this system, the ir volume cn be chnged from 1.9 to 7.6 m 3 /s in less thn 3 seconds. Air penetrtion ws quntified by mesuring ir velocity which reveled tht different settings of this ir regultory system cn produce significntly different ir chrcteristics cross ded nd medium density tree. Overll, increse in ir volume gve higher sptil distribution of spry mteril but for high volume ppliction chnging ir volume cn produce significnt high deposits in ner loctions when compred to fr loctions thus reducing off-trget wstge. xi

12 CHAPTER 1 INTRODUCTION 1.1 Justifiction Florid is the second lrgest producer of citrus in the world nd ccounts for more thn one-third of the world s grpefruit production. These trees re mostly spryed with ir-crrier spryers. These spryers use high volume nd velocity ir strems, produced by xil-, centrifugl-, or cross-flow fns, to trnsport the spry droplets to tree cnopies. The im is to replce the ir within the trees by strem of ir nd grochemicl droplets. Spry requirements of fully grown mture tree is substntilly different from those of medium or smll sized tree. In ddition, it is very common to find ded trees nd resets in the field. Efficient sprying in griculture involves optimum usge of the vilble resources. The idel sprying deposits the mteril on the intended trget, bsed on the type of tree nd minimizes ground deposit nd drift. In contrst, t full ir cpcity the spryers could deposit the mteril 2-3 rows beyond the row for which it ws intended. This results in considerble mount of wstge of grochemicls which hs economicl s well s environmentl concerns. Mny reserchers nd compnies hve successfully shown the dvntge of customizing the liquid output from nozzles bsed on the chrcteristics of the trees. However, there hve been fewer ttempts to modify the ir tht crries the spry droplets to the trees. The overll objective of this project is to evlute the ide of vrible rte spry ppliction by chnging the ir volume tht is used to trnsport these spry droplets. 1

13 2 1.2 Thesis Orgniztion My thesis is divided into five chpters. The second chpter, Bckground, dels with relevnt literture review nd pst work done in this re. The third chpter, Mterils nd Methods, describes the progressive steps tken to develop system to ddress the problem described in section 1.1. Subsequently, it lso describes the experiments conducted to test this system. The fourth chpter, Results nd Discussion, evlutes the control system, nd further, discusses the results tht were obtined from the experiments described in chpter three. The fifth chpter, Conclusions, gives n overll perspective of this reserch project.

14 CHAPTER 2 BACKGROUND This chpter gives the bckground nd lys foundtion for this project. It strts with sttistics of the griculturl pesticide usge in Florid to recognize the significnce of developing technology which cn mximize ppliction efficiency nd minimize wstge. Pesticide ppliction technology is then discussed to give generl overview of equipment used for tree crops. Focus hs been plced on irblst spryers s the im of this project ws to optimize ir output in such spryers. Finlly, control systems in irblst spryers used for vrible rte ppliction re discussed. 2.1 Pesticide Usge in Florid Pesticides re grochemicls tht re intended for controlling pests. Pesticides in the form of sprys re commonly used to control pests tht ffect citrus. In 1999, Florid Agriculturl Sttistics Services reported tht herbicides were used on 96.3% of the 316,840 h of citrus nd vocdo. Similrly, insecticides nd fungicides were pplied on 91.3 nd 81.9% of the crege respectively in the sme yer. This shows the widespred use of pesticides for tree crops. Though the intention is to spry the mteril to the trget, considerble mount of pesticide is wsted. These wstges rise primrily from three sources: Off-trget sprying Drift to ir Runoff from leves flling on the ground It ws reported tht bout 22% of the spry mteril is wsted to the ground nd, bout 21% drifted to ir with commonly used irblst spryer t ppliction rte 3

15 4 of 950 L/h (Miller et l., 2003). This not only hs economic disdvntges, but lso leds to ground nd surfce wter pollution, ir pollution, disturbnce in the neighboring ecosystem nd my lso present helth concern to humns. Hence, efforts re being mde to optimize the mount of pesticide usge in griculture. 2.2 Pesticide Appliction Technology for Tree Crops Pesticide ppliction technology refers to the equipments tht re used to dispense pesticides. Pesticide pplictors for tree crops re chrcterized by their high volume ppliction to cover the dense folige of trees. Pesticides re pplied in solid or liquid formultion with or without ssistnce of ir. Pesticides in solid formultion re pplied either in grnulr or powdered form. These pplictors re clled grnulr pesticide pplictors or dusters, respectively. Pesticides tht re used in liquid formultion re pplied using spryers. Commonly, spryers use ssistnce of ir to crry the spry droplets, nd for better cnopy penetrtion. These spryers re designted s ir-crrier spryers Spryers Spryers use hydrulic systems to trnsport nd spry pesticides to the trget. These hydrulic systems comprise tnk to store the liquid formultion, pump to develop the necessry pressure, mnifolds to trnsport the pressurized liquid nd nozzles to convert the liquid into droplets nd disseminte the pesticide. These spryers hve boom to support the nozzles t the outlet Air-Crrier Spryers Air-crrier spryers ccount for 89% of spry mchines for citrus production in Florid (Stover et l., 2002) nd hence, re importnt equipment for pesticide ppliction. These spryers re commonly used for sprying tree crops hving lrge

