�ݺ�ߣ

TOOL PATH PLANNING TECHNIQUE FOR ROBOT PLASMA SPRAY
COATING ON DUCT COMPONENT
Nattapon Aroonchote
TAIST Toyo Tech Automotive Engineering, Thailand / numar151@hotmail.com
Dr. Nirut Naksuk
National Metal and Materials Technology Center (MTEC), Thailand
Dr. Nattawoot Depaiwa
King Mongkut’s Institute of Technology Ladkrabang (KMITL), Thailand
Prof. Hiroshi Yamaura
Tokyo Institute of Technology, Japan
ABSTRACT
The internal plasma spray coating process for multi-
section shapes tubular, which has been applied by multi-
axis robot station, is difficult to generate collision-free
tool path manually. Robotic software is very helpful to
create a tool path. At this point, a plasma flame should be
perpendicular to the coating surface and there should be
adequate spraying distance, from 3D CAD model.
However, the constraints of both shape and size of spray
gun and the tubular cause the collision to occur in certain
positions on the motion path. The collision between the
tubular with robotic parts or any part of the system is
strictly prohibited. Therefore, this thesis studies on
collision protection system to detect spraying point on the
tool path which is the cause of collision. Furthermore, the
researchers find tool path adaptation algorithm in order to
automatically modify the spraying point to flee from
collision. Moreover, the application also retains a
spraying condition to keep the coating surface good as
well.
KEY WORDS
INTERNAL PLASMA SPRAY COATING, COLLISION
AVOIDANCE, TOOL PATH ADAPTATION
1. Introduction
Plasma Spray is the spraying of molten or heat softened
material onto a surface to provide a protective coating.
The plasma spray coating process is widely applied by
multi-axis industrial robot. Our process operates with a
plasma spray gun on a 6-axis industrial robot station and a
fixed transition piece on a 2-axis turntable as shown in
Fig.1. Generally, robot trajectories are generated by
manual jock and teaching at each via point. This practice
is convenient for coating a simple model but on the other
hand, it is a very difficult task for coating a complex
model. Obtaining complex robot motion by manual jock
is time-consuming, expensive and very difficult for robot
operators. Robotic software, which is tool path planning,
is very helpful to generate the robot motion path,
especially robot operating with complex 3D surface. The
robotic software generates tool path in Cartesian space
which usually comprises of position and orientation for
each point on the path in order to determine the
corresponding joint position.
Figure 1. Robot Thermal Spray Coating Station
In this research work, the inner surface plasma spray
coating of transition piece is considered as a case study.
The transition piece, which is a part in gas turbine engine,
always has a very hot air tubular flow. The inner surface
plasma spray coating is necessary to prevent the damage
of the transition piece. The transition piece is a multi-
section shapes tubular; therefore, it is difficult to generate
collision-free tool path manually. Spraying condition for
the best coating surface, location and direction of
spraying are important to consider. The location is a
spraying position and the direction is spraying
orientations which both are specified by transformation
matrix. Plasma flame should be perpendicular to the
coating surface and spraying distance should be adequate
throughout the process. Therefore, tool path has been
created from 3D CAD model. However, the constraints of
both shape and size of the spray gun and the transition
piece can cause the collision to occur in certain positions
on the motion path. In robot operation, the collision is a
big problem; especially, through a complex shape. The
collision between any parts of the system is strictly
prohibited. Since the transition piece and system
Proceedings of the IASTED International Conference
November 24 - 26, 2010 Phuket, Thailand
Robotics (Robo 2010)
DOI: 10.2316/P.2010.703-019 69

