DEVELOPING AN EMBEDDED CONTROL SYSTEM TO MINITURIZE AUTOMATIC NDT INSPECTION OF STEEL PLATES.

fangscaryAI and Robotics

Nov 13, 2013 (3 years and 9 months ago)

55 views


DEVELOPING AN EMBEDD
ED CONTROL SYSTEM TO

MINITURIZE
AUTOMATIC NDT INSPEC
TION OF STEEL PLATES
.



Suresh Sreedharan, Goran Bezanov, Bryan Bridge


South Bank University, London, U.K.


ABSTRACT


To Automate NDT Inspection of steel plates, a number of successf
ul attempts have been made in recent
years, however the resulting systems invariably rely on a heavy umbilical cord for supplying the drive
power to the inspection vehicle and delivering measurement data to the central computer. The weight of the
cord limi
ts the distance that an inspection system can cover. As a consequence there is room for
improvement by reducing the size as well as complexity of inspection systems. This paper describes an
novel approach to the design of an embedded control system, base
d on the 8051 microcontroller, in
developing a compact, automated steel plate NDT inspection system. Our results will show that it is
possible to reduce the size and complexity of NDT inspection systems through the use of advanced design
and programming te
chniques in embedded control applications. We will also show that a large number of
similar systems, deployed concurrently with centralized supervisory control would offer enhanced
performance in terms of area coverage and speed of NDT inspection.



INTRO
DUCTION


The non
-
destructive testing (NDT) of steel plates
is performed to detect defects. Ultrasound
techniques of Pulse
-
Echo and Time of Flight
Diffraction (TOFD) are proven methods of NDT
inspection [1]. These techniques utilise a piezzo
electric transd
ucer to generate an ultrasonic
pulse, which is transmitted at an angle and
propagated through the material under test and
which in turn interacts with any defects. This
interaction produces diffracted waves and
reflected waves as shown in figure 1 above.



Figure 1. Ultrasound wave Diffraction and
Reflection interaction with a defect.


Detection of the reflected wave is the basis for
Pulse
-
Echo method. In order to recognise the
various orientations of the defects; in practice a
number of transducers are re
quired with
ultrasound being propagated at a variety of
angles. Automated systems have evolved
offering comprehensive control and data
acquisition hardware [2]. This approach is very
effective, however it is recognised that the large
number of probes pla
ced on a dedicated robotic
vehicle and the associated signal processing
hardware, carry the penalty of large umbilical
chords dragged along inspection platforms
carrying power and data signals between the
inspection system and the signal processing
station
.


The number of probes can be reduced if
diffracted wave is analysed (TOFD). In contrast
to reflected waves the diffracted wave energy is
emitted in all directions and it is potentially
possible to perform complete inspection using
only two probes.


In b
oth cases probes are positioned on a vehicle
which is guided through a central control system.
Signals are processed in real
-
time and displayed
on a central console. Speed of NDT inspection is
largely governed by data processing, which in
turn requires con
siderable computer processing
power. Each probe requires its own high
resolution ADC and DAC for signal control, on
Transducer
Reflectio
n
Defect
Ultrasound wave
Diffraction
top of which shaft encoders and associated drive
hardware is required to provide orientation and
speed control. [2]


With continued advances

in embedded systems
hardware and software it is now possible to
produce a compact NDT inspection system
which is remotely controlled and which has an
on
-
board power supply and signal processing
capabilities.


As an example, the 16
-
bit high speed embedded
processor with 128 Kbytes on
-
chip memory [3],
offers possibilities to develop an on
-
board NDT
data acquisition and processing as well as
supervisory control over the Vehicle.


Microstepping and chopper circuitry using
stepper motors in place of servos con
siderable
reduce energy consumption to deliver lower
drive power and thus facilitate on
-
board power
sourcing. An efficient Li
-
ion Battery can
provide longer inspection periods for the vehicle
Radio modem technologies provide data
communication for supervi
sory control [5].


CONCEPTUAL DESIGN


In order to reduce size and drive power
requirements, the TOFD technique [7] has been
adopted rather than conventional Pulse
-
Echo
method, and uses angle probes for maximum
scan plane coverage [8]. Novel Hydro probes c
an
also be used as dry contact which eliminate steel
plate contamination by using water couplet for
wet contacts[9]. Additionally, wheel probes can
survive uneven plate surface for a certain degree
of elevation.




Figure 2 AMSPIR Prototype

1


Dual Stepp
er Motor driver unit combination has
been selected to avoid considerable power
consumption by using servomotors [4]. To
achieve maximum maneuverability of the
vehicle, the idea of two
-
Hybrid stepper motors
coupled together to perform a 360


turn on
central

axis by rotating the wheels in opposite
direction, has been used in the design. The
prototype layout of ARMSPIR
-

‘A Remote
Miniature Steel Plate Inspection Robot’ is shown
in figure 1. The vehicle measures 180mm x 150
mm x 80mm and accommodates onboard
circuitry, power source and ultrasound probes.


HARDWARE


16
-
Bit embedded processor with 40MHz
-
clock
speed has the ability to process data from flash
A/D of the ultrasonic cards for different scan
views in order to analyse faults in the plate.


Figure 2 s
hows the hardware architecture to
support the automated remotely controlled NDT
inspection system. The embedded processor and
adopted interfacing methods render the
production of the system and its operation
simpler compared to conventional systems.


The
system always sends the vehicle position
records to a nearby computer and when detection
occurs it sends the fault position as well. System
software further allows the computer to draw the
real time trajectory path of the vehicle together
with the fault p
osition. The complete data is
stored in a plugged
-
in on board memory for
further analysis.


The flowchart of figure 3 shows the overall
functionality of AMSPIR. Note that the system
will perform a thorough scan over the area of
fault detection to determ
ine the presence of
genuine flaws and also facilitate sizing, which is
a feature attracting increased popularity within
the NDT research community.