16 5 folige. They use strem of ir to trnsport spry droplets towrds the tree nd penetrte the cnopy. The ir in nd round the trees is thus replced with mixture of ir nd pesticide droplets. This method of ppliction is considered superior s it increses the deposition (Reichrd, 1977; Slyni, 1988). These spryers re vilble in different shpes, sizes, fn types, nozzles nd re operted with different volume rtes nd ground speeds. The working of n ir-crrier spryer system cn be brodly divided into two modules. The first one comprises of the components tht hndle the delivery of liquid nd is referred to s the liquid delivery system. The second module reltes to the delivery of ir nd is clled ir delivery system Liquid delivery system The tsk of liquid delivery system in n ir-crrier spryer is to store, pressurize nd produce droplets of the liquid pesticide formultion. This is chieved by hydrulic circuit which employs pump to pressurize nd force the liquid from the tnk to the nozzles. The nozzles ligned on the spryer cn be ctegorized to be in three sections nmely; top, middle nd bottom, reflecting the section of the tree tht the spry mteril ffects. These nozzles produce droplets of liquid pesticide which re trnsported to the trget due to the functioning of ir delivery system. The number of nozzles, nozzle size, nozzle pressure, nd ground speed of spryer ply n importnt role in the deposition efficiency of spryer. These hve to be djusted to mximize the efficiency of sprying. For instnce, it ws suggested by Slyni (2000) tht for better deposition t lower ppliction rtes, reducing the number of nozzles nd using smller nozzles would be dvntgeous rther thn driving the spryer t higher speed. On the other hnd, for higher volume pplictions

17 6 it would be helpful to use more nozzles nd higher speed rther thn using lrge disc nd core sizes. Investigtion on the effect of ground speed of irblst spryer (50 m 3 /s cpcity) on spry deposition by Slyni nd Whitney (1990) reveled tht incresing the speed does not necessrily reduce the deposition. Cnopy structure in grove is diverse nd hence, uniform sprying cn result in some spry losses. To optimize the usge of pesticide on tree crops ccording to cnopy mkeup, the generl trend mong reserchers nd commercil spryer mnufcturers hs been to develop spry systems which mtch the spry output with cnopy structure. Mny reserchers hve successfully shown the dvntge of sprying tree crops ccording to vegettion volume (Blsri nd Tmgnone, 1998; Blsri et l., 2003; Escol et l., 2003; Giles et l., 1987). Commercil systems like Roper s Tree-See TM (Roper Grower Coopertive, Winter Grden, FL), Durnd Wylnd s SmrtSpry TM (Durnd-Wylnd Inc., LGrnge, GA) nd AgTech s Tree-Sense TM (AgTech Inc., Mnhttn, KS) hve implemented the sme by using sensors tht trget specific zones on tree. The nozzles re then ctivted by electric solenoid vlves so tht only the zone detected by the sensors is spryed Air delivery system Air delivery system in ir-crrier spryers is responsible for producing high volume nd velocity irflow to trnsport liquid droplets from the nozzles to the trees. The min components in the ir delivery system include fn, irflow strightener nd ir deflectors. Vrious types of fns tht hve been used in irblst spryers re xil-, centrifugl- nd tngentil-flow fns. Axil flow fns re most populr for their lrge volume nd low pressure pplictions. Airblst spryers re type of ircrrier spryers tht use such fns. The fns consist of series of rdil bldes

18 7 ttched to rotting hub. This ssembly of bldes nd hub is termed impeller or rotor. Air drwn by the rotor is dischrged by the tngentil component of velocity. This results in swirling motion of the ir, commonly known s slipstrem rottion. The efficiency of the fn decreses with swirling s the ir encounters more resistnce. This swirl is removed by the sttor or strightener plced downstrem of the rotor. The dynmic pressure developed here is converted to sttic pressure rise. The ir is then deflected towrds the nozzles by set of deflector surfces (pltes), sometimes, by bout 90º to trget trees lterl to the spryer. It ws mnully observed tht the spry cn rech 2-3 rows beyond where it is intended. This ws prtilly due to the lrge mount of ir tht is uniformly used to trnsport the spry to the trget. Whitney et l. (1986) hve reported tht Power-Tke- Off (PTO) powered irblst spryers cn hve irflow rte from 3.77 to m 3 /s. It hs lso been suggested tht such high ir volume is justified only for lrge nd densely folited trees (Blsri et l., 2001; Slyni nd Frooq, 2003). Additionlly, such high volume cn led to drift of spry mteril into the ir. An experiment ws conducted by Slyni nd Frooq (2004) to quntify nd compre drift potentil of the commonly used irblst spryers. Due to the rdil dischrge of the PowerBlst spryer (used in this project), it hd highest bove cnopy drift compred to other spryers. However, it ws seen tht lower ground speeds (2.4 to 2.8 km/h) of the spryer produced higher spry deposition s compred to higher ground speeds (4.8 to 5.8 km/h) (Slyni et l., 2000). Slyni nd Hoffmnn (1996) reported tht, in generl, the velocity of ir reduced t incresing distnce from the spryer. Further, it ws lso found tht ir velocity from trveling spryer hd lesser mgnitude

