Force Control of a Duct Cleaning Robot Brush
using a Compliance Device
Wootae Jeong
, Seung-Woo Jeon
, Duckshin Park
and Soon-Bark Kwon
Eco-Transport Research Division, Korea Railroad Research Institute, Gyeonggi-do, Uiwang, Korea
Department of Virtual Engineering, University of Science and Technology, Daejeon, Korea
Keywords: Compliance Device, Service Robot, Air Duct Cleaning, Air Quality, Force Control.
Abstract: Conserving clean air and removing contaminants and particular matters accumulated in the ventilation
system of the subway stations are key issue for high air quality and green environment. Accumulated
various pollutants at inner duct surface can cause secondary air contamination and injure subway
passengers respiratory system and health. In fact, periodic duct cleaning works can improve indoor air
quality, but cleaning entire ventilation system takes high cost and manpower. This study proposes a newly
developed duct cleaning robot to provide autonomous air duct cleaning. In addition, effective cleaning
method with an automated robot device is developed. In particular, the new duct cleaning robot has
functionality that cleans four sides of inner duct surface simultaneously with a constant pressure by using a
force compliance brush. Control method with the compliant device has also been analysed. The proposed
design of autonomous duct cleaning robot is expected to save the operating cost of subway ventilation
system and sustain clean indoor air quality by providing easier and faster cleaning tools.
The main purpose of the ventilation system and air
duct is to supply fresh air into closed spaces such as
buildings and subway stations where people work
and spend most of their daily hours. The air duct and
ventilation system controls various air flows, i.e.,
outdoor air, supply air, return air, and exhaust air as
depicted in Figure 1. The ventilation system also
consists of mechanical components such as dampers,
fans, filters, and duct terminals. Various particulate
matters are initially filtrated with particle filters
installed at the side of the supply air duct. The filters
commonly used, however, are insufficient to prevent
the entrance of all the particulate matters from
outdoor air into the duct. Therefore, transported dust
and other impurities are accumulated at the duct
surface inside the ventilation system. Accumulated
dust inside air duct may also originate from the
facility construction phase or from ventilation duct
installation (Pasanen, 1998).
To provide fresh and clean supply air through the
ventilation system into the closed space such as
subway stations, eliminating source for the
pollutants and contaminants is the most cost
effective than cleaning and replacement of the air
duct. However, duct cleaning is essential for
maintenance after completion of the ventilation
system installation. Many countries have existing
regulation and guideline for ventilation system
cleaning intervals and specific guidelines of
ventilation system cleanliness (FiSIAQ, 2001). A
few countries do not still provide legal regulation,
which makes air quality from the ventilation system
severely contaminated.
Figure 1: Air duct and ventilation System.
In this paper, various duct cleaning technologies
are introduced and types of contaminants are
analysed. Based on the mechanical brushing
technology, a new autonomous air duct cleaning
Jeong W., Jeon S., Park D. and Kwon S..
Force Control of a Duct Cleaning Robot Brush using a Compliance Device.
DOI: 10.5220/0004119703720376
In Proceedings of the 9th International Conference on Informatics in Control, Automation and Robotics (ICINCO-2012), pages 372-376
ISBN: 978-989-8565-22-8
2012 SCITEPRESS (Science and Technology Publications, Lda.)
robot has been designed to improve the cleaning
efficiency and overcome restrictions of manual duct
cleaning. In particular, in order to control constant
brush pressure on the duct surface, a simple force
compliance device is designed and installed at the
brushing arm, which enables consistent cleaning
operation despite of irregular duct surface quality.
The ventilation system can be contaminated and
acted as another source of pollutants, such as
microbe, chemical compounds or odours, particulate
matters. Accumulated dust and particulate matters
inside duct can stick to the inner surface of duct and
scatter by air stream of the air duct. Several cleaning
technologies can be used to remove the dust and
other contaminants effectively from the duct surface.
