Kinematics of Pipe Crawling Robots


Oct 31, 2013 (4 years and 8 months ago)








Dynamics of Wheeled In
Pipe Inspection Robots


Paul W. Leu






6350 Mai
n St.


(713) 348

Houston, TX, 77005


Dr. Fathi H. Ghorbel


Mechanical Engineering


(713) 348





pipe inspection robots are needed with smaller size, longer range, and increased
maneuverability. Wheeled in
pipe inspection robots offer simplicity and efficiency and thus,
potential for miniaturization, long range, and good mane
uverability. The first step towards
optimizing the designs and control of such robots is through an analytical dynamic methodology.
This approach involves deriving both the kinematics and kinetics of the robot. Reference frames
are attached to the wheel
s and the body of the robot and combined to derive the robot’s overall
kinematics. Dynamics requirements are then incorporated into the model through analytical
dynamic methods.

This methodology is demonstrated on an actual robot with simulations and
ign optimization methods. Finally, some actual physical experiments are conducted on the
actual robot.

3. Introduction

Pipelines increasingly need to be inspected, maintained, and/or repaired in a wide range
of industries such as i
n the petroleum, chemical, nuclear, space/aeronautic, and waste fields. Not
only is pipe inspection important for optimizing flow efficiency, but also critical to prevent
failure. The effects of time, corrosion, and damage make pipeline failure an increa
sing concern
with some pipelines in use for 30 to 40 years [1]. Pipelines are subject to a variety of operating
conditions that can lead to aging, pitting, corrosion, cracking, material buildup, and mechanical
damages that may cause disastrous and costly
local failures. These failures may result in harm to
workers and the public, environmental damage or contamination, expensive cleanup and repair
costs, and court liabilities. Pipeline safety technologies have been propelled into the national
priority lis
t with recent pipeline accidents. The Pipeline Safety Act of 2000 and the Pipeline
Safety Improvement Act of 2001 now require the development of pipeline management
programs as well as periodic pipeline inspections.

A variety of nondestructive pipe tes
ting methods are currently used to evaluate pipe
conditions. Some methods such as corrosion coupon monitoring, spool pieces, and selective
metallurgical analysis involve removing a section of the pipe to provide information about its
general condition. H
owever, these methods require that the section of pipe be replaced and come
with an associated down time. In addition, underground pipes require excavation and backfilling
for removal and larger diameter pipes cannot be removed for testing. Pipe conditio
ns also vary
between different locations, so these tests do not provide substantive results justifying major or
complete pipe repair or replacements. Pipe failures are often due to an operating anomaly or
localized defect, which are not identified in such


Other methods such as visual outer pipe inspection, ultrasonic pipe testing, eddy current
sensing, and magnetic flux leakage (MFL) inspection do not require pipe shutdown and are
useful for evaluating the outer diameter (OD) surface of pipes a
nd providing wall thickness
dimensions. However, they provide limited insight into internal diameter (ID) defects and can
only be used when the outer pipe surface is readily accessible. Identifying and analyzing internal
defects is critical to pipe maint
enance and repair, as the ID is subject to various operating
conditions. ID inspection is also important where the buildup of scale and/or foreign deposits
inhibit flow.

Inline pipe inspection systems that traverse through the pipes are thus necessary
for pipe
inspection. They can provide visual inspection and documentation, answer specific questions,
identify specific defects, and identify microorganisms involved in certain types of corrosion.
Evaluation of these ID conditions is often desired before

performing corrective measures or
expending capitol. These systems also have the advantage of being able to traverse pipes in a
variety of inaccessible locations such as underground or in certain building locations.

Several inline pipe inspection syst
ems are dominate industry. The most common systems
used are “pigs” or “smart pigs” that utilize fluid pressure to flow along the pipeline and can reach
velocities of up to seven miles per hour depending on driving pressure. Different types of pigs
a variety of functions such as batching or separating dissimilar products, cleaning pipes,
and inspecting internal pipe surfaces. They have a simple structure, are economical to use, and
thus many large pipelines are currently designed with pigging mainte
nance in mind. However,
pigs must have high fluid pressure, cannot stop at arbitrary distances, and cannot maneuver
through various pipe configurations such as elbows. Inspection pigs utilize drive cups on their
front to transfer fluid pressure into a dr
iving force. Thus, inspection pigs must have a diameter
close to that of the pipe ID and they may get stuck due to material buildup or pipe deformations.
To maintain high enough fluid propulsion, inline inspection pigs are usually large (24 inches to
inches in diameter) with the smallest commercially available pig 6 inches in diameter [2].
They cannot get through smaller pipes, tight turns, and some valves. Other types of
commercially available inline pipe inspection systems include tethered robots,
which have
limited range and mobility.