19 8 compred to sttionry spryer. It should be noted tht these tests cn be ffected by wind speeds nd direction. Efforts to optimize the mteril trnsport from the nozzles of irblst spryers cn be complemented by system tht cn control the mount of ir tht trnsports this spry mteril bsed on the morphologicl chrcteristics of trees. In this regrd, Blsri et l. (2003) ttempted to chnge the ir output by using n djustble diphrgm t the xil fn inlet, but no rel experiments were reported. Also, this method is not suitble becuse equl mount of ir will be output from ech side of the spryer. This defets the purpose of providing vrible ir output for trees of different physicl chrcteristics on ech side of the spryer s it trvels through the grove. Lnders nd Gil (2006) tested n ir deflector system which directed the ir horizontlly into the cnopy on both sides of the spryer. It ws reported tht 25% improvement in deposition could be chieved using this system Testing Methodologies for Spryer Air nd Liquid Output An importnt spect of testing irblst spryer efficiency is to choose pproprite methodology to quntify the spry deposits nd ir chrcteristics. There re severl fctors tht cn ffect the results obtined from such experimentl design nd hence re reported long with the other results of the experiment. Some of these fctors re: ground speed of spryer; number of nozzles; nozzle type; nozzle pressure; nozzle orienttion; ir velocity nd volume; type nd loction of trgets; nd environmentl conditions like temperture, reltive humidity, wind speed nd direction. Spry pplictions hve been quntified using severl methods. Ech of these methods hs certin dvntges nd disdvntges. Hence, it is importnt to hve

20 9 some ide bout ech of them before selecting n pproprite method. Fluorescent trcers nd fluorometry hve lrgely been used to quntify deposition nd drift of spry mteril. A problem ssocited with this method is tht the commonly used wter soluble dyes re prone to degrdtion under solr rdition (Slyni, 1993). An lterntive could be to use metllic trcers such s copper which do not degrde under sunlight. The deposits in this cse re quntified using colorimetry. To ctch the trcer in spry, vriety of trgets hve been used. Though lef smples re idel for simulting the ctul trget, rtificil trgets such s pper, mylr, etc. cn provide certin dvntges in quntifying the spry deposition. Slyni nd Whitney (1988) conducted n experiment to compre the deposition on lef smples with mylr trgets using fluorometry nd colorimetry. They published correltion of R 2 =0.90 using colorimetry, nd R 2 =0.85 using fluorometry between lef nd mylr trget deposits. Slyni nd Fox (1999) compred oil nd wter sensitive ppers s trgets. They reported mjor chllenges in hndling these trgets becuse of their sensitivity to ir temperture, humidity nd opertor error. Additionlly, for high volume pplictions it might be difficult to quntify the spry mount s the trgets become over-covered with droplets. A comprison ws mde between string nd ribbon smplers by Slyni et l. (2006) in field pplictions. Spry mixtures in this experiment consisted of fluorescent trcer t different volume rtes nd ground speeds. It ws reported tht string smplers hd higher cpture efficiencies compred to ribbon smplers for ll smple loctions. 2.3 Control Systems in Pesticide Appliction Technology A control system is n integrtion of severl electricl/mechnicl components used to regulte desired output. Control systems hve been trditionlly used to

21 10 utomte processes in vrious res of the industry. In precision griculture, they hve been used to relize some of the gols of vrible rte ppliction technology. Control systems cn be brodly clssified s closed loop nd open loop systems. In open loop systems the controller directly gives commnds to n ctutor without receiving feedbck bout the ctutor s previous stte. This form of control cn be pplied when the ctution required is not very ccurte. Closed loop control systems continuously monitor the commnds sent bsed on the feedbck from the ctutor s previous stte nd informtion regrding the present stte. This form of control cn result in higher ccurcy nd fster response (Cugti et l., 2006, Gebhrdt et l., 1974). The min components of closed loop control system re plnt, computer/controller nd sensor. The plnt includes set of electromechnicl components which ct upon electricl signls sent by the controller to perform its function. Computer/controllers re electronic devices tht control the ctutor nd indirectly vry the ppliction rte of the products being pplied bsed on informtion from severl sources such s the ppliction equipment itself or other sensors. Controllers form the fundmentl component of ny vrible rte ppliction system (Clrk & McGuckin, 1996). These controllers re typiclly driven by microprocessor tht works bsed on set of rules or lgorithm. A sensor is trnsducer tht is used to mesure physicl quntity such s temperture, pressure, etc. nd convert it into n electricl signl. Sensors cn lso mesure prticulr stte of the plnt nd give feedbck signl to the computer/controller. Trditionlly, in spryers, control systems hve been used to utomticlly control nozzle dischrge rte in the liquid delivery system (Stover, 2002). The

