Citation
Controlling Airblast Sprayer Air for Variable Rate Application in Orchards

Material Information

Title:
Controlling Airblast Sprayer Air for Variable Rate Application in Orchards
Creator:
PAI, NARESH ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Canopy ( jstor )
Control systems ( jstor )
Ground speed ( jstor )
Lasers ( jstor )
Nozzles ( jstor )
Pesticides ( jstor )
Sensors ( jstor )
Spray deposition ( jstor )
Spraying ( jstor )
Velocity ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Naresh Pai. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
7/12/2007
Resource Identifier:
660033977 ( OCLC )

Downloads

This item has the following downloads:


Full Text





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


































2007 Naresh Pai

































To Archana and Anant Pai















ACKNOWLEDGMENTS

I would like to extend my thanks to the many individuals who have contributed to

make this project a success and my educational experience so enjoyable. I would like to

thank Dr. Masoud Salyani, my advisory committee chairman, for his faith in me, and for

his continued support and inspiration. His constant encouragement, timely critical

evaluation, and enthusiasm for my work have resulted in the successful completion of my

research. I am indebted to Mr. Roy Sweeb, Senior Engineering Technician, for his

insightful ideas, hands-on support, and training in the workshop and field, throughout the

work.

I would also like to thank Dr. Thomas Burks and Dr. John Schueller for giving me

valuable knowledge through their courses and serving on my supervisory committee. I

want to acknowledge the Agricultural and Biological Engineering (ABE) department for

providing me the opportunity, and the Citrus Research and Education Centre (CREC) for

the assistantship and technical resources to conduct my research. I am also grateful to Dr.

Reza Ehsani, Mr. Troy Gainey, and the staff of the CREC maintenance department for

letting me use equipment needed for the project.

On a personal note, I would like to thank my parents and brother whose support

was of inestimable value. A final word of thanks goes out to all my friends who have

directly or indirectly contributed to the successful completion of my work.
















TABLE OF CONTENTS



A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES ........... ..... ............. ........... .... ......... .............. .. vii

LIST OF FIGURES ............. ............. ........ ....... .......................... viii

A B ST R A C T .......... ..... ...................................................................................... x

CHAPTER

1 IN TR OD U CTION ............................................... .. ......................... ..

1.1 Ju stification ............................................. ........ ...... .... ........ ... 1
1.2 T hesis O organization .......... .. ...... .......................................... .. .......... .. 2

2 B A C K G R O U N D ....................................................... 3

2.1 Pesticide U sage in Florida ................................................ ........................... 3
2.2 Pesticide Application Technology for Tree Crops .........................................4
2.2.1 Sprayers ........................................................................... ....... ........ ..... ........ 4
2.2.2 A ir-Carrier Sprayers ............................................................................. 4
2.2.2.1 Liquid delivery system ........................................... ............... 5
2.2.2.2 A ir delivery system ........................ ................ ................... 6
2.2.3 Testing Methodologies for Sprayer Air and Liquid Output.....................
2.3 Control Systems in Pesticide Application Technology ..........................................9
2.4 Objectives .................. ........ ..................... ........ .... .... ........ 12

3 M ATERIALS AND M ETHOD S ........................................ ......................... 13

3.1 A irblast Sprayer D description ........................................ .......................... 13
3 .2 L aser S en sor......................................................................................... 14
3.3 Prelim inary E xperim ents ................................................... ........ ............... 15
3.3.1 Airblast Spray Distribution Pattern .........................................................15
3.3.2 Restricted A ir-Input Test................................ ......................... ........ 16
3.3.3 D eflecting A ir at O utput...................................... .................. .... ........... 18
3.4 Automation of Deflector Plate Movement ..........................................................21
3.5 Real-time Collection of Tree Parameters Using Laser Sensor .............................27
3.6 Experiment 1: Air Penetration through Tree Canopy..................................29



v









3.7 Experiment 2: Effect of Deflector Plate on Spatial Distribution of Spray
D position ......................................................................................................32

4 RE SU LTS A N D D ISCU SSION ...................................................... .....................36

4.1 Evaluation of the Electromechanical Control System ..........................................36
4.2 Experiment I: Air Velocity Measurements...................................................38
4.3 Experim ent II: Spray D position ................................... .................................... 42
4.4 D discussion ...................................................................... .. ....... ...... 43

5 CON CLU SION S ..................................... .......... ......... ........... 48

APPENDIX

A STEP MOTOR SIZING CALCULATION.....................................................49

B COM PONENT SPECIFICATION ................................... .............................. ...... 53

LIST OF REFEREN CES ............................................................ .................... 54

B IO G R A PH IC A L SK E TCH ..................................................................... ..................58
















LIST OF TABLES

Table p

3.1 Characteristic of the horizontally deflected air due to various plate positions at
sp ray er o u tlet ................................................................... ................ 19

3.2 Input output relations of the control system ........................................ .................26

A.1 Specifications of the actuation mechanism ........... ........................................49

B.1 Ball screw and nut assembly specifications .................................. ............... 53

B.2 Step motor technical specifications.................... ...... .............. ...............53

B .3 Step m otor controller features ...................................................................... .. .... 53
















LIST OF FIGURES


Figure p

3.1 Schematic of the PowerBlast airblast sprayer......................................................14

3.2 Schematic view of spraying application....................... ....................... 15

3.3 Airblast sprayer with different air intake area............................... ............... 17

3.4 Effect of fan inlet diameter on the air output of airblast sprayer ............................18

3.5 Schematic view of the deflector plate motion ......................................................20

3.6 Effect of deflector plate location on air output......................................................20

3.7 Components of the control system .............. .. ..................................... 21

3.8 The step motor controller, m STEP-407 ....................................... ............... 23

3.9 Schematic of actuation mechanism for the deflector plate ....................................25

3.10 Relationship between indexing value to the controller board required for a range
of laser sensor density reading ........................................... .......................... 27

3.11 Tree canopies of different densities...................................... ........................ 30

3.12 Experimental setup for measuring air velocity............... ..... ...............31

3.13 Schematic view of spray application experiment and sampling layout ...................34

4.1 Relation between actual plate position and tree density........................................37

4.2 Mean air velocity due to different deflector plate location at 2.15 and 4.73 km/h. .40

4.3 Maximum air velocity due to different deflector plate location at 2.15 and 4.73
km /h ......... .................. .................................... ...........................4 1

4.4 Maximum air velocity at two ground speeds of the airblast sprayer. ...................42

4.5 Effect of two application volume rate on total deposition at two spatial sections...45









4.6 Low application rate: spatial distribution of deposition at different deflector
plate p o sition .........................................................................4 6

4.7 High application rate: spatial distribution of deposition at different deflector
p late p o sition s ...................................... ............................................. 4 7

A. 1 Block diagram used for stepper motor calculations .............................................49















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science

CONTROLLING AIRBLAST SPRAYER AIR FOR VARIABLE RATE
APPLICATION IN ORCHARDS

By

Naresh Pai

May 2007

Chair: Masoud Salyani
Major: Agricultural and Biological Engineering

Spray requirements vary considerably throughout the grove on account of

variability in citrus canopy size and foliage density. Configuring sprayers to suit this tree

variability is vital for efficient spraying. Currently, crops are sprayed uniformly

throughout the field based on experience. Uniform application of agrochemicals not only

wastes chemicals but also has environmental implications.

At present, airblast sprayers account for majority of sprayers for tree crop

application in Florida. While moving across the grove, these sprayers rely on a stream of

air supply generated by fan(s) to carry the material from the nozzles to the canopy. The

air volume generated in these sprayers range between 3.7 46.7 m3/s. A fully grown tree

has a different spray requirement as compared to a small or medium sized tree. It has also

been observed that a typical airblast sprayer may deposit 2-3 rows beyond the immediate

row for which it was intended. Since the tree size and density distribution on the field are









usually non-uniform, uniform spraying could result in substantial material losses as

ground fallout and drift.

My work describes design modification of an existing airblast sprayer to test the

idea of variable rate spray application by adjusting the air output. My project involved the

design, implementation and testing of an electromechanical system to change the volume

of air going to the trees based on the tree density information from a laser sensor. This

process gives real time change in air output characteristics as the sprayer moves across a

row of trees. By using this system, the air volume can be changed from 1.9 to 7.6 m3/s in

less than 3 seconds. Air penetration was quantified by measuring air velocity which

revealed that different settings of this air regulatory system can produce significantly

different air characteristics across a dead and medium density tree. Overall, increase in

air volume gave higher spatial distribution of spray material but for high volume

application changing air volume can produce significant high deposits in near locations

when compared to far locations thus reducing off-target wastage.














CHAPTER 1
INTRODUCTION

1.1 Justification

Florida is the second largest producer of citrus in the world and accounts for more

than one-third of the world's grapefruit production. These trees are mostly sprayed with

air-carrier sprayers. These sprayers use high volume and velocity air streams, produced

by axial-, centrifugal-, or cross-flow fans, to transport the spray droplets to tree canopies.

The aim is to replace the air within the trees by a stream of air and agrochemical droplets.

Spray requirements of a fully grown mature tree is substantially different from

those of medium or small sized tree. In addition, it is very common to find dead trees and

resets in the field. Efficient spraying in agriculture involves optimum usage of the

available resources. The ideal spraying deposits the material on the intended target, based

on the type of tree and minimizes ground deposit and drift. In contrast, at full air capacity

the sprayers could deposit the material 2-3 rows beyond the row for which it was

intended. This results in considerable amount of wastage of agrochemicals which has

economical as well as environmental concerns.

Many researchers and companies have successfully shown the advantage of

customizing the liquid output from nozzles based on the characteristics of the trees.

However, there have been fewer attempts to modify the air that carries the spray droplets

to the trees. The overall objective of this project is to evaluate the idea of variable rate

spray application by changing the air volume that is used to transport these spray

droplets.









1.2 Thesis Organization

My thesis is divided into five chapters. The second chapter, Background, deals with

relevant literature review and past work done in this area. The third chapter, Materials

and Methods, describes the progressive steps taken to develop a system to address the

problem described in section 1.1. Subsequently, it also describes the experiments

conducted to test this system. The fourth chapter, Results and Discussion, evaluates the

control system, and further, discusses the results that were obtained from the experiments

described in chapter three. The fifth chapter, Conclusions, gives an overall perspective of

this research project.














CHAPTER 2
BACKGROUND

This chapter gives the background and lays a foundation for this project. It

starts with statistics of the agricultural pesticide usage in Florida to recognize the

significance of developing technology which can maximize application efficiency and

minimize wastage. Pesticide application technology is then discussed to give a

general overview of equipment used for tree crops. Focus has been placed on airblast

sprayers as the aim of this project was to optimize air output in such sprayers. Finally,

control systems in airblast sprayers used for variable rate application are discussed.

2.1 Pesticide Usage in Florida

Pesticides are agrochemicals that are intended for controlling pests. Pesticides

in the form of sprays are commonly used to control pests that affect citrus. In 1999,

Florida Agricultural Statistics Services reported that herbicides were used on 96.3%

of the 316,840 ha of citrus and avocado. Similarly, insecticides and fungicides were

applied on 91.3 and 81.9% of the acreage respectively in the same year. This shows

the widespread use of pesticides for tree crops. Though the intention is to spray the

material to the target, a considerable amount of pesticide is wasted. These wastages

arise primarily from three sources:

Off-target spraying
Drift to air
Runoff from leaves falling on the ground

It was reported that about 22% of the spray material is wasted to the ground

and, about 21% drifted to air with commonly used airblast sprayer at application rate









of 950 L/ha (Miller et al., 2003). This not only has economic disadvantages, but also

leads to ground and surface water pollution, air pollution, disturbance in the

neighboring ecosystem and may also present health concern to humans. Hence,

efforts are being made to optimize the amount of pesticide usage in agriculture.

2.2 Pesticide Application Technology for Tree Crops

Pesticide application technology refers to the equipment that are used to

dispense pesticides. Pesticide applicators for tree crops are characterized by their high

volume application to cover the dense foliage of trees. Pesticides are applied in solid

or liquid formulation with or without assistance of air. Pesticides in solid formulation

are applied either in granular or powdered form. These applicators are called granular

pesticide applicators or dusters, respectively. Pesticides that are used in liquid

formulation are applied using sprayers. Commonly, sprayers use assistance of air to

carry the spray droplets, and for better canopy penetration. These sprayers are

designated as air-carrier sprayers.

2.2.1 Sprayers

Sprayers use hydraulic systems to transport and spray pesticides to the target.

These hydraulic systems comprise a tank to store the liquid formulation, a pump to

develop the necessary pressure, manifolds to transport the pressurized liquid and

nozzles to convert the liquid into droplets and disseminate the pesticide. These

sprayers have a boom to support the nozzles at the outlet.

2.2.2 Air-Carrier Sprayers

Air-carrier sprayers account for 89% of spray machines for citrus production in

Florida (Stover et al., 2002a) and hence, are important equipment for pesticide

application. These sprayers are commonly used for spraying tree crops having large









foliage. They use a stream of air to transport spray droplets towards the tree and

penetrate the canopy. The air in and around the trees is thus replaced with a mixture

of air and pesticide droplets. This method of application is considered superior as it

increases the deposition (Reichard, 1977; Salyani, 1988). These sprayers are available

in different shapes, sizes, fan types, nozzles and are operated with different volume

rates and ground speeds.

The working of an air-carrier sprayer system can be broadly divided into two

modules. The first one comprises of the components that handle the delivery of liquid

and is referred to as the liquid delivery system. The second module relates to the

delivery of air and is called air delivery system.

2.2.2.1 Liquid delivery system

The task of liquid delivery system in an air-carrier sprayer is to store, pressurize

and produce droplets of the liquid pesticide formulation. This is achieved by a

hydraulic circuit which employs a pump to pressurize and force the liquid from the

tank to the nozzles. The nozzles aligned on the sprayer can be categorized to be in

three sections namely; top, middle and bottom, reflecting the section of the tree that

the spray material affects. These nozzles produce droplets of liquid pesticide which

are transported to the target due to the functioning of air delivery system.

The number of nozzles, nozzle size, nozzle pressure, and ground speed of

sprayer play an important role in the deposition efficiency of a sprayer. These have to

be adjusted to maximize the efficiency of spraying. For instance, it was suggested by

Salyani (2000a) that for better deposition at lower application rates, reducing the

number of nozzles and using smaller nozzles would be advantageous rather than

driving the sprayer at higher speed. On the other hand, for higher volume applications









it would be helpful to use more nozzles and higher speed rather than using large disc

and core sizes. Investigation on the effect of ground speed of airblast sprayer (50 m3/s

capacity) on spray deposition by Salyani and Whitney (1990) revealed that increasing

the speed does not necessarily reduce the deposition.

Canopy structure in a grove is diverse and hence, uniform spraying can result in

some spray losses. To optimize the usage of pesticide on tree crops according to

canopy makeup, the general trend among researchers and commercial sprayer

manufacturers has been to develop spray systems which match the spray output with

canopy structure. Many researchers have successfully shown the advantage of

spraying tree crops according to vegetation volume (Balsari and Tamagnone, 1998;

Balsari et al., 2003; Escola et al., 2003; Giles et al., 1987). Commercial systems like

Roper's Tree-SeeTM (Roper Grower Cooperative, Winter Garden, FL), Durand

Wayland's Sin.hlt prayTM (Durand-Wayland Inc., LaGrange, GA) and AgTech's

Tree-SenseTM (AgTech Inc., Manhattan, KS) have implemented the same by using

sensors that target specific zones on a tree. The nozzles are then activated by electric

solenoid valves so that only the zone detected by the sensors is sprayed.

2.2.2.2 Air delivery system

Air delivery system in air-carrier sprayers is responsible for producing high

volume and velocity airflow to transport liquid droplets from the nozzles to the trees.

The main components in the air delivery system include a fan, airflow straightener

and air deflectors. Various types of fans that have been used in airblast sprayers are

axial-, centrifugal- and tangential-flow fans. Axial flow fans are most popular for

their large volume and low pressure applications. Airblast sprayers are a type of air-

carrier sprayers that use such fans. The fans consist of a series of radial blades









attached to a rotating hub. This assembly of blades and hub is termed impeller or

rotor. Air drawn by the rotor is discharged by the tangential component of velocity.

This results in a swirling motion of the air, commonly known as slipstream rotation.

The efficiency of the fan decreases with swirling as the air encounters more

resistance. This swirl is removed by the stator or straightener placed downstream of

the rotor. The dynamic pressure developed here is converted to static pressure rise.

The air is then deflected towards the nozzles by a set of deflector surfaces (plates),

sometimes, by about 900 to target trees lateral to the sprayer.

It was manually observed that the spray can reach 2-3 rows beyond where it is

intended. This was partially due to the large amount of air that is uniformly used to

transport the spray to the target. Whitney et al. (1986) have reported that Power-Take-

Off (PTO) powered airblast sprayers can have airflow rate from 3.77 to 25.01 m3/s. It

has also been suggested that such high air volume is justified only for large and

densely foliated trees (Balsari et al., 2001; Salyani and Farooq, 2003). Additionally,

such high volume can lead to drift of spray material into the air. An experiment was

conducted by Salyani and Farooq (2004) to quantify and compare drift potential of

the commonly used airblast sprayers. Due to the radial discharge of the PowerBlast

sprayer (used in this project), it had highest above canopy drift compared to other

sprayers. However, it was seen that lower ground speeds (2.4 to 2.8 km/h) of the

sprayer produced higher spray deposition as compared to higher ground speeds (4.8

to 5.8 km/h) (Salyani et al., 2000). Salyani and Hoffmann (1996) reported that, in

general, the velocity of air reduced at increasing distance from the sprayer. Further, it

was also found that air velocity from a traveling sprayer had lesser magnitude









compared to a stationary sprayer. It should be noted that these tests can be affected by

wind speeds and direction.

Efforts to optimize the material transport from the nozzles of airblast sprayers

can be complemented by a system that can control the amount of air that transports

this spray material based on the morphological characteristics of trees. In this regard,

Balsari et al. (2003) attempted to change the air output by using an adjustable

diaphragm at the axial fan inlet, but no real experiments were reported. Also, this

method is not suitable because equal amount of air will be output from each side of

the sprayer. This defeats the purpose of providing variable air output for trees of

different physical characteristics on each side of the sprayer as it travels through the

grove. Landers and Gil (2006) tested an air deflector system which directed the air

horizontally into the canopy on both sides of the sprayer. It was reported that a 25%

improvement in deposition could be achieved using this system.