equipments are expensive and the crash may cause
damage.
This research focuses on tool path adaptation to flee
from collision. Traditionally, the modification of robot
trajectory to flee from collision is done manually. Manual
modification is a very time consuming work and it is very
difficult to control the coating quality which will keep
spray angle and length close to the initial path as much as
possible. For our work, we use software simulation to
simulate plasma spray coating process for testing collision
avoidance and tool path adaptation algorithm. Application
and adaptation algorithm detect a collision, then it will
modify tool path automatically. The automatic process on
software simulation makes a lot of loss-time reduction.
The automatic adaptation modifies tool path from initial
path which is generated by coating condition from 3D
CAD model. This can also make the application retain
spraying condition to keep the coating surface good as
well.
2. Thermal Spray Coating tool path
Transition piece plasma spray coating with separate spray
paths is a way to operate without collision but it is
difficult to control coating thickness on the coating joints
[7]. Bad coating joints might be a cause of problem. The
problem can occur due to the gap between the joints or
over-coating at the joints. The gap can cause the
effectiveness of heating protection to decrease. In this
case, during the usage where there is a lot of heat, the gap
can cause the piece to get damaged and has fewer
lifetimes. On the other hand, the over-coating increases a
risk for the piece to peel off and the peel can cause
damage when it is in use. For good coating surface
forming, not only spraying distance and spraying angle
are to be considered but also spraying should be done
continually so that the piece will not face any of the
troubles stated above.
Figure 2. Tool Path
In order to get a continual spraying, we create the
tool path which moves back and forth along the length of
the transition piece. Also, the returning point should be
away from the tip of the tubular piece in order to prevent
it from over-coating at the tip as well.
This tool path; however, faces a problem at the
bending side as shown in the figure3. If we keep the
actual distance, the collision might occur as shown in the
figure. The adaptation of the tool path should not only
avoid the collision but also retain the initial spraying
angle as much as possible. As we can see from the
picture, after we have modified the tool path, the spray
gun will be just far enough from the collision.
3. Collision Avoidance and Tool Path
Adaptation Algorithm
In order to perform tool path adaptation to flee from
collision for the inner surface thermal spray coating of the
transition piece, spraying distance and direction which
contain two variables parameters can be adjusted.
However, the adaptation must be within the acceptable
range because the tool path modification by spraying
position and orientation adjustment affect to the coating
quality. The acceptable spraying distance for the surface
coating in order to maintain the quality of the piece is on
60-75 mm. Also, the spraying direction, acceptable at
vertical deviation from the scene, is less than thirty
degrees. The tool path which is generated from 3D CAD
model contains a lot of spots lining one after another to
form a moving path. The starting point for each spot is far
enough from the coating surface to be suitable for
spraying, which is at 65 mm, and is also set at
perpendicular direction to the coating surface. When the
tool path is loaded into simulations, it starts computing
the inverse kinematics and then chooses the next robot
configuration by keeping all robots’ joints move at the
minimum from the previous one. After that the robot’s
movement is simulated along path points, simultaneously,
collision flee automatically performs on gun and the
transition piece.
Figure 3. Coating Orientation Modification
70

3.1 The gun adapted Technique
According to physical constraint of these, the inner space
of transition piece is freely av ailable place for gun
adaptation; however, the spraying direction and distance
have to be considered seriously in order to get the coating
surface as the specification. Consequently, the adjustment
principles of the spraying gun alignment to the center axis
of the transition piece are based on no-collision and
closing to the initially perfect tool path.
Figure 4. Dodge from Collision
The algorithm begins when robot simulator
detects collision while moving along a sequence of robot
poses on a predefined tool path. At the collision point, it
calculates modification angle required by creating two
vectors originated at the spray deposition location Pd .
The first vector gp directs through gun handle while the
second vector cp directs through the middle of the
circular transition piece mouth. The angle  between
these 2 vectors could be calculated by (1).
(1)
Figure 5. Rotating Axis
To adjust the initial spraying orientation in order to
rotate the spray gun away from the collision, we need to
generate an arbitrary axis a . This rotational vector a could
be computed as cross product c gp p . The rotational
matrix for the arbitrary axis a could be computed as (2).
(2)
Where cosc 
sins 
1-cosv 
For each collision, the incremental tool rotational
adjustment can be computed iteratively by (3) and noted
that 0     .
(3)
Where iR is a current tool rotational matrix and
1iR is a new rotational matrix. The equation (3 )is
repeated until collision is not detected. The result of this
algorithm can keep the spraying direction closest to the
surface normal without collision.
4. Robot Simulation
Figure 6. Off-line Tool Path Modification
Simulation software in this research work is developed on
visual C++. We create it according to the real spraying
station. This simulation consists of ABB robot model
6500, with a long type of spraying gun in order to reach
the inner part of the piece. As for the transition piece, it is
b)a)
c)
e)
d)
f)

a
gp
cp
 1i i    aR R R
Pd
1
cos
c g
c g
 

p p
p p
 
x x x y z x z y
x y z y y y z x
x z y y z x z z
a a v c a a v a s a a v a s
a a v a s a a v c a a v a s
a a v a s a a v a s a a v c
     
      
     
   
 
    
    