22 11 trditionl pproch to hndle this hs been to regulte the pressure cross the nozzle (Giles et l., 1996). But this form of control cn hve significntly delyed response time resulting in poor performnce (Hn et l., 2000). To counter this, pulse-width modultion (PWM) hs been used to control electriclly ctuted solenoid vlves tht re connected to the nozzles (Giles & Comino, 1990, Hn et l., 2000). Electricl solenoid vlves cn give considerble shorter response time compred to conventionl pressure-bsed flow control system. Vribility in ppliction system cn be initited by two pproches: mp bsed ppliction nd sensor bsed ppliction. Sensor bsed ppliction, used in this project, hs n dvntge over mp bsed ppliction due to higher ccurcy (Swyer, 1994) nd rel time control (Zhng et l., 2002). Gebhrdt et l. (1974) developed n utomtic spryer control system tht chnged the output from the nozzle bsed on the ground speed of the spryer. A tchometer genertor sensor provided dc voltge to ger motor which in turn controlled the metering vlve in rel time t the output of the spry tnk. Ghte & Perry (1994) developed similr system where rdr sensor ws used to sense ground speed which vried pesticide ppliction rte by controlling 12 V dc step motor. Tngwongkit et l. (2006) used softwre bsed mchine vision system tht sensed greenness level of weeds to spry herbicide ccordingly. The mchine vision sensor ws connected to lptop which sent commnds to PWM circuit which in turn controlled dc electric motor. A lser sensor hs been developed tht cn give the height, volume nd density of ech scn on lptop in rel time s it trvels in the grove (Wei nd Slyni, 2005). This sensor reportedly gve better results t low speed (1.6 km/h) s compred to high speed (3.2

23 12 km/h) due to higher number of scns it mde t the lower speed (Slyni nd Wei, 2005). This reserch project ims t developing control system to regulte the mount of ir in irblst spryers to complement some of the systems designed to control liquid flow. Regulting the mount of ir using control system in spryer introduces new scenrio nd provides different chllenges. The irflow from the spryer is t high velocity nd turbulent in nture. The high inerti of xil fn restricts ny possibility of reducing the speed in rel time for smller trees. Moreover, the correct mount of ir needed to spry prticulr tree my not be known. This necessittes the development of control system tht cn be reprogrmmed nd is dptble to chnging ir volume bsed on different sensor inputs relting to tree chrcteristics. Specific objectives of this project re: 2.4 Objectives To design nd fbricte n electromechnicl system by which the mount of ir going to the trees cn be regulted. To integrte the signl from lser sensor with the ir regultor system to enble vrible rte sprying in the field. Evlute the functionlity of the developed system through ir velocity mesurements cross citrus cnopies with different folige densities. Determine sptil distribution of the spry droplets t different ir volume output.

24 CHAPTER 3 MATERIALS AND METHODS This chpter documents mterils nd methods tht were used in the design nd implementtion nd testing of n electromechnicl ir control system for the irblst spryer. 3.1 Airblst Spryer Description The spryer used for this project ws PowerBlst irblst spryer (Model No. PB533ST, Rer s Mnufcturing Compny, Eugene, OR). A schemtic of this irblst spryer is shown in Figure 3.1. It is PTO-driven nd hs single xil flow fn. This fn hs 9 bldes with dimeter of 0.84 m nd pitch of 32º. The spryer uses the trctor Power Tke Off (PTO) power through Constnt Velocity (C.V.) joint nd 3-point hitch connection. The PTO driveline trnsfers hevy torque lods from the trctor to the xil flow fn, which is operted by n electricl clutch. The speed of the fn, t P.T.O speed of 540 rpm, is round 2160 rpm. The fn rottes in counterclockwise direction looking from the rer of the spryer. It is followed by 24 blde flow strightener unit. The ir outlet of the spryer hs n inverted U-shpe slot of x 12.7 cm on ech side long its periphery. There re 24 hydrulic nozzles on ech side of the spryer. For the fn configurtion mentioned bove nd for PTO speed of 540 rpm, ir psses over the nozzles t mximum speed of bout 188 km/h (PTO PowerBlst mnul, Rer s Mnufcturing Compny, Eugene, OR). Under stndrd settings of fn, the totl volume rte of ir output is bout 16m 3 /s (pproximtely 33,901 cfm). 13

25 m Lser sensor Flow strightener Axil flow fn Tnk 2.4 m PTO Air 3 point hitch Hydrulic nozzle Figure 3.1. Schemtic of the PowerBlst irblst spryer 3.2 Lser Sensor The sensor used in this reserch project ws developed by Wei nd Slyni (2004; 2005) to mesure tree height, cnopy volume nd folige density. It uses n infrred lser emitter with wvelength of 780 nm. A line scnner consisting of motor, with n incrementl encoder, rottes the mirror tht deflects the outgoing bem of the lser emitter by 90º nd sweeps it through 360º s it rottes. The returning bem from the trget is deflected off the mirror bck to photodiode in the sensor. The lser sensor hs two RS232 interfce cbles which perform different functions. One cble is used to clibrte the lser sensor using the COM1 port on lptop while the second cble connects to High Speed InterFce (HSIF) Crd nd gives pulse signls.