Current research focuses on developing new types of actuators and sensors for inline pipe
inspection robots. Main areas of current research include

1. miniaturization. These robots are being produced for smalle
r and smaller pipelines.
Currently, there are robots [1
3] under research for pipelines as small as from 3 cm to
less than 1 cm in diameter.


durability and endurance. Robots are being developed with higher efficiencies and
longer range.


ty. Pipes include a variety of configurations including elbows, T’s, U
shape bends, variable diameter, branches, and steps that robots must maneuver

The types of robots can be classified according to their method of propulsion

wheels, crawlers

(tracks), legs, inchworm, and contortion.

Among these drive mechanisms, wheeled robots are the simplest, most energy efficient,
and have the best potential for long range. Through loading the wheels with springs, these robots
also offer some advanta
ges in maneuverability with the ability to adapt to in
pipe unevenness,
move vertically in pipes, and stay stable without slipping in pipes. These type of robots also
have the advantage of potential of easier miniaturization.

Some examples of inline whe
eled pipe inspection robots have been developed and are
currently under study [5

9]. This includes the Theseuse robot series and a screw
principle robot
at Tokyo Institute of Technology, the MRINSPECT series at Sungkyunkwan University, and
micro inspect
ion robot at the Toshiba Corporation. The various types of wheels used in these
designs include conventional wheels and screw drive mechanisms. Within the development of
these different drive mechanisms, a dynamic methodology is needed for improved mecha
designs and control.

A fundamental part of dynamics is the discipline of kinematics, the study of the relative
motion of mechanical systems due to geometrical constraints, separate from the forces that
produce such motion. It deals with the positi
on, rotation, displacement, velocity, and
acceleration of these systems. A dynamic model can then be developed around a kinematic
model and then used for feedback control.

The three main types of kinematic modeling approaches are based on geometric,
based, and matrix transformations methods. Muir and Neuman developed a systematic
transformation approach to model the kinematics of wheeled mobile robots. This involves
extending the kinematic modeling of stationary manipulators to wheeled mobile

robots. Wheels
are unique in that they are closed
link chains, with the ground closing the link, as opposed to
manipulators, which are open, except when touching the ground. Wheels are also higher
joints, with point or line contacts, as opposed to
manipulators with surface to surface contact. In
addition, only some degrees of freedom of wheels are actuated, and only some have position or
velocity sensors. Muir and Neuman apply Sheth
Uicker notation, a modified form of Denavit
Hartenberg notation,
to deal with these particularities [7].

The main types of dynamic approaches available are Newton
Euler and Lagrangian.
Kane’s approach, Lagrange’s form of the d’Alembert’s principle, has also been developed for
holonomic systems [8]. Several kinemat
ic and dynamic models have been introduced [9].

The dynamic models can then be used in simulations for optimizing the design of the
surgical manipulators and testing control strategies.

4. Proposed Work and Methods

Model the kinematics of wheeled mobi
le robots inside pipes. This involves using a different
reference frames to relate the kinematics of the robot’s wheels to the motion of the overall
robot. The individual wheel kinematics are combined to obtain the composite robot

The dyn
amic equations of motion are derived by relating the kinematics of the robot to the
forces within the system. This dynamic model is validated through Matlab simulations.

An actual robot is then constructed, and tests are performed, and data is interpret


Kinematic and dynamic modeling by December

Simulations by March

Thesis and experiments by April

5. Qualifications for Work

Graduating May, 02

Rice University

Houston, TX


overall GPA


GPA in Mechanical Engin
eering Major (B.S.M.E.)


out of 2400; 770 Verbal, 800 Quantitative, 800 Analytical

Rice University Scholars Program

Classes: Analytical Dynamics, Feedback Control of Dynamic Systems, Mechanical Design
Applications, Computer Aided Design, Thermody
namics and Heat Transfer, Fluid Dynamics,
Electromechanical Devices and Systems, Material Deformation, Newtonian and Lagrangian
Mechanics, Materials Science, Vector and Multivariable Calculus, Differential Equations



Montville High School Montville, NJ

Spring, 2001 to current
Rice University

Houston, TX

Research Assistant C
ontrol Systems and Robotics Dr. Ghorbel, Dr. Dabney

“Kinematics of In
pipe Inspection Robots”; working on deriving kinematics and modeling screw
mechanism locomotion; heading design team that will build new generation of such robots

ULAP Mechanical Design Report”; designed and built a remotely controlled robotic pendulum
with multi
disciplinary team of engineers for research in kinematics, dynamics, and control; generated
and selected concepts, made mechanical analysis, built prototyp
es for designs, and drafted
manufacturing sketches

Summer, 2001
Caterpillar, Inc.

