Content uploaded by Samet Gul
Author content
All content in this area was uploaded by Samet Gul on Aug 14, 2018
Content may be subject to copyright.
International Conference on Engineering Technologies (ICENTE’17), Dec 07-09, 2017 Konya, Turkey
1
Abstract - The aim of this work is to design an ultrasonic
non-destructive testing (NDT) system that utilizes the
advanced features of smartphones and tablets. These superior
features of smart devices can be listed as wide and sensitive
touch screen, mobile internet connection, satisfactory CPU
performance, application development environment and more.
Integrating these features into a conventional NDT system
increases the capabilities of the system while decreasing
hardware costs. Therefore, an ultrasonic development card
was designed in order to obtain these advantages. This
development card is capable to measure thickness and detect
internal flaws. Designed NDT system was also supported with
encoder reading and ultrasonic data acquisition functions
which are needed in computer controlled ultrasonic scanning
systems. An android application was developed to send
configuration data to the ultrasonic development board and
visualize received signals captured from ultrasonic
transducers. Smart devices are paired to the development card
over Bluetooth communication channels. Ultrasonic signals
visualized by the smart device can be saved for future
investigations or shared via internet with different experts.
Access to saved ultrasonic data provides the results can be
evaluated in a variety of environments or monitor the
inspection of a test material remotely.
Keywords – Ultrasonic nondestructive testing, Smart
systems, Android, Smartphone, Tablet, Hardware design.
I. INTRODUCTION
ON destructive inspection (NDI) is a measurement and
evaluation technique that gives information about
inner structure of materials without harming or damaging
them. There are many NDI methods and ultrasonic testing
(UT) is one of them. However, it is apparent that ultrasonic
method is more versatile than the other methods. In most
cases, UT is used as a main or a complementary inspection
method [1]. Some important advantages of UT technique
can be listed such as easy to use, harmless to human and
environmental health and has high accuracy [2]. Despite the
fact that UT is a popular method, device prices are very
expensive.
Smartphones have many advanced features and its
powerful data processing capability makes them rapidly
evolving technology. Nowadays, this technology has
become an irreplaceable part of our lives [3].
Integrating smart devices into UT application has
included many useful advantages. In this application we
have designed our UT system extremely cheaper compared
to the market.
The main function of the ultrasonic card developed in
this study is to generate high voltage and high frequency
electric pulses, digitize the incoming ultrasonic echoes and
communicate with smartphone over Bluetooth. The
application that we developed for smart devices can
interpret and evaluate the ultrasonic data captured by the
UT board.
II. SYSTEM DESIGN
The method that producing and capturing ultrasonic
waves with the same single ultrasonic transducer is called
pulse-echo (PE) and it is one of the most commonly used
methods in ultrasonic NDI [2]. In this study, PE method
was also used in our application. Figure 1 illustrates the
basic idea of ultrasonic inspection process with our system.
The process of our inspection application is as follow.
First, the desired configuration values for the ultrasonic test
are sent from the Smartphone to the ultrasonic card via
Bluetooth. When the ultrasonic card received the
configuration data, the configuration of the card is
accomplished such as amplification rate, desired high
voltage amplitude value and other parameters. For example,
"Number of Pulses" parameter is written in the
corresponding address in FPGA because ultrasonic pulses
are generated by MOSFET bridge controlled by an FPGA.
When the configuration process is finished then the
initialization process is performed by microcontroller unit
(MCU). Ultrasonic waves are started to be generated by
ultrasonic transducer after the initialization process
completed. Reflected ultrasonic signals from the material
are captured by the same ultrasonic transducer and
amplified by amplifier units in ultrasonic card. Amplified
signals are transferred to the digitizer unit and converted to
10 bit binary values with Analog to Digital Converter
(ADC) unit. ADC outputs are connected to the FIFO
memory built in the FPGA. When FIFO memory becomes
full an external interrupt is created to activate MCU. FIFO
memory content is transferred to External Bus Interface
(EBI) of the MCU via Direct Memory Access (DMA) and
MCU sends this data to the smart device over Bluetooth
communication channel. The program developed for the
smart phone is used to evaluate and interpret the digitized
signal received.
Inspection results are displayed on smartphone screen as
A-Scan graphic representation, see Figure 1. In this graph,
the horizontal axis defines time and the vertical axis defines
the signal strength.
Smartphone controlled ultrasonic nondestructive
testing system design
S. GÜL1 and A.T. ÖZDEMİR2, 3
1 Bursa Technical University, Bursa/Turkey, samet.gul@btu.edu.tr
2 Erciyes University, Kayseri/Turkey, aturan@erciyes.edu.tr
3 Ultrasonar Defense and Aviation Technologies AS, Kayseri/Turkey, aturan@ultrasonar.com.tr
N
2
A. Hardware Design
This ultrasonic development card was previously
implemented on a TMSC6416 DSP development kit with
several auxiliary kits [4]. We designed our custom
ultrasonic card with the knowledge that we experienced our
previous implementation on a commercial development kit.