26 15 Distnce mesurements re mde by nlyzing these pulse signls. Since the lptop does not hve slots for dditionl crds, PCI to Crdbus dpter is used to estblish communiction between HSIF crd nd lptop using the PCMCIA port. The HSIF crd lso controls the motor nd records its position using n encoder through prllel (DB25) port. The lser sensor ws mounted to the front side of the spryer on verticl pole t bout 2.4 m from the ground (Figure 3.1). 3.3 Preliminry Experiments It ws discussed in the erlier chpter tht ir usge in irblst spryer ws not optimum. To support this clim, severl tests were conducted. These experiments nd their conclusions hve been detiled in the sections below to ssert the progressive nture of this reserch project Airblst Spry Distribution Pttern An initil visul ssessment of the sprying pttern of the irblst spryer ws mde. Blue nd lilc Albuz APT cone nozzles (Cermiques Techniques, Desmrquest, Frnce) were used on left nd right side of the spryer to observe the effect of high nd low volume rte sprying (Figure 3.2), respectively. High volume rte nozzles Low volume rte nozzles Figure 3.2. Schemtic view of sprying ppliction. (Adopted from Slyni (2003))

27 16 The following visul observtions were mde while sprying with the bove configurtion: 1. Deposition on trees ws very good on immedite row of trees. The spry evported very fst with lilc nozzles becuse of the smll size of droplets. 2. Spry from blue nozzles moved up to 2 rows beyond the immedite row while for the lilc nozzles it moved only by 1 row. 3. For both nozzles, considerble mount of mteril fell on the ground nd some droplets were sucked into the fn 4. Mny droplets spryed from the upper section of the nozzles went up in ir without hitting ny trget which could cuse drift Restricted Air-Input Test A possible solution to some of observtions bove ws to djust the ir output to the nozzles. Since droplets use ir s medium to trnsport them to trget, djusting ir flow cn reduce some of the errors in sprying. To test this ide, n experiment ws conducted to reduce the volume of ir t input in steps nd quntify the output t the ir outlet. Annulus shped wooden bords (Figure 3.3) of incresing dimeter were cut nd ttched to spryer so s to chnge the mount of ir input to the fn. The cuts were curved to reduce turbulence t the edges. The four holes in the concentric bords hd n intke re of 0.1, 0.21, 0.36 nd 0.55 m 2. A pitot tube mnometer ws used to mesure ir pressure cross virtul grid of 10 x 5 points on ech side of the spryer ir outlet. The ten points were cross the periphery of the spryer while five points long the width of the outlet. Mesurements were mde in three replictions while the spryer ws sttionry. An extr set of reding ws tken without ny obstruction. Figure 3.3 shows progressive pictures of the bords tht were used to tke ir mesurements.

28 17 () (b) (c) (d) Figure 3.3. Airblst spryer with different ir intke re () 0.1 m 2 (b) 0.21 m 2 (c) 0.36 m 2 nd (d) 0.55 m 2 It ws concluded tht restricting the mount of ir t fn input provided mens of regulting ir t output (Figure 3.4). But, nnulus shped wooden bords ws not vible option becuse it resulted in equl mount of ir volume on ech side of the spryer. In generl, trees my be of different shpes nd sizes on either side of the spryer. Hence, in order to chieve the gol of optimizing the ir output, it ws essentil to regulte the mount of ir independently on both side of the spryer.

29 18 No obstruction Intke re (m 2 ) Right Side Left Side Air Flow (m 3 /s) Figure 3.4. Effect of fn inlet dimeter on the ir output of irblst spryer Deflecting Air t Output A deflector plte ws fbricted (Figure 3.5) using sheet metl. It ws plced in the spce between the fn nd the ir outlet. Its horizontl positions from the outermost (1) to innermost (5) would djust the mount of ir output from minimum to mximum, respectively. The shpe of the deflector ws mde erodynmic to reduce the mount of energy loss in ir deflection. The height of the plte hd to be limited to llow sufficient horizontl motion of the plte nd lso to chieve resonble ir vribility t the curved periphery of the spryer. A similr experiment s described in previous section ws conducted, by fixing the deflector plte t different horizontl loctions nd mesuring the ir output t outlet. Five horizontl loctions were chosen to model the ir output t the periphery. Bsed on preliminry experiments described in section 3.3.2, it ws found tht the verge ir volume coming out t the spryer outlet on ech side ws bout 7.67 m 3 /s. When the deflector plte ws instlled on the left side of spryer, the ir output on this side ws

30 19 split in two prts. A portion of it ws dischrged verticlly while the other portion ws deflected horizontlly towrd the trees. Tble 3.1 shows the chrcteristics of the deflected ir t ech plte position when mesured t the outlet. Since this outlet is curved, s the plte position chnges horizontlly from 1-5, the deflected ir comes through n incresing outlet re. To ccount for this increse in re, the verticl (peripherl) grid points over which dt ws mesured ws incresed. It cn be seen tht ir volume nd velocity incresed s the deflector plte chnged its position from setting 1 to setting 5. The results from this experiment demonstrted tht the deflector plte could regulte the mount of ir t output of the spryer (Figure 3.6). Tble 3.1. Chrcteristic of the horizontlly deflected ir due to vrious plte positions t spryer outlet Deflector plte position Men ir velocity (m/s) Men ir volume (m 3 /s) Mesurement grid points (verticl x horizontl) x x x x x 5 Note: Redings were bsed upon experiments described in section with 3 replictions. Verticl points re long the periphery of the spryer while horizontl points re long the width of the ir outlet.

