View
21
Download
2
Category
Preview:
Citation preview
Preliminary Design of an Aircraft Automatic Painting andPaint Removal System
Umberto Morelli
Thesis to obtain the Master of Science Degree in
Aerospace Engineering
Supervisors: Prof. Filipe Szolnoky Ramos Pinto CunhaProf. Alexandra Bento Moutinho
Examination Committee
Chairperson: Prof. Fernando José Parracho LauSupervisor: Prof. Filipe Szolnoky Ramos Pinto Cunha
Member of the Committee: Prof. Full Name 3
October 2016
ii
Alla mia famiglia
unita nella tempesta
iii
iv
Resumo
Com o crescimento acelerado da industria aeroespacial decorrente dos ultimos anos e previsto para o
futuro, novas tecnologias e metodologias de producao tornam-se cada vez mais fundamentais. Uma
das areas mais carentes de inovacao e a manutencao do acabamento das aeronaves, incluindo os
processos de pintura e despintura. Atualmente. a manutencao e realizada manualmente, o que requer
muitas horas de mao de obra bracal num ambiente perigoso. Muitas solucoes para este problema
tem sido desenvolvidas, no entanto, um sistema eficaz seria pela automatizacao do processo, o qual
ainda nao esta disponıvel. Esta solucao podera acelerar drasticamente o processo, consequentemente
dimunuir o envolvimento direto da mao de obra, custos e riscos ambientais. Este trabalho tem como
objetivo aprofundar o tema apresentado e realizar uma proposta preliminar de projeto de uma solucao
automatizada respondendo a complexidade da questao a partir de uma solucao de baixo custo.
Palavras-chave: Pintura, Remocao de Tinta, Robo para Pintura, Sistema Automatico, Aeron-
aves.
v
vi
Abstract
The maintenance of the aircraft finish system is executed completely manually at present, involving
a big amount of manual labor for a long time and in a hazardous environment. The automation of
the process would be able to dramatically speed it up and to decrease manpower involved, with a
consequent contraction in costs and environmental risks. It is at the moment an important challenge
within the aerospace industry also because of the expectations of airplanes fleet growth over the coming
years. Several solutions are being developed, nevertheless, a system able to achieve the maintenance
process automatically is not yet available. Along this thesis, a preliminary design of an automatic system
for aircraft painting and paint removal has been carried out. The work points out that a low cost solution
for this complex problem is possible. As a preliminary study, this is intended to be a starting point for
further development on this subject.
Keywords: Aircraft finish system, Paint Removal, Spray Painting Robot, Automatic System,
Aircraft.
vii
viii
Contents
Resumo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii
1 Introduction 1
1.1 Finish System Maintenance Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Paint Removal Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1 Chemical Removers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.2 Mechanical Removal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.3 Optical Removal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Painting Methods and Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.1 High Volume Low Pressure Spray Method . . . . . . . . . . . . . . . . . . . . . . . 6
1.3.2 Airless Spray Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.3 Hot Spray Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.4 Air-Assisted Airless Spray Method . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.5 Electrostatic Spray Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3.6 Spray Painting Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4 Aircraft Painting Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.4.1 Priming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4.2 Topcoating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.5 Thesis Motivation and Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.6 Existing Automated Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.7 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2 Finish System Automatic Maintenance Solutions 17
2.1 Specifications and Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.1.1 Aircraft impact on the design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.1.2 Painting requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
ix
2.1.3 Coating removal requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2 Possible Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2.1 Multirotor UAV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2.2 Rail Mounted Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.2.3 Mast Mounted Robot on AGV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2.4 Multi DoF Structure on AGV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2.5 Lifting Structure on AGV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3 Design Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3 System Design 29
3.1 Robotic Arm Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.2 Structure Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.2.1 Horizontal Beam Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.2.2 Beam Support Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.2.3 Lifting System Structure Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.3 Lifting System Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.3.1 Linear Guides Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.3.2 Lifting Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.4 Subsystems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.5 AGV Platform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.5.1 AGV Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.5.2 Subsystems Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.6 Cost Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4 Conclusions 55
4.1 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Bibliography 59
x
List of Tables
2.1 Geometrical features of the C-130 H. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.2 Trade-off between different possible solutions. . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.1 Different robotic arms specifications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.2 Aluminium alloy Al 6061-T6 properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.3 Sample of the bean cross section properties table. . . . . . . . . . . . . . . . . . . . . . . 32
3.4 Beam cross section properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.5 SSETWNO M16-CR linear bearing specifications. . . . . . . . . . . . . . . . . . . . . . . 43
3.6 Vertical load on the lift actuator. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.7 Angular velocity, lead and Vertical velocity for different screws. . . . . . . . . . . . . . . . 45
3.8 KGS-4040-023-RH ball screw specifications. . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.9 FANUC R-30iATMMate Controller features. . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.10 Paints specifications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.11 Specifications of different paint tanks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.12 AGV payload. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.13 Cost estimation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
xi
xii
List of Figures
1.1 Finish system scheme. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Cleaning of the tail os a F16 aircraft. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Aircraft masking before painting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Chemical stripping of helicopter coating. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.5 Scuff sanding a KC-10 aircraft. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.6 Paint removal by PMB method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.7 Portable handheld laser stripping device. . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.8 Primer application on aircraft fuselage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.9 Topcoat application on aircraft fuselage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.10 Aircraft painter at work. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.11 Hyundai Alabama robotic painting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.12 Rotor blades for wind power systems coating achieved by ABB’s painting robots IRB 5400. 12
1.13 ABB’s robotic mining truck-washing system in Brazil. . . . . . . . . . . . . . . . . . . . . . 13
1.14 Robotic system cleaning up the Sydney Harbor Bridge. . . . . . . . . . . . . . . . . . . . 13
1.15 UltraStrip Systems, Inc.’s M-2000 removing paint from the hull of a ship. . . . . . . . . . . 14
1.16 Advanced Robotic Laser Coating Removal System. . . . . . . . . . . . . . . . . . . . . . 14
1.17 Laser Coating removal Robot. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.18 Robotic Aircraft Finishing System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.19 Robotic system coating the B-777 wing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.1 A C-130E Hercules from the 43rd Airlift Wing, Pope Air Force Base, N.C. . . . . . . . . . 18
2.2 C-130H side and front views with dimensions. . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3 C-130H top view with dimensions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4 Multirotor UAV system sketch. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.5 Rail mounted robotic system sketch. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.6 Mast mounted robotic system on AGV sketch. . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.7 Multi DoF robotic system on AGV sketch. . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.8 Arm and lifting robotic system on AGV sketch. . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.1 Grafic representation of a human arm workspace. . . . . . . . . . . . . . . . . . . . . . . 30
3.2 FANUC PaintMate 200iA/5L and its workspace. . . . . . . . . . . . . . . . . . . . . . . . . 31
xiii
3.3 Forces acting on the horizontal beam. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.4 Scheme of the I-beam cross section. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.5 Beam weight parametric study. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.6 Beam forces at the root section. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.7 Stresses in the beam root section. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.8 Horizontal beam support structure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.9 Horizontal beam and support structure assembly. . . . . . . . . . . . . . . . . . . . . . . . 36
3.10 Beam support free body diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.11 Beam support parametric study. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.12 Beam support forces diagram and section. . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.13 Lifting system exploded top view. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.14 Lift structure plate cross section. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.15 Lift structure detail. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.16 Lift structure base. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.17 Lift base approximate structure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.18 Lift structure parametric studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.19 Lift structure forces diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.20 Lift structure truss scheme. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3.21 SSETWNO M16-CR linear bearing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.22 Linear guides: round rails. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.23 Lift system view with beam support. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.24 40MMx40MM ball screw with nut. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.25 PK599BE-N7.2 stepper motor torque vs. speed graph. . . . . . . . . . . . . . . . . . . . . 47
3.26 FANUC R-30iATMMate Controller. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.27 Automatic system overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.28 Top view of the system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.29 System sideview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.1 Complete system drawing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
xiv
Nomenclature
Greek symbols
ρ Material density
σ Normal stress
τ Shear stress
Roman symbols
A Area
db Horizontal beam tip deflection
E Modulus of elasticity
F Force
H Lift structure height
I Moment of inertia
lb Horizontal beam length
M Bending moment
nresonance Ball screw angular velocity at which resonance occurs
Sp Ultimate tensile strength
Sy Tensile yield strength
t Thickness
V Vertical shear force
W Weight
w Width
Subscripts
arm Robotic arm
xv
beam Horizontal beam
bear Linear bearings
cr Critical
house Beam support
leg Base column of the lift structure
lift Lifting system structure
s Ball screw
x, y, z Cartesian components
xvi
Acronyms
AGV Automatic Guided Vehicle
ARLCRS Advanced Robotic Laser Coating Removal
System
ASM Automated Spray Method
ATEX ATmosphere EXplosibles
CG Center of Gravity
CTC Concurrent Technologies Corporation
DoF Degree of Freedom
HVLP High Volume Low Pressure
JPL NASA’s Jet Propulsion Laboratory
LARPS Large Aircraft Robotic Paint Stripping
LCR Laser Coating removal Robot
MPW Medium Pressure Water
NREC National Robotics Engineering Consortium
PMB Plastic Media Blasting
RAFS Robotic Aircraft Finishing System
SLAM Simultaneous Localization and Mapping
UAV Unmanned Aerial Vehicle
USAF United States Air Force
iGPS indoor Ground Positioning System
xvii
xviii
Chapter 1
Introduction
The aircraft external structure is covered by a number of paint layers generally referred as finish sys-
tem. It has a large number of functions that differ between military and civil aviation. The primary
purposes of coating military aircraft are reducing radar and infrared signature, protecting the structure
(from corrosion, abrasion, chemicals . . . ) and appearance. Civilian aircraft coatings are used for struc-
ture protection, company identification and aesthetics appearance [1].
In both military and civilian cases, all along the aircraft life, the coating system is applied and removed
several times for a variety of reasons, a replacement of frayed coatings or changing in the livery are the
most common. Presently, the finish system is mainly removed by chemical stripping. This method
removes the paint through toxic chemicals and workers labor. If some paint resists to the chemicals, the
laborers strip it with wire brush or sandpaper. This task exposes them to health hazard because of the
highly toxic paint dust and residual chemicals. There is also a risk of damaging the aircraft structure in
case of error [2].
Beside the worker health hazard, both painting and paint removal processes present a critical envi-
ronmental pollution threat due to the chemicals used, especially in the paint stripper. Consequently, the
application, storage and waste disposal of these chemicals is highly regulated [3].
Other paint removal methods are mechanical and optical. The latter is still under development and
is being implemented just recently [4]. These methods are generally less polluting, due to the absence
of chemical products. They also tend to be more efficient but require higher skilled workers as well as
higher initial investment for training and equipment [2].
After the removal of the finish system, an aircraft structure inspection follows to detect cracks or
corrosion damages, and finally the aircraft is repainted [5].
Usually, the finish system consists of three layers: a pretreatment, a primer and a topcoat (see
Figure 1.1). The finish system is generally deposited with spray guns by highly skilled and experienced
workers. During both painting and paint removal, the workers use protective garments for the body,
hoods with visor and respirators [6].
1
Figure 1.1: Finish system scheme.
1.1 Finish System Maintenance Process
The maintenance of the finish system is a long-lasting process that follows a specific routine indicated
by the regulations. According to the United States Air Force (USAF) technical manual [5] the process
is divided into the following stages: (i) preparation for paint removal, (ii) paint removal, (iii) surface
preparation for painting and surface treatment, (iv) painting and (v) coating inspection.
First of all the aircraft is de-energized. Then, before the paint is removed, all the surfaces to be
worked are cleaned from grease, oil and dirt. These materials would act as a barrier protecting the
coating to be removed [5]. After washing the surfaces (Figure 1.2), all special areas, equipment and
material are protected by masking. The masking is essential to protect delicate areas of the aircraft, like
electronic equipment and windows, during the paint removal.
Figure 1.2: Cleaning of an F16 aircraft [7].
The finish system removal is a complex and critical process. To accomplish this task it is possible
to use different methods: chemical, mechanical and optical [2]. Each one has its advantages and
disadvantages (see Section 1.2).
After the finish system is removed the surface of the aircraft is prepared for the application of new
coatings. This is the most important stage for ensuring proper adherence and performance of the new
finish system. The life of a coating system, its effectiveness and appearance depend more on the
condition of surfaces receiving it than any other factor [5].
The surface to be painted is carefully cleaned and then inspected for corrosion and damages. Before
the painting operations begin, the aircraft is masked again (Figure 1.3). The mask is changed because
2
during the painting the parts to mask are different from the ones masked during the paint removal. The
masking material used is also different for the decoating and painting processes. The masking operation
generally consumes more man-hours than the actual painting [5].
Figure 1.3: Aircraft masking before painting1.
When the aircraft is entirely cleaned, treated and masked, it is ready to be painted. Painting can
involve many skilled workers at the same time depending on the aircraft size. Finally the coating system
is inspected to ensure its effectiveness all over the aircraft.
1.2 Paint Removal Methods
The objective of the paint removal process is the complete removal of the coating system of the aircraft
without damaging the surfaces on which it is applied. In order to do it, a variety of methods are used.
In this section they are individually described, pointing out advantages and disadvantages of each one
according to the USAF technical manual [5].
The removal methods can be divided into chemical, mechanical and optical. While the chemical and
mechanical methods are widespread, optical coating stripping is a young technology applied just since
few years.