2.2.3 Testing Methodologies for Sprayer Air and Liquid Output

An important aspect of testing airblast sprayer efficiency is to choose

appropriate methodology to quantify the spray deposits and air characteristics. There

are several factors that can affect the results obtained from such experimental design

and hence are reported along with the other results of the experiment. Some of these

factors are: ground speed of sprayer; number of nozzles; nozzle type; nozzle pressure;

nozzle orientation; air velocity and volume; type and location of targets; and

environmental conditions like temperature, relative humidity, wind speed and

direction.

Spray applications have been quantified using several methods. Each of these

methods has certain advantages and disadvantages. Hence, it is important to have









some idea about each of them before selecting an appropriate method. Fluorescent

tracers and fluorometry have largely been used to quantify deposition and drift of

spray material. A problem associated with this method is that the commonly used

water soluble dyes are prone to degradation under solar radiation (Salyani, 1993). An

alternative could be to use metallic tracers such as copper which do not degrade under

sunlight. The deposits in this case are quantified using colorimetry. To catch the

tracer in spray, a variety of targets have been used. Though leaf samples are ideal for

simulating the actual target, artificial targets such as paper, mylar, etc. can provide

certain advantages in quantifying the spray deposition. Salyani and Whitney (1988)

conducted an experiment to compare the deposition on leaf samples with mylar

targets using fluorometry and colorimetry. They published a correlation of R2=0.90

using colorimetry, and R2=0.85 using fluorometry between leaf and mylar target

deposits. Salyani and Fox (1999) compared oil and water sensitive papers as targets.

They reported major challenges in handling these targets because of their sensitivity

to air temperature, humidity and operator error. Additionally, for high volume

applications it might be difficult to quantify the spray amount as the targets become

over-covered with droplets. A comparison was made between string and ribbon

samplers by Salyani et al. (2006) in field applications. Spray mixtures in this

experiment consisted of fluorescent tracer at different volume rates and ground

speeds. It was reported that string samplers had higher capture efficiencies compared

to ribbon samplers for all sample locations.

2.3 Control Systems in Pesticide Application Technology

A control system is an integration of several electrical/mechanical components

used to regulate a desired output. Control systems have been traditionally used to









automate processes in various areas of the industry. In precision agriculture, they

have been used to realize some of the goals of variable rate application technology.

Control systems can be broadly classified as closed loop and open loop systems. In

open loop systems the controller directly gives commands to an actuator without

receiving feedback about the actuator's previous state. This form of control can be

applied when the actuation required is not very accurate. Closed loop control systems

continuously monitor the commands sent based on the feedback from the actuator's

previous state and information regarding the present state. This form of control can

result in higher accuracy and faster response (Cugati et al., 2006, Gebhardt et al.,

1974). The main components of a closed loop control system are plant,

computer/controller and sensor. The plant includes a set of electromechanical

components which act upon electrical signals sent by the controller to perform its

function. Computer/controllers are electronic devices that control the actuator and

indirectly vary the application rate of the products being applied based on information

from several sources such as the application equipment itself or other sensors.

Controllers form the fundamental component of any variable rate application system

(Clark & McGuckin, 1996). These controllers are typically driven by a

microprocessor that works based on a set of rules or algorithm. A sensor is a

transducer that is used to measure a physical quantity such as temperature, pressure,

etc. and convert it into an electrical signal. Sensors can also measure a particular state

of the plant and give feedback signal to the computer/controller.

Traditionally, in sprayers, control systems have been used to automatically

control nozzle discharge rate in the liquid delivery system (Stover, 2002). The









traditional approach to handle this has been to regulate the pressure across the nozzle

(Giles et al., 1996). But this form of control can have significantly delayed response

time resulting in poor performance (Han et al., 2000). To counter this, pulse-width

modulation (PWM) has been used to control electrically actuated solenoid valves that

are connected to the nozzles (Giles & Comino, 1990, Han et al., 2000). Electrical

solenoid valves can give considerable shorter response time compared to

conventional pressure-based flow control system.

Variability in application system can be initiated by two approaches: map based

application and sensor based application. Sensor based application, used in this

project, has an advantage over map based application due to higher accuracy (Sawyer,

1994) and real time control (Zhang et al., 2002). Gebhardt et al. (1974) developed an

automatic sprayer control system that changed the output from the nozzle based on

the ground speed of the sprayer. A tachometer generator sensor provided dc voltage

to a gear motor which in turn controlled the metering valve in real time at the output

of the spray tank. Ghate & Perry (1994) developed a similar system where a radar

sensor was used to sense ground speed which varied pesticide application rate by

controlling a 12 V dc step motor. Tangwongkit et al. (2006) used a software based

machine vision system that sensed greenness level of weeds to spray herbicide

accordingly. The machine vision sensor was connected to a laptop which sent

commands to a PWM circuit which in turn controlled a dc electric motor. A laser

sensor has been developed that can give the height, volume and density of each scan

on a laptop in real time as it travels in the grove (Wei and Salyani, 2005). This sensor

reportedly gave better results at low speed (1.6 km/h) as compared to high speed (3.2









km/h) due to higher number of scans it made at the lower speed (Salyani and Wei,

2005).

This research project aims at developing a control system to regulate the

amount of air in airblast sprayers to complement some of the systems designed to

control liquid flow. Regulating the amount of air using control system in a sprayer

introduces a new scenario and provides different challenges. The airflow from the

sprayer is at high velocity and turbulent in nature. The high inertia of axial fan

restricts any possibility of reducing the speed in real time for smaller trees. Moreover,

the correct amount of air needed to spray a particular tree may not be known. This

necessitates the development of a control system that can be reprogrammed and is

adaptable to changing air volume based on different sensor inputs relating to tree

characteristics.

2.4 Objectives

Specific objectives of this project are:

* To design and fabricate an electromechanical system by which the amount of
air going to the trees can be regulated.

* To integrate the signal from a laser sensor with the air regulator system to
enable variable rate spraying in the field.

* Evaluate the functionality of the developed system through air velocity
measurements across citrus canopies with different foliage densities.

* Determine spatial distribution of the spray droplets at different air volume
output.














CHAPTER 3
MATERIALS AND METHODS

This chapter documents materials and methods that were used in the design and

implementation and testing of an electromechanical air control system for the airblast

sprayer.

3.1 Airblast Sprayer Description

The sprayer used for this project was PowerBlast airblast sprayer (Model No.

PB533ST, Rear's Manufacturing Company, Eugene, OR). A schematic of this airblast

sprayer is shown in Figure 3.1. It is PTO-driven and has a single axial flow fan. This fan

has 9 blades with a diameter of 0.84 m and pitch of 32. The sprayer uses the tractor

Power Take Off (PTO) power through a Constant Velocity (C.V.) joint and 3-point hitch

connection.

The PTO driveline transfers heavy torque loads from the tractor to the axial flow

fan, which is operated by an electrical clutch. The speed of the fan, at P.T.O speed of 540

rpm, is around 2160 rpm. The fan rotates in a counterclockwise direction looking from

the rear of the sprayer. It is followed by a 24 blade flow straightener unit. The air outlet

of the sprayer has an inverted U-shape slot of 144.0 x 12.7 cm on each side along its

periphery. There are 24 hydraulic nozzles on each side of the sprayer. For the fan

configuration mentioned above and for PTO speed of 540 rpm, air passes over the

nozzles at a maximum speed of about 188 km/h (PTO PowerBlast manual, Rear's

Manufacturing Company, Eugene, OR). Under standard settings of fan, the total volume

rate of air output is about 16m3/s (approximately 33,901 cfm).
















2.7 m

S Laser Axial flow fan
sensor
Flow straightener

Tank







TO 11 Air





3 point Hydraulic nozzle
/ hitch



Figure 3.1. Schematic of the PowerBlast airblast sprayer

3.2 Laser Sensor

The sensor used in this research project was developed by Wei and Salyani (2004;

2005) to measure tree height, canopy volume and foliage density. It uses an infrared laser

emitter with a wavelength of 780 nm. A line scanner consisting of a motor, with an

incremental encoder, rotates the mirror that deflects the outgoing beam of the laser

emitter by 90 and sweeps it through 3600 as it rotates. The returning beam from the

target is deflected off the mirror back to a photodiode in the sensor.

The laser sensor has two RS232 interface cables which perform different functions.

One cable is used to calibrate the laser sensor using the COM1 port on a laptop while the

second cable connects to a High Speed InterFace (HSIF) Card and gives pulse signals.









Distance measurements are made by analyzing these pulse signals. Since the laptop does

not have slots for additional cards, a PCI to Cardbus adapter is used to establish

communication between HSIF card and laptop using the PCMCIA port. The HSIF card

also controls the motor and records its position using an encoder through a parallel

(DB25) port. The laser sensor was mounted to the front side of the sprayer on a vertical

pole at about 2.4 m from the ground (Figure 3.1).

3.3 Preliminary Experiments

It was discussed in the earlier chapter that air usage in airblast sprayer was not

optimum. To support this claim, several tests were conducted. These experiments and

their conclusions have been detailed in the sections below to assert the progressive nature

of this research project.

3.3.1 Airblast Spray Distribution Pattern

An initial visual assessment of the spraying pattern of the airblast sprayer was

made. Blue and lilac Albuz APT cone nozzles (Ceramiques Techniques, Desmarquest,

France) were used on left and right side of the sprayer to observe the effect of high and

low volume rate spraying (Figure 3.2), respectively.


High volume rate nozzles .. -"' m- 11u 1-
Figure 3.2. Schematic view of spraying application. (Adopted from Salyani (2003))









The following visual observations were made while spraying with the above

configuration:

1. Deposition on trees was very good on immediate row of trees. The spray
evaporated very fast with lilac nozzles because of the small size of droplets.

2. Spray from blue nozzles moved up to 2 rows beyond the immediate row while for
the lilac nozzles it moved only by 1 row.

3. For both nozzles, considerable amount of material fell on the ground and some
droplets were sucked into the fan

4. Many droplets sprayed from the upper section of the nozzles went up in air without
hitting any target which could cause drift.

3.3.2 Restricted Air-Input Test

A possible solution to some of observations above was to adjust the air output to

the nozzles. Since droplets use air as a medium to transport them to target, adjusting air

flow can reduce some of the errors in spraying. To test this idea, an experiment was

conducted to reduce the volume of air at input in steps and quantify the output at the air

outlet. Annulus shaped wooden boards (Figure 3.3) of increasing diameter were cut and

attached to sprayer so as to change the amount of air input to the fan. The cuts were

curved to reduce turbulence at the edges. The four holes in the concentric boards had an

intake area of 0.1, 0.21, 0.36 and 0.55 m2

A pitot tube manometer was used to measure air pressure across a virtual grid of 10

x 5 points on each side of the sprayer air outlet. The ten points were across the periphery

of the sprayer while five points along the width of the outlet. Measurements were made in

three replications while the sprayer was stationary. An extra set of reading was taken

without any obstruction. Figure 3.3 shows progressive pictures of the boards that were

used to take air measurements.



















(a) (b)


(c) (d)
Figure 3.3. Airblast sprayer with different air intake area (a) 0.1 m2 (b) 0.21 m2 (c) 0.36
m2 and (d) 0.55 m2
It was concluded that restricting the amount of air at fan input provided a means of
regulating air at output (Figure 3.4). But, annulus shaped wooden boards was not a viable
option because it resulted in equal amount of air volume on each side of the sprayer. In
general, trees may be of different shapes and sizes on either side of the sprayer. Hence, in
order to achieve the goal of optimizing the air output, it was essential to regulate the
amount of air independently on both side of the sprayer.


T:.=-


T -e











No
obstruction

S0.55

o 0.36


S0.21 Right Side

1 Left Side
0.10

0 2 4 6 8 10
Air Flow (m3/s)


Figure 3.4. Effect of fan inlet diameter on the air output of airblast sprayer

3.3.3 Deflecting Air at Output

A deflector plate was fabricated (Figure 3.5) using sheet metal. It was placed in the

space between the fan and the air outlet. Its horizontal positions from the outermost (1) to

innermost (5) would adjust the amount of air output from minimum to maximum,

respectively. The shape of the deflector was made aerodynamic to reduce the amount of

energy loss in air deflection. The height of the plate had to be limited to allow sufficient

horizontal motion of the plate and also to achieve reasonable air variability at the curved

periphery of the sprayer.

A similar experiment as described in previous section was conducted, by fixing the

deflector plate at different horizontal locations and measuring the air output at outlet.

Five horizontal locations were chosen to model the air output at the periphery. Based on

preliminary experiments described in section 3.3.2, it was found that the average air

volume coming out at the sprayer outlet on each side was about 7.67 m3/s. When the

deflector plate was installed on the left side of sprayer, the air output on this side was










split in two parts. A portion of it was discharged vertically while the other portion was

deflected horizontally toward the trees. Table 3.1 shows the characteristics of the

deflected air at each plate position when measured at the outlet. Since this outlet is

curved, as the plate position changes horizontally from 1-5, the deflected air comes

through an increasing outlet area. To account for this increase in area, the vertical

(peripheral) grid points over which data was measured was increased. It can be seen that

air volume and velocity increased as the deflector plate changed its position from setting

1 to setting 5. The results from this experiment demonstrated that the deflector plate

could regulate the amount of air at output of the sprayer (Figure 3.6).

Table 3.1. Characteristic of the horizontally deflected air due to various plate positions at
sprayer outlet
Deflector plate position Mean air velocity Mean air volume Measurement grid points
(m/s) (m3/s) (vertical x horizontal)
1 17.48 1.91 9x5
2 27.43 4.13 11x5
3 34.84 5.81 13 x 5
4 38.15 7.02 15 x 5
5 44.15 7.63 17 x 5
Note: Readings were based upon experiments described in section 3.3.3 with 3
replications. Vertical points are along the periphery of the sprayer while horizontal points
are along the width of the air outlet.


























Deflected air / '""



Shaft fix
deflectoi



\ Plate positions

Figure 3.5. Schematic view of the deflector plate motion

10


y = 0.02x + 2.43
S8


6 Std. Dev.


4
- frc
Gi th,
N -
zz


-Nozzles


0 50 100 150 200 25C
Plate position from outermost location (mm)


Figure 3.6. Effect of deflector plate location on air output










3.4 Automation of Deflector Plate Movement

An electromechanical system was designed and implemented to automate the

horizontal movement of the deflector plate only on the left side of the sprayer to test this

idea. The control objective for this system was to have horizontal motion of the plate as a

function of the density reading obtained from the laser sensor for each tree. The

procedure for real-time tree parameter data collection using laser sensor is discussed in

section 3.5. The system designed (Figure 3.7) consists of an actuator, which converts

electromagnetic energy into mechanical energy; a controller, which is the heart of any

control system; and a mechanical linkage, which takes an input and produces a different

output by changing the motion, velocity or acceleration of the input.


GPS












Controller Step Motor


Figure 3.7. Components of the control system


Laser sensor









Actuator: The actuator used in this electromechanical system was a step motor (AMH23-

258-3, Advanced Micro System, Inc., Nashua, NH). This type of motor provides precise

positioning of the deflector plate. Like conventional motors, a step motor also converts

electromagnetic energy into mechanical energy but the difference being that it is done in

steps. This essentially means that power to this motor can be sent in pulses which results

in precise motion of the shaft. The motor used here was a 1.80 or 200 steps per revolution

motor. This step motor received signal commands from a step motor controller (mStep-

407, Advanced Micro Systems, Inc. Nashua, NH) which is described below. Detailed

specifications of the motor are listed in Appendix-B. The step motor was connected to a

36 V power supply (three 12 V batteries in series). Motor sizing calculations have been

provided in Appendix-A.

Controller: The mSTEP-407 (Figure 3.8) is an on-board intelligent step motor

controller. The choice of this controller was based on two main reasons. The first reason

was the necessity for it to integrate easily with the existing system. The laser sensor

developed for sensing the tree parameters used a laptop to calibrate and collect data.

Since, the controller includes a serial link communication port, it was convenient and

cost-effective to accept commands from the laser sensor, process and send pulse signals

to the step motor. The second reason was to have a controller that not only improved the

accuracy but also made motion of the step motor smooth. This can be achieved by micro-

stepping which involves sending pulses that will rotate the motor in fractions of its steps.

The controller has the feature of one-tenth micro-stepping resulting in higher accuracy

and smoother rotary motion of motor. Specifications of the mSTEP-407 controller board










are provided in Appendix-B. The controller board needed an 8-15 V logic power supply

which was obtained from one of the three batteries used to power the motor.









Power
Input



Serial Switch
port



Figure 3.8. The step motor controller, mSTEP-407 (Advanced Micro Systems, Inc)

Mechanical Linkage: Mechanical linkages are a fundamental part of machine design.

The function of the mechanical linkage designed here was to convert rotary motion of the

step motor to horizontal motion of deflector plate. Figure 3.9 shows a schematic of this

actuation mechanism. Due to the limited space availability in the existing sprayer and its

ability to handle high torques, a ball screw and a nut assembly (Part number:

HL5134M20452, Techno Inc., New Hyde Park, NY) were used. Also, ball screw is a

very efficient and cost effective choice to position moving parts accurately. The

specifications of the ball screw have been given in Appendix-B. A cross bar connected

the nut on the ball screw with two shafts (Figure 3.5) on deflector plate. Two guide

rollers (Part number: VW-1, Modern Linear Inc., Corte Madera, CA) on shafts, fixed at

12 and 69 cm measured from the bottom, bear the weight of the deflector plate by riding

along a track (Part number: T-4, Modern Linear Inc., Corte Madera, CA) fixed on to the

sprayer. The step motor was coupled with ball screw using a right angle drive (Part









number: A 2Z28MC1010, 1:1 precision right angle drive,) and a cross joint type flexible

coupling (Part number: S50MCTM25P08P10).

It is important to estimate the expected loads and apply sufficient safety factor to

select linkages. Based on air measurements, it was expected that highest horizontal load

on the plate was about 221 N. This is based on a maximum air reading of 2.73 kPa and an

area of 0.08 m2 on the rear of deflector plate. To account for overloads, a safety factor of

10-15% was considered and hence, a maximum load of 250 N was used to design other

components of this system. A typical procedure used in designing mechanical motion

systems is to have a weak component in the mechanical linkage. This is done so that, in

case of heavy overload, the system breaks at that link thus protecting the more expensive

parts. In order to transmit the motion in normal circumstances, a coupling that had torque

limit of 2 N-m was used in this system. In situations when the deflector plate can get

stuck the coupling, which costs lesser, would break and stop the transmission but protect

the other expensive components from damage. Teflon sheets were placed on the sprayer

wall to reduce the friction that resulted when the plate rides along the wall.






