31 20 57 cm 80 cm Nozzles Deflected ir Fn ir Shft fixed on the deflector plte Plte positions Figure 3.5. Schemtic view of the deflector plte motion 10 Horizontl ir output (m 3 /s) Std. Dev. y = 0.02x Horizontl irflow from lower prt of the ir outlet Plte position from outermost loction (mm) Figure 3.6. Effect of deflector plte loction on ir output

32 Automtion of Deflector Plte Movement An electromechnicl system ws designed nd implemented to utomte the horizontl movement of the deflector plte only on the left side of the spryer to test this ide. The control objective for this system ws to hve horizontl motion of the plte s function of the density reding obtined from the lser sensor for ech tree. The procedure for rel-time tree prmeter dt collection using lser sensor is discussed in section 3.5. The system designed (Figure 3.7) consists of n ctutor, which converts electromgnetic energy into mechnicl energy; controller, which is the hert of ny control system; nd mechnicl linkge, which tkes n input nd produces different output by chnging the motion, velocity or ccelertion of the input. GPS Controller Step Motor Lser sensor Figure 3.7. Components of the control system

33 22 Actutor: The ctutor used in this electromechnicl system ws step motor (AMH , Advnced Micro System, Inc., Nshu, NH). This type of motor provides precise positioning of the deflector plte. Like conventionl motors, step motor lso converts electromgnetic energy into mechnicl energy but the difference being tht it is done in steps. This essentilly mens tht power to this motor cn be sent in pulses which results in precise motion of the shft. The motor used here ws 1.8º or 200 steps per revolution motor. This step motor received signl commnds from step motor controller (mstep- 407, Advnced Micro Systems, Inc. Nshu, NH) which is described below. Detiled specifictions of the motor re listed in Appendix-B. The step motor ws connected to 36 V power supply (three 12 V btteries in series). Motor sizing clcultions hve been provided in Appendix-A. Controller: The mstep-407 (Figure 3.8) is n on-bord intelligent step motor controller. The choice of this controller ws bsed on two min resons. The first reson ws the necessity for it to integrte esily with the existing system. The lser sensor developed for sensing the tree prmeters used lptop to clibrte nd collect dt. Since, the controller includes seril link communiction port, it ws convenient nd cost-effective to ccept commnds from the lser sensor, process nd send pulse signls to the step motor. The second reson ws to hve controller tht not only improved the ccurcy but lso mde motion of the step motor smooth. This cn be chieved by microstepping which involves sending pulses tht will rotte the motor in frctions of its steps. The controller hs the feture of one-tenth micro-stepping resulting in higher ccurcy nd smoother rotry motion of motor. Specifictions of the mstep-407 controller bord

34 23 re provided in Appendix-B. The controller bord needed n 8-15 V logic power supply which ws obtined from one of the three btteries used to power the motor. Power Input Seril port Switch Figure 3.8. The step motor controller, mstep-407 (Advnced Micro Systems, Inc) Mechnicl Linkge: Mechnicl linkges re fundmentl prt of mchine design. The function of the mechnicl linkge designed here ws to convert rotry motion of the step motor to horizontl motion of deflector plte. Figure 3.9 shows schemtic of this ctution mechnism. Due to the limited spce vilbility in the existing spryer nd its bility to hndle high torques, bll screw nd nut ssembly (Prt number: HL5134M20452, Techno Inc., New Hyde Prk, NY) were used. Also, bll screw is very efficient nd cost effective choice to position moving prts ccurtely. The specifictions of the bll screw hve been given in Appendix-B. A cross br connected the nut on the bll screw with two shfts (Figure 3.5) on deflector plte. Two guide rollers (Prt number: VW-1, Modern Liner Inc., Corte Mder, CA) on shfts, fixed t 12 nd 69 cm mesured from the bottom, ber the weight of the deflector plte by riding long trck (Prt number: T-4, Modern Liner Inc., Corte Mder, CA) fixed on to the spryer. The step motor ws coupled with bll screw using right ngle drive (Prt

35 24 number: A 2Z28MC1010, 1:1 precision right ngle drive,) nd cross joint type flexible coupling (Prt number: S50MCTM25P08P10). It is importnt to estimte the expected lods nd pply sufficient sfety fctor to select linkges. Bsed on ir mesurements, it ws expected tht highest horizontl lod on the plte ws bout 221 N. This is bsed on mximum ir reding of 2.73 kp nd n re of 0.081m 2 on the rer of deflector plte. To ccount for overlods, sfety fctor of 10-15% ws considered nd hence, mximum lod of 250 N ws used to design other components of this system. A typicl procedure used in designing mechnicl motion systems is to hve wek component in the mechnicl linkge. This is done so tht, in cse of hevy overlod, the system breks t tht link thus protecting the more expensive prts. In order to trnsmit the motion in norml circumstnces, coupling tht hd torque limit of 2 N-m ws used in this system. In situtions when the deflector plte cn get stuck the coupling, which costs lesser, would brek nd stop the trnsmission but protect the other expensive components from dmge. Teflon sheets were plced on the spryer wll to reduce the friction tht resulted when the plte rides long the wll.