1.2.1 Chemical Removers
The paint strippers are a mixture of five chemical components: organic solvents, thickeners, corrosion
inhibitors, surfactants and evaporation retardants. The components have to be mixed immediately before
use as chemicals tend to separate on standing.
The chemical remover is selected relatively to the finish system to be removed. If the component
to be stripped is small, it is possible to immerse it in a pool of hot removal (approximately 85 oC). This
method is obviously not applicable to the whole aircraft but is generally applied to the landing gear and
other small components.
1URL http://www.european-coatings.com/Homepage-news/New-coatings-to-enhance-range-for-aircraft [Ac-cessed:3 September 2016]
3
Alternatively, the mixture can be applied directly on the surface using sprayer, brush or roller. The
thickness of the stripper layer has to be light to medium as thick coats of it slow down the removal
rate and increases the operational cost. When the finish system is stripped as in Figure 1.4, or the
chemicals dwell time is exceeded, the whole area is agitated with a brush Immediately after all loosened
finish system is scraped from the surface. Chemical removers are reapplied in spots where the finish
system has not been removed. The complete process is repeated a maximum of three times. If after
the third application the finish system is not completely removed, the process has to be completed with
the mechanical removal of the coatings. Finally, the workers rinse the area thoroughly with hot water
between 37 oC and 49 oC at a pressure of 1-1.7 MPa.
Figure 1.4: Chemical stripping of helicopter coating2.
This is an old but still widespread method because of its effectiveness and economy. It requires few
equipment investments and low skilled workers.
Chemical paint stripping was developed for metallic surfaces and it is not possible to use it on com-
posite material surfaces because the chemicals would react with the composite structure [2]. This prob-
lem has to be taken into account because of the trend of aerospace industry to use more and more
composite materials. Moreover, it requires a long time and it is heavily influenced by the environment
temperature and humidity.
1.2.2 Mechanical Removal
Mechanical removal methods include the use of motor- or hand-driven wire brushes, abrasive paper and
mats, as well as abrasive blasting. These methods are recommended if chemical stripping is impractical
due to structural complexity, environmental restriction and working difficulties. Mechanical removal is
generally very effective. Nevertheless, it can cause severe damage to the structure if improperly used.
Mechanical removal consists of a simple mechanical abrasion of the finish system (see Figure 1.5).
It produces highly toxic dust, requiring workers and environment protections. The mechanical removal
is generally made by abrasive blasting of grit or sand (effective on iron and steel alloys), or hand or
motor-driven abrasive equipment.
2URL http://www.aviationpros.com/product/10472836/solvent-kleene-inc-aircraft-paint-stripper [Accessed:14September 2016]
3U.S. Air Force photo/Margo Wright - URL http://www.af.mil/shared/media/photodb/photos/081002-F-4094W-807.jpg
[Accessed:14 September 2016]
4
Figure 1.5: Scuff sanding a KC-10 aircraft3.
Other mechanical methods are Plastic Media Blasting (PMB) and Medium Pressure Water (MPW)
methods. PMB consists of blasting polyester plastic particles at a pressure in the range of 0.28 to
0.41 MPa with the nozzle tip at a distance within the range of 30 to 60 cm, while the angle of incidence
should be within the range of 30 to 60 degrees (see Figure 1.6). In the end, the plastic media can be
collected and reused. This method can not be used on metal structures having a thickness less than
0.4 mm. It is an efficient and rapid method both for metallic and composite structures, but a proper
waste management must be ensured for economic and environmental reasons. This method requires
specialized workers and high initial investment It is more environmental friendly than chemical stripping
and has a high removal rate (about 7 cm2/min).
Figure 1.6: Paint removal by PMB method4.
MPW method is the blasting of water and sodium bicarbonate. The injection system is a positive
feed control system (computer controlled). The water pressure is 100 MPa with a flow rate of 11 liters
per minute. The nozzle distance from the surface is within the range of 5 to 10 centimeter and the angle
of incidence in the 40 - 60 degrees range [5]. The speed on the nozzle across the surface should be 10
4URL http://www.fus.de/uploads/images/Gallery/FUS-DS/5D.jpg [Accessed:14 September 2016]
5
cm/s. As the PMB, this technique requires specialized workers and high initial investment. Nonetheless,
it is relatively environmental friendly and has a high removal rate.
1.2.3 Optical Removal
The optical removal of the finish system is made by a laser wave. The possibility to remove the coating
on a substrate by laser wave is being investigated since the early 90’s [8]. The results obtained show
that it is possible to remove the coating from inorganic as well as organic substrate. Many handheld
laser removal systems have been designed (see Figure 1.7) but they are heavy, dangerous and the final
result is really affected by the human-factor. With the finish system removal automation trend, in the last
years, laser removal is the subject of many new projects [9, 10].
Figure 1.7: Portable handheld laser stripping device [11].
The interest on laser removal is motivated by advantages it has when compared to the other tech-
niques. It does not involve the use of raw materials like chemicals or water and because of that the
waste produced is just the coating removed that can be easily vacuumed This leads to a cleaner envi-
ronment and a easier disposal of waste. This method can be used on aluminum as well as composite
substrate without damage of the structure. Furthermore, it is possible not to operate on delimited areas
and therefore the masking of the aircraft can be partly avoided [9].
Beside the stated advantages, a laser system involves a really high initial investment, and being a
relatively new technique, in many aircraft maintenance manuals it is not contemplated or allowed.
1.3 Painting Methods and Techniques
This section describes the different painting methods and techniques according to the USAF technical
manual [5].
Spray application is the standard for painting aircraft. It is a fast procedure and produces films of
good uniformity and quality. Other methods are brush or roller applications. These are useful in special
cases, particularly in non-aeronautical or less critical applications, but are not described here.
1.3.1 High Volume Low Pressure Spray Method
The standard spray method in the aerospace industry is the High Volume Low Pressure (HVLP) method
Using a high volume of low pressure air, the coating material is atomized through the spray gun nozzle.
6
The spray equipment generally utilizes low pressure gun cups to assist in the delivery of the coating
material to the gun nozzle, while low pressure air is used to atomize the coating material at the spray
head. A high volume of air pushes the coating material, forming a very soft, low velocity pattern. The
soft spray generally provides more consistent coverage and a better overall finish.
With this spray method, the gun is held closer to the surface (15 to 25 cm) than with other methods
because of the lower speed of the paint particles. Moreover, this method applies, in a single coat, a
thicker film than any other spray method here described.
1.3.2 Airless Spray Method
In the airless spray system no air pressure is used. Instead, hydraulic pressure is used to deliver the
coating material to the gun head. The paint is atomized by ejection from special spray nozzles that
increase the pressure by a factor of about 100. The paint droplets, moving toward the surface by their
momentum, are appreciably slowed down by air resistance. This method produces less bounce of the
coating material on arrival at the work surface and, therefore, less over-spray.
The paint is not cooled by the expansion of the air as in the conventional spray method, so the only
heat loss is through solvent evaporation.
1.3.3 Hot Spray Method
Hot spraying is the application of coatings with HVLP or airless spraying system using heat as a sub-
stitute for all or a portion of the thinner, generally used to reduce viscosity of the coating material. It is
most frequent and efficiently used with airless spray systems.
The hot paint is cooled rapidly when atomized but retains sufficient heat to still be close to the
ambient air temperature when it reaches the work surface. So the possibility of blushing, that is due to
condensation of moisture, is reduced and it is possible to spray under high humidity conditions. However,
heating the paint reduces its pot life (period in which the chemicals remain usable when mixed).
1.3.4 Air-Assisted Airless Spray Method
The coating material is atomized by hydraulic pressure as in the airless spray system but at a much
lower pressure. Low pressure air is added at the gun head through jets at the nozzle and directed at the
paint mist to control and form the spray pattern. This allows the operator to control the atomized coating
pattern which cannot be done with standard airless systems.
This method offers the same advantages of the airless method while being safer and requiring lower
maintenance on pumps, due to the lower hydraulic pressure. Moreover, the appearance of the coating
applied is better as the tendency to obtain a bumpy surface (usually called orange peel) is lessened.
7
1.3.5 Electrostatic Spray Method
This method is a modification of the spray methods previously described. It adds the feature of electro-
static charging of the paint material which is attracted by the grounded work piece. The coating material
can be charged either inside the gun or at a fine metal probe at the gun nozzle exit, where the latter is
the one generally used. This method achieves the best result with airless spray because the low velocity
of the paint particles and the electrostatic attraction produces a high transfer efficiency rate.
This method reduces over-spray and allows a better painting of hard-to-coat areas such as edges.
The Faraday, however, effect limits the effectiveness of this method in coating interior corners, crevices
and cavities.
1.3.6 Spray Painting Techniques
Different techniques are important when handling a spray gun. The coating quality is highly dependent
on how well these techniques are applied. This section describes the most important spray techniques
and theirs effects on the final result.
When spraying a surface, the distance of the gun to the aircraft surface depends on the width of the
spray pattern desired and on the type of gun used. If the spray gun is too far from the surface, it will
result in a dry spray, called dusting, and an excessive over-spray. Contrarily, if the spray gun is kept too
close to the surface, the coat will be too heavy, developing a tendency for sags or runs.
Stroking is the movement of the spray gun along the surface to paint. It is essential to maintain the
spray gun at the same distance to the surface, move it at the same speed, and hold it at the same time
perpendicular to the surface throughout the pass. Generally, the painters, when in an uncomfortable
position or fatigued, tend to arc the stroke, causing an uneven application with a thicker coat in the
middle of the stroke than at the ends. The rate of stroke should be uniform to produce an even wet coat.
Each stroke should be parallel to the other with a 50% overlap.
The technique that ensures the best coating integrity and coverage is the cross coating. It consists of
applying two layers of coat, one with a stroke perpendicular to the other. This technique should always
be used when applying multiple coats, except for high solid primers and topcoats with HVLP, airless, or
air-assisted airless equipment.
Another important technique is the triggering. It is the pressure and release of the spray gun trigger
during the stroking. This is a difficult technique that requires a long experience to be acquired. Generally,
the painter should press the trigger after the beginning of the stroke and release it before the end.
1.4 Aircraft Painting Process
This section presents a description of the aircraft painting process according to the technical manuals [5,
12].
The painting of an aircraft is accomplished by at least two painters supported by helpers. For large
aircraft, it may be necessary to increase the number of painters. Small aircraft are generally positioned
8
with the tail towards the exhaust filter bank of the painting hangar and vice-versa for larger aircraft. The
process will be described for small aircraft. For large aircraft the sequence of events presented has to
be inverted
1.4.1 Priming
The first coat to apply is the primer (see Figure 1.8). The priming starts at the end of the aircraft near
the exhaust filter bank and moves towards the air supply.
Figure 1.8: Primer application on aircraft fuselage5.
Starting from the tail, the priming begins from the higher surface (horizontal stabilizer for ”T” tail
aircraft, vertical stabilizer(s) for different configurations). The horizontal stabilizer is primed starting from
the upper surface at the center moving outboard to the tip stroking perpendicularly to the leading edge.
Then, the lower surface is primed, and finally the primer is applied to the edge from the outboard to the
junction with the vertical stabilizer. The priming of the vertical stabilizer(s) starts at the top and leading
edge of each side moving down and aft with vertical strokes. Finally the leading edge is primed from the
top down.
After the tail, a full wet coat of primer is applied at the aft section of the fuselage starting from the
aft end and the top moving forward and down with vertical strokes to the junction with the wing trailing
edge.
The wings are primed starting from the lower surface at the tip moving inboard with a stroke perpen-
dicular to the leading edge. Then the lower fuselage between the wings is primed. A full wet coat of
primer is applied to the upper surface of the wings and the upper fuselage with the same technique as
of the lower surface.
Finally the primer is applied to the forward section of the fuselage starting at the wing leading edge
and the top moving down and forward to the nose with a vertical stroke.
5URL http://www.flyingcolorspaintandinterior.com/paint/ [Accessed: 03 September 2016]
9
1.4.2 Topcoating
Topcoats can be applied in one coat as well as two coat system. The one coat system is applied with a
stroke in one direction followed by a stroke in the perpendicular direction working small areas at a time
(see Figure 1.9). The two coat system is applied stroking the two coat in perpendicular directions.
The topcoating follows the same process sequence of the priming but backwards. It starts from the
nose of the fuselage and ends on the upper part of the tail.
After topcoating, it is important to allow the paint to cure in a dust-free atmosphere.
Figure 1.9: Topcoat application on aircraft fuselage6.
1.5 Thesis Motivation and Objective
In many industries the painting process is a completely automatic procedure for over 50 years. This
is especially true for the automotive industry [13]. The aerospace industry has an automation delay
compared to other industries for many reasons (complexity, regulatory, materials, etc.). Dan Friz, director
of business development, KUKA Systems (Shelby Township, MI), said ‘The major automation challenge
within the aerospace industry is simply the aircraft was never designed for an automated process.’
The fleet of commercial airplanes is growing. Boeing, for example, forecasts that by 2032 more than
35,000 new airplanes will be built [14]. To manufacture and maintain all these airplanes, the whole
process has to get faster and cheaper, reason for which many aerospace manufacturers, especially in
commercial aviation, are looking for solutions to automate the aircraft finish system maintenance.