*-Pnng CUU llll- Air
Ball nut Ball crew drive ir




-- ----- Track ,r







12.7 cm



Figure 3.9. Schematic of actuation mechanism for the deflector plate









Input Output Relations of the Control System. The electromechanical system

can be characterized by five mathematical (Table 3.2) functions which provide a relation

between the input and output at various points.

Table 3.2. Input output relations of the control system

Nm Input Output y
Number -
(x) (y) x

Density reading from the Indexing value to the 240000
laser sensor controller board 40

Indexing value to the Rotation of motor shaft 1
2 controller board (revolutions) 2000

Rotation of motor shaft Rotation of screw rod 1
3 (revolutions) (revolutions) 1

Rotation of screw rod Horizontal movement of 5
4 (revolutions) plate (mm) 1

Horizontal movement of Air output to the tree 2.43
S plate (mm) (m/s) 0.02 +



Figure 3.10 shows the relation between indexing commands that were given by the

controller after getting density signal from laser sensor. The range 0.6-0.8 % for tree

density was chosen, based on trial runs of laser sensor for a particular row, to obtain

higher resolution of plate movement. Density readings less than 0.6 were assigned plate

location 0 mm (horizontal location 1) while readings greater than 0.8 were given 200 mm

(horizontal location 5). A linear relation was chosen for the purpose of simplicity. The

step motor indexing commands are based on 200 steps (1.80) per revolution step motor

and one-tenth resolution of the controller board. Hence, an indexing command of 2000

would rotate the motor shaft by one revolution.










Once the tree density (x,) from the laser sensor is obtained (explained in section


3.4), the value of transfer function number one in the above table can be calculated. This

is multiplied by transfer function numbers two, three and four in table 3.2 to obtain the

horizontal movement of the plate (x, ). An estimate of the horizontal airflow from the


lower part of the output can be obtained from Figure 3.6. Again, a linear equation was

used to estimate this for simplicity purpose. Based on this model, we can relate the

change in density value obtained from the laser sensor to change in horizontal airflow

output from the sprayer.

100
100 240000---------------------'----------------------
CO
S80 y = 400000x -240000
(D

o) 60

-D
S40
0

20 -

0
0.6 0.65 0.7 0.75 0.8 0.85
Tree Density from Laser Sensor (%)


Figure 3.10. Relationship between indexing value to the controller board required for a
range of laser sensor density reading




3.5 Real-time Collection of Tree Parameters Using Laser Sensor

Among the various ways to measure the characteristics of tree canopy, a laser

sensor is by far the most accurate. Salyani and Wei (2005) have shown the algorithm to

measure the height, volume and density of each scan based on mathematical approach.

The values of these three features which, henceforth, will be referred to as tree









parameters from each scan, along with their individual GPS location values were needed

for post-processing to get information about a tree. To eliminate the manual post-

processing, lasers sensor algorithm was enhanced to accurately collect tree parameter

data depending on tree spacing of that particular row. Tree parameter data collected over

a tree was later used to decide the amount of air necessary for that particular tree.

In order to get the start and end of each tree, latitude-longitude information from

GPS input signal was used to continuously calculate distance traveled by laser sensor

from its starting position. This distance was constantly compared with tree spacing data,

until tree spacing length was reached, to decide the start time and stop time for collecting

data for a particular tree. This is the default method for data collection.

Sometimes under dense canopy cover GPS signal might be unreliable. In order to

ensure validity of the GPS signal, a check was included in the laser program. To ensure

that the tree parameter data is accurate regardless of unreliable GPS signal, the user is

prompted to input expected nominal travel speed initially. Then the time required to

travel each tree is calculated. An in-built software timer counts in steps equal to the

nominal time required to travel successive trees, signaling that the laser sensor has

reached the end of tree. If GPS signal is unreliable, the system switches to this method of

data collection, thus, ensuring continuity. The program is restored back to its default

method as soon as the GPS signal is valid.

Another feature added to the program was collection of data when resolution of

GPS is not fine enough. At times, it is possible that distance calculated might not be equal

to the exact tree spacing distance. For example, for a 4.6 m tree spacing the travel

distance calculated have readings of 3.7, 4.0, 4.3 and 4.9 m thus skipping the 4.6 m









reading at which the program should have completed collection for that particular tree

spacing. This situation was handled by providing a range of 0.3 m to ensure that correct

data was collected for each tree spacing. A parallel time-based check was also

implemented if the resolution degraded to more than 0.3 m. This can be understood more

clearly with an example. If for a 4.6 m tree spacing row, the GPS did not log distances

between 4.3 and 4.9 m, then a time based check would complete the collection of data for

the tree using GPS speed information and tree spacing data. The procedure used is the

same as the one described above for the case when GPS signal is unreliable.

At the end of each tree spacing an average of the density readings is calculated.

This average density is used to obtain a corresponding motor indexing command. A serial

port communication feature was added to the laser program to send this motor indexing

command to the step motor with accurate amount of turns to move the deflector plate.

With these improvements to the laser program, the electromechanical system and

laser sensor were integrated with negligible delay between detecting the density of tree

and the deflector plate movement.

3.6 Experiment 1: Air Penetration through Tree Canopy

The objective of this experiment was to test the effectiveness of the plate locations

on air penetration through tree canopies with different densities. The velocity of air

passing through trees with four different densities at five horizontal locations of the

deflector plate was measured. The trees were visually selected based on observation in

increasing order of densities and are termed as dead canopy (D), low density (L), medium

density (M) and high density (H) trees (Figure 3.11).






































(c) (d)
Figure 3.11. Tree canopies of different densities. (a) Dead canopy (b) Low density (c)
Medium density (d) High density

Mean and maximum air velocity measurements were made with a hot film

anemometer (FlowMaster, Type 54N60, Dantec Measurement Technology, Denmark) at

a distance of 4.8 m from the center of the sprayer and at a height of 1 m from the ground

(Figure 3.12). The mean and max air velocities were taken over a period of 10 s while the

sprayer traveled across a 4.6 m tree. Measurements were taken while the sprayer was

drawn by a tractor at PTO speed of 540 rpm and at ground speeds of 2.15 and 4.73 km/h

heading in the east direction on a row that was aligned in the east-west direction.

Measurements were made in four replications.

























S 1.8m 3.0 m
Figure 3.12. Experimental setup for measuring air velocity. Note: Anemometer was held
perpendicular to the holder and going in to the page.

Weather data was obtained at a height of 2 m from the ground from FAWN

(Florida Automated Weather Network) for interval of the test. Ambient air temperature,

relative humidity, wind speed and direction during the experiment were 9.9-21.7C, 46-

92%, 1.34-4.02 m/s and 44-118o (0 represents north and 900 represents east),

respectively.

Data Analysis. The experiment was conducted as a Randomized Complete Block

Design (RCBD). Mean and maximum air velocities were analyzed using analysis of

variance. Interaction between factors was analyzed by considering this design as a three

factor split-split plot experiment and data was analyzed using MIXED procedure in SAS

(Freund et al., 1986). The three factors considered were sprayer ground speed, deflector

plate setting and tree density. Two sprayer ground speeds (2.15 and 4.73 km/h) divided

each of the four blocks replicationss) into 8 whole plots. Each plot was further divided in

five split plots by randomly assigning five deflector plate settings. Each split plot was

divided into four split-split plots, and four tree densities (dead, low, medium and high

density) were randomly assigned. A grand total of 160 measurements were available for









analysis. Means were separated using LSMEANS with PDIFF option at 5% level of

significance.

3.7 Experiment 2: Effect of Deflector Plate on Spatial Distribution of Spray
Deposition

The objective of this experiment was to determine the effect of sprayer air volume

rate on spatial distribution of spray droplets. Spray deposition was quantified by having

paper targets (Fisherbrand filter paper, Fisher Scientific, Pittsburgh, PA) at nine spatial

locations perpendicular to the direction of the travel of the sprayer. The sprayer was

operated with six open nozzles on the left side since they were directly affected by the

change in air volume due to the horizontal position of the deflector plate. Two types of

nozzles: lilac and blue Albuz APT cone nozzles (Ceramiques Techniques, Desmarquest,

France) were used to see the effect of low and high volume application, respectively. The

measured discharge rates of the six nozzles, at about 1000 kPa pressure, were 2.9 and

21.4 L/min respectively. These volumes corresponded to application rates of 215 and

1585 L/ha based on row spacing of 6 m and ground speed of 2.7 km/h. Five locations of

deflector plate, labeled 1, 2, 3, 4 and 5 corresponded to an air volume rate 1.9, 4.1, 5.8,

7.0 and 7.6 m3/s, respectively, of air deflected towards the targets at PTO speed of 540

rpm. The test structure made from PVC piping consisted of nine horizontal locations

labeled A to I at distances of 2.4, 3, 3.9, 4.8, 6.0, 7.2, 8.4, 9.6 m from the point of

discharge and at a height of 1.5 m from the ground (Figure 3.13). The travel direction

was towards west on a row that was oriented in the east-west direction.

Spray solutions contained Pyranine-1OG fluorescent dye (Keystone Aniline, Inc.,

Chicago, IL) as deposition tracer at a constant rate of 566 mg/L (ppm).Water sensitive

paper (Spraying Systems Co., Wheaton, IL) targets were also placed on each location to









visually compare it with results from paper targets. These targets were held by target

holders to keep them steady and perpendicular to spray direction during each sprayer

pass. The exposed area on the paper targets was 42.77 cm2. After each sprayer pass,

targets were immediately placed in sealable plastic bags, and stored in an enclosed

container for further laboratory analysis. Experiments were made in four replications. In

the laboratory, spray (dye) deposits on each target were quantified by fluorometry

(Salyani, 2000b). The deposits were normalized for differences in the application volume

rate (L/ha).

Ambient air temperatures and relative humidity were monitored at a height of 1.8

m, using a temperature/RH indicator (Model 870H, General Eastern, Watertown, MA).

Wind speed was measured using a vane anemometer (Model HH-30, Omega

Engineering, Stamford, CT) at the same height. A white ribbon tied to a pole was used to

visually note the wind direction. The ranges of temperature, relative humidity and wind

speed during the experiment were 4.2 26.0C, 24.3 52.2% and 0.2 2.5 m/s,

respectively. The wind direction was primarily from the north and north-east direction

and did not seem to have any significant effect on spraying as magnitude of winds were

low during the entire experiment.









6.1 m


I H G F


paper


Top view


PVC


Side view

Figure 3.13. Schematic view of spray application experiment and sampling layout



Data Analysis. The experiment was conducted as a Randomized Complete Block Design

(RCBD). Mean tracer depositions at near (ABC), far (DEFGHI) and at each target location,

were analyzed using analysis of variance. Analysis was done by considering this

experiment as a three factor split-split plot experiment and using MIXED procedure in






35


SAS (Freund et al., 1986). The three factors considered were nozzle type, deflector plate

setting and target distances. Two nozzle types (lilac and blue, with application rates of

215 and 1585 L/ha, respectively) divided each of the four blocks replicationss) into 8

whole plots. Each plot was further divided in five split plots by randomly assigning five

deflector plate settings. Each split plot was divided into two split-split plots by

considering two target locations (near and far). Means were separated using LSMEANS

with PDIFF option at 5% level of significance.














CHAPTER 4
RESULTS AND DISCUSSION

This chapter begins with an evaluation of the designed control system, followed by

a discussion of the results from the experiments described in the previous chapter.

4.1 Evaluation of the Electromechanical Control System

The electromechanical system designed in the previous chapter successfully

performed its function of adjusting the output of air from the sprayer. This system was

evaluated under the following topics:

Performance: The control system was able to automate the plate movement

satisfactorily. Figure 4.1 shows the data from one of the runs made in a row of 51 trees.

On the right Y-axis are the density values (shown as dots on graph) of 51 trees obtained

from laser sensor. The range of densities obtained for this trial run was 0.65 to 0.86%.

The left Y-axis shows actual plate positions (shown as bars on graph), ranging from 0 to

200 mm that were prescribed based on the corresponding density values. Density

readings above 0.80% were assigned plate position 5 which corresponds to 200 mm. It

can be observed from the graph that the change in density reading data corresponds

closely to change in plate movements. Further implications of the change in plate

movement on the air volume and spray deposition will be discussed in detail in sections

4.2 and 4.3. The response time of the plate to move between extreme ends, that is, when

moving from high density to low density tree, was seen to be less than 3 seconds.











250 0.9
0 Plate position
0 Tree density 0.85
200

0.8
150
0
Si0.75


. 0.7

50
0 0.65



1 5 10 15 20 25 30 35 40 45 50
Tree number

Figure 4.1. Relation between actual plate position and tree density

User-friendliness: The control program prompts the user to enter the tree spacing

distance and a nominal speed initially. After this no operator intervention is required

while spraying. A software switch is provided within the Graphical User Interface (GUI)

of the program to stop the process, should the operator choose to do so. Also the program

makes a log of each tree parameter data, plate movement and time information. These can

be used in further analysis. A hardware switch (Figure 3.8) was also installed which made

it easy to start or stop the controller manually.

Integration: The control system that was implemented for this project was easily

integrated with the existing sprayer. In cases where a control system is to be implemented

to test a particular idea, it is important to design with reversibility in mind. In other

words, if the automation of a particular system does not prove a certain idea it should be

fairly easy to revert to the original system. The airblast sprayer at plate setting 5 gives

almost the same air output as the sprayer would in normal circumstances. This allows the









sprayer to be used with the current method, but also enables testing with different air

volumes.

4.2 Experiment I: Air Velocity Measurements

The aim of this experiment was to quantify the effect of deflector plate position on

mean and maximum air velocity measurements through trees of different densities at two

different ground speeds.

Figure 4.2 shows the mean air velocities recorded over a 10 s interval while the

sprayer passed trees of varying densities with five plate settings and two speeds. It must

be noted that the dead tree chosen for this test had a dense tree preceding it. The foliage

of this dense tree was extending into the dead tree spacing. While taking measurement it

was likely that mean velocity averaged over this tree spacing would result in lower

values. Nonetheless, it can give a fair idea of dynamics of the plate in modifying the

mean air output. Figure 4.3 shows the maximum air velocity profile for trees with varying

density due to change in deflector plate setting. In general as the density of the tree

increased, mean and maximum air velocity across them decreased. However, for a

particular tree, as the deflector plate position changed from 1 to 5 the mean and

maximum velocities increased. For example it was found that, for the dead tree at 2.15

km/h, there was significant increase in maximum air velocity between plate setting pairs

1-2, 1-3, 1-4 and 1-5. It may be concluded that, when the sprayer encounters a dead tree it

may be beneficial to have plate setting 1. At 4.73 km/h, there was significant increase in

air velocity between the pairs 1-5, 2-5 and 3-5 for medium tree. For the dead tree, mean

air velocities at plate settings 1, 2 and 3 were significantly lesser compared to plate

settings 4 and 5. It may be concluded that by using lower plate setting at 4.73 km/h,

significantly lesser air velocities can be obtained while spraying. Change in air velocity









also leads to change in volume as can be seen from Table 3.1. It was published that there

was little or no relation between air velocity and spray deposition on leaf samples

(Salyani and Hoffmann, 1996) while, results from Balsari et al. (2001) showed that

reducing the air volume lead to a better spray deposition. Nonetheless, an

electromechanical system like this can help in efficiently testing spray deposition with

different air volume and velocity.

Figure 4.4 shows the maximum air velocity averaged over all plate settings for each

type of tree to see the effect of ground speed. Overall for both the speeds, maximum air

velocity decreased as the density of the tree increased. Lower ground speed resulted in

greater maximum air velocity which agrees with results from Salyani and Hoffmann

(1996). There was also interaction effect between the sprayer ground speed and deflector

plate setting. Higher deflector plate setting resulted in increased reduction in maximum

air velocity from lower to higher ground speeds. But this interaction was inconsistent

with the high density tree as it resulted in almost same air velocities at both speeds. This

could be attributed to the dense foliage of the tree which substantially blocked the air

from coming onto the hot-wire anemometer, leading to the inconsistent result. It may also

be noted that some high volume of air may be necessary to move the heavy foliage and

transport the spray droplets for high density trees (Salyani and Farooq, 2003).

















2.5


2.0


1.5


1.0


-0.5


t 0,0
o

>2.5


S2.0


1.5


1.0


0.5


0.0


1 2 3 4 5 1 2 3 4 5
Plate position Plate position


Figure 4.2.


Mean air velocity due to different deflector plate location at 2.15 and 4.73 km/h. Lowercase bold, lowercase italics and
uppercase letters show mean separation between plate settings at 2.15, 4.73 km/h and combined ground speeds,
respectively

















6- A
AB AB


4-
B B T I- mI
I I 1




2-
b h ab a
A a
b ab b a ab
0 -


_6 E02 2.15 km/hr
SMedium Densitv I Hih Density Tree
El 4.73 km/hr

B A
-B AB r Mean separation

2. A A A A A



1 2 34 5 1 2 3.. 4 5
Plate position Plate position


Figure 4.3. Maximum air velocity due to different deflector plate location at 2.15 and 4.73 km/h. Lowercase bold, lowercase italics
and uppercase letters show mean separation between plate settings at 2.15, 4.73 km/h and combined ground speeds,
respectively








8

A0 2.15 km/-
/ -E 0 4.73 km/-
"- 6 Mean separation letter
"Std. dev.
O B

4- -

E1 C



2I T I

a b a br

dead low medium high
Tree density

Figure 4.4. Maximum air velocity at two ground speeds of the airblast sprayer. Note:
Averaged over five plate settings. Lowercase bold, lowercase italics and
uppercase letters show mean separation between different tree densities at
2.15, 4.73 km/h and combined ground speeds, respectively

4.3 Experiment II: Spray Deposition

The aim of this experiment was to quantify the effect of changing air volume and

application volume rate on the spatial distribution of spray droplets.

Plate setting 1 resulted in lower total deposition recovery from targets compared to

other plate settings due to its low air velocity profile. The air volume output decreases by

about 75% from plate setting 5 to 1. Hence, most of the spray droplets dropped to the

ground even before reaching the first target. In general, mean deposition decreased as the

target distance from the sprayer increased (Figure 4.6 and 4.7). Total depositions were

calculated as near (A-C) for the first three locations and far (D-I) for the last six

locations. These also signified spray material going to the tree and away from the tree,

respectively. Figure 4.5 shows the deposition at these two spatial sections for two

application volume rates due to different deflector plate settings. In all replications, the









deposition was higher in the near location than the far locations except for plate setting 1.