Peoria, IL

Summer Intern Analysis Tool Development

Tom Allen

“Machine Learning Methods for Product Design and Development”; researched and investigated the
application of two machine learning techniques, locally weighted regression (LWR) and multivariate
adaptive regre
ssion spines (MARS), to system design and optimization; analyzed and assessed
techniques by designing experimental methodologies, conducting tests, analyzing gear churning loss
datasets, and drafting preliminary conclusions; installed MARS technique into C
aterpillar’s Data Analysis
Toolkit; presented findings in front of Analysis Tool Development group

Fall, 1999 to current
Rice University

Houston, TX

1999 Champion

Lego Design Lab Dr. Bennett, Dr. Young

“Millenium Ladybug Design Report”; led a three
person team that designed and assembled an
autonomous robot that won Rice’s ’99 foam cube
collecting competit
ion; identified competition needs,
brainstormed and selected concepts, tested and refined robot through iterative process; finished 1st
place out of 48 teams

Summer, 2000

Rochester, MI

Vehicle Engineering College Intern

Large Car Platform

Evaluated car assembly and manufacturing processes and participated in improvin
g and refining the
industrial processes; implemented and managed timelines for various engine programs; updated weight
databases and summarized various weight packages

Summer, 1998
Covista Communications, Inc.

Little Falls, NJ



ted off
traffic bills and mapped out conversions between bills; coordinated database of
various business accounts

Quality Function Deployment (QFD)

Programming Languages: Visual BASIC, C, C++, MATLAB, Html, Fortran

Software: IDEAS, OptdesX, Excel, Mic
rosoft Project

Operating Systems: Windows, UNIX, Macintosh OS.

Experienced in gas and arc
welding; using a mill, lathe, and other machining equipment






American Society of Mechanical Engineers (ASME)

Rice Chapter

Vice President

InterVarsity (IV) Small Gr
oup Leader

YMCA Camp Counselor

Rice Student Volunteer Program (RSVP) and Key Club

President’s Honor Roll

every semester

Tau Beta Pi National Engineering Honor Society

Rice Engineering Alumni Junior Merit Award

Sartwelle Scholarship

W. L. Moody, Jr
. Scholarship in Engineering, Edward J. Bloustein Distinguished Scholar, Louis J. Walsh
Scholarship in Engineering

National AP Scholar; National Merit Scholar
National Honors Society



There is an increasing need for pipe inspection systems
with smaller size, longer range, and
better maneuverability. Wheeled mobile robots satisfy many of the needs for such systems with
high efficiency, good maneuverability, and potential miniaturization. The first step to the better
design and control of su
ch robots is through a dynamic methodology. These equations can be
applied to a variety of configurations and then optimized for design and control.

7. References

[1] Suzumori, K.; Hori, K.; Miyagawa, T. “A direct
drive pneumatic stepping motor fo
r robots:
designs for pipe
inspection microrobots and for human
care robots” Robotics and Automation,
1998. Proceedings. 1998 IEEE International Conference on , Volume: 4 , 1998 Page(s): 3047
3052 vol.4

[2] Idogaki, T.; Kanayama, H.; Ohya, N.; Su
zuki, H.; Hattori, T. “Characteristics of
piezoelectric locomotive mechanism for an in
pipe micro inspection machine” Micro Machine
and Human Science, 1995. MHS '95., Proceedings of the Sixth International Symposium on ,
1995 Page(s): 193




[5] Design of in
pipe inspection vehicles for /spl phi/25, /spl phi/50, /spl phi/150 pipes

Hirose, S.; Ohno, H.; Mitsui, T.; Suya
ma, K.

Robotics and Automation, 1999. Proceedings. 1999 IEEE International Conference on , Volume:
3 , 1999

Page(s): 2309
14 vol.2

[6] “In
pipe inspection robot system with active steering mechanism”

Ryew, S.M.; Baik, S.H.; Ryu, S.W.; Jung, K.M.; Roh, S
.G.; Choi, H.R.

Intelligent Robots and Systems, 2000. (IROS 2000). Proceedings. 2000 IEEE/RSJ International
Conference on , Volume: 3 , 2000

Page(s): 1652
1657 vol.3

[7] An in
pipe operation microrobot based on the principle of screw

development of a
prototype for running in long and bent pipes

Hayashi, I.; Iwatsuki, N.; Morikawa, K.; Ogata, M.


Micromechatronics and Human Science, 1997. Proceedings of the 1997 International

Symposium on , 1997

Page(s): 125


[8] “Micro inspection robot for 1
in p

Suzumori, K.; Miyagawa, T.; Kimura, M.; Hasegawa, Y.

Mechatronics, IEEE/ASME Transactions on , Volume: 4 Issue: 3 , Sept. 1999

Page(s): 286


[9] “Kinematic Modeling of Wheeled Mobile Robots”, Muir, PF; Neuman, CP; Journal of
Robotic Systems 4
(2): 281
340 APR 1987

[10] “A Generic Kinematic Formulation for Wheeled Mobile Robots” Rajagopalan R. Journal
of Robotic Systems 14 (2): 77
91 FEB 1997

8. Detailed Budget for Work

$1,000 towards computer for the lab

$ 200 for books

$ 500 for robot m