The automation of the aircraft painting and paint removal procedures leads to many benefits for both
workers and companies. The robotic application of paint on the aircraft can lead to material savings
of between 30 to 50%, which means a 30-50% savings in airplane weight as well [14]. This weight
saving interests the airplane manufacturers because it leads to a fuel saving for airline customers, with
economic and environmental benefits.
6URL http://www.csiro.au/en/Research/MF/Areas/Chemicals-and-fibres/Materials-for-industry-and-environment/
Coatings-and-surfaces/TopCoat [Accessed: 26 September 2016]
10
On the other hand, both painting and paint removal processes require a number of highly skilled
workers operating for a long time on each aircraft, in a polluted environment (see Figure 1.10) and often
without the possibility to pause the ongoing process. The introduction of an automatic system reduces
the number of workers involved and their exposition to health hazards. It also ensures the possibility to
process an aircraft as soon as it is possible reducing the time waste due to the employees working time.
OGMA Industria Aeronautica de Portugal, S.A founded in 1918 in Alverca, is an international player
in aerospace maintenance, repair and overhaul, and manufacturing business. The company is nowa-
days accomplishing the whole finish system maintenance of both military and civil aircraft completely
manually. The subject of this study is to make the preliminary design of an automatic painting and paint
removal system on behalf of this company.
Figure 1.10: Aircraft painter at work7.
1.6 Existing Automated Solutions
The first painting robot was presented in 1967 by the Norwegian company Trallfa that was producing
wheelbarrows. It was an electro-hydraulic robot which could perform continuous movements and that
was meant for their internal use only. It was developed into a commercial success as in 1985 ASEA
(later ABB) took over Trallfa [15].
From the 1960s on, robots have been taking the place of human workers in carrying out many
manufacturing tasks increasing the productivity in many industries. This transition is still going on, the
market research company Forrester says in a report that robots will eliminate 6% of all US jobs by
2021 [16].
Many projects of interest for this study have been developed. The majority accomplish only the
painting of surfaces, a process involved in many manufacturing processes, while only few concern the
paint removal and none was found about the masking and demasking.
All systems developed to paint surfaces always involve the use of a robotic arm spraying on the
surface and, except for small parts, any other technique (brush, roller, etc.) has not been experimented
so far.
The automotive industry completely automated the painting of almost all car parts not only because of
the production volume but also because of the quality and consistency aspects [17]. Chad Henry, North7URL http://excelaviation.ie/jobs/aircraft-spray-painter/ [Accessed: 26 September 2016]
11
American Sales Manager at Staubli Corporation in Duncan, South Carolina, speaking about robotic
painting said ‘You simply cannot get repetitive performance with manual operation.’
Generally, in car manufacturing, the robot base stays steady while the part to be painted moves along
the assembly line as in Figure 1.11. An exception is the system developed by Aerobotix, that using two
rail mounted FANUC P-250iA/15 robots, is able to paint camouflage pattern on military vehicles [18].
Figure 1.11: Hyundai Alabama robotic painting8.
The robotic leader industry ABB automated the coating of 80 meters long wind turbine blades using
two rail mounted IRB 5400 robotic arms (Figure 1.12), achieving a reduction of energy consumption by
up to 60 percent and of paint consumption of 25 percent compared to the standard paint application [19].
Figure 1.12: Rotor blades for wind power systems coating achieved by ABB’s painting robots IRB 540010.
In another field, Figure 1.13 shown the mining trucks cleaning system designed by ABB. It uses high
pressure waterjet from a IRB 6650S robot mounted on movable rails, saving 60 percent of the time
compared to the manual process [20].
Another interesting project is the system developed by the University of Technology of Sydney for
steel bridge maintenance. It is able to remove the paint and rust from steel bridges with a grit-blasting
technique. This system avoids that workers stay in partially closed spaces with highly polluted air or in
dangerous freefall conditions [21]. The system is composed of a robotic arm able to sense the workable
8URL http://www.hotrod.com/articles/14-steps-great-paint/ [Accessed: 26 September 2016]10URL http://www.abb.com/cawp/seitp202/90c5c8ab0a1fd46ac125759a003ec090.aspx [Accessed: 27 September 2016]11URL http://www.pngindustrynews.net/commonlib/ImageEnlarge.asp?strImageFileName=ABB_robots.jpg [Accessed:
27 September 2016]
12
Figure 1.13: ABB’s robotic mining truck-washing system in Brazil11.
surface inside its workspace and complete the paint removal task autonomously. However, it is not able
to move its base by itself.
Figure 1.14: Robotic system cleaning up the Sydney Harbor Bridge12.
In the naval industry there are many projects for the robotic removal of paint from the ships hulls
while few studies have investigated the paint application [22]. Differently from the previews projects, all
the ships paint removal robots found are able to attach their structure to the surface and move along it.
The most popular naval paint remover is the M-2000 in Figure 1.15. It is a semi-autonomous robotic
paint removal system, built out of a partnership between NASA’s Jet Propulsion Laboratory (JPL), the
National Robotics Engineering Consortium (NREC) at Carnegie Mellon University, and UltraStrip, that
strips paint from ships hulls [23]. It attaches itself magnetically to the hull of a ship, with a vacuum hose
running from it. A controller helps navigate the robot along the surface of the ship with 360 degrees
of movement. The M-2000’s high-pressure water jet generates 270 MPa of pressure to blast away the
paint right down to the ship’s steel substrate. The water and the stripped paint are then captured by the
vacuum system.
Research papers about automatic painting and/or paint removal systems in the aerospace industry
go back to the early 90’s [24] but just during the last years the aerospace companies began implementing12URL http://www.pulse-pr.co.uk/service-robots-revolutionize-clean-up-of-the-sydney-harbor-bridge-85.asp
[Accessed: 27 September 2016]13URL https://spinoff.nasa.gov/spinoff2000/er1.htm [Accessed: 7 October 2016]
13
Figure 1.15: UltraStrip Systems, Inc.’s M-2000 removing paint from the hull of a ship13.
these systems in their manufacturing process [14]. Most of the projects are still under development and,
with few exceptions, concern only the paint application or the paint removal.
During the 90’s the NASA and the USBI Company developed a robotic high pressure waterjet system
for the paint stripping of the Space Shuttle solid rocket boosters. Years later, this project leaded to
the Large Aircraft Robotic Paint Stripping (LARPS) designed by the Pratt & Whitney Waterjet System,
a commercial robot able to remove the finish system from civil and military aircraft by high pressure
waterjet [25], but the project never came to an end.
The Carnegie Mellon University’s NREC and Concurrent Technologies Corporation (CTC) of John-
stown developed, and are now testing, the Advanced Robotic Laser Coating Removal System (ARLCRS)
in Figure 1.16. It uses a continuous wave laser mounted on a state-of-the-art mobile robot to remove the
coating system from medium to small size military aircraft. It is also able to avoid selected areas limiting
the masking needed [10].
Figure 1.16: Advanced Robotic Laser Coating Removal System14.
Another similar project under development by STRATAGEM is the Laser Coating removal Robot
(LCR). This system uses a 20 kW CO2 laser to evaporate and combust the paint that is immediately
14URL http://www.nrec.ri.cmu.edu/projects/ctc/ [Accessed: 27 September 2016]
14
vacuumed from the surface and passed through a filtration system. The laser is mounted on a eight
Degree of Freedom (DoF) robotic arm and four DoF mobile platform (see Figure 1.17). On the end-
effector is also mounted a forward scanner to check the aircraft geometry in real-time [9]. STRATAGEM
expects 50% reduction in processing time and 90% labor reduction. Moreover, STRATAGEM is planning
to implement on the same structure paint spraying tools to build also an aircraft painting system.
Figure 1.17: Laser Coating removal Robot15.
Regarding the coating of aircraft there are two military projects both for stealth coating of fighting
aircraft: the Sandia F117 robot for the coating of the F117 Nighthawk and the Robotic Aircraft Finishing
System (RAFS) developed by Lockheed Martin for the F-35 coating (see Figure 1.18). The first system
has three robotic arms: two rail-mounted and one floor-mounted, all with seven DoF. In 1999 it has
successfully painted the first F-117 Nighthawk fighter [26]. The latter coated its first aircraft in 2008, it
comprises three FANUC R2000iA 125L robots, each with six DoF, mounted on auxiliary axis rails [27].
Figure 1.18: Robotic Aircraft Finishing System [27].
Another painting system, shown in Figure 1.19, was developed by Boeing to automatically coat the
B-777 wings. It has two robotic arms rail mounted and is called Automated Spray Method (ASM). The
system is replacing 35 to 40 painters and is able to apply two different paints at two different thickness
at the same time. The company is now planning to extent the process to other parts of the aircraft [28].
15URL http://www.lcrsystem.com/ [Accessed: 27 September 2016]16URL http://mashable.com/2013/06/04/boeing-777-robots [Accessed: 27 September 2016]
15
Figure 1.19: Robotic system coating the B-777 wing16.
1.7 Thesis Outline
Along the present thesis the design of an aircraft automatic painting and paint removal system is ex-
plained in detail.
Specifically, in Chapter 2 the system requirements are expounded. Then, different possible solutions
are illustrated and evaluated. Between these, a trade-off process is carried out to select the solution to
be developed.
In Chapter 3, the system components are designed and a cost estimation is performed. Finally, in
Chapter 4, a general overview of the present work is carried out giving a path for the future work to be
executed for the completion of the project.
16
Chapter 2
Finish System Automatic Maintenance
Solutions
Along this chapter the problem is analyzed in detail and a set of possible solutions is presented.
The subject of this study is the automation of aircraft painting and paint removal. The automation
of the other phases of the finish system maintenance (i.e. masking, cleaning etc.) has not been taken
under consideration for two specific reasons: there is a lack in the technology development (masking)
and the advantages for a further development into automation compared to the development time and
investments required have been considered neglectable.
To satisfy the client requirements neither a painting nor a paint removal method is selected. The
present study has been confined to describe the possible solutions available and leave the painting and
paint removal method selection to the client.
2.1 Specifications and Requirements
The automation of painting and paint removal is influenced primarily by three factors: the aircraft size
and shape, the paint application requirements and the paint removal requirements. In this section these
factors are analyzed in detail.
2.1.1 Aircraft impact on the design
The automatic system has to process any aircraft the company is presently maintaining and possibly the
ones it will maintain in the future, in order to provide a real economic benefit. Generally the airplanes
are different in size and shape. While the dimension of the aircraft influences the size of the system
workspace, its shape determines the dexterity the system needs to process all the required surfaces.
The system will be sized on the Lockheed C-130 Hercules (Figure 2.1). This is the bigger aircraft
OGMA is maintaining as well as the one with the lower distance between the fuselage and the ground.
The geometrical features of this aircraft are set out in Table 2.1 where the heights are measured with the
17
retracted landing gear. To ensure the possibility to process differently shaped aircraft, the system will be
oversized.
Figure 2.1: A C-130E Hercules from the 43rd Airlift Wing, Pope Air Force Base, N.C.1
Table 2.1: C-130 H geometrical features [29, 30].
[m]
Length 29.3Height 11.4Wingspan 39.7Wing root chord 4.9Fuselage height 4.6Fuselage width 4.3Landing gear height 0.52
The four dimensions in Figure 2.2 are used to size the system workspace. Moreover, to size a system
composed of a moving vertical structure another dimension is required. This is computed as the lenght
of the hypotenuse of a right triangle having as legs half the wing root chord and half the fuselage width
(see Figure 2.3). Then, a system composed of a vertical structure to reach any point of the surface has
to locate its end-effector 3.25 m away from it. Applying a conservative approach, the above length is
multiplied by a factor 1.25, obtaining 4 m.
Figure 2.2: C-130H side and front views with dimensions.
The dimensions of the system workspace are here defined in accordance with the previews results.
The minimum height the system has to reach is 13 m while the width and length are 45 m and 35 m1URL http://www.af.mil/shared/media/photodb/photos/990101-F-5502B-002.jpg [Accessed: 28 September 2016]
18
Figure 2.3: C-130H top view with dimensions.
respectively. Furthermore, it also has to process the bottom of a fuselage located 0.5 m over the ground.
This initial sizing of the automatic system is useful in order to carry out a preliminary analysis on different
designs.
2.1.2 Painting requirements
To paint an aircraft the system has to handle one of the methods described in Section 1.3. To do it the
required equipment is: an air compressor, a paint tank and a spray gun. The system has to handle
this equipment and to apply the paint with the required tickness following the technical prescriptions [5].
Moreover, the system has to clean the hoses and the spray gun, and to be able to change paint.
To paint an aircraft, at least one painter at each side of it is needed. This is because the overlap
of the coating has to happen while it is still wet. With only one painter it would not be possible to do it
because the paint would dry on the middle line of the fuselage causing a thickness discontinuity. For the
same reason, stopping and restarting the painting process reduces the quality of the final result.
Because of the solvents inside the paint spread in the air, in the painting area the antiexplosion
regulation ATmosphere EXplosibles (ATEX) is applied [31]. The ATEX directory divides the working
area into different zones depending on the explosion risk. According to it all equipment has to be ATEX
certified.
2.1.3 Coating removal requirements
The requirements for the automation of the paint removal depend upon the removal method selected
by the client among those described in Section 1.2. In this section, each method requirements are
described.
To automatize the chemical paint stripping it is possible to use the same equipment used to paint
changing the spray gun, because to atomize the paint stripper is not allowed. The thickness control and
precision required is lower compared to the paint application [5]. The system has also to agitate the
whole surface and scrape all loosened coatings with a squeegee. At the end it has to clean the surface
19
with hot water.