Plate setting l's air distribution was such that some of the output from the top most

nozzle (Figure 3.17) was exposed to high volume air that was vertically directed from

behind the deflector plate. This contributed to the excess deposition at far sections.

However, at plate settings 2-5 the trend reversed and there were more depositions at near

locations. Thus, depending on the spray requirement, particular plate setting can be

chosen to obtain variable deposition to the tree.

P-values of 0.09 at low application rate and 0.0001 at high application rate were

obtained between far and near location, respectively. It can be inferred that effect of

change in air volume on spray deposition is significant for high volume application

compared to low volume application. There was interaction effect between plate setting 2

and application rate which produced lower deposition at far location during high

application rate compared to other plate settings. Also, there were some interaction

effects between application volume rate and target location. For a particular plate setting,

it is natural to expect more deposition at near location compared to far location. However,

plate settings 3 and 4 resulted in lesser difference between near and far locations at high

application rate compared to other plate settings.

4.4 Discussion

The results presented in this research demonstrate that the modified sprayer could

customize the deposition by changing air output. Further testing can be done on the spray

deposition across a row of trees of variable densities while the plate moves in real time to

see the effect of deflector plate location on penetration of spray inside a canopy.

The distance between the laser sensor and the sprayer air outlet was 2.7 m (Figure

3.1). For a 3 m tree spacing, by the time the laser has obtained the tree data the air outlet









has already sprayed 10% of the tree spacing with different air volume. This can result in

inaccurate spraying for a portion of the tree. This can be improved by having the laser

sensor at a distance at least equal to the tree spacing of the block to be sprayed.

A 3-second travel time was achieved for plate travel between extreme ends. For a 3

m tree spacing and travel speed 1.61 km/h, it would take about 6 seconds to travel the

tree spacing. For the worst case, wherein, the sprayer has to move across a high density

tree to a dead tree, the travel time of the plate will result in 50% of the tree being in

transition period from full to least air volume. This 3 second travel time can be reduced

by using a superior motor or higher power, thus improving the design.











4--




3




2




01



0 -

CF )



3




2






0-





Figure 4.5.


Effect of two application volume rate on total deposition at two spatial
sections. Lowercase bold, lowercase italics and uppercase letters show mean
separation between plate settings at near (A-C), far (D-I) and combined (A-I)
locations, respectively


1 2 3 4 5







46
1.5


1.0
Plate 1
Std. dev.
0.5


0.0


1.0

Plate 2
0.5



0 .0 ,. ,










0.0

1.0


SPlate 4
0.5

0.0




1.0 A Target location
1.0












B Plate 5
0.5










0.5 C D

G H T
0.0
2 4 6 8 10 12
Distance from sprayer nozzles (m)
Figure 4.6. Low application rate: spatial distribution of deposition (see Figure 3.13) at
different deflector plate position







47


1.5


1.0
Std. dev. Pate 1
0.5


0.0


1.0

Plate 2
0.5


O 0.0

0
: 1.0
O
0.Plate 3
ePlate
B 0.5


0.0


1.0

Plate 4
0.5


0.0 ,- -
0 .0-. i.. .. i .,

1.0 Target location
1.0- A

B C Plate 5
0.5 D
E
C -H T
0.0, ,
2 4 6 8 10 12
Distance from sprayer nozzles (m)

Figure 4.7. High application rate: spatial distribution of deposition (see Figure 3.13) at
different deflector plate positions














CHAPTER 5
CONCLUSIONS

One of the objectives of my project was to be able to spray a particular row based

on the morphological characteristics of the tree. The electromechanical system developed

modifies the air characteristic and thus gives variable spray deposition. The density

values from the laser sensor were used as an input to achieve this variability. Evaluation

of my system shows that different settings of the deflector plate results in changing air

velocity through the canopy of different densities. A 61% reduction in air velocity was

obtained at the outlet of sprayer from outermost to innermost deflector plate position.

Further, spatial distribution of the spray droplets at different output volume was

determined. A change of plate setting from outermost to innermost yields about 37%

reduction in deposition at far location at high application rate. This results in some

reduction of off-target spraying. Since the volume of air coming out of sprayer depends

upon commands sent by the computer, it can be customized based on the requirement.

This is important as the idea of changing air volume in sprayer for variable rate

application is relatively new and hence may need fine tuning. A complementary system

that shuts off sections of nozzles based on tree height information from sensor will

further improve this system.















APPENDIX A
STEP MOTOR SIZING CALCULATION

The step motor used in this research project was selected based on calculations that

have been described below (Oriental Motor General Catalog, 2004) (Figure A-1).

Variables used in the equations are listed in Table A-1.

im





Programmable Step motor PB
controller _.r


Figure A. 1. Block diagram used for stepper motor calculations (Adopted from Oriental
Motor General Catalog, 2004)


Table A.1. Specifications of the actuation mechanism
Total mass of the plate m = 2
Ball screw efficiency 7 = (
Internal frictional coefficient of ball nut /o = C
Ball screw shaft diameter DB = 1
Total length of travel LB =2
Material of ball screw Steel
Density of steel p = 7
Pitch of the ball screw PB = 5
Step size of motor s = 1
Maximum load F =
Motor rotor inertia Jo = 3


.9 kg (6.5 lb)
).9
).3
6 mm (0.63 inch)
00 mm (7.87 inch)

'.8 g/cm3 (4.64 oz/in3)
mm (0.2 inch)
.80
250 N
43.9 g-cm2 (1.88 oz-in2)









Inertia:
Inertia of the ball screw (JB) is given by,


J = -pL BD = x 4.64 x 7.87 x 0.634 = 0.56 oz-in2
32 32


Inertia of the plate (Jp) is given by,
2
Jp= = 6.51b x 16 x 0 ev 0.11 oz-in
lb 2i

Total Inertia (Jr)

JT = J, + Jp = 0.56 + 0.11 = 0.67 oz-in2 = 122.5 g-cm2

The motor rotor inertia (JM) should not be very small compared to total inertia (Jr)

Jr
< 30 (Clarence, 2005)
JM
where JM = Motor inertia

J > 4.1 g-cm2

Operating Parameters:

Number of operating pulses (A) required for the entire travel length of the deflector plate

is given by,

Travellength (LB) 360" 7.87 3600
A = x = x = 7870 pulses
Ball screw pitch (PB) Step angle (8,) 0.2 1.80

This implies, 7870 pulses are required to move the deflector plate between extreme

positions. Based on several design iterations on the type of motor available, its size and

cost, it was decided to select a motor using a 3 s time interval (t) required to complete the

entire travel length (LB) The pulse frequency (f) is determined based on this positioning

period, t.









Number of Operating Pulses (A) 7870
f = 2623 Hz
Positioning Period (t) 3

Thus, the operating speed (N) of the step motor required to achieve this motion is given

by,

0, 1.8
N= fx x60 =2623 x x 60 = 787 rpm
360 360

Load Torque (TL):

The load torque is the maximum torque encountered by the motor while moving the

deflector plate. The deflector plate faces maximum air resistance when moving from

outermost position towards the innermost position. Based on the air measurements made

at sprayer outlet, it was found that the maximum load (F) on the plate was about 250 N.

The load torque is a summation of this load, F and the pilot load, Fp, due to friction of

ball nut bearings. A rule of thumb is to consider this pilot load to be approximately a third

F
of the load F. Hence, F = = 83.3 N.
3

FPB PoFpPB 250(5) 0.3(83.3)(5)
TL P + + =0.24 N-m
2ir7 2r 2((0.9) 2i

Acceleration torque (Ta)

Acceleration torque is the torque required by the motor to initially step up to the required

operating speed of 787 rpm. This is calculated based on the rotor inertia (Jo), total inertia

(Jr) and the pulse frequency (/). An acceleration time (tl) equal to 25% of positioning

time, t = 3 s is allowed to step up to the required operating speed.

SJo +Jr 7XTO f 1.88+0.67 ;(1.8) 7870 N-
Tx- x-- x- =2.17 oz-in. = 0.015 N-m
g 180 t, 386 180 0.75









Required Torque (T):

The total torque required by the step motor will be summation of the load torque (TL) and

the acceleration torque (Ta) considering a safety factor of 2.

T= (TL + Ta) x Safety factor = (0.24 + 0.015) x 2 = 0.5 N-m

Based on this requirement, availability and cost, the step motor with model number

AMH23-258-3 (Advanced Micro Systems, Inc., Nashua, NH) which had a torque

capacity of 1.82 N-m was chosen for the project. Further specifications of this motor have

been provided in Appendix-B.















APPENDIX B
COMPONENT SPECIFICATION


Table B.1. Ball screw and nut assembly specifications
Properties Details
Material Cf 53, induction hardened to HRC 62 + 2.
Pitch accuracy 0.1 mm/300 mm
Diameter 16 mm
Pitch 5 mm
Features Ball screw designed for greater than 90% efficiency in
converting rotary to linear motion. Ball nut is designed
to prevent backlash


Table B.2. Step motor technical specification
Properties
Model no.
No. Of leads
Phase connection
Holding torque
Inertia
Weight


Details
AMH23-258-3
8
Parallel
258 oz-in
1.88 oz-in2
2.2 lb


Table B.3. Step motor controller features
Properties Details
Model No. mStep-407
Current 7 A
Voltage 24 V (dc) to 80 V (dc)
Resolution 1/10
Communication RS232/Serial Port
Memory 2 kb non-volatile
















LIST OF REFERENCES


Balsari, P. and M. Tamagnone. 1998. The necessity to determine the correct amount of
air to use in airblast sprayer. Paper 98-A-075, Proceedings ofInternational
Conference on Agricultural Engineering, Oslo, Norway.

Balsari, P., G. Oggero and M. Tamagnone. 2001. The choice of air volume rate for
mistblowers in apple orchards. ATWSymposium, Stuttgart, Germany.

Balsari, P., M. Tamagnone and P. Marucco. 2003. Innovative technologies for
orchard/vineyard sprayers. VII Workshop on Spray Application Techniques in Fruit
Growing. 35-42.

Clark, R. L. and R. L. McGuckin. 1996. Variable rate application technology: an
overview. Proc. 3rd International Conference on Precision Agriculture WI:
ASA/CSSA/SSSA.

Clarence, W. S. 2005. Mechatronics: an integrated approach, CRC Press LLC.

Cugati, S. A., W. M. Miller, A. W. Schumann and J. K. Schueller. 2006. Dynamic
characteristics of two commercial hydraulic flow-control valves for a variable-rate
granular fertilizer spreader. ASABE Paper No. 061071. St. Joseph, Mich.: ASABE.

Escola, A., F. Camp, F. Solanelles, S. Planas and J. R. Rosell. 2003. Tree crop
proportional spraying according to the vegetation volume. First results. VII
Workshop on Spray Application Techniques in Fruit Growing. 43-49.

Freund, R. J., R. C. Littell and P. C. Spector. 1986. SAS systems for linear models. Cary,
NC: SAS Inst. Inc.

Gebhardt, M. R., C. L. Day, C. E. Goering and L. E. Bode. 1974. Automatic sprayer
control system. Transactions of the ASAE 17(6):1043-1047.

Ghate, S. R. and C. D. Perry. 1994. Ground speed control of pesticide application rates in
a compressed air direct injection sprayer. Transactions of the ASAE 37(1):33-38.

Giles, D. K. and J. A. Comino. 1990. Droplet size and spray pattern characteristics of an
electronic flow controller for spray nozzles. J. Agric. Engng. Res. 47:249-267.

Giles, D. K., M. Delwiche and R. Dodd. 1987. Control of orchard spraying based on
electronic sensing of target characteristics. Transactions of the ASAE 30:1624-
1630.









Giles, D. K., G. W. Henderson and K. Funk. 1996. Digital control of flow rate and spray
droplet size from agricultural nozzles for precision chemical application.
Proceedings of the ThirdInternational Conference on Precision Agriculture, ASA,
CSSA, SSSA, Madison, WI. 729-738.

Han, S., L. Hendrickson and B. Ni. 2000. A variable rate application system for sprayers.
Proceedings of the Fifth International Conference on Precision Agriculture, ASA,
CSSA, SSSA, Madison, WI.

Landers, A. J. and E. Gil. 2006. Development and validation of a new deflector system to
improve pesticide application in New York and Pennsylvania grape production
areas. ASABE Paper No. 061001. St. Joseph, Mich.: ASABE.

Miller, D. R., E. W. Huddlestin, J. B. Ross and W. E. Steinke. 2003. Airblast spray
partitioning in a mature pecan orchard. Transactions of the ASAE 46(6):1495-1501.

Oriental Motor U.S.A Corp. 2004. General Catalog. F2 F6.
http://www.orientalmotor.com/products/pdfs/F_TecRef/TecMtrSiz.pdf(Last
accessed: March 4th, 2007).

Reichard, D. L., H. J. Retzer, L. A. Liljedahl and F. R. Hall. 1977. Spray droplet
distribution delivered by air blast orchard sprayers. Transactions of the ASAE
20(1):232-237,242.

Salyani, M. 1988. Droplet size effect on spray deposition efficiency of citrus leaves.
Transactions of the ASAE 31:1680-1684.

Salyani, M. 1993. Degradation of fluorescent tracer dyes used in spray applications.
ASTMInternational, West Conshohocken, PA: ASTM STP 1183.13:215-226.

Salyani, M. 2000a. Optimization of deposition efficiency for airblast sprayers.
Transactions of the ASAE 43(2):247-253.

Salyani, M. 2000b. Methodologies for assessment of spray deposition in orchard
applications. ASAE Paper No. 00-1031, St. Joseph, Mich.: ASAE.

Salyani, M. 2003. Calibration of airblast sprayers. EDIS AE238.
http://edis.ifas.ufl.edu/AE238 (Last accessed: March 11th, 2007).

Salyani, M. and M. Farooq. 2003. Sprayer air energy demand for satisfactory spray
coverage in citrus applications. Proc. Fla. Stat Hort. Soc. 116:298-304.

Salyani, M. and M. Farooq. 2004. Drift potential of citrus air-carrier sprayers. Proc. Fla.
State. Hort. Soc. 117:130-135.

Salyani, M. and R. D. Fox. 1999. Evaluation of spray quality by oil- and water-sensitive
papers. Transactions of the ASAE 42(1):37-43.









Salyani, M. and W. C. Hoffmann. 1996. Air and spray distribution from an air-carrier
sprayer, Applied Engineering in Agriculture 12(5):539-545.

Salyani, M. and J. Wei. 2005. Effect of travel speed on characterizing citrus canopy
structure with a laser scanner. Precision Agriculture '05. J. V. Stafford, Ed. 185-
192.

Salyani, M. and J. D. Whitney. 1988. Evaluation of methodologies for field studies of
spray deposition. Transactions of the ASAE 31(2):390-395.

Salyani, M. and J. D. Whitney. 1990. Ground speed effect on spray deposition inside
citrus trees. Transactions of the ASAE 33(2):361-366.

Salyani, M. and J. D. Whitney. 1991. Effect of oscillators on deposition characteristics of
an airblast sprayer. Transactions of the ASAE 34(4):1618-1622.

Salyani, M., E. BenSalem and J. D. Whitney. 2002. Spray deposition and abscission
efficacy of CMN-pyrazole in mechanical harvesting of valencia oranges.
Transactions of the ASAE 45(2):265-271.

Salyani, M., Y. M. Koo and R. D. Sweeb. 2000. Spray application variables affect air
velocity and deposition characteristics of a tower sprayer. Proc. Fla. State Hort.
Soc. 113:96-101.

Salyani, M., R. D. Sweeb and M. Farooq. 2006. Comparison of string and ribbon
samplers in orchard spray applications. Transactions of the ASAE 49(6): 1705-1710.

Sawyer, J. E. 1994. Concepts of variable rate technology with considerations for fertilizer
application. Journal of Production Agriculture 7:195-201.

Stover, E. 2002. Sensor-controlled spray systems for florida citrus. EDIS HS-140.
http://edis.ifas.ufl.edu/HS140 (Last accessed: March 10th, 2007).

Stover, E., D. Scotto and J. Salvatore. 2002a. Spray applications to citrus: survey of
current practices in the Indian river area. EDIS HS-127.
http://edis.ifas.ufl.edu/HS127 (Last accessed: March 10th, 2007).

Stover, E., D. Scotto, C. Wilson and M. Salyani. 2002b. Spray applications to citrus:
overview of factors influencing spraying efficacy and off-target deposition. EDIS
HS-128. http://edis.ifas.ufl.edu/HS128 (Last accessed: March 10th, 2007).

Tangwongkit, R., V. Salokhe and H. Jayasuriya. 2006. Development of a real time,
variable rate herbicide applicator using machine vision for between-row weeding of
sugarcane fields. Agricultural Engineering International: the CIGR Ejournal.
Manuscript PM 06 009. Vol. VIII.






57


Wei, J. and M. Salyani. 2004. Development of a laser scanner for measuring tree canopy
characteristics: phase 1. prototype development. Transactions of the ASAE
47(6):2101-2107.

Wei, J. and M. Salyani. 2005. Development of a laser scanner for measuring tree canopy
characteristics: phase 2. foliage density measurement. Transactions of the ASAE
48(4):1595-1601.

Whitney, J. D., S. L. Hedden, D. B. Churchill and R. P. Cromwell. 1986. Performance
characteristics of PTO airblast sprayers for citrus. Proc. Fla. Stat Hort. Soc 99:59-
65.

Zhang, N., M. Wang and N. Wang. 2002. Precision agriculture-a worldwide overview.
Computers and electronics in agriculture. 36:113-132.















BIOGRAPHICAL SKETCH

Naresh Pai was born in 1981, in Mumbai in Maharashtra, India, to Archana and

Anant Pai. He received his Bachelor of Engineering degree in mechanical engineering

from the Manipal Institute of Technology, India, in June 2003. He then enrolled in

the Graduate School of the University of Florida. He expects to receive a Master of

Science degree in agricultural engineering and a concurrent Master of Science

in mechanical engineering degree in May 2007. After completing his Master of Science

degrees, he plans to work in the field of controls and automation using the technical skills

he acquired during his studies.