36 Rer wll slot Flow strightener Fn blde Deflector plte Rer wll of the ir outlet Teflon lining Guide roller 57 cm Bll nut Cross br Bll screw Flexible coupling Right ngle drive Step Motor Torque limiting coupling Air 25 Trck 12.7 cm Figure 3.9. Schemtic of ctution mechnism for the deflector plte

37 26 Input Output Reltions of the Control System. The electromechnicl system cn be chrcterized by five mthemticl (Tble 3.2) functions which provide reltion between the input nd output t vrious points. Tble 3.2. Input output reltions of the control system Number Input ( x ) Output y ( y ) x 1 Density reding from the lser sensor Indexing vlue to the controller bord x 1 2 Indexing vlue to the controller bord Rottion of motor shft 1 (revolutions) Rottion of motor shft (revolutions) Rottion of screw rod (revolutions) Rottion of screw rod (revolutions) Horizontl movement of plte (mm) Horizontl movement of plte (mm) Air output to the tree (m 3 /s) x 5 Figure 3.10 shows the reltion between indexing commnds tht were given by the controller fter getting density signl from lser sensor. The rnge % for tree density ws chosen, bsed on tril runs of lser sensor for prticulr row, to obtin higher resolution of plte movement. Density redings less thn 0.6 were ssigned plte loction 0 mm (horizontl loction 1) while redings greter thn 0.8 were given 200 mm (horizontl loction 5). A liner reltion ws chosen for the purpose of simplicity. The step motor indexing commnds re bsed on 200 steps (1.8º) per revolution step motor nd one-tenth resolution of the controller bord. Hence, n indexing commnd of 2000 would rotte the motor shft by one revolution.

38 27 Once the tree density ( x 1 ) from the lser sensor is obtined (explined in section 3.4), the vlue of trnsfer function number one in the bove tble cn be clculted. This is multiplied by trnsfer function numbers two, three nd four in tble 3.2 to obtin the horizontl movement of the plte ( x 5 ). An estimte of the horizontl irflow from the lower prt of the output cn be obtined from Figure 3.6. Agin, liner eqution ws used to estimte this for simplicity purpose. Bsed on this model, we cn relte the chnge in density vlue obtined from the lser sensor to chnge in horizontl irflow output from the spryer. 100 Step Motor Indexing Vlue x y = x Tree Density from Lser Sensor (%) Figure Reltionship between indexing vlue to the controller bord required for rnge of lser sensor density reding 3.5 Rel-time Collection of Tree Prmeters Using Lser Sensor Among the vrious wys to mesure the chrcteristics of tree cnopy, lser sensor is by fr the most ccurte. Slyni nd Wei (2005) hve shown the lgorithm to mesure the height, volume nd density of ech scn bsed on mthemticl pproch. The vlues of these three fetures which, henceforth, will be referred to s tree

39 28 prmeters from ech scn, long with their individul GPS loction vlues were needed for post-processing to get informtion bout tree. To eliminte the mnul postprocessing, lsers sensor lgorithm ws enhnced to ccurtely collect tree prmeter dt depending on tree spcing of tht prticulr row. Tree prmeter dt collected over tree ws lter used to decide the mount of ir necessry for tht prticulr tree. In order to get the strt nd end of ech tree, ltitude-longitude informtion from GPS input signl ws used to continuously clculte distnce trveled by lser sensor from its strting position. This distnce ws constntly compred with tree spcing dt, until tree spcing length ws reched, to decide the strt time nd stop time for collecting dt for prticulr tree. This is the defult method for dt collection. Sometimes under dense cnopy cover GPS signl might be unrelible. In order to ensure vlidity of the GPS signl, check ws included in the lser progrm. To ensure tht the tree prmeter dt is ccurte regrdless of unrelible GPS signl, the user is prompted to input expected nominl trvel speed initilly. Then the time required to trvel ech tree is clculted. An in-built softwre timer counts in steps equl to the nominl time required to trvel successive trees, signling tht the lser sensor hs reched the end of tree. If GPS signl is unrelible, the system switches to this method of dt collection, thus, ensuring continuity. The progrm is restored bck to its defult method s soon s the GPS signl is vlid. Another feture dded to the progrm ws collection of dt when resolution of GPS is not fine enough. At times, it is possible tht distnce clculted might not be equl to the exct tree spcing distnce. For exmple, for 4.6 m tree spcing the trvel distnce clculted hve redings of 3.7, 4.0, 4.3 nd 4.9 m thus skipping the 4.6 m

40 29 reding t which the progrm should hve completed collection for tht prticulr tree spcing. This sitution ws hndled by providing rnge of ± 0.3 m to ensure tht correct dt ws collected for ech tree spcing. A prllel time-bsed check ws lso implemented if the resolution degrded to more thn 0.3 m. This cn be understood more clerly with n exmple. If for 4.6 m tree spcing row, the GPS did not log distnces between 4.3 nd 4.9 m, then time bsed check would complete the collection of dt for the tree using GPS speed informtion nd tree spcing dt. The procedure used is the sme s the one described bove for the cse when GPS signl is unrelible. At the end of ech tree spcing n verge of the density redings is clculted. This verge density is used to obtin corresponding motor indexing commnd. A seril port communiction feture ws dded to the lser progrm to send this motor indexing commnd to the step motor with ccurte mount of turns to move the deflector plte. With these improvements to the lser progrm, the electromechnicl system nd lser sensor were integrted with negligible dely between detecting the density of tree nd the deflector plte movement. 3.6 Experiment 1: Air Penetrtion through Tree Cnopy The objective of this experiment ws to test the effectiveness of the plte loctions on ir penetrtion through tree cnopies with different densities. The velocity of ir pssing through trees with four different densities t five horizontl loctions of the deflector plte ws mesured. The trees were visully selected bsed on observtion in incresing order of densities nd re termed s ded cnopy (D), low density (L), medium density (M) nd high density (H) trees (Figure 3.11).