This removal method needs a long time because of the chemical reactions involved, thus the main
advantage of a robotic execution is not to have workers doing an hazardous task. Moreover, to make the
chemical stripping completely automatic, the system must be able to handle many different tools (spray
gun, squeegee and water gun) increasing the complexity of the system as well as its cost.
The paint removal by mechanical methods requires a higher system precision and accuracy because
there is a risk to damage the aircraft structure and equipment. Moreover, the system has to know when
the paint is removed from the area it is working.
If an electric sander is used, the system has to hold the sander, keep it in contact with the surface
and react to the friction forces acting on it. On the other hand, the PMB, MPW and abrasive blasting
methods require the system to handle a blasting equipment and a media or water tanks. Moreover, it
has to react to the forces generated by the blasting as well as keep the end effector at the right distance
and orientation with respect to the surface.
Finally, the laser removal requires the system to supply and handle with high precision and accuracy
a laser equipment. It also has to sense when the paint is removed from the area it is working.
2.2 Possible Solutions
Along this section various automatic solutions for both painting and paint removal are analyzed and
evaluated.
Taking into account the systems described in Section 1.6 as well as the above specifications and re-
quirements, two different general solutions are evaluated: a robotic arm mounted on a movable structure
and a multirotor Unmanned Aerial Vehicle (UAV).
Relatively to the robotic arm solution, there are many possible structures able to position the arm
in the space. In the present study four different designs are presented and evaluated. Assuming the
system is able to work in any painting hangar, the proposed designs do not produce any load on the
hangar structure besides the system weight on the ground.
2.2.1 Multirotor UAV
This design is made up of a multirotor UAV able to fly in the painting hangar. The UAV holds a three
rotational DoF robotic wrist that controls the orientation of its end-effector. In Figure 2.4 the system is
illustrated with a sketch.
The flying platform can position the end-effector in any point of the workspace while the robotic wrist
can control its orientation with respect to the work surface. In this way it is possible to process the whole
surface of the aircraft.
During the painting, the UAV carries the electric power storage, the paint tank, the air compres-
sor, while the end-effector is a spray gun. The paint tank, the spray gun, and the compressor can be
substituted for the paint removal with the dedicated equipment.
20
Figure 2.4: Multirotor UAV system sketch.
The implementation of most of the removal methods of Section 1.2 on this system is an issue. To
install a motor-driven abrasive remover can be difficult due to the torque the propellers have to counter-
act. Also the MPW, PMB and grit blasting methods are hard to implement because of the big amount of
media the vehicle has to lift and the force due to the blasting. The laser removal is difficult to implement
too because it needs a big amount of electrical power, thus heavy batteries, and very high precision in
positioning with respect to the work surface.
This solution presents some benefits mainly due to its fly ability. The UAV can reach any point of
the hangar, thus it can process aircraft of any size. Moreover, it is able to paint with stroke as long as
needed achieving a better final result. Furthermore, the multirotor structure allows an easy maintenance
of the system because the maintainer has clear access to all the component of the system.
The use of UAVs gives advantages as well as many disadvantages. Due to their design, they produce
a lot of noise while flying. They have low flying autonomy especially in heavy payload applications as this
one [32]. Moreover, they have poor positioning and accuracy skills as well as vibration problems [33].
A common problem with the multirotor platform is the failure of one of the engine causing the failure
of the whole system. In this application the problem is even bigger because a failure of the system can
cause serious damages to the aircraft structure if it falls on breakable parts. Not least, the air flow due
to the propellers can affect the painting final result.
The precision control of multirotor UAVs is an hard task, and in this application the weight change due
to the paint (paint remover) consumption increases the complexity of the problem. The more accuracy
and precision required, the more work has to be done on the control of the UAV, thus a long development
time is expected.
21
2.2.2 Rail Mounted Robot
A robotic arm is mounted on a three linear DoF rail able to move it everywhere in the workspace defined
in Section 2.1.1 (see Figure 2.5). Because the rail structure is big and heavy, its motion is supposed to
be slow. Therefore, it locates the robotic arm close to the surface to process and the arm positions and
directs the end-effector, ensuring a proper final result. Thus, the robotic arm shall have at least 6 DoF
for both the position and orientation of the end-effector along the aircraft surfaces.
Figure 2.5: Rail mounted robotic system sketch.
This design allows an accurate positioning of the end-effector because it is possible to determine the
position of the arm along the rails with precision.
The rails ensure a energy supply line to the robot as well as a supply of compressed air, therefore
there is no need of placing batteries and air compressor close to the robotic arm. Thanks to this, the
laser paint removal method, that requires a big amount of electric energy, is easy to implement. In this
way, all the paint removal methods of section 1.2 can be implemented on this design.
Due to the big workspace, the rail structure has to be really long, heavy and expensive. A limit of this
solution is that it is impossible to use the present system to process aircraft bigger than the workspace.
For this system, positioning the aircraft with precision inside the workspace is essential. Furthermore,
it has a problem processing some part of the aircraft especially if complex shapes are involved. For
example, to reach the nose of the aircraft the rail gets as close as possible to the fuselage but then a big
robotic arm is needed to go from the rail to the nose increasing the cost of the whole system.
Because of the paint requirement of subsection 2.1.2 at least two robots at opposite sides of the
aircraft are needed. In this architecture, the motion of the two robots is not completely independent one
from the other but the elevation of the rail must be the same to share part of the structure.
22
2.2.3 Mast Mounted Robot on AGV
A robotic arm is mounted on a mast that is located on an omnidirectional Automatic Guided Vehicle
(AGV) able to move in any direction as well as to perform zero radius turns [34]. The mast rises and
lowers the robotic arm while the AGV positions and rotates it [34]. The robotic arm is then required for
the motion and orientation of the end-effector along the surface, requiring at least 6 DoF. A sketch of this
solution is shown in Figure 2.6.
Figure 2.6: Mast mounted robotic system on AGV sketch.
From Section 2.1.1 considerations, the length of the robotic arm workspace has to be of 4 m in a
ground parallel plane. Painting robotic arms of this size are not available on the market, so it has to be
made ad hoc for this system, increasing the cost of the system and the development time. A bigger arm
for equal DoF has generally a lower dexterity, which can be a problem with complexly shaped aircraft.
Some features of this design are due to the use of an omnidirectional AVG and are common to all the
following solutions. It adds to the system 3 DoF, i.e. the system can move everywhere on the ground as
well as rotate around a vertical axis. Therefore, a high precision in positioning the aircraft in the hangar
is not required. However, complex systems to localize and determine the attitude of the system are
required [35, 36].
23
All the heavy and cumbersome parts of the system, i.e. tanks, batteries, compressor, etc., are
located on the AGV platform. Nevertheless, because each wheel has a maximum load the positioning
of the center of gravity on the AGV is important.
This design can accommodate any of the finish system removal methods described in section 1.2
but, depending of the material of the AGV wheels, the chemical stripping can be unsuitable, reacting with
the wheels material. Furthermore, the paint removal can be accomplished by just one robot because it
is able to move around the whole aircraft while for the coating always two robots are needed.
2.2.4 Multi DoF Structure on AGV
A robotic arm is located at the tip of a beam which is supported and moved up and down by a lifting
system mounted on an AGV. The beam is able to move back and forward along its axis as well as to
rotate around the joint with the lifting structure (see Figure 2.7).
Figure 2.7: Multi DoF robotic system on AGV sketch.
In accordance with Section 2.1.1, the length of the beam plus the length of the extended arm has to
be at least four meters. While the height of the lifting system depends on the height of the AGV, on the
maximum height reached by the robotic arm and on the beam length and maximum rotation angle.
24
The idea behind this solution is to use a small and light robotic arm with a structure able to position
it at any point of the work surface. However, the robotic arm has to have 6 DoF at least to control the
end-effector orientation with respect to the surface as well as to control the stroke of it because the other
motions of the structure are supposed to be too slow to control it adequately.
With a six DoF robotic arm, this architecture has 12 DoF. It allows the robot to work on complexly
shaped aircraft. Nevertheless, it needs a sophisticated control software increasing the development
time and system cost. Moreover, it is possible to have precision and accuracy problems due to the many
movable joints involved. For the same reasons, the reliability of this system is supposed to be low.
Mainly not to overload the beam, most of the equipment of the system should be located on the
AGV. So does the paint (paint remover) tank. The paint (paint remover), the electric power and the
compressed air have to be supplied to the robotic arm. Because the lifting structure and the beam can
move and rotate, the design of the supply lines is expected to be difficult.
2.2.5 Lifting Structure on AGV
A robotic arm is located at one end of a beam which has its longitudinal axis in a ground parallel plane.
The beam is supported on the other end and moved up and down by a lifting system positioned on a
omnidirectional AGV (see Figure 2.8).
The AGV and the lift, providing the system with 4 DoF, position the robotic arm with respect to
the work surface. The arm has to position and direct its end-effector, thus it needs at least 6 DoF.
Accordingly, the system has 10 DoF.
Compared to the previous solution, the lifting structure is higher and therefore heavier. However, the
lack of the two movable joints between the beam and the lifting structure makes the control of this robot
and the design of the supply lines to the robotic arm easier. It also increases the reliability of the system
as well as its precision and accuracy.
2.3 Design Selection
Along this section a trade-off between the designs expounded in Section 2.2 is carried out. According
to the K. Otto and K. Antonsson trade-off strategy [37], the most important features for the final design
are also selected. A weight from 1 to 5 is assigned to each feature according to the importance it has in
the project. Then, a mark per feature is assigned to each design of Section 2.2. The marks, from -5 to
5, follow the analysis made in the previous Section 2.2.
To accomplish a selection between the five solutions, each mark is multiplied by its weight and then
summed to the other marks of the same solution. The chosen design is the one with the higher final
score. This trade-off process is resumed in Table 2.2. The selection criteria are rewarding the design
with the lowest cost and complexity from both the mechanical and control point of view. A higher weight
has been given to features like cost and control but also development time and reliability. From the
performance point of view, accuracy and repeat ability have been considered important as well as the
25
Figure 2.8: Arm and lifting robotic system on AGV sketch.
adaptability of the system to different aircraft shapes and sizes.
Table 2.2: Possible solutions trade-off.Weight Multirotor UAV Rail Mounted Mast on AGV Multi DoF on AGV Lift on AGV
Control 5 -3 3 3 0 3Cost 5 3 -2 -4 2 3Simplicity 4 1 3 3 -1 4Accuracy 4 -4 4 3 1 2Repeatability 4 -3 4 2 2 3Development time 4 -1 3 -4 2 3Adaptability 4 5 -4 3 4 3Multiple tasks 3 0 2 4 2 2Reliability 3 -1 4 2 1 3Maintenance 3 5 2 2 2 3Rapidity 2 3 2 2 0 2Payload 1 -4 4 2 2 3
Result 6 77 59 53 120
The solution selected by the trade-off analysis is the lifting structure on AGV. It has also been selected
26
because between the AGV mounted systems it is the cheapest and the simplest solution. The UAV
solution has big control problems and this impairs its selection. The rail mounted system, on the other
hand, has costs and adaptability problems.
27
28
Chapter 3
System Design
In this section the preliminary design of the automatic system is presented.
From the trade-off analysis of Chapter 2 the solution selected is composed of a small robotic arm
located at the tip of a cantilever beam moved along a vertical axis by lifting system mounted on an AGV.
This system has to reach any point of the workspace in accordance with Section 2.1.1. Moreover, it has
to ensure a proper paint application and removal.
As already mentioned, the system is designed in order to allow the client to choose the painting and
paint removal methods to implement.
To start designing the components of the structure it is necessary, first of all, to select the robotic arm.
This is essential to know the load the structure has to support as well as the dimensions of the other
parts of the system. Once the weight and the workspace of the robotic arm are known, the horizontal
beam and the lifting system (lift) are designed. With the selection of the other subsystems required, the
weight the AGV has to support is known, then it is selected. Finally, the location of the parts on the AGV
is decided to avoid an overload of its wheels.
3.1 Robotic Arm Selection
In this project an off-the-shelf robotic arm is used for the following reasons: the design and manufacture
of a robotic arm would increase the development time and the system costs (especially because of the
safety certification involved), secondly the arm control system is already developed and implemented.
Many robotic arms are available on the market. To select one of these, the following criteria are
applied:
• Lightness
• Workspace equal to or bigger than a human painter
• ATEX certification
• Production company able to ensure spare parts supply in the next decades
29
• Different end-effector tools
The arm has to be light to reduce the weight of the whole structure on the AGV. To be light the arm
has to be small, but to ensure a proper length of the strokes it needs a workspace equal to or bigger than
an human arm. Moreover, by regulation the equipment handled by the workers has to be ligher than 5 kg,
thus, to install these tools on the arm, its maximum payload can not be less that 5 kg [38]. The human
arm workspace, shown in Figure 3.1, has a maximum reach in the horizontal plane of 60 cm as well
as on the vertical plane. Its shape is approximately three quarter of a sphere. Following the preceding
criteria different robotic arms are selected from the catalog of the most solid robotic industries.
Figure 3.1: Human arm workspace [39].
The specifications of different robotic arms chosen following the above criteria are shown in Table 3.1.
To have an approximate indication of the workspace dimensions, in Table 3.1 reach means the maximum
extension in a horizontal plane.
Table 3.1: Different robotic arms specifications.