Full Text

PAGE 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 FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2007

PAGE 2

2007 Naresh Pai

PAGE 3

To Archana and Anant Pai

PAGE 4

iv ACKNOWLEDGMENTS I would like to extend my thanks to the many individuals who have contributed to make this project a success and my educationa l experience so enjoya ble. I would like to thank Dr. Masoud Salyani, my advisory committee chairman, for his faith in me, and for his continued support and inspiration. His constant encouragement, timely critical evaluation, and enthusiasm for my work have resulted in the successful completion of my research. I am indebted to Mr. Roy Swee b, Senior Engineering Technician, for his insightful ideas, hands-on support, and traini ng in the workshop and field, throughout the work. I would also like to thank Dr. Thomas Burks and Dr. John Schueller for giving me valuable knowledge through their courses and serving on my supervisory committee. I want to acknowledge the Agricultural and Biological Engineering (ABE) department for providing me the opportunity, and the Citrus Re search and Education Centre (CREC) for the assistantship and technical resources to cond uct my research. I am also grateful to Dr. Reza Ehsani, Mr. Troy Gainey, and the staff of the CREC maintenance department for letting me use equipment needed for the project. On a personal note, I would like to th ank my parents and brother whose support was of inestimable value. A final word of thanks goes out to all my friends who have directly or indirectly co ntributed to the successful completion of my work.

PAGE 5

v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT....................................................................................................................... ..x CHAPTER 1 INTRODUCTION........................................................................................................1 1.1 Justification.............................................................................................................1 1.2 Thesis Organization................................................................................................2 2 BACKGROUND..........................................................................................................3 2.1 Pesticide Usage in Florida.....................................................................................3 2.2 Pesticide Application Technology for Tree Crops................................................4 2.2.1 Sprayers.......................................................................................................4 2.2.2 Air-Carrier Sprayers....................................................................................4 2.2.2.1 Liquid delivery system......................................................................5 2.2.2.2 Air delivery system...........................................................................6 2.2.3 Testing Methodologies for Sprayer Air and Liquid Output........................8 2.3 Control Systems in Pestic ide Application Technology..........................................9 2.4 Objectives.............................................................................................................12 3 MATERIALS AND METHODS...............................................................................13 3.1 Airblast Sprayer Description................................................................................13 3.2 Laser Sensor..........................................................................................................14 3.3 Preliminary Experiments......................................................................................15 3.3.1 Airblast Spray Distribution Pattern............................................................15 3.3.2 Restricted Air-Input Test............................................................................16 3.3.3 Deflecting Air at Output.............................................................................18 3.4 Automation of Deflector Plate Movement...........................................................21 3.5 Real-time Collection of Tree Pa rameters Using Laser Sensor.............................27 3.6 Experiment 1: Air Pene tration through Tree Canopy...........................................29

PAGE 6

vi 3.7 Experiment 2: Effect of Deflector Plate on Spatial Distribution of Spray Deposition...............................................................................................................32 4 RESULTS AND DISCUSSION.................................................................................36 4.1 Evaluation of the Electromechanical Control System..........................................36 4.2 Experiment I: Air Velocity Measurements...........................................................38 4.3 Experiment II: Spray Deposition..........................................................................42 4.4 Discussion.............................................................................................................43 5 CONCLUSIONS........................................................................................................48 APPENDIX A STEP MOTOR SIZING CALCULATION................................................................49 B COMPONENT SPECIFICATION.............................................................................53 LIST OF REFERENCES...................................................................................................54 BIOGRAPHICAL SKETCH.............................................................................................58

PAGE 7

vii LIST OF TABLES Table page 3.1 Characteristic of the horizontally deflected air due to various plate positions at sprayer outlet............................................................................................................19 3.2 Input output relations of the control system.............................................................26 A.1 Specifications of the actuation mechanism..............................................................49 B.1 Ball screw and nut assembly specifications.............................................................53 B.2 Step motor technical specifications..........................................................................53 B.3 Step motor controller features..................................................................................53

PAGE 8

viii LIST OF FIGURES Figure page 3.1 Schematic of the PowerBlast airblast sprayer..........................................................14 3.2 Schematic view of spraying application...................................................................15 3.3 Airblast sprayer with different air intake area..........................................................17 3.4 Effect of fan inlet diameter on th e air output of airblast sprayer.............................18 3.5 Schematic view of the deflector plate motion..........................................................20 3.6 Effect of deflector pl ate location on air output.........................................................20 3.7 Components of the control system...........................................................................21 3.8 The step motor controller, mSTEP-407...................................................................23 3.9 Schematic of actuation mechanism for the deflector plate......................................25 3.10 Relationship between indexing value to the controller board required for a range of laser sensor density reading.................................................................................27 3.11 Tree canopies of different densities..........................................................................30 3.12 Experimental setup for measuring air velocity.........................................................31 3.13 Schematic view of spray applica tion experiment and sampling layout...................34 4.1 Relation between actual plate position and tree density...........................................37 4.2 Mean air velocity due to different defl ector plate location at 2.15 and 4.73 km/h..40 4.3 Maximum air velocity due to different deflector plate lo cation at 2.15 and 4.73 km/h..........................................................................................................................4 1 4.4 Maximum air velocity at two ground speeds of the airblast sprayer.......................42 4.5 Effect of two application volume rate on total deposition at two spatial sections...45

PAGE 9

ix 4.6 Low application rate: spatial distributi on of deposition at different deflector plate position............................................................................................................46 4.7 High application rate: spatial distributi on of deposition at different deflector plate positions...........................................................................................................47 A.1 Block diagram used for stepper motor calculations ................................................49

PAGE 10

x Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science CONTROLLING AIRBLAST SPRAYER AIR FOR VARIABLE RATE APPLICATION IN ORCHARDS By Naresh Pai May 2007 Chair: Masoud Salyani Major: Agricultural and Biological Engineering Spray requirements vary considerably throughout the grove on account of variability in citrus canopy size and foliage de nsity. Configuring sprayers to suit this tree variability is vital for e fficient spraying. Currently, crops are sprayed uniformly throughout the field based on experience. Unif orm application of agrochemicals not only wastes chemicals but also has environmental implications. At present, airblast sprayers account for majority of sprayers for tree crop application in Florida. While moving across th e grove, these sprayers rely on a stream of air supply generated by fan(s) to carry the material from the nozzles to the canopy. The air volume generated in these sp rayers range between 3.7 46.7 m3/s. A fully grown tree has a different spray requirement as compared to a small or medium sized tree. It has also been observed that a typical airblast sprayer may deposit 2-3 rows beyond the immediate row for which it was intended. Since the tree si ze and density distri bution on the field are

PAGE 11

xi usually non-uniform, uniform spraying could result in substantial material losses as ground fallout and drift. My work describes design modification of an existing airblast sprayer to test the idea of variable rate spray a pplication by adjusting the air ou tput. My project involved the design, implementation and testing of an electro mechanical system to change the volume of air going to the trees based on the tree dens ity information from a laser sensor. This process gives real time change in air output characteristics as the sprayer moves across a row of trees. By using this system, the ai r volume can be changed from 1.9 to 7.6 m3/s in less than 3 seconds. Air pene tration was quantified by meas uring air velocity which revealed that different settings of this air regulatory system can produce significantly different air characteristics across a dead and medium density tree. Overall, increase in air volume gave higher spatial distributi on of spray material but for high volume application changing air volume can produce sign ificant high deposits in near locations when compared to far locations t hus reducing off-target wastage.

PAGE 12

1 CHAPTER 1 INTRODUCTION 1.1 Justification Florida is the second largest producer of c itrus in the world and accounts for more than one-third of the worlds grapefruit produ ction. These trees are mostly sprayed with air-carrier sprayers. These sprayers use hi gh volume and velocity air streams, produced by axial-, centrifugal-, or cross-flow fans, to transport the spray dr oplets to tree canopies. The aim is to replace the air within the trees by a stream of air and agrochemical droplets. Spray requirements of a fully grown mature tree is substantially different from those of medium or small sized tree. In addi tion, it is very common to find dead trees and resets in the field. Efficient spraying in agriculture involves op timum usage of the available resources. The ideal spraying deposits the material on the intended target, based on the type of tree and minimizes ground deposit a nd drift. In contrast, at full air capacity the sprayers could deposit the material 2-3 rows beyond the row for which it was intended. This results in cons iderable amount of wastage of agrochemicals which has economical as well as environmental concerns. Many researchers and companies have successfully shown the advantage of customizing the liquid output from nozzles based on the characteristics of the trees. However, there have been fewer attempts to mo dify the air that carries the spray droplets to the trees. The overall objectiv e of this project is to evalua te the idea of variable rate spray application by changing the air volume that is used to transport these spray droplets.

PAGE 13

2 1.2 Thesis Organization My thesis is divided into five chapters The second chapter, Background, deals with relevant literature review and past work done in this area. The third chapter, Materials and Methods, describes the progressive steps taken to develop a system to address the problem described in section 1.1. Subsequent ly, it also describes the experiments conducted to test this system. The fourth chapter, Results and Discussion, evaluates the control system, and further, discusses the resu lts that were obtained from the experiments described in chapter three. The fifth chapter, Conclusions, gives an overall perspective of this research project.

PAGE 14

3 CHAPTER 2 BACKGROUND This chapter gives the background and la ys a foundation for this project. It starts with statistics of the agricultural pesticide usage in Florida to recognize the significance of developing t echnology which can maximize application efficiency and minimize wastage. Pesticide application t echnology is then discussed to give a general overview of equipment used for tree crops. Focus has been placed on airblast sprayers as the aim of this project was to optimize air output in such sprayers. Finally, control systems in airblast sprayers used fo r variable rate appli cation are discussed. 2.1 Pesticide Usage in Florida Pesticides are agrochemicals that are in tended for controlling pests. Pesticides in the form of sprays are commonly used to control pests that affect citrus. In 1999, Florida Agricultural Statistics Services reported that he rbicides were used on 96.3% of the 316,840 ha of citrus and avocado. Sim ilarly, insecticides and fungicides were applied on 91.3 and 81.9% of the acreage resp ectively in the same year. This shows the widespread use of pesticides for tree crops. Though the intention is to spray the material to the target, a c onsiderable amount of pesticid e is wasted. These wastages arise primarily from three sources: Off-target spraying Drift to air Runoff from leaves falling on the ground It was reported that abou t 22% of the spray material is wasted to the ground and, about 21% drifted to air with commonly us ed airblast sprayer at application rate

PAGE 15

4 of 950 L/ha (Miller et al., 2003). This not only has economic disadvantages, but also leads to ground and surface water pollu tion, air pollution, disturbance in the neighboring ecosystem and may also presen t health concern to humans. Hence, efforts are being made to optimize the amount of pesticide usage in agriculture. 2.2 Pesticide Application Technology for Tree Crops Pesticide application tec hnology refers to the equipments that are used to dispense pesticides. Pesticide applicators fo r tree crops are characterized by their high volume application to cover the dense foliage of trees. Pesticides are applied in solid or liquid formulation with or without assistance of air. Pest icides in solid formulation are applied either in granular or powdered form. These applicators are called granular pesticide applicators or dusters, respectiv ely. Pesticides that are used in liquid formulation are applied using sprayers. Co mmonly, sprayers use a ssistance of air to carry the spray droplets, and for bette r canopy penetration. These sprayers are designated as air-carrier sprayers. 2.2.1 Sprayers Sprayers use hydraulic systems to transpor t and spray pesticides to the target. These hydraulic systems comprise a tank to store the liquid formulation, a pump to develop the necessary pressure, manifolds to transport the pre ssurized liquid and nozzles to convert the liquid into droplet s and disseminate the pesticide. These sprayers have a boom to suppor t the nozzles at the outlet. 2.2.2 Air-Carrier Sprayers Air-carrier sprayers account for 89% of sp ray machines for citrus production in Florida (Stover et al., 2002a ) and hence, are importan t equipment for pesticide application. These sprayers are commonly used for spraying tree crops having large

PAGE 16

5 foliage. They use a stream of air to tran sport spray droplets towards the tree and penetrate the canopy. The air in and around the trees is thus replaced with a mixture of air and pesticide droplets. This method of application is consider ed superior as it increases the deposition (Rei chard, 1977; Salyani, 1988). These sprayers are available in different shapes, sizes, fan types, no zzles and are operated with different volume rates and ground speeds. The working of an air-carrier sprayer sy stem can be broadly divided into two modules. The first one comprises of the com ponents that handle the delivery of liquid and is referred to as the liquid deliver y system. The second module relates to the delivery of air and is ca lled air delivery system. 2.2.2.1 Liquid delivery system The task of liquid delivery system in an ai r-carrier sprayer is to store, pressurize and produce droplets of the liquid pesticid e formulation. This is achieved by a hydraulic circuit which employs a pump to pressurize and force the liquid from the tank to the nozzles. The nozzles aligned on th e sprayer can be categorized to be in three sections namely; top, middle and bottom, reflecting the section of the tree that the spray material affects. These nozzles produce droplets of liquid pesticide which are transported to the target due to the functioning of air delivery system. The number of nozzles, nozzle size, no zzle pressure, and ground speed of sprayer play an important role in the deposit ion efficiency of a sprayer. These have to be adjusted to maximize the efficiency of spraying. For instance, it was suggested by Salyani (2000a) that for be tter deposition at lo wer application rates, reducing the number of nozzles and using smaller nozzl es would be advantageous rather than driving the sprayer at higher speed. On th e other hand, for higher volume applications

PAGE 17

6 it would be helpful to use more nozzles and higher speed rather than using large disc and core sizes. Investigation on the effect of ground speed of ai rblast sprayer (50 m3/s capacity) on spray deposition by Salyani and Whitney (1990) revealed that increasing the speed does not necessarily reduce the deposition. Canopy structure in a grove is diverse a nd hence, uniform spraying can result in some spray losses. To optimize the usage of pesticide on tree crops according to canopy makeup, the general trend among re searchers and commercial sprayer manufacturers has been to develop spray systems which match the spray output with canopy structure. Many researchers have successfully shown the advantage of spraying tree crops according to vegeta tion volume (Balsari and Tamagnone, 1998; Balsari et al., 2003; Escola et al., 2003; Giles et al., 1987). Commercial systems like Ropers Tree-SeeTM (Roper Grower Cooperative, Winter Garden, FL), Durand Waylands SmartSprayTM (Durand-Wayland Inc., LaGrange, GA) and AgTechs Tree-SenseTM (AgTech Inc., Manhattan, KS) have implemented the same by using sensors that target specific zones on a tree. The nozzles ar e then activated by electric solenoid valves so that only the zone detected by the sensors is sprayed. 2.2.2.2 Air delivery system Air delivery system in air-carrier spra yers is responsible for producing high volume and velocity airflow to transport liquid droplets from the nozzles to the trees. The main components in the air delivery sy stem include a fan, airflow straightener and air deflectors. Various types of fans that have been used in airblast sprayers are axial-, centrifugaland tangential-flow fa ns. Axial flow fans are most popular for their large volume and low pressure applica tions. Airblast sprayers are a type of aircarrier sprayers that use such fans. The fa ns consist of a series of radial blades

PAGE 18

7 attached to a rotating hub. This assembly of blades and hub is termed impeller or rotor. Air drawn by the rotor is discharged by the tangential component of velocity. This results in a swirling motion of the air, commonly known as slipstream rotation. The efficiency of the fan decreases w ith swirling as the air encounters more resistance. This swirl is removed by the st ator or straightener placed downstream of the rotor. The dynamic pressure developed here is converted to static pressure rise. The air is then deflected towards the nozzl es by a set of deflector surfaces (plates), sometimes, by about 90 to target trees lateral to the sprayer. It was manually observed that the spra y can reach 2-3 rows beyond where it is intended. This was partially due to the large amount of air that is uniformly used to transport the spray to the target. Whitney et al. (1986) have reported that Power-TakeOff (PTO) powered airblast sprayers can have airflow rate from 3.77 to 25.01 m3/s. It has also been suggested that such high air volume is justif ied only for large and densely foliated trees (Balsari et al., 2001; Salyani and Farooq, 2003). Additionally, such high volume can lead to drift of spray material into the air. An experiment was conducted by Salyani and Farooq (2004) to quantify and compare drift potential of the commonly used airblast sprayers. Due to the radial discharge of the PowerBlast sprayer (used in this project), it had hi ghest above canopy drift compared to other sprayers. However, it was seen that lo wer ground speeds (2.4 to 2.8 km/h) of the sprayer produced higher spra y deposition as compared to higher ground speeds (4.8 to 5.8 km/h) (Salyani et al., 2000). Salyan i and Hoffmann (1996) reported that, in general, the velocity of air reduced at increasing distance fr om the sprayer. Further, it was also found that air velocity from a traveling sprayer had lesser magnitude

PAGE 19

8 compared to a stationary spra yer. It should be noted that these tests can be affected by wind speeds and direction. Efforts to optimize the material transport from the nozzles of airblast sprayers can be complemented by a system that can control the amount of air that transports this spray material based on the morphological characteristics of tr ees. In this regard, Balsari et al. (2003) attempted to change the air output by using an adjustable diaphragm at the axial fan inlet, but no r eal experiments were reported. Also, this method is not suitable because equal amount of air will be output from each side of the sprayer. This defeats the purpose of pr oviding variable air output for trees of different physical characteristics on each si de of the sprayer as it travels through the grove. Landers and Gil (2006) te sted an air deflector system which directed the air horizontally into the canopy on both sides of the sprayer. It was reported that a 25% improvement in deposition could be achieved using this system. 2.2.3 Testing Methodologies for Sprayer Air and Liquid Output An important aspect of testing airbla st sprayer efficiency is to choose appropriate methodology to quantify the spra y deposits and air characteristics. There are several factors that can affect the resu lts obtained from such experimental design and hence are reported along with the other re sults of the experiment. Some of these factors are: ground speed of sprayer; number of nozzles; nozzle ty pe; nozzle pressure; nozzle orientation; air velocity and volu me; type and location of targets; and environmental conditions like temperat ure, relative humidity, wind speed and direction. Spray applications have been quantifie d using several methods. Each of these methods has certain advantages and disadva ntages. Hence, it is important to have