2.1 Contaminants in Duct
The estimated annual accumulation rate for dust in
commercial supply air ducting is normally set at
. However, an average accumulation level in
supply air ducts of building occupied less than a year
was 5.1 g/m
(Pasanen, 1998) while dust level in
new constructions was measured as high as 4.9 g/m
(Holopainen et al., 2002). In general, if dust
accumulation on an inner duct surface exceeds
upon the class of ventilation system the
duct has to be cleaned. Amount of dust accumulated
inside duct is related to the types of facility,
complexity of duct and its components and age of
building. In addition, dust concentration is affected
by factors that interrupt air flow of duct such as
surface roughness, air velocity, humidity, number of
dampers and diffusers installed.
Accumulation of dust also provides suitable
conditions that microbes, bacteria and other
microorganism can propagate. In fact, it has been
reported that 400 times of penicillium and 9.5 times
of aspergillus exists within 1 micro particulate
matter (Morey, 1988). In addition, amount of fungi
in contaminated indoor air is 10 times higher than
that of outdoor air. Many other particulate matters
accumulated at a duct terminal may cause secondary
infection or contamination of indoor air through the
air supply duct.
2.2 Cleaning Technologies
The air duct cleaning methods can be either dry or
wet (HVCA, 1998). The most commonly used dry
duct cleaning methods are mechanical brushing,
compressed air cleaning and vacuuming. Dry
cleaning methods use a rotating brush, a powerful air
jet or a suction force to detach the dust from the duct
surface mechanically. The loose dust is carried out
of duct by airflow. Wet cleaning methods include
water jet or chemically sterilizing process to
eliminate microorganism and bacteria. However, wet
cleaning methods are seldom used to clean air ducts
because the ductworks are not normally watertight.
To remove accumulated dust and oil in duct,
mechanical brushing is faster and more effective
than compressed air cleaning. Cleaning methods are
summarized in Table 1.
Table 1: Summary of duct cleaning techniques.
Cleaning Techniques
A brushing or mechanical action is used to
dislocate dust from surfaces and transferred
to a vacuum collector. The most commonly
used is the rotating brushes.
air cleaning
Dust is dislodged from surfaces using
airflow movement (via air nozzle) and
collected using a vacuum collector.
Suction and brushing using a brush head to
transfer dirt to a collection point.
Cleaning components surfaces by hand using
tools such as brushes, sponges and a source
of water with a cleaning agent.
Water jet
Liquid solutions are sprayed or wet-fogged
to adhere, bond, or fibre- fixed particles that
were not removed mechanically.
The use of biocides and sealants to coat and
encapsulate duct surfaces. Some duct
cleaning contractors introduce ozone as part
of the disinfection process.
Figure 2: Duct cleaning with mechanical brushing and
In this study, we have focused on mechanical
brushing method with autonomous cleaning robot to
improve cleaning efficiency for all types of
contaminants. As illustrated in Figure 2, the air duct
cleaning robot system also utilizes the push-pull
Force Control of a Duct Cleaning Robot Brush using a Compliance Device
technique; the rotating brush removes dust and
debris from the inner surface of the duct. The dust is
then drawn into a negative collector. Compressed air
is often used to push the dust and debris.
Increased interests in air duct cleaning technology
stimulate development of various duct cleaning
robots. Most of duct cleaning robots is based on the
dry cleaning method with mechanical brush.
As an example, the Articulated Nimble
Adaptable Trunk (ANAT) robot has been developed
to clean and inspect HVAC ducts. The
ANATROLLER ARI-100 duct cleaning robot rolls
on tracks or wheels and will continue to operate
even if flipped upside down. It is composed of two
modules containing the air jet, lighting and camera
rotates through 180 degrees, which allows the robot
to change its shape to get around or over obstacles
(Robotics Design Group). However, the ARI-100
still needs manpower to operate.
The XPW-series duct cleaning robot is of height
adjustable rotary brush and rotating camera for
inspection and remote control. Its rotary brush
mounted at machine bed is oscillatable within a
prescribed angle range in a vertical direction
(Hanlim Mechatronics Co. Ltd.).