Robotic arm Reach [mm] Weight [kg] Payload [kg]
FANUC Paint Mate 200iA [40] 704 35 5FANUC Paint Mate 200iA/5L [41] 892 37 5Motoman EPX 1250 [42] 1250 110 5ABB IRB 52 [43] 1200 250 7
The two robots by FANUC are the lightest thanks to their reduced dimensions and their aluminum
structure. Between these two is selected the FANUC Paint Mate 200iA/5L (Figure 3.2) because of its
higher reach compared to the FANUC Paint Mate 200iA that weights 2 kg less but has a reach lower by
188 mm. The bigger reach allows a reduction of the supporting beam length as well as the possibility to
work larger surface without moving the whole system.
Important features of this robotic arm are its avarage power consumption of 0.5 kW, the dimention
of its footprint (260 x 265 mm), and the possibility of mounting it in any position with respect to the
ground [41].
30
(a) FANUC PaintMate 200iA/5L. (b) FANUC PaintMate 200iA/5L workspace in mm.
Figure 3.2: FANUC PaintMate 200iA/5L and its workspace [41].
3.2 Structure Design
According to Section 2.1.1, the system has to reach a height of 13 m and be able to locate its end-effector
at least 4 m away from the lift in a horizontal plane.
The robotic arm can extend up to 1267 mm from its base in the vertical plane (see Figure 3.2(b)).
To make a conservative design, the lift structure height is computed without take into account the AGV
height. So the height of the lifting system is Hlift = 13− 1.3 = 11.7 m. Following the same process, the
length of the beam should be 3.1 m but is chosen to use a 3.5 m long beam to oversize the system.
3.2.1 Horizontal Beam Design
The horizontal cantilevered structure supporting the robotic arm can be approximated as a uni-dimensional
structure, then easy computations about its deformation are done using the Euler-Bernoulli beam theory.
The weight of the robotic arm and the weight of the beam itself cause a bending moment on the
cantilever beam. It is important to decide how much the tip of the beam is allowed to move downward
from its ideal position taking into account that the final position of the robotic arm depends also from the
deformation of the lifting structure.
The forces acting upon the beam are the weight of the robotic arm and the weight of the beam itself
as shown in Figure 3.3. The weight of the robotic arm Warm is known (37 kg) and is applied at the
tip of the beam, this is an approximation as the arm center of gravity changes position with its motion.
However, the robotic arms are made with a structure getting lighter from the base to the tip [13], thus
this is a good approximation. On the other side, the weight of the beam is applied in the beam center of
gravity and depends by its material and shape.
31
Figure 3.3: Forces acting on the horizontal beam.
The materials generally used for structural beams are steel and aluminum alloys. In this project an
aluminum alloy beam is used to have a lighter structure. Between the aluminum alloys the Al 6061-T6
is selected. It is the most commonly used aluminum alloy for structural applications, has above average
corrosion resistance, good machinability, and is excellent for welding [44]. Its properties are shown in
Table 3.2.
Table 3.2: Aluminium alloy Al 6061-T6 properties [44].
Density ρ 2.7·10-6 kg/mm3
Ultimate Tensile Strength Sp 310 MPaTensile Yield Strength Sy 276 MPaModulus of Elasticity E 68.9 GPa
It is decided to use a constant section beam to have easy structure computations and a cheap off-
the-shelf part. The cross section of the beam determines how the beam deforms under a load. To have
a pure bending, a symmetrical cross section is selected and specifically a I-section beam that has high
strength to weight ratio. I-sections beams are generally rolled or extruded, thus are produced in large
quantities at economic prices [45].
To select the dimensions of the beam section an iterative computation is used. The American Society
for Testing and Materials developed a standard for the section of metal beams, the ASTM A6 [46],
where standard dimensions and properties of I-section beams are listed. It is possible to make a table
of different I-sections sorted according to the cross section moment of inertia Ix about the x axis of
Figure 3.4. A sample of such table is shown in Table 3.3 where h, s, and w are the dimensions in
Figure 3.4.
Table 3.3: Part of the bean cross section properties table.
Metric Depth Width Web Thickness Sectional Area Weight Ix Iymm x mm x kg/m h [mm] w [mm] s [mm] [cm2] [kg/m] [cm4] [cm4]
W 310 x 310 x 44.2 318.0 308.0 13.1 165.0 44.22 30770.0 10040.0W 310 x 310 x 48.8 323.0 309.0 14.0 182.0 48.78 34760.0 11270.0W 310 x 310 x 53.9 327.0 310.0 15.5 201.0 53.87 38630.0 12470.0W 310 x 310 x 61.1 333.0 313.0 18.0 228.0 61.10 44530.0 14380.0
The bending moment Mx acting about the x axis of the beam is given by Equation 3.1 where Wbeam
32
Figure 3.4: Beam cross section.
is the beam weight, Warm the robotic arm weight and lb the length of the beam. As a result of the Euler-
Bernoulli theory the tip deflection dbeam of a cantilever beam is given by Equation 3.2, where E is the
Modulus of Elasticity of the beam material and Ix is the moment of inertia about the x axis.
Mx =Warm · lb +Wbeam · lb/2 . (3.1)
dbeam =Mx · l2b2 · E · Ix
=⇒ Ix =Mx · l2b
2 · E · dbeam. (3.2)
The iterative computation calculates the bending moment starting from a zero weight beam, then
computes the moment of inertia needed to have the imposed tip deflection from Equation 3.2. Known
the moment of inertia, it looks into the ASTM A6 sorted table for the first section with a moment of inertia
bigger than the required one. From the weight per meter column of the table it computes the weight of
the beam, then the process restarts. It stops when it selects twice the same section.
A parametric study of the beam weight as a function of the maximum tip deflection has been per-
formed. The result of the study is set out in Figure 3.5. The graph illustrates that the weight drops to
53.2 kg for a displacement of 2.75 mm, then it remains stable. Because the weight decreases with an
average gradient of 170 kg/mm, to have a light structure the beam section of Table 3.4 is selected. Its
weight is 53.2 kg corresponding to 2.75 mm of tip displacement. This displacement is constant during
the whole operative life of the system, is then possible take it into account during the system control
design restricting the error introduction.
To have the total weight of the beam is necessary to add 500 mm, i.e. 7.6 kg, that is the part of the
beam inside its support, designed in Section 3.2.2. Finally, the total beam weight is 60.8 kg.
Table 3.4: Beam cross section properties.
Metric Depth Width Web Thickness Sectional Area Weight Ix Iymm x mm x kg/m h [mm] w [mm] s [mm] [cm2] [kg/m] [cm4] [cm4]
W 310 x 165 x 15.2 313,00 166,00 6,60 56,70 15,20 9934,00 854,70
A stress analysis is now carried out to ensure that the stress on the beam is lower then the material
33
Figure 3.5: Beam weight vs. tip deflection.
tensile yield strength. The beam section with the higher stresses is the cantilever section, the shear
force and moment in this section are computed by Equations 3.3 according to Figure 3.6.
Vy =Wbeam +Warm = 884.9 N , (3.3a)
Mx = 3500 ·Warm + 1750 ·Wbeam = 2.184 · 106 N ·mm. (3.3b)
Figure 3.6: Beam forces at the root section.
The normal stress in the section is given by Equation 3.4a. For I-sections, the shear load is given by
Equation 3.4b and is assumed to be evenly applied only to the section web [47]. The total stress σ is
computed using the Von Mises yield criterion in Equation 3.4c [48]. To find the maximum stress in the
section, it is computed for points A and B of Figure 3.7. The maximum total stress is 3.44 N/mm2 at
point A. It ensures safety working condition to the structure being 80 times lower than the tensile yield
strength of the material (276 MPa).
σz =Mxy
Ix, (3.4a)
τzy =Vy
Areaweb, (3.4b)
σ =√σ2z + 3τ2zy . (3.4c)
34
Figure 3.7: Stresses in the beam root section.
3.2.2 Beam Support Design
The support of the horizontal beam consists of four aluminum alloy plates welded together to form the
structure in Figure 3.8, the aluminum alloy of Table 3.2 is used. The beam is held and bolted to the
support as in Figure 3.9, where also the bearings and the screw nut that connect the support to the
lifting structure are drawn. The bearings are able to transmit to the lifting structure only forces normal to
their axis, while the vertical load is transmitted to the ball screw through the screw nut (their functions
are explained in details in Section 3.3).
Figure 3.8: Horizontal beam support structure.
The forces and moment applied to the beam support are shown in Figure 3.10. The forces are
assumed symmetrical with reference to the z-y plane where the z axis coincides with the beam axis of
Figure 3.3. In Figure 3.10, Vy and Mx are the force and moment due to the weight of the robotic arm
and the portion of the beam outside the support computed by Equations 3.3. Whouse is the weight of the
portion of the beam inside the support. Fs is the lifting screw force, while Fbear1,Fbear2,Fbear3 and Fbear4
are the forces on the bearings. All the computations are made in a steady situation but a conservative
approach is adopted to take into account dynamic loads.
According to Figure 3.10, the equilibrium Equations 3.5 are written. In these equations the distance
between the bearing axis and the support edge is neglected (wbear ' whouse).
There are three equilibrium equations and seven unknowns, then in Equations 3.6 the following
35
Figure 3.9: Horizontal beam and support structure assembly.
Figure 3.10: Beam support free body diagram.
assumptions are made: Fbear4 = 0 and Fbear3 = Fbear2. Finally, the parametric analysis of Figure 3.11
is executed to choose wbear and hbear.
To select wbear and hbear has been taken into account the load on the bearings and the general
dimensions of the beam support. The load on the bearing influences their rails dimensions and weight,
on the other hand the beam support can not be too cumbersome. Thus, wbear is chosen equal to 500
mm and hbear equal to 700 mm, it implies a design force of 3436 N on the bearings and of 959.5 N on the
screw. It is noted that, throughout these computations, the beam support weight has been neglected.
Fs =Whouse + Vy = 0.149 N/mm · wbear + 884.9 N , (3.5a)
Fbear4 + Fbear3 − Fbear1 − Fbear2 = 0 , (3.5b)
(2Fbear1 + 2Fbear2 + 2Fbear3 + 2Fbear4)hbear2
=Mx + Vywbear
2. (3.5c)
36
Fs =Whouse + Vy = 0.149 N/mm · wbear + 884.9 N , (3.6a)
Fbear1 = 0 N , (3.6b)
Fbear2hbear =Mx + Vywbear
2. (3.6c)
Figure 3.11: Beam support parametric study.
To compute the plates thickness, it is assumed that all the load acts only on one of the two horizontal
plates. The stress on this plate is computed studying it as the cantilever beam of Figure 3.12.
The maximum stress is located in the cantilever section, the shear force Vr and moment Mr in this
section are computed by Equations 3.7. The normal stress due to the bending is σz = MryIx
where
Ix = 166t3
12 is the section moment of inertia about the x axis of Figure 3.12 where t is the plate thickness.
The shear stress is τyz = ( 6Vr
166t3 )(t2
4 − y2). Then, applying the Von Mises yield criterion of Equation 3.4c,
the maximum stress is σ(y = t/2) = Mr
27.7t2 . Equaling this value with the material tensile yield strength,
a minimum thickness of 18.6 mm is computed. A 19 mm plate is then selected for the beam support
structure which weights 44 kg.
Vr =Whouse + Vy = 959.5 N , (3.7a)
Mr =Mx + 500 · Vy + 250 ·Whouse = 2.645 · 106 N ·mm. (3.7b)
3.2.3 Lifting System Structure Design
Along the present section the lifting system structure design is described. The lift is 11.7 m tall and its
functions are: holding the horizontal beam support and moving it in a vertical direction. The system is
mounted on an AGV platform, then it is important to limit its weight.
The first lift concept, shown in Figure 3.13, was composed of two continuous metal plates, with two
37
Figure 3.12: Beam support forces diagram and section.
linear rails and a ball screw mounted on each side of the beam support. The function of the linear
bearings that connect the beam support with the lift is to transmit the moment due to the robotic arm
and the beam weight to the lift structure (see Figure 3.19), while the screw carries the vertical loads and
transmits the motion to the horizontal beam through the screw nut. Then, in first approximation, the only
loads on the lift structure are: the moment transmitted by the bearings and the weight of the structure
itself.
Figure 3.13: Lifting system exploded top view.
Being a tall column, the lift structure is designed to avoid buckling. According to the Euler bucking
theory, maximum height for a free-standing, vertical column, loaded by its own weight, is given by Equa-
tion 3.8 [48]. Where E is the Young’s modulus, I is the minimum moment of inertia of the beam cross
section, g is the acceleration due to gravity, A the cross section area and ρ the material density.
Hlift = 7.84EI
ρgA. (3.8)
tmin =
√12Hliftρg
7.84E= 31 mm. (3.9)
The lift plate has the rectangular cross section in Figure 3.14. The minimum moment of inertia is
Iz = t3 w12 , while the area is A = t w. Where t is the plate thickness and w its width. To minimize the
lift weight, the plates width w is selected equal to 600 mm, this is the minimum width for the lift due to
the bearings rails spacing selected in Section 3.2.2. Substituting the moment of inertia and the area
38
equations into Equation 3.8, the minimum thickness to avoid buckling is given by Equation 3.9, where
the material is the aluminum alloy which properties are in Table 3.2. It corresponds to a weight of 588 kg
per plate. To reduce the lift weight, a different lift structure has been designed.