PAGE 20

9 some idea about each of them before selecting an appropriate method. Fluorescent tracers and fluorometry have largely been used to quantify deposition and drift of spray material. A problem associated with this method is that the commonly used water soluble dyes are prone to degradati on under solar radiation (Salyani, 1993). An alternative could be to use metallic tracers such as copper which do not degrade under sunlight. The deposits in this case are quantified using colorimetry. To catch the tracer in spray, a variety of targets have been used. Though leaf samples are ideal for simulating the actual target, artificial targets such as paper, mylar, etc. can provide certain advantages in qu antifying the spray deposition. Salyani and Whitney (1988) conducted an experiment to compare th e deposition on leaf sa mples with mylar targets using fluorometry and colorime try. They published a correlation of R2=0.90 using colorimetry, and R2=0.85 using fluorometry between leaf and mylar target deposits. Salyani and Fox (1999) compared oil and water sensitive papers as targets. They reported major challenges in handling these targets because of their sensitivity to air temperature, humidity and opera tor error. Additionally, for high volume applications it might be difficult to quantif y the spray amount as the targets become over-covered with droplets. A compar ison was made between string and ribbon samplers by Salyani et al. (2006) in fiel d applications. Spray mixtures in this experiment consisted of fluorescent trace r at different volume rates and ground speeds. It was reported that string samplers had higher capture efficiencies compared to ribbon samplers for all sample locations. 2.3 Control Systems in Pestic ide Application Technology A control system is an integration of several electrical/mech anical components used to regulate a desired output. Control systems have been traditionally used to

PAGE 21

10 automate processes in various areas of th e industry. In precision agriculture, they have been used to realize some of the goa ls of variable rate application technology. Control systems can be broadly classified as closed loop and open loop systems. In open loop systems the controller directly gives commands to an actuator without receiving feedback about the ac tuators previous state. Th is form of control can be applied when the actuation required is not very accurate. Closed loop control systems continuously monitor the commands sent ba sed on the feedback from the actuators previous state and information regarding the present state. This form of control can result in higher accuracy and faster re sponse (Cugati et al., 2006, Gebhardt et al., 1974). The main components of a clos ed loop control system are plant, computer/controller and sensor. The plan t includes a set of electromechanical components which act upon electrical signals sent by the contro ller to perform its function. Computer/controllers are electronic devices that control the actuator and indirectly vary the applica tion rate of the products bei ng applied based on information from several sources such as the appli cation equipment itself or other sensors. Controllers form the fundamental component of any variable rate application system (Clark & McGuckin, 1996). These cont rollers are typically driven by a microprocessor that works based on a set of rules or algorithm. A sensor is a transducer that is used to measure a physic al quantity such as temperature, pressure, etc. and convert it into an electrical signal. Sensors can also measure a particular state of the plant and give feedback si gnal to the computer/controller. Traditionally, in sprayers, control systems have been used to automatically control nozzle discharge rate in the li quid delivery system (Stover, 2002). The

PAGE 22

11 traditional approach to handle this has been to regulate the pressure across the nozzle (Giles et al., 1996). But this form of cont rol can have significantly delayed response time resulting in poor perfor mance (Han et al., 2000). To counter this, pulse-width modulation (PWM) has been used to control electrically actu ated solenoid valves that are connected to the nozzles (Giles & Comino, 1990, Han et al., 2000). Electrical solenoid valves can give considerable shorter response time compared to conventional pressure-based flow control system. Variability in application system can be initiated by two approaches: map based application and sensor based application. Sensor based application, used in this project, has an advantage over map based a pplication due to higher accuracy (Sawyer, 1994) and real time control (Zhang et al., 2002). Gebhardt et al. (1974) developed an automatic sprayer control system that ch anged the output from the nozzle based on the ground speed of the sprayer. A tachomet er generator sensor provided dc voltage to a gear motor which in turn controlled th e metering valve in real time at the output of the spray tank. Ghate & Pe rry (1994) developed a sim ilar system where a radar sensor was used to sense ground speed whic h varied pesticide application rate by controlling a 12 V dc step motor. Tangwongk it et al. (2006) used a software based machine vision system that sensed gree nness level of weeds to spray herbicide accordingly. The machine vision sensor was connected to a laptop which sent commands to a PWM circuit which in turn controlled a dc electric motor. A laser sensor has been developed that can give th e height, volume and density of each scan on a laptop in real time as it travels in the grove (Wei and Salyani, 2005). This sensor reportedly gave better results at low speed (1.6 km/h) as compared to high speed (3.2

PAGE 23

12 km/h) due to higher number of scans it made at the lower speed (Salyani and Wei, 2005). This research project aims at deve loping a control system to regulate the amount of air in airblast sprayers to co mplement some of the systems designed to control liquid flow. Regulating the amount of air using control system in a sprayer introduces a new scenario and provides diffe rent challenges. The airflow from the sprayer is at high velocity and turbulent in nature. The high inertia of axial fan restricts any possibility of re ducing the speed in real time for smaller trees. Moreover, the correct amount of air needed to spray a particular tree may not be known. This necessitates the development of a control system that can be reprogrammed and is adaptable to changing air volume based on di fferent sensor inputs relating to tree characteristics. 2.4 Objectives Specific objectives of this project are: To design and fabricate an electromech anical system by which the amount of air going to the trees can be regulated. To integrate the signal from a laser sens or with the air regulator system to enable variable rate spraying in the field. Evaluate the functionality of the de veloped system through air velocity measurements across citrus canopies with different foliage densities. Determine spatial distribution of the sp ray droplets at different air volume output.

PAGE 24

13 CHAPTER 3 MATERIALS AND METHODS This chapter documents materials and met hods that were used in the design and implementation and testing of an electromechan ical air control system for the airblast sprayer. 3.1 Airblast Sprayer Description The sprayer used for this project was PowerBlast airblast sprayer (Model No. PB533ST, Rears Manufacturing Company, Eugene OR). A schematic of this airblast sprayer is shown in Figure 3.1. It is PTO-driven and has a single axial flow fan. This fan has 9 blades with a diameter of 0.84 m and pitch of 32. The sprayer uses the tractor Power Take Off (PTO) power through a Constant Velocity (C.V.) joint and 3-point hitch connection. The PTO driveline transfers heavy torque loads from the tractor to the axial flow fan, which is operated by an electrical clutc h. The speed of the fan, at P.T.O speed of 540 rpm, is around 2160 rpm. The fan rotates in a counterclockwise direction looking from the rear of the sprayer. It is followed by a 24 blade flow straightener unit. The air outlet of the sprayer has an inverted U-shape slot of 144.0 x 12.7 cm on each side along its periphery. There are 24 hydraulic nozzles on each side of the sprayer. For the fan configuration mentioned above and for PT O speed of 540 rpm, air passes over the nozzles at a maximum speed of about 188 km/h (PTO PowerBlast manual, Rears Manufacturing Company, Eugene, OR). Under st andard settings of fan, the total volume rate of air output is about 16m3/s (approximately 33,901 cfm).

PAGE 25

14 Figure 3.1. Schematic of the PowerBlast airblast sprayer 3.2 Laser Sensor The sensor used in this research proj ect was developed by Wei and Salyani (2004; 2005) to measure tree height, canopy volume and fo liage density. It uses an infrared laser emitter with a wavelength of 780 nm. A line s canner consisting of a motor, with an incremental encoder, rotates the mirror that deflects the outgoi ng beam of the laser emitter by 90 and sweeps it through 360 as it rotates. The return ing beam from the target is deflected off the mirror ba ck to a photodiode in the sensor. The laser sensor has two RS232 interface cab les which perform different functions. One cable is used to calibrate the laser se nsor using the COM1 port on a laptop while the second cable connects to a High Speed InterFac e (HSIF) Card and gives pulse signals. Laser sensor 2.7 m Flow straightene r Axial flow fan Tan k PTO 3 point hitch Hydraulic nozzle 2.4 m Air

PAGE 26

15 Distance measurements are made by analyzing these pulse signals. Since the laptop does not have slots for additional cards, a PCI to Cardbus adapter is used to establish communication between HSIF card and laptop using the PCMCIA port. The HSIF card also controls the motor and records its pos ition using an encoder through a parallel (DB25) port. The laser sensor was mounted to the front side of the sprayer on a vertical pole at about 2.4 m from the ground (Figure 3.1). 3.3 Preliminary Experiments It was discussed in the earlier chapter th at air usage in airblast sprayer was not optimum. To support this claim, several te sts were conducted. These experiments and their conclusions have been deta iled in the sections below to assert the progressive nature of this research project. 3.3.1 Airblast Spray Distribution Pattern An initial visual assessment of the spra ying pattern of the ai rblast sprayer was made. Blue and lilac Albuz APT cone nozzl es (Ceramiques Techniques, Desmarquest, France) were used on left and right side of the sprayer to observe the effect of high and low volume rate spraying (F igure 3.2), respectively. Figure 3.2. Schematic view of spraying a pplication. (Adopted from Salyani (2003)) High volume rate nozzles Low volume rate nozzles

PAGE 27

16 The following visual observations were made while spraying with the above configuration: 1. Deposition on trees was very good on immediate row of trees. The spray evaporated very fast with lilac nozzles because of the small size of droplets. 2. Spray from blue nozzles moved up to 2 rows beyond the immediate row while for the lilac nozzles it moved only by 1 row. 3. For both nozzles, considerable amount of material fell on the ground and some droplets were sucked into the fan 4. Many droplets sprayed from the upper section of the nozzles went up in air without hitting any target whic h could cause drift. 3.3.2 Restricted Air-Input Test A possible solution to some of observations above was to adjust the air output to the nozzles. Since droplets use air as a medium to transport them to target, adjusting air flow can reduce some of the errors in spra ying. To test this idea an experiment was conducted to reduce the volume of air at input in steps and quantify the output at the air outlet. Annulus shaped wooden boards (Figure 3.3) of increasing diameter were cut and attached to sprayer so as to change the am ount of air input to the fan. The cuts were curved to reduce turbulence at the edges. The four holes in the concentric boards had an intake area of 0.1, 0.21, 0.36 and 0.55 m2. A pitot tube manometer was used to measur e air pressure across a virtual grid of 10 x 5 points on each side of the sprayer air out let. The ten points were across the periphery of the sprayer while five points along the widt h of the outlet. Measurements were made in three replications while the sprayer was sta tionary. An extra set of reading was taken without any obstruction. Figure 3.3 shows progressive pictures of the boards that were used to take air measurements.

PAGE 28

17 (a) (b) (c) (d) Figure 3.3. Airblast sprayer with di fferent air intake area (a) 0.1 m2 (b) 0.21 m2 (c) 0.36 m2 and (d) 0.55 m2 It was concluded that restri cting the amount of air at fa n input provided a means of regulating air at output (Figure 3.4). But, a nnulus shaped wooden boards was not a viable option because it resulted in equal amount of air volume on each side of the sprayer. In general, trees may be of different shapes and si zes on either side of the sprayer. Hence, in order to achieve the goal of optimizing the air output, it was essent ial to regulate the amount of air independently on both side of the sprayer.

PAGE 29

18 Figure 3.4. Effect of fan inlet diameter on the air output of airblast sprayer 3.3.3 Deflecting Air at Output A deflector plate was fabricated (Figure 3.5) using sheet metal. It was placed in the space between the fan and the air outlet. Its horizontal positions from the outermost (1) to innermost (5) would adjust the amount of air output fr om minimum to maximum, respectively. The shape of the deflector wa s made aerodynamic to reduce the amount of energy loss in air deflection. The height of th e plate had to be limited to allow sufficient horizontal motion of the plate and also to achie ve reasonable air variability at the curved periphery of the sprayer. A similar experiment as described in pr evious section was conducted, by fixing the deflector plate at different hor izontal locations and measuri ng the air output at outlet. Five horizontal locations were chosen to model the air output at the periphery. Based on preliminary experiments described in sect ion 3.3.2, it was found that the average air volume coming out at the sprayer outlet on each side was about 7.67 m3/s. When the deflector plate was installed on the left side of sprayer, th e air output on this side was 0246810 0.10 0.21 0.36 0.55Intake area (m2 )Air Flow (m3/s) Right Side Left Side No obstruction

PAGE 30

19 split in two parts. A portion of it was discha rged vertically while the other portion was deflected horizontally toward the trees. Table 3.1 shows the ch aracteristics of the deflected air at each plate position when meas ured at the outlet. Since this outlet is curved, as the plate position changes horiz ontally from 1-5, the deflected air comes through an increasing outlet area. To account for this increase in area, the vertical (peripheral) grid points over which data was measured was increased. It can be seen that air volume and velocity increased as the de flector plate changed its position from setting 1 to setting 5. The results from this experi ment demonstrated that the deflector plate could regulate the amount of air at output of the sprayer (Figure 3.6). Table 3.1. Characteristic of the horizontally defl ected air due to various plate positions at sprayer outlet Deflector plate position Mean air velocity (m/s) Mean air volume (m3/s) Measurement grid points (vertical x horizontal) 1 17.48 1.91 9 x 5 2 27.43 4.13 11 x 5 3 34.84 5.81 13 x 5 4 38.15 7.02 15 x 5 5 44.15 7.63 17 x 5 Note: Readings were based upon experime nts described in section 3.3.3 with 3 replications. Vertical points ar e along the periphery of the sp rayer while horizontal points are along the width of the air outlet.

PAGE 31

20 Figure 3.5. Schematic view of the deflector plate motion Figure 3.6. Effect of deflector plate location on air output y = 0.02x + 2.43 0 2 4 6 8 10 05010015020025 0 Horizontal airflow from lower part of the air outlet Horizontal air output (m3/s) Plate position from out ermost location (mm) Std. Dev. Shaft fixed on the deflector p late 80 cm 57 cm Fan air Deflected air N ozzles Plate p ositions

PAGE 32

21 3.4 Automation of Deflector Plate Movement An electromechanical system was desi gned and implemented to automate the horizontal movement of the deflect or plate only on the left side of the sprayer to test this idea. The control objective for this system wa s to have horizontal motion of the plate as a function of the density reading obtained fr om the laser sensor for each tree. The procedure for real-time tree parameter data co llection using laser sens or is discussed in section 3.5. The system designed (Figure 3.7) consists of an actua tor, which converts electromagnetic energy into mechanical energy; a controller, which is the heart of any control system; and a mechanical linkage, wh ich takes an input and produces a different output by changing the motion, velocity or acceleration of the input. Figure 3.7. Components of the control system Controlle r Ste p Moto r GPS Laser senso r

PAGE 33

22 Actuator : The actuator used in this electromech anical system was a step motor (AMH23258-3, Advanced Micro System, Inc., Nashua, NH). This type of motor provides precise positioning of the deflector plate. Like conven tional motors, a step motor also converts electromagnetic energy into mechanical energy but the difference being that it is done in steps. This essentially means that power to th is motor can be sent in pulses which results in precise motion of the shaft. The motor us ed here was a 1.8 or 200 steps per revolution motor. This step motor received signal comm ands from a step motor controller (mStep407, Advanced Micro Systems, Inc. Nashua, NH) which is described below. Detailed specifications of the motor are listed in Appe ndix-B. The step motor was connected to a 36 V power supply (three 12 V batteries in se ries). Motor sizing cal culations have been provided in Appendix-A. Controller: The mSTEP-407 (Figure 3.8) is an on-board intelligent step motor controller. The choice of this controller wa s based on two main reasons. The first reason was the necessity for it to integrate easily with the existing system. The laser sensor developed for sensing the tree parameters us ed a laptop to calibrate and collect data. Since, the controller include s a serial link communication port, it was convenient and cost-effective to accept commands from the la ser sensor, process and send pulse signals to the step motor. The second reason was to have a controller that not only improved the accuracy but also made motion of the step motor smooth. This can be achieved by microstepping which involves sending pulses that will ro tate the motor in fractions of its steps. The controller has the feature of one-tenth micro-stepping resulting in higher accuracy and smoother rotary motion of motor. Specifi cations of the mSTEP-407 controller board

PAGE 34

23 are provided in Appendix-B. The controller board needed an 8-15 V logic power supply which was obtained from one of the three batteries used to power the motor. Figure 3.8. The step motor controller, mS TEP-407 (Advanced Micro Systems, Inc) Mechanical Linkage: Mechanical linkages are a fundame ntal part of machine design. The function of the mechanical linkage designed here was to convert rotary motion of the step motor to horizontal motion of deflector plate. Figure 3.9 shows a schematic of this actuation mechanism. Due to the limited space av ailability in the existing sprayer and its ability to handle high torques, a ball screw and a nut assembly (Part number: HL5134M20452, Techno Inc., New Hyde Park, NY) were used. Also, ball screw is a very efficient and cost effective choice to position moving parts accurately. The specifications of the ball screw have been given in Appendix-B. A cross bar connected the nut on the ball screw with two shafts (F igure 3.5) on deflector plate. Two guide rollers (Part number: VW-1, Modern Linear Inc., Corte Madera, CA) on shafts, fixed at 12 and 69 cm measured from the bottom, bear the weight of the de flector plate by riding along a track (Part number: T-4, Modern Linear Inc., Corte Madera, CA) fixed on to the sprayer. The step motor was coupled with ball screw using a ri ght angle drive (Part Serial port Switch Power Input

PAGE 35

24 number: A 2Z28MC1010, 1:1 precision right angle drive,) and a cross joint type flexible coupling (Part number: S50MCTM25P08P10). It is important to estimate the expected lo ads and apply sufficient safety factor to select linkages. Based on air measurements, it was expected that highest horizontal load on the plate was about 221 N. This is base d on a maximum air reading of 2.73 kPa and an area of 0.081m2 on the rear of deflector plate. To acc ount for overloads, a safety factor of 10-15% was considered and hence, a maximu m load of 250 N was used to design other components of this system. A typical proce dure used in designing mechanical motion systems is to have a weak component in the mechanical linkage. This is done so that, in case of heavy overload, the system breaks at th at link thus protecti ng the more expensive parts. In order to transmit the motion in nor mal circumstances, a coup ling that had torque limit of 2 N-m was used in this system. In situations when the deflector plate can get stuck the coupling, which costs lesser, would break and stop the transmission but protect the other expensive components from damage. Teflon sheets were placed on the sprayer wall to reduce the friction that resulte d when the plate rides along the wall.