41 30 () (b) (c) (d) Figure Tree cnopies of different densities. () Ded cnopy (b) Low density (c) Medium density (d) High density Men nd mximum ir velocity mesurements were mde with hot film nemometer (FlowMster, Type 54N60, Dntec Mesurement Technology, Denmrk) t distnce of 4.8 m from the center of the spryer nd t height of 1 m from the ground (Figure 3.12). The men nd mx ir velocities were tken over period of 10 s while the spryer trveled cross 4.6 m tree. Mesurements were tken while the spryer ws drwn by trctor t PTO speed of 540 rpm nd t ground speeds of 2.15 nd 4.73 km/h heding in the est direction on row tht ws ligned in the est-west direction. Mesurements were mde in four replictions.

42 31 Direction of ir Deflector plte 1.0 m Figure Experimentl setup for mesuring ir velocity. Note: Anemometer ws held perpendiculr to the holder nd going in to the pge. Wether dt ws obtined t height of 2 m from the ground from FAWN (Florid Automted Wether Network) for intervl of the test. Ambient ir temperture, reltive humidity, wind speed nd direction during the experiment were ºC, 46-92%, m/s nd 44º-118º (0º represents north nd 90º represents est), respectively. 1.8 m Dt Anlysis. The experiment ws conducted s Rndomized Complete Block Design (RCBD). Men nd mximum ir velocities were nlyzed using nlysis of vrince. Interction between fctors ws nlyzed by considering this design s three fctor split-split plot experiment nd dt ws nlyzed using MIXED procedure in SAS (Freund et l., 1986). The three fctors considered were spryer ground speed, deflector plte setting nd tree density. Two spryer ground speeds (2.15 nd 4.73 km/h) divided ech of the four blocks (replictions) into 8 whole plots. Ech plot ws further divided in five split plots by rndomly ssigning five deflector plte settings. Ech split plot ws divided into four split-split plots, nd four tree densities (ded, low, medium nd high density) were rndomly ssigned. A grnd totl of 160 mesurements were vilble for 3.0 m

43 32 nlysis. Mens were seprted using LSMEANS with PDIFF option t 5% level of significnce. 3.7 Experiment 2: Effect of Deflector Plte on Sptil Distribution of Spry Deposition The objective of this experiment ws to determine the effect of spryer ir volume rte on sptil distribution of spry droplets. Spry deposition ws quntified by hving pper trgets (Fisherbrnd filter pper, Fisher Scientific, Pittsburgh, PA) t nine sptil loctions perpendiculr to the direction of the trvel of the spryer. The spryer ws operted with six open nozzles on the left side since they were directly ffected by the chnge in ir volume due to the horizontl position of the deflector plte. Two types of nozzles: lilc nd blue Albuz APT cone nozzles (Cermiques Techniques, Desmrquest, Frnce) were used to see the effect of low nd high volume ppliction, respectively. The mesured dischrge rtes of the six nozzles, t bout 1000 kp pressure, were 2.9 nd 21.4 L/min respectively. These volumes corresponded to ppliction rtes of 215 nd 1585 L/h bsed on row spcing of 6 m nd ground speed of 2.7 km/h. Five loctions of deflector plte, lbeled 1, 2, 3, 4 nd 5 corresponded to n ir volume rte 1.9, 4.1, 5.8, 7.0 nd 7.6 m 3 /s, respectively, of ir deflected towrds the trgets t PTO speed of 540 rpm. The test structure mde from PVC piping consisted of nine horizontl loctions lbeled A to I t distnces of 2.4, 3, 3.9, 4.8, 6.0, 7.2, 8.4, 9.6 m from the point of dischrge nd t height of 1.5 m from the ground (Figure 3.13). The trvel direction ws towrds west on row tht ws oriented in the est-west direction. Spry solutions contined Pyrnine-10G fluorescent dye (Keystone Aniline, Inc., Chicgo, IL) s deposition trcer t constnt rte of 566 mg/l (ppm).wter sensitive pper (Sprying Systems Co., Wheton, IL) trgets were lso plced on ech loction to

44 33 visully compre it with results from pper trgets. These trgets were held by trget holders to keep them stedy nd perpendiculr to spry direction during ech spryer pss. The exposed re on the pper trgets ws cm 2. After ech spryer pss, trgets were immeditely plced in selble plstic bgs, nd stored in n enclosed continer for further lbortory nlysis. Experiments were mde in four replictions. In the lbortory, spry (dye) deposits on ech trget were quntified by fluorometry (Slyni, 2000b). The deposits were normlized for differences in the ppliction volume rte (L/h). Ambient ir tempertures nd reltive humidity were monitored t height of 1.8 m, using temperture/rh indictor (Model 870H, Generl Estern, Wtertown, MA). Wind speed ws mesured using vne nemometer (Model HH-30, Omeg Engineering, Stmford, CT) t the sme height. A white ribbon tied to pole ws used to visully note the wind direction. The rnges of temperture, reltive humidity nd wind speed during the experiment were ºC, % nd m/s, respectively. The wind direction ws primrily from the north nd north-est direction nd did not seem to hve ny significnt effect on sprying s mgnitude of winds were low during the entire experiment.