Figure 3.14: Lift structure plate cross section.
To make the lift light, a truss structure is used. It is manufactured from only one aluminum alloy plate
bent and cut as in Figure 3.15. Again, the material selected is the aluminum alloy Al 6061-T6 whose
properties are summarized in Table 3.2. To allow preliminary computations, all its structural elements
have the same thickness and width.
Figure 3.15: Lift structure detail.
To size the structure, it is studied as composed of jointed beams. The idea is to analyze the buckling
of the column at the lift base, pointed out in Figure 3.16. As shown in Figure 3.17(a), the concentric
axial load due to the structure weight Wleg is applied to the column. It is approximated by Equation 3.10,
where Hlift is the lift height, ρ the material density, g the gravitational acceleration and Aleg the column
area.
Wleg = Hlift · ρ · g ·Aleg . (3.10)
The column has the boundaries conditions shown in Figure 3.17(a), i.e. one end fixed and the other
supported. To size the structure the following dimensions have to be selected: the structure section
thickness tleg and width Lleg (shown in Figure 3.17(b)), and the column height hleg.
According to the Euler buckling theory [48], the critical load for the described column is given by
Equation 3.11. Where E is the material Young’s modulus, hlegcr is the column height at which buck-
39
Figure 3.16: Lift structure base.
(a) Base support forces and constraints. (b) Base support section.
Figure 3.17: Lift base approximate structure.
ling occurs and Ileg is the column section smallest moment of inertia about its principal axes given by
Equation 3.12 (for equation clarity, tleg and Lleg are abbreviated to t and L respectively).
Pcr =π2EIleg
(1.2 hlegcr )2. (3.11)
Ileg =t(2L4 − 4L3t+ 8L2t2 − 6Lt3 + t4
12(2L− t). (3.12)
Substitution of Equation 3.10 into Equation 3.11 gives Equation 3.13, where a safety factor f = 3
is added. The parametric studies in Figure 3.18 have been performed to size the column section. In
Figure 3.18(a), the behavior of the column critical height hlegcr as a function of tleg and Lleg is shown,
while Figure 3.18(b) shows the behavior of the lift structure weight (approximated as ρ Aleg Hlift) as a
function of the same parameters.
hlegcr =
√π2 E Ileg
1.44 f Hlift ρ g Aleg. (3.13)
From Figure 3.18, it is clear that there is not bucking problem under this load for the column and that,
to have the lightest structure, the section has to be as small as possible. Then, to have enough material
to bolt the bearings rails, the dimensions selected are hleg = 780 mm, Lleg = 60 mm and tleg = 5 mm.
40
(a) Column critical height. (b) Lift structure weight.
Figure 3.18: Lift structure parametric studies.
Consequently, the lift structure weight is 103.2 kg.
On the lift structure, beside its own weight, are acting the forces transmitted by the linear bearings.
To make a structural analysis on the structure, the load acting on it is supposed to be the moment Mlift
computed by Equation 3.14, according to Figure 3.19. Where Warm is the robotic arm weight located at
the beam tip and Wbeam is the horizontal beam weight acting in the beam Center of Gravity (CG).
Figure 3.19: Lift structure forces diagram.
Mlift = 3750mm·Warm+2000mm·Wbeam = 3750mm·521.9N+2000mm·363.0N = 2.684·106 N ·mm.
(3.14)
To compute the stresses along the structure, Ftool is used [49]. It is a simple two-dimensional frame
analysis tool that solves forces equations of truss structures. The lift structure has been analyzed as the
trass pillar in Figure 3.20 loaded by its weight and half of the moment computed by Equation 3.14.
The software computed that the maximum load acts on the root pillar (red in Figure 3.20) and it is
a compression load of 2.7 kN. Then, because the pillar critical load, computed by Equation 3.11, is
41
Figure 3.20: Lift structure truss scheme.
62 kN, it is stated that there is not buckling in the structure. The compression stress related to the load
is 4.696 N/mm2. It is 59 times lower than the material compressive yield strength (assumed equal to the
material tensile yield strength [50]) ensuring enough structure stiffness.
As a simulation result, the lift tip deflection is 17 mm along the z axis of Figure 3.19. The structure
deformation decreases almost linearly to zero at the root of the column. This deformation has to be taken
into account during the control design because it affects the position of the robotic arm end-effector.
3.3 Lifting System Design
The lifting system moves the horizontal beam along the lift, carries its vertical load, and transmits the
moment, due to the beam and robot weight, to the lift structure. The first two tasks are demanded to a
screw-nut system, while the latter is demanded to two sets of linear guides. The system is symmetrical
about the vertical plane in order to split equally the load between two systems and to avoid torsion in the
lift structure. Along this section the linear bearings, the rails and the lifting mechanism are selected.
3.3.1 Linear Guides Selection
Each linear guide is composed of a rail and two linear bearings. In Section 3.2.2 a bearings design load
equal to 3436 N has been computed. Along this section the linear bearings and the rails are selected.
To ensure a spare parts supply along the entire life of the system, all the linear motion systems in this
project are selected from the catalog of Thomson Industries, it is a linear systems leader manufacturer .
The linear rails can be round or square, and end or continuously supported. In this application, the
rails have to transmit a moment to the lift structure, then, continuously supported rails are selected. They
ensure a reduced bending on the rail itself and do not present bucking problems. Between round and
square rails, round rails are selected because they present self-alignment, i.e. the friction increases
much less than for square rails when the lift structure, and then the rails, deforms [51].
For the bearings selection, the following criteria is applied: between the bearings able to support
the load required (3436 N) the one with the smaller shaft diameter is selected. Moreover, it has to be
corrosion resistant to ensure a long working life in a polluted environment. The rail shaft diameter is
42
important because the rail weight increases approximately with the squared diameter of its shaft.
The SSETWNO M16-CR bearing in Figure 3.21 has beam selected, its specifications are summa-
rized in Table 3.5.
Figure 3.21: SSETWNO M16-CR linear bearing1.
Table 3.5: SSETWNO M16-CR linear bearing specifications [51].
Shaft Nominal Diameter 16 mmLoad Capacity 4400 NWeight 0.37 kgBearing Type Ball Bushing BearingMax. Operation Temperature 85o C
Two different continuously supported rails can be coupled to the selected bearings. The rail in Fig-
ure 3.22(a) is an aluminum alloy rail, the height from the base to the mean shaft center is 30 mm, has
a weight of 4.7 kg/m and its attachment bolts are from above. On the other hand the LSRM16 rail in
Figure 3.22(b) is a steel rail, the height from the base to the mean shaft center is 18 mm, has a weight
of 2 kg/m and its attachment bolts are from underneath. The latter is selected to have a lighter and more
compact system. Moreover, the attachment of the rails to the lift structure is easier because the access
to the bolts is from outside the structure. According to the previews selections, the weight of the linear
guiding system is 96.56 kg.
Because the rail material is steel while the lift structure is made of aluminum alloy, a different rail
choice can be done if the system is installed in an environment with considerable temperature gradients.
3.3.2 Lifting Mechanism
The lifting mechanism moves the horizontal beam and the robotic arm as well as supports their weight.
Many linear actuators are able to perform this task, they can be divided into hydraulic, pneumatic and
electromechanical systems.
Hydraulic actuators, using a pressurized fluid to generate thrust, are generally heavier than others
actuators and require continuous electric power to hold the load. Furthermore, their positioning accuracy
1URL http://www.thomsonlinear.com/en/product/SSETWNOM16DD [Accessed: 06 October 2016]
43
(a) SRM16 rail. (b) LSRM16 rail.
Figure 3.22: Linear guides: round rails [51].
is low and require more maintenance than both pneumatic and electromechanical actuators. Pneumatic
actuators are the cheapest and the most powerful actuators but, as the hydraulic actuators, they have
poor positioning accuracy and high maintenance costs. Electromechanical systems are the most ex-
pensive nevertheless they have low maintenance costs, high accuracy and easy control. Moreover, they
hold the load without consuming power [52].
In view of the above, an electromechanical linear actuator is installed. Specifically, it is composed
of screws driven by electric motors. The screws have an end mounted on the AGV, where the motor is
located, while the other is supported by the lift structure as in Figure 3.23. A screw nut is bolted to the
beam support structure, when the screw rotates the horizontal beam moves along the screw axis. In the
present section, screws, screw nuts and electric motors are selected.
Figure 3.23: Lift system view with beam support.
Two type of screws are available: lead screws or ball screws. For this application, ball screws
are used because of their higher precision and efficiency, lower vibrations and longer operative life;
nevertheless they are more expensive [48].
44
According to Table 3.6, the load on the screws is 144.8 kg. The load is supposed to be equally
divided between two screws, one on each side of the beam.
Table 3.6: Vertical load on the lift actuator.Part Weight [kg]
Robotic arm 37Horizontal beam 60.8Bearings 3Beam support 44
Total 144.8
Each screw has approximately the same length of the lift structure and is loaded by concentric axial
load. To find the screw diameter dscrew, a buckling analysis has been carried on.
The minimum axial load causing the bucking of the screw Fbuckling is given by the Euler’s column
buckling formula in Equation 3.15, where n is a factor accounting for the end conditions, E [MPa] the
modulus of elasticity, I [mm4] the screw section moment of inertia and L [mm] the screw length [48, 53].
Fbuckling[N ] = n · π2E · IL2
= 4 · 9.687 · 104 d4screw
L2⇒ dscrewmin
= 26.6 mm. (3.15)
Equation 3.16 gives the screw angular velocity at which resonance occurs [53]. Finally, the screw
lead is computed by Equation 3.17. In this application, the screw is mounted with both the ends fixed to
minimize bucking problems.
nresonance[RPM ] = 1.2 · 108C dscrewL2
= 2.23 · 1.2 · 108 dscrewL2
. (3.16)
Screw Lead =V ertical V elocity
Angular V elocity. (3.17)
From Equation 3.15 the minimum screw diameter is 26.6 mm. On the Thomson catalog are available
ball screws with 32 mm and 40 mm diameters [54]. For these screws the maximum angular velocity
is computed multiplying for a 0.8 safety factor the natural frequency. Each screw has a maximum lead
available, thus, using Equation 3.17, the maximum vertical velocity for each one is computed. The
results are shown in Table 3.7.
Table 3.7: Angular velocity, lead and Vertical velocity for different screws.
Diameter [mm] 32 40
Max.angular velocity [RPM] 50.0 62.6Max. screw lead [mm] 40 40Vertical velocity [m/min] 2.0 2.5
Finally, the 40 mm diameter screw in Figure 3.24 is selected because, in spite of being heavier, it
allows a vertical velocity of 2.5 m/min that is comparable with the vertical velocity of commonly used
human lifts [55, 56]. Screw specifications are summarized in Table 3.8.
45
Figure 3.24: 40MMx40MM ball screw with nut2.
Table 3.8: Ball screw specifications [54].
Model KGS-4040-023-RHDiameter 40 mmLead 40 mmStandard lead accuracy ±23 µm/300 mmMax. backlash 0.041 mmMax. dynamic load 35 kNMax static load 101.9 kNWeight per meter 9.0 kg/mWeight 105 kg
The screw buckling load, computed by Equation 3.15 is 7250 N. It is five times bigger than the total
vertical load on the system (1420 N). Then, to limit weight and costs, only one screw is used. This
design induce an torsion on the beam support structure, in this preliminary study it is neglected because
the arm of the screw thrust with respect to the beam support axis is small.
Between the possible nuts to couple with the ball screw, the flanged nut in Figure 3.24 is selected
because it can be easily bolted to the beam support while is less cumbersome than a round flanged nut.
To drive the screw an electric motor is required. Its minimum torque and power are computed re-
spectively by Equation 3.18 and Equation 3.19, where ε is the screw efficiency equal to 0.9 and the Load
and Angular velocity are the same of the screw.
Torque = Load · lead2πε
= 9.7 Nm. (3.18)
Power = Torque ·Angular velocity = 63.6W . (3.19)
To drive screw linear actuators, generally three type of electric motor are used: Direct Current (DC),
stepper and servo motors. DC motors are continuous rotation motors, they generally run at high speed
2URL http://www.thomsonlinear.com/en/product/7115-448-076 [Accessed: 06 October 2016]
46
and, due to their poor accuracy, are rarely used for accurate positioning. Between servo and step-
per motors, the latter are selected because they are cheaper, can work in an open loop, have higher
performance at low speeds, and require less maintenance (stepper motors are brushless) [57].
Knowing speed and torque ranges, the motor is selected on order to connect it directly with the
screw without adding a gear, this design reduces the system weight and transmission looses. The
stepper motor is then selected from the Oriental Motor catalog that provides the torque-speed graph for
each motor. The motor used is, finally, the PK599BE-N7.2 with the torque-speed graph in Figure 3.25.
Figure 3.25: Stepper motor torque vs. speed graph3.
3.4 Subsystems
In this section are listed, described and selected the required subsystems not covered in the previews
sections.
To control the robotic arm, it has to be linked to its controller, the R-30iATMMate Controller in Fig-
ure 3.26 whose specifications are summarized in Table 3.9.
To paint the aircraft, paint and compressed air are supplied to the robotic arm. The air pressure and
flow rate depend from the technique used. In the manual process the spray guns are fed by long hoses
linked to one or two common air compressor.