PAGE 36

25 Figure 3.9. Schematic of actuation m echanism for the deflector plate Guide roller Ball screw Flexible coupling Track Ball nut Right angle drive Cross bar Rear wall of the air outlet Torque limiting coupling Step Motor Teflon lining Deflector plate Air 12.7 cm 57 cm Flow straightener Fan blade Rear wall slot

PAGE 37

26 Input Output Relations of the Control System. The electromechanical system can be characterized by five mathematical (T able 3.2) functions wh ich provide a relation between the input and out put at various points. Table 3.2. Input output relati ons of the control system Number Input ( x ) Output (y) x y 1 Density reading from the laser sensor Indexing value to the controller board 1240000 400000 x 2 Indexing value to the controller board Rotation of motor shaft (revolutions) 2000 1 3 Rotation of motor shaft (revolutions) Rotation of screw rod (revolutions) 1 1 4 Rotation of screw rod (revolutions) Horizontal movement of plate (mm) 1 5 5 Horizontal movement of plate (mm) Air output to the tree (m3/s) 543 2 02 0 x Figure 3.10 shows the relation between i ndexing commands that were given by the controller after getting density signal from laser sensor. The range 0.6-0.8 % for tree density was chosen, based on trial runs of la ser sensor for a particular row, to obtain higher resolution of plate movement. Density r eadings less than 0. 6 were assigned plate location 0 mm (horizontal location 1) while r eadings greater than 0.8 were given 200 mm (horizontal location 5). A linear relation was chosen for the purpose of simplicity. The step motor indexing commands are based on 20 0 steps (1.8) per revolution step motor and one-tenth resolution of the controller board. Hence, an indexing command of 2000 would rotate the motor shaft by one revolution.

PAGE 38

27 Once the tree density (1x) from the laser sensor is obtained (explained in section 3.4), the value of transfer function number one in the above table can be calculated. This is multiplied by transfer function numbers two, three and four in table 3.2 to obtain the horizontal movement of the plate (5x). An estimate of the horizontal airflow from the lower part of the output can be obtained from Figure 3.6. Again, a linear equation was used to estimate this for simplicity purpose. Based on this model, we can relate the change in density value obtained from the laser sensor to change in horizontal airflow output from the sprayer. Figure 3.10. Relationship between indexing va lue to the controller board required for a range of laser sensor density reading 3.5 Real-time Collection of Tree Parameters Using Laser Sensor Among the various ways to measure the characteristics of tree canopy, a laser sensor is by far the most accurate. Salyani and Wei (2005) have shown the algorithm to measure the height, volume and density of each scan based on mathematical approach. The values of these three features which, henceforth, will be referred to as tree 0 20 40 60 80 100 0.60.650.70.750.80.85Tree Density from Laser Sensor (%)Step Motor Indexing Value x 10 3 240000 400000 x y

PAGE 39

28 parameters from each scan, along with their individual GPS location values were needed for post-processing to get information about a tree. To eliminate the manual postprocessing, lasers sensor algorithm was enha nced to accurately collect tree parameter data depending on tree spacing of that particul ar row. Tree parameter data collected over a tree was later used to decide the amount of air necessary for th at particular tree. In order to get the start and end of each tree, latitude-longitude information from GPS input signal was used to continuously ca lculate distance traveled by laser sensor from its starting position. This distance wa s constantly compared with tree spacing data, until tree spacing length was reached, to decide the start time and st op time for collecting data for a particular tree. This is the default method for data collection. Sometimes under dense canopy cover GPS signa l might be unreliable. In order to ensure validity of the GPS signal, a check wa s included in the laser program. To ensure that the tree parameter data is accurate regard less of unreliable GPS signal, the user is prompted to input expected nominal travel speed initially. Then the time required to travel each tree is calculated. An in-built so ftware timer counts in steps equal to the nominal time required to travel successive trees, signaling that the laser sensor has reached the end of tree. If GPS signal is unrelia ble, the system switches to this method of data collection, thus, ensuring continuity. Th e program is restored back to its default method as soon as the GPS signal is valid. Another feature added to th e program was collection of data when resolution of GPS is not fine enough. At times, it is possible that distance calculated might not be equal to the exact tree spacing distance. For exam ple, for a 4.6 m tree spacing the travel distance calculated have read ings of 3.7, 4.0, 4.3 and 4.9 m thus skipping the 4.6 m

PAGE 40

29 reading at which the program should have comp leted collection for that particular tree spacing. This situation was handled by providing a range of 0.3 m to ensure that correct data was collected for each tree spaci ng. A parallel time-based check was also implemented if the resolution degraded to mo re than 0.3 m. This can be understood more clearly with an example. If for a 4.6 m tree spacing row, the GPS did not log distances between 4.3 and 4.9 m, then a time based check would complete the co llection of data for the tree using GPS speed information and tree spacing data. The procedure used is the same as the one described above for the case when GPS signal is unreliable. At the end of each tree spacing an average of the density readings is calculated. This average density is used to obtain a corresponding motor indexing command. A serial port communication feature was added to the laser program to send this motor indexing command to the step motor with accurate amount of turns to move the deflector plate. With these improvements to the laser program, the electromechanical system and laser sensor were integrated with negligible delay between detecti ng the density of tree and the deflector plate movement. 3.6 Experiment 1: Air Penetration through Tree Canopy The objective of this experiment was to test the effectiveness of the plate locations on air penetration through tree canopies with different densities. The velocity of air passing through trees with four different dens ities at five horizont al locations of the deflector plate was measured. The trees were visually selected based on observation in increasing order of densities and are termed as dead canopy (D), low density (L), medium density (M) and high density (H) trees (Figure 3.11).

PAGE 41

30 (a) (b) (c) (d) Figure 3.11. Tree canopies of different densities. (a) Dead canopy (b) Low density (c) Medium density (d) High density Mean and maximum air velocity measurements were made with a hot film anemometer (FlowMaster, Type 54N60, Dant ec Measurement Technology, Denmark) at a distance of 4.8 m from the cen ter of the sprayer and at a he ight of 1 m from the ground (Figure 3.12). The mean and max air velocitie s were taken over a period of 10 s while the sprayer traveled across a 4.6 m tree. Measur ements were taken while the sprayer was drawn by a tractor at PTO speed of 540 rpm and at ground speeds of 2.15 and 4.73 km/h heading in the east direction on a row that was aligned in the east-west direction. Measurements were made in four replications.

PAGE 42

31 Figure 3.12. Experimental setup for measuri ng air velocity. Note: Anemometer was held perpendicular to the holder and going in to the page. Weather data was obtained at a height of 2 m from the ground from FAWN (Florida Automated Weather Network) for interval of the test. Ambient air temperature, relative humidity, wind speed and direction during the experiment were 9.9-21.7C, 4692%, 1.34-4.02 m/s and 44-118 (0 represents north a nd 90 represents east), respectively. Data Analysis. The experiment was conducted as a Randomized Complete Block Design (RCBD). Mean and maximum air veloci ties were analyzed using analysis of variance. Interaction between factors was anal yzed by considering this design as a three factor split-split plot experiment and data was analyzed using MIXED procedure in SAS (Freund et al., 1986). The three factors consid ered were sprayer ground speed, deflector plate setting and tree density. Two spraye r ground speeds (2.15 and 4.73 km/h) divided each of the four blocks (replica tions) into 8 whole plots. Each plot was further divided in five split plots by randomly assigning five deflector plate settings. Each split plot was divided into four split-split plots, and four tree densities (dead, low, medium and high density) were randomly assigned. A grand tota l of 160 measurements were available for 1.0 m Deflector plate 3.0 m Direction of air 1.8 m

PAGE 43

32 analysis. Means were separated using LSME ANS with PDIFF option at 5% level of significance. 3.7 Experiment 2: Effect of Deflector Plate on Spatial Distribution of Spray Deposition The objective of this experiment was to de termine the effect of sprayer air volume rate on spatial distribution of spray droplet s. Spray deposition was quantified by having paper targets (Fisherbrand filter paper, Fisher Scientific, Pittsburgh, PA) at nine spatial locations perpendicular to the direction of the travel of the sprayer. The sprayer was operated with six open nozzles on the left side since they we re directly affected by the change in air volume due to the horizontal po sition of the deflector plate. Two types of nozzles: lilac and blue Albuz APT cone no zzles (Ceramiques Techniques, Desmarquest, France) were used to see the effect of low and high volume applica tion, respectively. The measured discharge rates of the six nozzles at about 1000 kPa pre ssure, were 2.9 and 21.4 L/min respectively. These volumes co rresponded to application rates of 215 and 1585 L/ha based on row spacing of 6 m and ground speed of 2.7 km/h. Five locations of deflector plate, labeled 1, 2, 3, 4 and 5 corre sponded to an air volume rate 1.9, 4.1, 5.8, 7.0 and 7.6 m3/s, respectively, of air deflected to wards the targets at PTO speed of 540 rpm. The test structure made from PVC pi ping consisted of nine horizontal locations labeled A to I at distances of 2.4, 3, 3.9, 4.8, 6.0, 7.2, 8.4, 9.6 m from the point of discharge and at a height of 1.5 m from the ground (Figur e 3.13). The travel direction was towards west on a row that was orie nted in the east-west direction. Spray solutions contained Pyranine-10G fluorescent dye (Keystone Aniline, Inc., Chicago, IL) as deposition tracer at a cons tant rate of 566 mg/L (ppm).Water sensitive paper (Spraying Systems Co., Wheaton, IL) ta rgets were also placed on each location to

PAGE 44

33 visually compare it with results from paper targets. These targets were held by target holders to keep them steady and perpendicu lar to spray direction during each sprayer pass. The exposed area on the paper targets was 42.77 cm2. After each sprayer pass, targets were immediately placed in sealable plastic bags, and stored in an enclosed container for further laboratory analysis. Expe riments were made in four replications. In the laboratory, spray (dye) deposits on each target were quantified by fluorometry (Salyani, 2000b). The deposits were normalized for differences in th e application volume rate (L/ha). Ambient air temperatures and relative humid ity were monitored at a height of 1.8 m, using a temperature/RH indicator (Model 870H, General Eastern, Watertown, MA). Wind speed was measured using a va ne anemometer (Model HH-30, Omega Engineering, Stamford, CT) at the same hei ght. A white ribbon tied to a pole was used to visually note the wind direction. The ranges of temperature, relative humidity and wind speed during the experiment were 4.2 26.0C, 24.3 52.2% and 0.2 2.5 m/s, respectively. The wind direct ion was primarily from the north and north-east direction and did not seem to have any significant eff ect on spraying as magnitude of winds were low during the entire experiment.

PAGE 45

34 ` 6.1 m N S E Top view 10.8 8.4 6.0 3.9 2.4m Direction of spray Paper target Water sensitive paper 3.0 m PVC Pole I H G F E D C B A W 3.0 m I H G F E D C B A 1.5 m 3.6 m Side view Figure 3.13. Schematic view of spray application experiment and sampling layout Data Analysis. The experiment was conducted as a Randomized Complete Block Design (RCBD). Mean tracer depositions at near (ABC), far (DEFGHI) and at each target location, were analyzed using analysis of variance. Analysis was done by considering this experiment as a three factor split-split plot experiment and using MIXED procedure in

PAGE 46

35 SAS (Freund et al., 1986). The three factors co nsidered were nozzle type, deflector plate setting and target distances. Two nozzle type s (lilac and blue, with application rates of 215 and 1585 L/ha, respectively) divided each of the four blocks (replications) into 8 whole plots. Each plot was further divided in five split plots by randomly assigning five deflector plate settings. Each split plot was divided into two split-split plots by considering two target locations (near and far). Means were separated using LSMEANS with PDIFF option at 5% level of significance.

PAGE 47

36 CHAPTER 4 RESULTS AND DISCUSSION This chapter begins with an evaluation of the designed control system, followed by a discussion of the results from the experime nts described in the previous chapter. 4.1 Evaluation of the Electromechanical Control System The electromechanical system designed in the previous chapter successfully performed its function of adjusting the output of air from the sprayer. This system was evaluated under the following topics: Performance: The control system was able to automate the plate movement satisfactorily. Figure 4.1 shows the data from one of the runs made in a row of 51 trees. On the right Y-axis are the density values (shown as dots on graph) of 51 trees obtained from laser sensor. The range of densities obt ained for this trial run was 0.65 to 0.86%. The left Y-axis shows actual plate positions (shown as bars on graph), ranging from 0 to 200 mm that were prescribed based on the corresponding density values. Density readings above 0.80% were assigned plate position 5 whic h corresponds to 200 mm. It can be observed from the graph that the ch ange in density reading data corresponds closely to change in plate movements. Furt her implications of the change in plate movement on the air volume and spray deposition will be discussed in detail in sections 4.2 and 4.3. The response time of the plate to move between extreme ends, that is, when moving from high density to low density tr ee, was seen to be less than 3 seconds.

PAGE 48

37 Figure 4.1. Relation between actual plate position and tree density User-friendliness: The control program prompts the user to enter the tree spacing distance and a nominal speed initially. Afte r this no operator inte rvention is required while spraying. A software switch is provide d within the Graphical User Interface (GUI) of the program to stop the process, should th e operator choose to do so. Also the program makes a log of each tree parameter data, plate movement and time information. These can be used in further analysis. A hardware switch (Figure 3.8) was also installed which made it easy to start or stop the controller manually. Integration: The control system that was implem ented for this project was easily integrated with the existing sprayer. In cases where a control system is to be implemented to test a particular idea, it is important to design with reversibility in mind. In other words, if the automation of a particular syst em does not prove a cer tain idea it should be fairly easy to revert to the original system The airblast sprayer at plate setting 5 gives almost the same air output as the sprayer would in normal circumstances. This allows the 1 5 10 15 20 25 30 35 40 45 50 0 50 100 150 200 250 Tree numberPlate position (mm)0.6 0.65 0.7 0.75 0.8 0.85 0.9Tree density (%) Plate position Tree density

PAGE 49

38 sprayer to be used with the current method, but also enables testi ng with different air volumes. 4.2 Experiment I: Air Ve locity Measurements The aim of this experiment was to quantify the effect of deflector plate position on mean and maximum air velocity measurements through trees of different densities at two different ground speeds. Figure 4.2 shows the mean air velocities recorded over a 10 s interval while the sprayer passed trees of varying densities with five plate settings and two speeds. It must be noted that the dead tree chosen for this test had a dense tree preceding it. The foliage of this dense tree was extending into the d ead tree spacing. While taking measurement it was likely that mean velocity averaged over this tree spacing would result in lower values. Nonetheless, it can give a fair id ea of dynamics of the plate in modifying the mean air output. Figure 4.3 shows the maximum ai r velocity profile for trees with varying density due to change in deflector plate se tting. In general as the density of the tree increased, mean and maximum air velocity across them decreased. However, for a particular tree, as the deflector plate position changed from 1 to 5 the mean and maximum velocities increased. For example it was found that, for the dead tree at 2.15 km/h, there was significant increase in maximu m air velocity between plate setting pairs 1-2, 1-3, 1-4 and 1-5. It may be concluded th at, when the sprayer encounters a dead tree it may be beneficial to have plate setting 1. At 4.73 km/h, there was significant increase in air velocity between the pairs 1-5, 2-5 and 3-5 for medium tree. For the dead tree, mean air velocities at plate setti ngs 1, 2 and 3 were significantly lesser compared to plate settings 4 and 5. It may be concluded that by using lower plate setting at 4.73 km/h, significantly lesser air velocitie s can be obtained while sprayi ng. Change in air velocity

PAGE 50

39 also leads to change in volume as can be s een from Table 3.1. It wa s published that there was little or no relation between air velo city and spray deposition on leaf samples (Salyani and Hoffmann, 1996) while, results from Balsari et al. (2001) showed that reducing the air volume lead to a bett er spray deposition. Nonetheless, an electromechanical system like this can help in efficiently testing spray deposition with different air volume and velocity. Figure 4.4 shows the maximum air velocity averaged over all plate settings for each type of tree to see the eff ect of ground speed. Overall fo r both the speeds, maximum air velocity decreased as the density of the tr ee increased. Lower ground speed resulted in greater maximum air velocity which agrees with results from Salyani and Hoffmann (1996). There was also interaction effect be tween the sprayer ground speed and deflector plate setting. Higher deflector plate setting resulted in increased reduction in maximum air velocity from lower to higher ground speed s. But this interaction was inconsistent with the high density tree as it resulted in almost same air velocities at both speeds. This could be attributed to the dense foliage of the tree which substantially blocked the air from coming onto the hot-wire anemometer, lead ing to the inconsistent result. It may also be noted that some high volume of air may be necessary to move the heavy foliage and transport the spray droplets for high de nsity trees (Salyani and Farooq, 2003).

PAGE 51

40 Figure 4.2. Mean air velocity due to di fferent deflector plate loca tion at 2.15 and 4.73 km/h. Lo wercase bold, lowercase itali cs and uppercase letters show mean separation between plate settings at 2.15, 4.73 km/h and combined ground speeds, respectively Mean air velocit y ( m/s ) Plate position Plate position 0.0 0.5 1.0 1.5 2.0 2.5 3.0 A 30 A B B B Dead Tree c abc c c bc bc ab a a ab Low Density Tree ab ab b bc ab b a a a ac AC A B B B C 0.0 0.5 1.0 1.5 2.0 2.5 3.0 12345 Medium Density Tree b c b bc ab bc ab ab a a A A B B B B 12345 2.15 km/hr 4.73 km/hr Std. dev. Mean separation letter High Density Tree a a a a a a a a a a A A A A A

PAGE 52

41 Figure 4.3. Maximum air velocity due to different deflector plate location at 2.15 and 4.73 km/h. Lowercase bold, lowercase i talics and uppercase letters show m ean separation between plate settings at 2.15, 4.73 km/h and combined ground speeds, respectively Maximum air velocit y ( m/s ) Plate position Plate position 0 2 4 6 8 12345 Medium Densit y A B B B b b b ab ab b ab ab a a AB 12345 2.15 km/hr 4.73 km/hr Std. dev. Mean separation l A A A A A a a a a a a a a a a Hi g h Densit y Tree 0 2 4 6 8 c b b b 8 D ead A A B C C ab b a a a a Low Densit y Tree A AB B B AB ab b b ab ab b a a a ab

PAGE 53

42 Figure 4.4. Maximum air velocity at two ground speeds of the airblast sprayer. Note: Averaged over five plate settings. Lowercase bold, lowercase italics and uppercase letters show mean separation between differe nt tree densities at 2.15, 4.73 km/h and combined ground speeds, respectively 4.3 Experiment II: Spray Deposition The aim of this experiment was to quan tify the effect of changing air volume and application volume rate on the spatia l distribution of spray droplets. Plate setting 1 resulted in lower total deposition recovery from targets compared to other plate settings due to its low air velo city profile. The air vol ume output decreases by about 75% from plate setting 5 to 1. Hence, most of the spray droplets dropped to the ground even before reaching the first target. In general, mean deposit ion decreased as the target distance from the sprayer increased (Figure 4.6 and 4.7). Total depositions were calculated as near (A-C) for the first thr ee locations and far (D-I) for the last six locations. These also signified spray material going to the tree and away from the tree, respectively. Figure 4.5 show s the deposition at these two spatial sections for two application volume rates due to different defl ector plate settings. In all replications, the 0 2 4 6 8 dead low medium high 2.15 km/h 4.73 km/hMaximum air velocit y ( m/s ) Tree density Std. dev. a Meanseparationletter b a c b d a c A B C D

PAGE 54

43 deposition was higher in the near location than the far locations except for plate setting 1. Plate setting 1Â’s air distribution was such th at some of the output from the top most nozzle (Figure 3.17) was exposed to high volum e air that was verti cally directed from behind the deflector plate. This contributed to the excess deposition at far sections. However, at plate settings 2-5 the trend reve rsed and there were more depositions at near locations. Thus, depending on the spray requir ement, particular plate setting can be chosen to obtain variable deposition to the tree. P-values of 0.09 at low application rate and 0.0001 at high app lication rate were obtained between far and near location, respec tively. It can be inferred that effect of change in air volume on spray deposition is significant for high volume application compared to low volume application. There wa s interaction effect be tween plate setting 2 and application rate which produced lo wer deposition at far location during high application rate compared to other plate se ttings. Also, there were some interaction effects between application volume rate and targ et location. For a particular plate setting, it is natural to expect more deposition at near location compared to far location. However, plate settings 3 and 4 resulted in lesser difference between near and far locations at high application rate compared to other plate settings. 4.4 Discussion The results presented in this research de monstrate that the modified sprayer could customize the deposition by cha nging air output. Further test ing can be done on the spray deposition across a row of trees of variable dens ities while the plate moves in real time to see the effect of deflector plate location on penetration of spray inside a canopy. The distance between the laser sensor a nd the sprayer air outlet was 2.7 m (Figure 3.1). For a 3 m tree spacing, by the time the la ser has obtained the tree data the air outlet

PAGE 55

44 has already sprayed 10% of the tree spacing wi th different air volume. This can result in inaccurate spraying for a portion of the tree. This can be improved by having the laser sensor at a distance at least equal to the tree spacing of the block to be sprayed. A 3-second travel time was achieved for plat e travel between extreme ends. For a 3 m tree spacing and travel speed 1.61 km/h, it would take about 6 seconds to travel the tree spacing. For the worst case, wherein, th e sprayer has to move across a high density tree to a dead tree, the travel time of the pl ate will result in 50% of the tree being in transition period from full to least air volume. This 3 sec ond travel time can be reduced by using a superior motor or highe r power, thus improving the design.