Recently developed cleaning robots can be fitted
with spinning brushes, directional air nozzles and
whips, sampling devices, and spraying attachments
for spraying sanitizing solutions for various coatings.
Figure 3: Duct cleaning robot with rotary brush arms.
In this research, the duct cleaning robot designed
will be unique in its functionality and usability. The
robot has equipped with adjustable brush arms at
which rotary brushes are attached to clean four duct
surfaces autonomously. In addition, a force
compliance device was designed and installed at the
end effector of the robot brush arm. Since the
compact compliance device can recognize the
pressure between brush and duct surface, the robot
brushes can clean up the surface as a constant
pressure even if the cleaning surface is of
Figure 4: Brush and compliance device.
3.1 Mobile Platform and Arm
A newly developed duct cleaning robot is depicted
in Figure 3. The robot consists of a base body,
continuous tracked wheels, a camera module, LED
light and two brush arms; upper and low arm. Six
motors have been used to provide proper torques for
ductworks. The rubber-based tracked wheel was
selected not only to decrease slippage between
driving wheels and duct surface, but also to reduce
weight of the robot platform (Wang et al., 2006).
The lower brush arm is attached at the front of
platform body and the upper brush is located at the
end effector of the upper arm. The upper arm with
R-R-P joints is height adjustable for handling
various duct sizes.
The compact force compliance device is installed
between two cylindrical rotary brushes at the end of
the upper arm. The compliance device consists of
two springs and linear sensors to read deflection of
the spring. The two springs are connected each other
at the end point to read two directional force. The
brush and compliance device are depicted in Figure
3.2 Force Control with a Compliance
The compliance device attached at the end of the
upper arm enable robot arm to detect force between
brush and surface of duct, which make it possible to
control the constant pressure. Figure 5 presents a
model of force compliance brush illustrated in
Figure 4. Figure 5(a) represent a model when the
brush is contacting only with the upper surface of
duct, and Figure 5(b) shows that the brush is
ICINCO 2012 - 9th International Conference on Informatics in Control, Automation and Robotics
contacting at two points, i.e., upper and side surface
of duct. During the brush is under pressure, normal
force and tangential force are acting on the
compliance device.
(a) (b)
Figure 5: Force compliance brush models; one point
contacting model(a) and two points contacting model(b).
The force acting on the compliance device can be
simply calculated by using spring constant K, which
is given by
where Δx is a spring deformation and K is a spring
constant. The two spring constants are identical
when same spring is used. It should be noted that the
deformation of spring (Δx) is a nonlinear because
the stiffness of brush is, in general, empirically
derived and deflection of brush is also affected by
many factors such as bristle modulus, the number of
tufts, and the trim length of bristles (Rawls et al.,
1990). Therefore, value of the Δx has to be found by
calculating the stiffness of brush empirically.
Assuming that 2 α is the angle between two
springs each other, the forces acting on the brush
are calculated as
is a normal force action on the brush
from the upper surface. The friction force is written
The dynamic friction coefficient
can be
derived from Eq. (3) and (4) and given by
Therefore, normal force acting the brush can be
controlled by using robot arm parameters, i.e.,
translation length d and rotation angle θ in case of
one point contact.
When the rotary brush is contacting at two
surfaces as shown in Figure 5(b), the force
equilibrium equation from Eq.(3) and (4) can be
rewritten as
is a normal force between the brush and
side surface and
is a dynamic friction coefficient
at side surface of duct.
Assuming that dynamic friction coefficients are
can be simplified as
Consequently, the cleaning pressure of robot
brush can be constantly controlled with an actuating
arm at both models.
3.3 Control Scheme of Robot System
In addition to control the cleaning pressure of the
upper arm brush, it is required to control the driving
motion of the mobile platform. The mobile platform
can be either controlled by joystick or automatically
by using ultrasonic sensors to adjust position and
orientation of the robot platform. The overall control
scheme is shown in Figure 6.