To limit the weight and the cost of the system, the air is supplied by hoses linked to an external
compressor. This solution also avoids the air compressor to introduce vibrations into the system. To
get out of heavy and expensive batteries on-board, also the electric power is supplied to the system by
cable linked to an external power source. Then, the system does not have to stop to recharge or change
batteries.
3URL http://catalog.orientalmotor.com/?plpver=11 [Accessed: 06 October 2016]
47
Figure 3.26: FANUC R-30iATMMate Controller.
Table 3.9: Robotic arm controller specifications [40].
Model FANUC R-30iATMMate ControllerWidth 470 mmDepth 320 mmHeight 950 mmAmbient temperature 0 - 45◦ CPower supply 200 - 230 VAC Single PhasePower consumption 0.5 KW (average), 1.2 KW (maximum)Controlled axis 24Built-in I/O Up to 28 digital inputs / 24 digital outputs 24VDCWeight 56 kg
It has to be ensured that the AGV platform does not run over the supply line. Thus, the electric cable
and the air hose are mounted on a retractable reel. Two type of retractable reel are possible: spring or
motor driven. For this system, electric motor driven retractable reel are used in order to control actively
the tension force on the cables and hoses.
48
The paint tank is located on the AGV platform. To size it, the paint usage is estimated. The Lockheed
C-130 Hercules has a wetted area of 2323 m2 [58]. Assuming two robots painting it, each robot paints
1163 m2, the half of it. In Table 3.10 the specifications for different coatings used are shown, they may
vary in accordance with color, manufacturer and environment conditions.
Table 3.10: Paints specifications.Paint Theoretical Coverage [m2/l] Volume [l] Density [kg/l] Weight [kg]
Epoxy Primer [59] 24 48.5 1.32 63.7Elastomeric Polyurethane Primer [60] 11.9 97.7 1.32 128.8High Solid Polyurethan Coating [61] 20 58.1 1.28 74.5Topcoat [62] 82 14.2 1.28 18.2
Multiplying the maximum paint volume needed by a 1.2 factor to take into account coatings with
lower coverage, a 120 liter paint tank is needed. A tank of this size is not required when painting smaller
aircraft or with higher coverage paints, it is difficult to clean and handle, heavy, expensive and increase
the paint waste due to the paint left-over.
In Table 3.11 different paint tanks specifications are shown. To reduce the weight, the cost and the
paint waste a 60 liter tank is selected. This selection implies that painting larger aircraft a refill of the
tank may be required. The total tank weight is then 128 kg (assuming 80 kg of paint).
Table 3.11: Paint tanks spacifications [63].
Volume [l] Weight [kg] Max. pressure [bar] Cost [e]
22 31 6 278045 44 6 310060 48 3 - 6 3600 - 3700
120 72 1.5 - 6 6700 - 8000
The subsystems required for the paint removal depend from the method selected. Chemical stripping
does not requires any additional subsystem, while to optical remove the paint a laser equipment is
necessary. The mechanical removal by water or media blasting requires a dedicated tank and blasting
equipment on the robotic arm tip. On the other hand, if the paint is removed by a motor driven abrasive
equipment, it is only required to replace the arm manifold. Therefore, if the customer requires laser,
MPW or PMB paint removal ad hoc studies are are required to select and locate the subsystem.
3.5 AGV Platform
In the present section an omnidirectional AGV is selected. Then the location of the subsystems on the
vehicle is analyzed.
3.5.1 AGV Selection
To select an AGV it is necessary to know the load it supports. According to Table 3.12, the payload on
the vehicle is 635.6 kg. The weight of the parts not yet designed (electric power, paint and air supply
49
lines, bolts etc.) is unknown. Thus, to select conservatively the vehicle, a minimum payload of 1000 kg
is selected.
Table 3.12: AGV payload.
Part Weight [kg]
Robotic arm 37Arm controller 56Horizontal beam 60.8Beam support 44Ball screw 105Linear bearings 3Rails 93.6Stepper motor 5Lift structure 103.2Paint tank 48Paint 80
Total 635.6
For the present project, it is required a zero turning radius vehicle and the possibility to guide it with
high accuracy. Moreover, it has to be possible to locate the robotic arm at an height of 0.5 m above the
ground. The majority of AGVs are used in warehouses or in assembly lines and are designed for much
higher payload [64].
Only one AGV that satisfies all the requirements has finally been found. It is the RoboMate 17
by Vetex, the vehicle is equipped with four omnidirectional wheels that support up to 1000 kg each.
Unloaded, its maximum speed is 67 m/min. Moreover, it is a modular system and the vehicle structure
is designed meating the client requests [65].
The company only sells the vehicle, the navigation control has to be developed ad hoc for the ap-
plication and implemented. Many different navigation systems have been developed, they include ultra
wide band indoor Ground Positioning System (iGPS), laser, and vision based Simultaneous Localization
and Mapping (SLAM). Typical accuracy of an omnidirectional vehicle under autonomous global naviga-
tion can be from +/- 20 mm to +/- 10 mm, depending on the type of system used. Often, when greater
accuracies are required, a combination of sensing systems can be used, such as switching from ultra
wide band iGPS indoors, to vision based localization at a micro level [66, 67, 68]. The development of
the navigation systems is demanded to future works.
With the selection of the AGV, the main components of the system have been designed or selected
off-the-shelf. In Figure 3.27, an overview of the whole system is shown.
3.5.2 Subsystems Location
Locating the subsystem on the vehicle, an even distribution of load has to be ensured. There are two
main problem to take into account: paint consumption and positioning of the lift axis in the center of the
vehicle. The paint consumption causes a system center of gravity shift, thus the load on each wheel
50
Figure 3.27: Automatic system overview.
changes during the painting. This problem could be solved positioning the paint tank in the wheels
centroid (from here on called only centroid), but this is not possible because the lift axis has to be
located coincident with the centroid. It ensures that, when the vehicle is performing a zero radius turn,
the result is only a rotation about the lift axis without any translation.
The CG positioning problem is divided into lateral and longitudinal positioning. Along this section,
the AGV is supposed to have a longitudinal and a lateral plane of symmetry.
The lateral positioning of the subsystems is easier because the CG of most of the parts is located
in the vehicle plane of symmetry (z-y plane), as shown in Figure 3.28. The only two parts with the CG
away from the z-y plane are the screw and the robotic arm controller. The latter is used to balance the
screw moment and locate the system CG in the z-y plane. The controller position is then computed
by Equation 3.20, where Wscrew and Wcontroller are the ball screw and the robotic arm controller weights
respectively.
Figure 3.28: Top view of the system.
51
xcontroller = −150Wscrew
Wcontroller= −216 mm. (3.20)
The longitudinal positioning of the subsystem is shown in Figure 3.29. From the system drawings,
it is measured that the paint tank CG can be located at the minimum distance of 540 mm from the y
axis. Neglecting the paint weight, to balance the system only with the arm controller it shall be located
4 m away from the y axis. Considering an AGV length of 2.5 m , the controller CG is located at the
distance of 1000 mm from the y axis. The position of system CG is then, 318 mm along the z axis with
the empty tank and 210 mm with the tank fulled with 80 kg of paint. It corresponds to a load per wheel
of 211 kg on the front wheels and of 109 kg on the real wheels. The weight of the AGV itself and of
the yet-to-design parts have to be added. Locating them in the rear part of the vehicle, it is possible to
reduce the unbalance between front and rear wheels load.
Figure 3.29: System sideview.
3.6 Cost Estimation
In the present section a system cost estimation is carried out. The estimated cost is related to the parts
of the system already designed or selected. The labor cost for the development of the missing parts and
for the assembly has not been considered. Thus, the estimated cost is simply the sum of the costs of
the parts described along the present chapter.
In Table 3.13 the cost for each part is shown. Most of this data comes from a direct contact with
the manufacturers. The total cost for the system is estimated to be 86.6 ke. To this value the price
of the parts not yet designed, as the system sensors, has to be added as well as the control system
development costs and assembly costs.
52
Table 3.13: Cost estimation.Part Cost [e]
Robotic arm w/ controller 20000Horizontal beam 900Beam support 100Lifting system 4000Lift structure 2000Stepper motor 1000Paint tank 3600AGV 55000
Total 86600
53
54
Chapter 4
Conclusions
The fleet of commercial aircraft is growing, and in the future years more aircraft will require maintenance.
A phase of the aircraft maintenance is the removal and application of the finish system to check the
substrate integrity and protect it from corrosion.
The whole process is, nowadays, achieved completely manually. It requires a big amount of time
and labor. Furthermore, it has to be accomplished inside a dedicated hangar for environmental safety
reasons, thus at the same time only one aircraft per painting hangar is processed. To tackle the aircraft
number growth, there are two ways: either to increase the painting hangars number and, consequently,
the number of workers, or to increase the finish system maintenance rate.
The automation of the aircraft finish system maintenance is nowadays a big challenge in the aerospace
industry. The implementation of an automatic system induces great advantages under many aspects.
The maintenance rate increases as well as the quality of the final result. Meanwhile, the whole process
cost drops because a big amount of highly skilled labor is no more required and the material waste is
reduced. Moreover, the introduction of an automatic system drastically reduces the workers exposition
to a toxic environment during the painting and paint removal of airplanes. Not least, the environmental
impact of the process is reduced thanks to the waste optimization.
Consequently, as an aerospace maintenance leader company, OGMA required the preliminary de-
sign of an automatic finish system maintenance procedure. This has been the subject of the present
work.
Presently, there is no robotic system able to achieve the whole coatings system maintenance. Al-
though several projects are under development especially for the paint removal automation, only a few
involve the painting of aircraft. Beside painting and paint removal, the automation of the aircraft masking
would give big benefits but the technology to do it is not yet available and/or the system would get too
complex and expensive. Therefore, along the present thesis a solution for the automation of aircraft
painting and paint removal has been designed.
There are many different methods and technologies for both painting and paint removal. The design
criteria of this thesis is to study a system able to implement as many of them as possible, in order to give
the client the opportunity to choose methods and technologies taking into account the advantages and
55
disadvantages of each one.
The system has been sized in order to process airplanes with a maximum height of 13 m and a
minimum height of 0.5 m. The airplanes OGMA maintains are within these dimensions as most of the
medium to small commercial and military aircraft.
Five possible systems have been evaluated and, through a trade-off process, the one in Figure 4.1
has been selected. The selection criteria were rewarding the design with the lowest cost and complexity.
Specifically, the system consists of a robotic arm mounted on the tip of a cantilever beam which is held
by a structure able to move it along a vertical axis. This structure is mounted on an omnidirectional AGV
where all the required subsystems are located.
Figure 4.1: Complete system drawing.
To compete with the state-of-the-art robots under development, the system has to ensure high per-
formance and rapid development at the lowest investment and operating costs. Thus, the driving design
criteria has been simplicity.
To ensure high performance, high accuracy selection criteria were adopted during the design. More-
over, the system has been designed to process as many aircraft as possible taking into account different
shapes and sizes. Finally, to reduce the costs, it is composed mostly by off-the-shelf components. They
do not require to be designed and manufactured but are mass-produced by specialized companies at
a lower cost. Furthermore, it is possible to have a spare parts supply during the operative life of the
system.
Because this is a complex problem, the development time for a project of this kind is generally
long. Namely, the developing and validation of the control system require a long time. To reduce the
development time, the system has been designed as simple as possible both from the structural and the
control point of view.
In detail, the system is composed of a FANUC Paint Mate 200iA/5L robotic arm installed at the tip of
an aluminum alloy I-beam 4 m long. The beam is held by a support that connects it to a lifting system
56
composed by an aluminum truss structure, four linear guides and a ball screw driven by a stepper motor.
The whole structure is then mounted on a Vetex RoboMate onmidirectional AGV on which the paint tank
and the robotic arm controller are installed. The AGV receives electric power and compressed air from
external sources through cables and hoses.
Except for the lift truss structure and the beam support, all the components listed above are off-the-
shelf.
4.1 Future Work
Being a preliminary design study, this is a stepping stone for future works. First of all, some components
of the system are yet do be designed, i.e. the electric power and air supply lines, AGV chassis and ball
screw support.
To lead preliminary evaluations, simplified theories and big safety coefficients have been used. This
led to a conservative design from the structural and mechanical point of view. An interesting future work
would be to carry on more precise analysis to reduce the structure weight and increase its stiffness.
From the structural point of view also a dynamic analysis has to be done. Especially vibration a analysis
is required to avoid resonance in the structure.
Based on the present work, an active collaboration with the client is necessary to define in detail the
system features. Specifically, painting and paint removal methods have to be selected to advance on
the design.
Once the system equipment, structure and mechanics are completely defined, the design of the
control system begins. First of all, the accuracy and precision required have to be computed. To do that,
tests and simulations for both painting and paint removal processes have to be done. Then, the general
approach to the problem has to be selected.
Generally, there are two possibilities: open-loop and closed-loop control. The first one does not use
feedback and would process the aircraft based on the drawings and process software without checking
its real position with respect to the surface and the result accomplished. The second control technique
uses the feedback loop, based on sensors, to check the system error with respect to the required position
and nullify it [13]. Beside the control of the system itself, a navigation system has to be designed as well
as a collision avoidance system to ensure the aircraft integrity.
Not last, the system cost estimation of Section 3.6 has to be improved including the control develop-
ment and implementation costs as well as the operative costs.
As already said, this work is a stepping stone, this complex problem is far from being solved. But, as
Mattie J.T. Stepanek said, ’even though the future seems far away, it is actually beginning right now’.