aR
71

placed on a turntable by the fixers, particularly for each
model. In order to show the work result of the robot, we
use OpenGL to demonstrate the graphic in simulation.
Lastly, for the collision detection, we use ColDet Library
to help detecting the collision within robot simulation.
The initial tool path has 4500 points. Each of the point has
spray distance at 65 mm.
The adaptation of tool path in form of off-line path
planning in robot simulation software is shown in figure6.
The adaptation begins whenever there is a collision
between the gun and the piece anywhere on the tool path
as shown in figure 6-a. The robot slowly adjusts the
spraying direction by rotating the spray gun to flee from
collision as in figure 6-f. Then, the robot remembers the
new adjusted direction and also calculates the changing
degree. Those are used in order to consider whether it is
acceptable. After that, the program will continue on its
work at the next position.
Figure 7. Collision Free Tool Path in Robot Simulation
However, it still causes problems at some positions.
Since there are some positions on the tool path that the
collision cannot be fled by only one technique, we have
performed the adjustment manually in advance in order to
succeed the procedure. The problem that occurs is that the
tip of the gun crashes with the transition piece. Manual
adjustment is very easy to perform. We just need to twist
the handle of the gun for a few degrees and we are able to
flee from collision just in time before using the same
adjustment technique as the other positions. This case
only occurs at a few positions at the critical areas like at
the angle of the square
For the result of adapted technique, the spray angles
have been modified and all constrained by the acceptable
criteria. The robot can successfully perform the spraying
within the transition piece, going along the length of the
piece and returning back without collision. Figure7 shows
the movement of the robot along the tool path which has
been successful adjusted. This algorithm can reduce time-
consuming more than all manual tool path modification.
Also, it can steadily retain coating condition better than
trial manual adjustment.
5. Experimental Real Coating Process
The tool path which has been modified and examined by
robot simulation software will examine the next step with
the real spraying process on a robot station. As illustrated
in figure8, the robot station is dry running before
performing the spraying process. The robot can perform
its work effectively.
Figure 8. Dry Run on Robot Station
The tool path generated from the simulation is free
from collision. However, in the real robot station, the path
position may not be accurate as in the model. Therefore,
robot calibration is necessarily done for getting more
accurate similarities of simulation model and real robot
station.
The finished surface is illustrated in figure 9. It must
be inspected It must be inspected with microscope. The
rectangle sample plates will be examined for surface
quality.
b)a)
c)
e)
d)
f)
g) h)
72

Figure 9. Internal Coating Surface
6. Conclusion
The adaptation technique helps to improve a tool path to
avoid collision automatically, to spend less setup time,
and to perform more systematically than manual
operation. Also, the technique efficiently controls the
spray direction to be close to the initial path in order to
control coating surface quality. Moreover, it bases on
fundamental concept which can be additionally applied
with internal thermal spray coating on other models.
Acknowledgements
This work is supported by a research grant from
Electricity Generating Authority of Thailand.
References
[1] Okada, K.; Inaba, M.; Inoue, H.; “Real-time and Precise
Self Collision Detection System for Humanoid Robots”
Robotics and Automation, ICRA Proc IEEE Int Conf, April
2005 Page 1060 - 1065
[2] ColDet - Free 3D Collision Detection Library.
http://coldet.sourceforge.net/.
[3] S.H. Leigh, C.C. Berndt; “Evaluation of off-angle
thermal spray” Surface and Coatings Technology, Volume
89, 1 March 1997, Pages 213-224
[4] Anila Jahangiri, Richard Allor; “Thermal Spray Control
System”
[5] Kuu-Young Young, Chun-Yuh Huang “Collision-free
path planning and modification based on task requirements”
Mechanism and Machine Theory, Volume 32, January 1997,
Pages 21-37
[6] A. Candel and R. Gadow “Trajectory Generation and
Coupled Numerical Simulation for Thermal Spraying
Applications on Complex Geometries” Journal of Thermal
Spray Technology, Published online: 21 July 2009
[7] S. Ahmaniemi, P. Vuoristo, T. Mäntylä, J. Latokartano,
Tampere/FIN, and I. Salonen, Linnavuori/ FIN
“Optimisation of the robot controlled plasma spraying of
thermal barrier coating for gas turbine transition duct” ITSC
Int Thermal Spray Conf 2002, Pages 208-212
[8] S. Deng, H. Li, H. Liao, C. Coddet “New Functions of
Thermal Spray Toolkit A Software Developed for Off-line
and Rapid Robot Programming” Proc Int Thermal Spray
Conf (ASM International), 2006 , Pages 1437-1442
[9] Liu, K., et al., “Steward-platform-based inlet duct
painting system” in Proc IEEE Int Conf Rob Autom, 1993,
Pages 106-113.
[10] Yan Chen; Junyi Shao; Chuanqing Zhang; Chenglong
Fu; Chen, K.; “Redundant-robot-based painting system for
variable cross-section S-shape pipe” Reconfigurable
Mechanisms and Robots, ASME/ IFToMM Int Conf, June
2009, Pages 743 – 749
[11] Pichler, A.; Vincze, M.; Andersen, H.; Madsen, O.;
Hausler, K.; “A Method for Automatic Spray Painting of
Unknown Parts” Robotics and Automation, ICRA Proc IEEE
Int Conf, 2002, Pages 444 – 449
[12] Mahadevan Balasubramaniam, Stephen Ho, Sanjay
Sarma, Yoshitaka Adachi “Generation of collision-free 5-
axis tool paths using a haptic surface” Computer-Aided
Design, Volume 34, Issue 4, 1 April 2002, Pages 267-279
73

�ݺ�ߣ

703-019

More Related Content

703-019