The CCD camera and LED light enables
operators to monitor and control the system
manually if needed. Thus, the duct cleaning robot
can be controlled automatically or manually through
the user interface system.
Figure 6: Control scheme of the duct cleaning robot.
Force Control of a Duct Cleaning Robot Brush using a Compliance Device
In this study, a new type of duct cleaning robot has
been designed and prototyped. In particular, the
robot has functionality that cleans four sides of inner
duct simultaneously with a constant pressure by
using a force compliance device. For more dedicated
control of brush pressure, the stiffness of brush
should be empirically achieved. The adjustable arm
brush can make it possible to use in different size of
However, there are still more room for
improvement, especially on the autonomous system.
In a real ductwork, there are many components at
the inner duct such as dampers, fans, joints and
curves that make it difficult to operate the robot
consistently and autonomously. Therefore, more
intelligent operating system has to be implemented
to improve the cleaning process effectively. The
developed prototype robot will be continuously
upgraded and tested at the test-bed of air duct
This research was carried out as a part of the subway
air duct cleaning robot project (Eco-Innovation, No.
E211-40002-0003-0) funded by the Ministry of
Environment in Korea.
Brosseau, L. M., Vesley, D., Kuehn, T. H., Melson, J.,
Han, H. S. 2000. ‘Duct cleaning: A review of
associated health effects and results of company and
expert surveys’, ASHRAE Trans, 106, 180-187.
Finnish Society of Indoor Air Quality and Climate
(FiSIAQ). 2001. Classification of indoor climate 2000,
Espoo, Finland.
Foarde, K. K., Menetrez, M. Y. 2002. ‘Evaluating the
potential efficacy of three antifungal sealants of duct
liner and galvanized steel as used in HVAC systems’,
J Int Microbiol Biotech, 29, 38 43.
Hanlim Mechatronics Co. Ltd., XPW-601 Duct Robot,, Korea
Holopainen, R., Tuomainen, M., Asikainen, V., Pasanen,
P, Säteri, J., Seppänen, O. 2002. ‘The effect of
cleanliness control during installation work on the
amount of accumulated dust in ducts of new HVAC
installations, Indoor Air, 12, 191-197.
Holopainen, R., Asikainen, V., Tuomainen, M., Björkroth,
M., Pasanen, P., Seppänen, O. 2003. ‘Effectiveness of
duct cleaning methods on newly installed duct
surfaces’, Indoor Air, 13, 212-222.
HVCA, 1998. Cleanliness of Ventilation System, Guide to
Good Practice Cleanliness of Ventilation Systems.
Heating and Ventilating Contractors Association,
London, HVCA Publications.
Jung, Y. H., Ahn, B. W., 2003. “Measurements on
Contamination in Air Duct and Air Handling Unit,”
Journal of the Korean Society of Living Environment
System, Vol. 10, No. 1, pp. 41-46.
Morey, P. R. 1988. ‘Experience on the contribution of
structure to environmental pollution’. In R. B.
Kundisin (ed.), Architectural design and indoor
microbial pollution. Oxford University Press, New
York: 40-79.
Pasanen, P. 1998. ‘Emissions from the filters and hygiene
of air ducts in the ventilation systems of office
buildings’. Doctoral dissertation, Department of
Environmental Sciences, University of Kuopio,
Kuopio, Finland.
Robotics Design, ANATROLLER ARI-100, http://www., Canada
Rawls, H. R., Mkwayi-Tulloch, N. J., Krull, M. E., 1990.
A mathematical model for predicting toothbrush
stiffness. Dental Materials, Vol. 6(2), pp. 111-117.
Wang, Y., Zhang, J., 2006. Autonomous Air Duct
Cleaning Robot System, Proc. of International
Midwest Symposium on Circuits And System, pp. 510-
513, 2006
ICINCO 2012 - 9th International Conference on Informatics in Control, Automation and Robotics