The embedded processor has build in CAN
-

Controller Area Network [6] and it allows
multitasking with other
similar robotic vehicle
inspection systems. In this manner, multiple
units can be used to facilitate coverage of large
inspection areas. The processors can
communicate with each other by using CAN and
an integrated supervisory control system can be
program
med to plan the inspection trajectories of
each unit for their concurrent inspection. The
CAN network can be connected via radio
modems thus providing remote data
communications features.




Figure 3 Hardware architecture


The 16 Bit flash A/D converter
of NDT data
acquisition converter rate requires a fast
processor to handle the output data. The
processor also provides on
-
line data processing
to distinguish faults and other anomalies from
the captured data


To miniaturize NDT process the system should
be comparatively small in size and should
perform better than conventional methods.
Several concepts has been adopted and tested to
miniaturize the drive system.
Hybrid stepper
motors with microstepping and chopper circuitry
must be the solution because co
mpared to
servomotors, minimum control hardware is
required.


Microstepping is a way of moving the stator flux
of a stepper more smoothly than in full
-

or half
-
step drive modes. The Torque (T) developed by
the motor is a function of the holding torque




(
H
T
) and the distance between the stator flux

(
s
f
) and the rotor position (
r
f
).

)
sin(
r
s
H
f
f
T
T






(2)

Where (
s
f
) and (
r
f
) are given in elect
rical
degrees.

mech
el
f
n
f


)
4
/
(



(3)




Where
n

is the number of full
-
steps per
revolution.



Figure 4 Flow Chart of Overall system.


The chopper circuitry reduces power
consumption considerably for the drive power
around 450mA compared to 3A in c
onventional
servomotors, which means that the system can be
battery powered.


RESULTS AND DISCUSSION


The Prototype development shown in figure 8
incorporates the ideas behind the miniaturization
of steel plate NDT inspection. The software has
been develop
ed in embedded C and the overall
function is shown in Figure 7. The Vehicle has
been subjected to several tests. Figure 11 shows
one of the B
-
Scan view of Steel plate Inspection.
The prototype has been controlled remotely, and
the trajectory plan as shown
in figure 5.




Figure 5 Trajectory Plan for Automation


Affiliated research work in the Center of
Automated and Robotic NDT at South Bank
University has established that TOFD is a viable
NDT inspection method [7] that can be
incorporated in this design.
See Figures 9 and 10
which proving advantages of Automated NDT.


Initial trials have therefore concentrated on
vehicle control.


CONCLUSIONS


It has been stated that to automate NDT
inspection of steel plates, a miniature Robotic
NDT inspection Vehicle can

be designed by
using high speed novel embedded processors.
The Robust design offers less complex NDT
inspection for finding faults and other anomalies
in steel plates. The development of an embedded
control system offers enhanced possibilities for
an inte
grated, compact NDT inspection system
that can function as one of many units in a

centrally co
-
ordinated NDT environment.


The developed system is completely
independent, can be programmed to be self
-
navigational and is designed to facilitate on
-
board rea
l
-
time data processing and is battery
powered. The built
-
in CAN of the chosen
processor allows multitasking of similar systems
working concurrently to inspect large surface
area.


Figure 7 Prototype Function Flow Chart


REFERENCES


1.

Bridge,B, Khaled,A, Y
ochev,B, Non
-
Destructive Testing, An Eastern
-
Western
European Perspective, Published by The
Institute of Non
-
Destructive Testing, Sept.
1998.

2.

Rakocevic,M, Wang,X, “Automated
Inspection System for NDT of Steel Plates
”,
British Institute of Non
-
Destructive T
esting,

September 1998
.

Changed
scan side

3.

Siemens

C163
-
16F, “
Highest flexibility
with 128 KByte on chip memory
-


September ', Press Release, 1997.

4.

Lin ,T,T, TECHNICAL PROGRAM, “High
Resolution Motor For
Microstepping
Control Lin Engineering (U.S.A.) 19962.
,
Oct. 1995, ISI
R/IA 95.

5.

Robert Fitzgerald (CEO)., Falls Church,
Virginia and Columbia, Maryland, Press
Release,
“YOUNG DESIGN, INC. Acquires
ZEUS WIRELESS, INC”
, May 7, 2001.


6.

Thomas Dedelmahr, Thomas Schmid., 1997,
“The C167CR adds gateway functions to the
automotive c
omputer
”,
This article was
released in the Hitex UK Ltd. C51/166
newsletter.

7.

Hecht, A, Time of Flight Diffraction
Technique(TOFD)
-
An Ultrasonic Testing
Method for all Applications?, NDT.net
-
Vol.2 No.09, September 1997.

8.

Shyamal Mondal, “An overview TOFD
m
ethod and its Mathematical Model”,
NDT.net


Vol.5 No. 04, April 2000.

9.

S. Bourne
-

Sonatest PLC 'United Kingdom;
M. Newborough, D. Highgate
-

Cranfield
University' UK

Roma (Italy)., 15
-
21 October 2000, “Novel
Solid Contact Ultrasonic Couplants Based
on Hy
drophilic Polymers”, 15th World
Conference on Nondestructive Testing.




Figure 8 The Remote Controlled
Prototype.




Figure 9 Manual TOFD Scan of Steel plate. A,
B
-
Scan View.



Figure 10 Automated Scan Result
Compared to Manu
al Scan .



Figure 11 B
-

scan View highlighting defects
in steel Plate by using AMSPIR Prototype 1.


AUTHORS ADDRESS


Suresh Sreedharan

CART


Department


Embedded Lab, T518,

South Bank University,

103, Borough Road,

London,

UK.

Email: sreedhs@sbu.ac.u
k