PAGE 56

45 Figure 4.5. Effect of two application volume rate on to tal deposition at two spatial sections. Lowercase bold, lowercase ita lics and uppercase letters show mean separation between plate settings at n ear (A-C), far (D-I ) and combined (A-I) locations, respectively Tracer deposition Low Volume Application (215 L/ha)0 1 2 3 4 Near Far Std. dev. Mean separation letters b a a a a a a a a a B A A AB AB Plate Position High Volume Application (1585 L/ha)0 1 2 3 4 12345 b a a a a a a a a a B A A A A

PAGE 57

46 Figure 4.6. Low application rate: spatial di stribution of deposition (see Figure 3.13) at different deflector plate position Plate 50.0 0.5 1.0 24681012 Tracer deposition (g/cm2) A B C D E F G H I Plate 20.0 0.5 1.0 Plate 3 0 .0 0 .5 1.0 Plate 40.0 0.5 1.0 Distance from sprayer nozzles (m) Plate 1 0 .0 0 .5 1.0 1.5 Std. dev. Target location

PAGE 58

47 Figure 4.7. High application ra te: spatial distribution of de position (see Figure 3.13) at different deflector plate positions Plate 20.0 0.5 1.0 Plate 50.0 0.5 1.0 24681012 Plate 10.0 0.5 1.0 1.5 Plate 30.0 0.5 1.0 Plate 40.0 0.5 1.0 Tracer deposition (g/cm2) Distance from sprayer nozzles (m) A B C D E F G H I Std. dev. Target location

PAGE 59

48 CHAPTER 5 CONCLUSIONS One of the objectives of my project was to be able to spray a particular row based on the morphological characterist ics of the tree. The electrom echanical system developed modifies the air characteristic and thus gi ves variable spray deposition. The density values from the laser sensor were used as an input to achieve this variability. Evaluation of my system shows that different settings of the deflector plate results in changing air velocity through the cano py of different densities. A 61% reduction in air velocity was obtained at the outlet of sprayer from outer most to innermost deflector plate position. Further, spatial distribution of the spra y droplets at different output volume was determined. A change of plate setting from outermost to innermost yields about 37% reduction in deposition at far location at hi gh application rate. This results in some reduction of off-target spraying. Since the volume of air coming out of sprayer depends upon commands sent by the computer, it can be customized based on the requirement. This is important as the idea of changing air volume in sprayer for variable rate application is relatively new and hence may need fine tuni ng. A complementary system that shuts off sections of nozzles based on tree height informati on from sensor will further improve this system.

PAGE 60

49 APPENDIX A STEP MOTOR SIZING CALCULATION The step motor used in this research proj ect was selected based on calculations that have been described below (Oriental Mo tor General Catalog, 2004) (Figure A-1). Variables used in the equations are listed in Table A-1. Figure A.1. Block diagram used for steppe r motor calculations (A dopted from Oriental Motor General Catalog, 2004) Table A.1. Specifications of the actuation mechanism Total mass of the plate m = 2.9 kg (6.5 lb) Ball screw efficiency = 0.9 Internal frictional coefficient of ball nut o = 0.3 Ball screw shaft diameter DB = 16 mm (0.63 inch) Total length of travel LB = 200 mm (7.87 inch) Material of ball screw Steel Density of steel = 7.8 g/cm3 (4.64 oz/in3) Pitch of the ball screw PB = 5 mm (0.2 inch) Step size of motor s = 1.8 Maximum load F = 250 N Motor rotor inertia Jo = 343.9 g-cm2 (1.88 oz-in2) F m D PB Programmable controlle r Step motor

PAGE 61

50 Inertia: Inertia of the ball screw (JB) is given by, 432B B BD L J 463 0 87 7 64 4 32 = 0.56 oz-in2 Inertia of the plate ( Jp) is given by, Jp = 22 BP m= 22 2 0 16 5 6 rev in lb oz lb = 0.11 oz-in2 Total Inertia ( JT) 11 0 56 0p B TJ J J 0.67 oz-in2 = 122.5 g-cm2 The motor rotor inertia ( JM) should not be very small compared to total inertia ( JT) 30 M TJ J (Clarence, 2005) where JM = Motor inertia 1 4 MJ g-cm2 Operating Parameters: Number of operating pulses ( A ) required for the entire travel length of the deflector plate is given by, ) ( 360 ) ( ) (s B Bangle Step P pitch screw Ball L length Travel A o o8 1 360 2 0 87 7 = 7870 pulses This implies, 7870 pulses are required to m ove the deflector plate between extreme positions. Based on several design iterations on the type of motor available, its size and cost, it was decided to select a motor using a 3 s time interval ( t) required to complete the entire travel length ( LB) The pulse frequency ( f ) is determined based on this positioning period, t

PAGE 62

51 ) ( ) ( t Period g Positionin A Pulses Operating of Number f Hz 2623 3 7870 Thus, the operating speed ( N ) of the step motor required to achieve this motion is given by, 60 360 sf N 60 360 8 1 2623 = 787 rpm Load Torque ( TL): The load torque is the maximum torque encountered by the motor while moving the deflector plate. The deflector plate faces maximum air resistance when moving from outermost position towards the innermost posit ion. Based on the air measurements made at sprayer outlet, it was found that the maximum load ( F ) on the plate was about 250 N. The load torque is a summation of this load, F and the pilot load, Fp, due to friction of ball nut bearings. A rule of thumb is to consider this pilot load to be approximately a third of the load F Hence, Fp = 3 F = 83.3 N. 2 2B p o B LP F FP T 2 ) 5 )( 3 83 ( 3 0 ) 9 0 ( 2 ) 5 ( 250 = 0.24 N-m Acceleration torque ( Ta) Acceleration torque is the torque required by th e motor to initially step up to the required operating speed of 787 rpm. This is cal culated based on the rotor inertia ( Jo), total inertia ( JT) and the pulse frequency ( f ). An acceleration time ( t1) equal to 25% of positioning time, t = 3 s is allowed to step up to the required operating speed. 1180 t f g J J Ts T o a 75 0 7870 180 ) 8 1 ( 386 67 0 88 1 = 2.17 oz-in. = 0.015 N-m

PAGE 63

52 Required Torque ( T ): The total torque required by the step motor will be summation of the load torque ( TL) and the acceleration torque ( Ta) considering a safety factor of 2. T = ( TL + Ta) x Safety factor = (0.24 + 0.015) x 2 = 0.5 N-m Based on this requirement, availability and cost, the step motor with model number AMH23-258-3 (Advanced Micro Systems, In c., Nashua, NH) which had a torque capacity of 1.82 N-m was chosen for the project. Further specifications of this motor have been provided in Appendix-B.

PAGE 64

53 APPENDIX B COMPONENT SPECIFICATION Table B.1. Ball screw and nut assembly specifications Properties Details Material Cf 53, induction hardened to HRC 62 2. Pitch accuracy 0.1 mm/300 mm Diameter 16 mm Pitch 5 mm Features Ball screw designed for gr eater than 90% efficiency in converting rotary to linear motion. Ball nut is designed to prevent backlash Table B.2. Step motor technical specifications Properties Details Model no. AMH23-258-3 No. Of leads 8 Phase connection Parallel Holding torque 258 oz-in Inertia 1.88 oz-in2 Weight 2.2 lb Table B.3. Step motor controller features Properties Details Model No. mStep-407 Current 7 A Voltage 24 V (dc) to 80 V (dc) Resolution 1/10 Communication RS232/Serial Port Memory 2 kb non-volatile

PAGE 65

54 LIST OF REFERENCES Balsari, P. and M. Tamagnone. 1998. The neces sity to determine th e correct amount of air to use in airblast sprayer. Paper 98-A-075, Proceedings of International Conference on Agricultural Engineering Oslo, Norway. Balsari, P., G. Oggero and M. Tamagnone. 2001. The choice of air volume rate for mistblowers in apple orchards. ATW Symposium Stuttgart, Germany. Balsari, P., M. Tamagnone and P. Maru cco. 2003. Innovative technologies for orchard/vineyard sprayers. VII Workshop on Spray Applic ation Techniques in Fruit Growing 35-42. Clark, R. L. and R. L. McGuckin. 1996. Variable rate application technology: an overview. Proc. 3rd International Conferen ce on Precision Agriculture WI: ASA/CSSA/SSSA. Clarence, W. S. 2005. Mechatronics: an integrated approach CRC Press LLC. Cugati, S. A., W. M. Miller, A. W. Schumann and J. K. Schueller. 2006. Dynamic characteristics of two commercial hydrauli c flow-control valves for a variable-rate granular fertilizer spreader. ASABE Paper No. 061071. St. Joseph, Mich.: ASABE. Escola, A., F. Camp, F. Solanelles, S. Planas and J. R. Rosell. 2003. Tree crop proportional spraying according to the vegetation volume. First results. VII Workshop on Spray Application Techniques in Fruit Growing 43-49. Freund, R. J., R. C. Littell and P. C. Spector. 1986. SAS systems for linear models Cary, NC: SAS Inst. Inc. Gebhardt, M. R., C. L. Day, C. E. Goer ing and L. E. Bode. 1974. Automatic sprayer control system. Transactions of the ASAE 17(6):1043-1047. Ghate, S. R. and C. D. Perry. 1994. Ground speed control of pesticide application rates in a compressed air direct injection sprayer. Transactions of the ASAE 37(1):33-38. Giles, D. K. and J. A. Comino. 1990. Droplet size and spray pattern characteristics of an electronic flow controller for spray nozzles. J. Agric. Engng. Res. 47:249-267. Giles, D. K., M. Delwiche and R. Dodd. 1987. Control of orchard spraying based on electronic sensing of ta rget characteristics. Transactions of the ASAE 30:16241630.

PAGE 66

55 Giles, D. K., G. W. Henderson and K. Funk. 1996. Digital control of flow rate and spray droplet size from agricultural nozzles for precision chemical application. Proceedings of the Third International Conference on Precision Agriculture ASA, CSSA, SSSA, Madison, WI. 729-738. Han, S., L. Hendrickson and B. Ni. 2000. A vari able rate application system for sprayers. Proceedings of the Fifth International Conference on Precision Agriculture ASA, CSSA, SSSA, Madison, WI. Landers, A. J. and E. Gil. 2006. Development and validation of a new deflector system to improve pesticide application in New York and Pennsylvania grape production areas. ASABE Paper No. 061001. St. Joseph, Mich.: ASABE. Miller, D. R., E. W. Huddlestin, J. B. Ross and W. E. Steinke. 2003. Airblast spray partitioning in a mature pecan orchard. Transactions of the ASAE 46(6):1495-1501. Oriental Motor U.S.A Corp. 2004. General Catalog F2 – F6. http://www.orientalmotor.com/pr oducts/pdfs/F_TecRef/TecMtrSiz.pdf (Last accessed: March 4th, 2007). Reichard, D. L., H. J. Retzer, L. A. Li ljedahl and F. R. Hall. 1977. Spray droplet distribution delivered by ai r blast orchard sprayers. Transactions of the ASAE 20(1):232-237,242. Salyani, M. 1988. Droplet size effect on spra y deposition efficiency of citrus leaves. Transactions of the ASAE 31:1680-1684. Salyani, M. 1993. Degradation of fluorescent tracer dyes used in spray applications. ASTM International West Conshohocken, PA: ASTM STP 1183.13:215-226. Salyani, M. 2000a. Optimization of deposit ion efficiency for airblast sprayers. Transactions of the ASAE 43(2):247-253. Salyani, M. 2000b. Methodologies for asse ssment of spray deposition in orchard applications. ASAE Paper No. 00-1031, St. Joseph, Mich.: ASAE. Salyani, M. 2003. Calibration of ai rblast sprayers EDIS AE238. http://edis.ifas.ufl.edu/AE238 (Last accessed: March 11th, 2007). Salyani, M. and M. Farooq. 2003. Spraye r air energy demand for satisfactory spray coverage in citrus applications. Proc. Fla. Stat Hort. Soc. 116:298-304. Salyani, M. and M. Farooq. 2004. Drift pot ential of citrus ai r-carrier sprayers. Proc. Fla. State. Hort. Soc. 117:130-135. Salyani, M. and R. D. Fox. 1999. Evaluation of spray quality by oiland water-sensitive papers. Transactions of the ASAE 42(1):37-43.

PAGE 67

56 Salyani, M. and W. C. Hoffmann. 1996. Air and spray distribution from an air-carrier sprayer, Applied Engineering in Agriculture 12(5):539-545. Salyani, M. and J. Wei. 2005. Effect of travel speed on characterizing citrus canopy structure with a laser scanner. Precision Agriculture '05. J. V. Stafford, Ed. 185192. Salyani, M. and J. D. Whitney. 1988. Evalua tion of methodologies fo r field studies of spray deposition. Transactions of the ASAE 31(2):390-395. Salyani, M. and J. D. Whitney. 1990. Ground speed effect on spray deposition inside citrus trees. Transactions of the ASAE 33(2):361-366. Salyani, M. and J. D. Whitney. 1991. Effect of oscillators on depositi on characteristics of an airblast sprayer. Transactions of the ASAE 34(4):1618-1622. Salyani, M., E. BenSalem and J. D. Wh itney. 2002. Spray deposition and abscission efficacy of CMN-pyrazole in mechanical harvesting of valencia oranges. Transactions of the ASAE 45(2):265-271. Salyani, M., Y. M. Koo and R. D. Sweeb. 2000. Spray application variables affect air velocity and deposition characte ristics of a tower sprayer. Proc. Fla. State Hort. Soc. 113:96-101. Salyani, M., R. D. Sweeb and M. Farooq. 2006. Comparison of string and ribbon samplers in orchard spray applications. Transactions of the ASAE 49(6):1705-1710. Sawyer, J. E. 1994. Concepts of variable rate technology with consider ations for fertilizer application. Journal of Production Agriculture 7:195-201. Stover, E. 2002. Sensor-controlled spray sy stems for florida citrus. EDIS HS-140. http://edis.ifas.ufl.edu/HS140 (Last accessed: March 10th, 2007). Stover, E., D. Scotto and J. Salvatore. 2002a Spray applications to citrus: survey of current practices in the Indi an river area. EDIS HS-127. http://edis.ifas.ufl.edu/HS127 (Last accessed: March 10th, 2007). Stover, E., D. Scotto, C. Wilson and M. Sa lyani. 2002b. Spray app lications to citrus: overview of factors influencing spraying efficacy and off-target deposition. EDIS HS-128. http://edis.ifas.ufl.edu/HS128 (Last accessed: March 10th, 2007). Tangwongkit, R., V. Salokhe and H. Jayasu riya. 2006. Development of a real time, variable rate herbicide applicator usi ng machine vision for be tween-row weeding of sugarcane fields. Agricultural Engineering In ternational: the CIGR Ejournal Manuscript PM 06 009. Vol. VIII.

PAGE 68

57 Wei, J. and M. Salyani. 2004. Development of a laser scanner fo r measuring tree canopy characteristics: phase 1. prototype development. Transactions of the ASAE 47(6):2101-2107. Wei, J. and M. Salyani. 2005. Development of a laser scanner fo r measuring tree canopy characteristics: phase 2. fo liage density measurement. Transactions of the ASAE 48(4):1595-1601. Whitney, J. D., S. L. Hedden, D. B. Chur chill and R. P. Cromwell. 1986. Performance characteristics of PTO airblast sprayers for citrus. Proc. Fla. Stat Hort. Soc 99:5965. Zhang, N., M. Wang and N. Wang. 2002. Pr ecision agriculture-a worldwide overview. Computers and electron ics in agriculture 36:113-132.

PAGE 69

58 BIOGRAPHICAL SKETCH Naresh Pai was born in 1981, in Mumbai in Maharashtra, India, to Archana and Anant Pai. He received his Bach elor of Engineering degree in mechanical engineering from the Manipal Institute of Technology, India, in June 2003. He then enrolled in the Graduate School of the Univer sity of Florida. He expects to receive a Master of Science degree in agricultural en gineering and a concurrent Master of Science in mechanical engineering degree in May 2007. After comple ting his Master of Science degrees, he plans to work in the field of c ontrols and automation using the technical skills he acquired during his studies.