57
58
Bibliography
[1] J. Koleske. Paint and Coating Testing Manual: Fifteenth Edition of the Gardner-Sward Hand-
book. ASTM manual series. ASTM International, 2012. URL https://books.google.pt/books?
id=5SlHpwAACAAJ.
[2] V. S. Agarwala, K. Rajeshwar, et al. Paint and coating removal technologies: past, present and
future. In CORROSION 2001. NACE International, 2001.
[3] J. T. C. Michael. The future of aircraft paint removal methods. Master’s thesis, Air Institute of
Technology, Wright-Palterson Air Force Base, Ohio, September 1989.
[4] P. Lovoi. Laser paint stripping offers control and flexibility. Laser focus world, 30(11):75–82, 1994.
[5] USAF. TO 1-1-8 Technical Manual: Application and Removal of Organic Coatings, Aerospace and
non-Aerospace Equipment. Secretary of the Air Force, March 2016.
[6] EASA. Part 145. Annex II (Maintenance Organisation Approvals).
[7] M. Jeffers. Wash-rack keeps F-16s clean. 20th Fighter Wing Public Affairs, August 2012.
[8] J. Head and J. P. Niedzielski. Laser paint stripping. Technical report, DTIC Document, 1991.
[9] STRATAGEM. Laser coating removal robot. URL http://www.stratagemgroup.nl/
project-laser-coating.html. [Accessed: 12 April 2016].
[10] J. Arthur. Robotic laser system to strip paint from aircraft. Advanced Coatings & Surface Technol-
ogy, 26(1):2–3, 2013.
[11] K. Wolf, R. Krincher, and J. Ermalovich. Laser strip: a portable hand-held laser stripping device
for reducing voc, toxic and particulate emissions. Institute for Research and Technical Assistance,
IRTA, USA, 2009.
[12] S. P. Andrews. Aircraft Paint Application Manual. Products Techniques, Inc., Bloomington, CA
92316, USA, 1 edition, October 2013.
[13] B. Siciliano and O. Khatib. Springer handbook of robotics. Springer Science & Business Media,
2008.
[14] P. Waurzyniak. Expanding the horizons of aerospace automation. MANUFACTURING ENGINEER-
ING, 156(2):59–67, 2016.
59
[15] J. Walen. The history of the industrial robot. Technical Report 2853, Linkoping University, The
Institute of Technology, 2008.
[16] O. Solon. Robots will eliminate 6% of all us jobs by 2021, report
says, 2016. URL https://www.theguardian.com/technology/2016/sep/13/
artificial-intelligence-robots-threat-jobs-forrester-report. [Accessed: 27 September
2016].
[17] T. Anandan. The art of industrial painting with robots, 2013. URL https://
www.robotics.org/content-detail.cfm/Industrial-Robotics-Industry-Insights/
The-Art-of-Industrial-Painting-with-Robots/content_id/4447#top. [Accessed: 27
September 2016].
[18] FANUC. Painting robots paint camouflage patterns on a vehicle, 2009. URL http://robot.
fanucamerica.com/robot-applications/painting_camouflage_patterns_on_vehicle.aspx.
[Accessed: 27 September 2016].
[19] ABB Communications. ABB robotics makes wind turbine manufacturing more efficient, 2010. URL
http://www.abb.com/cawp/seitp202/470db65362e52de1c12578000060a9c8.aspx. [Accessed:
12 April 2016].
[20] J. Small. Clean machines. ABB Robotics Customer Magazine, 2(13):17, December 2013.
[21] G. e. a. Paul. A robotic system for steel bridge maintenance: field testing. In Proceedings of the
Australasian Conference on Robotics and Automation, 2010.
[22] D. Tolliver and S. Morton. Computer-controlled robotic marine coatings application system. Journal
of ship production, 21(3):195–202, 2005.
[23] B. Ross, J. Bares, and C. Fromme. A semi-autonomous robot for stripping paint from large vessels.
The international journal of robotics research, 22(7-8):617–626, 2003.
[24] J. E. Baker, J. V. Draper, F. G. Pin, A. H. Primm, and S. Shekhar. Conceptual design of an aircraft
automated coating removal system. Technical report, Oak Ridge National Lab., TN (United States),
1996.
[25] S. Hofacker. Large aircraft robotic paint stripping. In Proceedings of the SME Conference on
Maintaining and Supporting an Aircraft Fleet, pages 9–11, 1992.
[26] P. Garcia. Painting the stealth. Sandia Technoligy Journal, 2(2):8–10, 2000.
[27] N. A. Seegmiller, J. A. Bailiff, and R. K. Franks. Precision robotic coating application and thickness
control optimization for F-35 final finishes. SAE International Journal of Aerospace, 2(2009-01-
3280):284–290, 2009.
[28] P. Waurzyniak. Picking up the pace in aerospace production. Manufacturing Engineering, pages
69–79, March 2014.
60
[29] C. Reed. Lockheed C-130 Hercules and Its Variants: (Schiffer Military History). Schiffer Publishing,
Ltd., 1999.
[30] Y. Lage. Structural Analysis for Lockheed Martin C-130H Wing. PhD thesis, Universidade Tecnica
de Lisboa, Lisboa, PORTUGAL, 2009.
[31] N. H. Versloot, A. Klein, and M. De Maaijer. The ATEX directives: Explosion safety and regu-
lation, the european approach. In Center for Chemical Process Safety-19th Annual International
Conference: Emergency Planning Preparedness, Prevention & Response (6/29/04-7/1/04 Orlando,
Florida), page 389. John Wiley & Sons, 2010.
[32] K. Chang, P. Rammos, S. Wilkerson, M. Bundy, and S. A. Gadsden. LiPo battery energy studies
for improved flight performance of unmanned aerial systems. In SPIE Defense+ Security, pages
98370W–98370W. International Society for Optics and Photonics, 2016.
[33] U. Morelli. Dynamic analysis of propeller induced vibrations in multirotor AGV, July 2014. Bachelor’s
thesis, Politecnico di Torino, Torino, ITALY.
[34] O. Diegel, A. Badve, G. Bright, J. Potgieter, and S. Tlale. Improved mecanum wheel design for omni-
directional robots. In Proc. 2002 Australasian Conference on Robotics and Automation, Auckland,
pages 117–121, 2002.
[35] H. Durrant-Whyte, D. Rye, and E. Nebot. Localization of autonomous guided vehicles. In Robotics
Research, pages 613–625. Springer, 1996.
[36] K. Yousif, A. Bab-Hadiashar, and R. Hoseinnezhad. An overview to visual odometry and visual
slam: Applications to mobile robotics. Intelligent Industrial Systems, 1(4):289–311, 2015.
[37] K. N. Otto and E. K. Antonsson. Trade-off strategies in engineering design. Research in engineering
Design, 3(2):87–103, 1991.
[38] Health and Safety Authority. Guide to the Safety, Health and Welfare at Work (General Application)
Regulations 2007. Health and Safety Authority, 2007.
[39] N. Klopcar and J. Lenarcic. Kinematic model for determination of human arm reachable workspace.
Meccanica, 40(2):203–219, 2005. doi: 10.1007/s11012-005-3067-0.
[40] Paint Mate 200iA, MDS 00084. FANUC Robotics America, Inc., December 2009. URL http://www.
fanuc.eu/ch/en/robots/robot-filter-page/paint-series/paint-mate-200ia. [Accessed: 13
October 2016].
[41] Paint Mate 200iA/5L, MDS 00085. FANUC Robotics America, Inc., December 2009. URL http://
www.fanuc.eu/ch/en/robots/robot-filter-page/paint-series/paint-mate-200ia-5l. [Ac-
cessed: 13 October 2016].
[42] EPX1250, DS 523. Yaskawa America, INC., January 2012. URL https://www.motoman.com/
industrial-robots/epx1250. [Accessed: 13 October 2016].
61
[43] IRB 52, ROB0037EN C. ABB, April 2009. URL http://new.abb.com/products/robotics/
industrial-robots/irb-52. [Accessed: 13 October 2016].
[44] ASM International and ASM International. Handbook Committee and ASM International. Alloy
Phase Diagram Committee. Metals Handbook: Properties and selection, volume 2. ASM Inter-
national, 1990.
[45] M. Bakhoum. Structural mechanics, volume 1. Mourad Bakhoum, 1992.
[46] A. ASTM. 6–standard specification for general requirements for rolled structural steel bars. Plates,
Shapes, and Sheet Piling, 2009.
[47] D. W. Rees. Mechanics of solids and structures, volume 2. Imperial College Press, London WC2H
9HE, UK, 2000.
[48] R. C. Juvinall and K. M. Marshek. Fundamentals of machine component design, volume 83. John
Wiley & Sons New York, 2006.
[49] L. F. Martha. Ftool–two-dimensional frame analysis tool. Educational version, 2, 2001.
[50] W. D. Callister and D. G. Rethwisch. Materials science and engineering: an introduction, volume 7.
Wiley New York, 2007.
[51] RoundRail Linear Guides and Components, CTEN-0002-03A. Thomson Industries, Inc., Au-
gust 2016. URL http://www.thomsonlinear.com/downloads/bearings_guides/RoundRail_
LinearGuides_Components_cten.pdf. [Accessed: 16 October 2016].
[52] I. Boldea and S. A. Nasar. Linear electric actuators and generators. In Electric Machines and Drives
Conference Record, 1997. IEEE International, pages MA1–1. IEEE, 1997.
[53] R. Thomas. Selecting and Applying Rolling Element Linear Bearings and Guides. Thomson Indus-
tries, Inc., 1500 Mittel Blvd, Wood Dale, IL.
[54] Lead Screws, Ball Screws and Ball Splines, CTEN-0006-02A — 20151030TJ. Thomson Indus-
tries, Inc., August 2015. URL http://www.thomsonlinear.com/downloads/screws/Leadscrews_
Ballscrews_Splines_cten.pdf. [Accessed: 16 October 2016].
[55] MODEL LPCE: 3-Axis Lifts, M:11 Multi-Axis. LPI, Inc., March 2013. URL http://en.lpi-inc.
com/documents/products/en/lpce.pdf. [Accessed: 16 October 2016].
[56] Overhead Lift System, L-1111. LPI, Inc., October 2015. URL http://en.lpi-inc.com/
documents/products/en/l-1111-overhead-lift.pdf. [Accessed: 16 October 2016].
[57] J. E. Carryer, R. M. Ohline, and T. W. Kenny. Introduction to mechatronic design, volume 1. Prentice
Hall, 2011.
[58] E. E. Covert. Progress In Astronautics and Aeronautics: Thrust and Drag: Its Prediction and
Verification, volume 98. AIAA, 1985.
62
[59] DeSoto R© 513X419 Military Epoxy Polyamide Primer, Lit: 4121. PRC-DeSoto In-
ternational, Inc., August 2015. URL http://www.ppgaerospace.com/getmedia/
d7916ba9-6155-4809-b53d-3cc891a2906f/513X419_DeSoto_Military_Epoxy_Polyamide_
Primer_1.pdf.aspx?ext=.pdf. [Accessed: 16 October 2016].
[60] Koroflex R© 823X439 Conventional Solids Primer, Lit: 4123. PRC-DeSoto In-
ternational, Inc., August 2015. URL http://www.ppgaerospace.com/getmedia/
e4236ab5-4717-4343-a68e-04429933287c/823X439_Koroflex_Conventional_Solids_Primer.
pdf.aspx?ext=.pdf. [Accessed: 16 October 2016].
[61] Desothane R© HS CA8200 Military and Defense Topcoats, Lit: 0520. PRC-DeSoto
International, Inc., August 2016. URL http://www.ppgaerospace.com/getmedia/
69d6f992-8fd8-4895-b5db-9bf7cec0da03/ca8200_desothane_hs_military_topcoat_series_
1.pdf?ext=.pdf. [Accessed: 16 October 2016].
[62] 01F Series HS Epoxy Topcoat, Lit: 4411. PRC-DeSoto International, Inc., January 2016. URL
http://www.ppgaerospace.com/getmedia/3756d525-fee6-4d67-9b3e-9d432cbc13b2/01F_
Series_Epoxy_Topcoat.pdf?ext=.pdf. [Accessed: 16 October 2016].
[63] ESSKA.de GmbH. Pressure tanks. URL https://www.esska-tech.co.uk/esska_eng_s/
Paint_Spraying_Coating_Systems/Paint_Distribution/Pressure_Tanks.html. [Accessed: 13
September 2016].
[64] R. Bloss. Ease of programming and sophisticated sensors see robots advancing in transport lo-
gistics, palletizing, order picking and assembly. Industrial Robot: An International Journal, 40(5):
420–424, 2013.
[65] Vetex. RoboMate. URL http://www.vetexinc.com/vehicles/robomate.html. [Accessed: 7 Oc-
tober 2016].
[66] G. Blanc, Y. Mezouar, and P. Martinet. Indoor navigation of a wheeled mobile robot along visual
routes. In Proceedings of the 2005 IEEE international conference on robotics and automation,
pages 3354–3359. IEEE, 2005.
[67] S. Thrun. Learning metric-topological maps for indoor mobile robot navigation. Artificial Intelligence,
99(1):21–71, 1998.
[68] G. N. DeSouza and A. C. Kak. Vision for mobile robot navigation: A survey. IEEE transactions on
pattern analysis and machine intelligence, 24(2):237–267, 2002.
63
64
Recommended