The general block diagram of our custom ultrasonic card is
given in Figure 2. The block diagram of the system consists
of four main units, Ultrasonic Analog Front End (UAFE),
voltage regulation, FPGA and MCU units.
Figure 1: General overview of ultrasonic inspection process of
the developed system
This card is able to produce adjustable high amplitude
voltage between +40 V and +250 V. Our ultrasonic card
has a programmable high-frequency ultrasonic pulse
generator, 0.5 MHz to 10 MHz, in ±250 V range.
Parameters of ultrasonic pulses such as amplitude,
frequency, repetition frequency and number of pulses can
be set by user. This ultrasonic card has also an adjustable
72 dB amplifier and a fast 100 MHz digitizer (ADC).
FPGA unit is employed for data acquisition and signal
processing. The MCU unit has a high-speed (200 MHz)
DSP microcontroller that manages peripherals by
controlling LCD screen, keypad, buzzer, buttons and set
parameters of regulators, amplifiers and ADC. MCU unit
also communicates with internal memories and USB port.
The ultrasonic card developed in this study supports both
PE and Through Transmission (TT) ultrasonic inspection
methods. Ultrasonic data acquisition system consists of two
channels, A and B Channels. This means two different
frequencies can be used at the same time in a single
inspection. It can be operated with up to four single-element
(two TT channels) and two dual-elements (two PE
channels) ultrasonic piezoelectric transducers.
This development card provides ultrasonic inspection
with a smartphone over Bluetooth and with a PC over USB
communication channels.
Analog to digital conversion, data acquisition and data
processing operations were made at very high frequencies
by a 100 MHz ADC, a 50 MHz Cyclone FPGA and a 200
MHz PIC32MZ family microcontroller respectively.
B. Software Design
In this study, LG G5 H850 smartphone is used as
android software development platform for ultrasonic
inspection applications. This smart device has a Qualcomm
MSM8996 Quad-core CPU chipset which is powerful
enough to deliver satisfactory computing performance.
There are two of 2.15 GHz Kryo and two of 1.6 GHz Kryo
embedded CPUs in this chipset [5]. The android application
that developed in this study also has the ability to save the
inspection data to a SQLite database. This database can be
transferred from smartphone to PC. This function gives user
an opportunity to work with saved inspection data on
MATLAB environment.
Structure of the data received via Bluetooth in the
Android application is described in Table 1. Each package
starts with hexadecimal '0xFF' value means open a new
transfer. Then the package length is transmitted with two
bytes, the first bit of these two byte is the most significant.
The forth byte defines the package topic, and the next bytes
are the data of the topic.
Table 1: The structure of the data package used in communication
between the smart phone and the card
Data No
Byte0
Byte1
Byte2
Byte3
Byte4
…
Data
0xFF
0x00
0x03
0x01
0x01
…
Meaning
Start the
Transfer
Package length
Package
Topic
Topic
Data
The program of the Android interface consists of two
main parts, which are configuration and analysis. Some
features in the configuration settings menu are given as
follows. "Buffer size" is the parameter that specifies the
memory size of captured ultrasonic signals. This feature can
be set between 200 and 16.000 samples (each sample is
represented with eight bits). It limits the amount of data to
be acquired after an ultrasonic wave package is sent to
material. The acquisition duration of ultrasonic echoes
varies depending on the ADC sampling frequency. For
example, when the sampling frequency is set to 100 MHz,
then sampling process will be performed by 10 ns steps. In
this test if the buffer size is set to 16.000 then ultrasonic
data acquisition will be performed in 160 µs (10 ns ×
16.000 samples). When the sampling frequency decreased
to 50 MHz, sampling will be made with 20 ns steps, and the
data acquisition will be completed within 320 µs. "Pulse
frequency" is used to change the frequency of high
amplitude pulses depending on ultrasonic transducers.
“High voltage” digitally adjusts the pulser voltage with 0.5
V steps. Depending on different types of ultrasonic tests,
the number of pulses applied to the material may vary and
this parameter can be changed with “The number of pulses”
submenu. "Sampling frequency" is an important system
parameter and defines the resolution of the acquired
ultrasonic signal. The system capacity is maximum 100
MHz however 50, 25 and 12.5 MHz options are available.
Mode of UT method can be switched between TT and PE
by "Ultrasonic method" parameter. In TT mode channels
can be set as pulser or receiver by labeling them as “Select
Read Channel" or "Select Pulse Channel" options.
International Conference on Engineering Technologies (ICENTE’17), Dec 07-09, 2017 Konya, Turkey
3
Figure 2: General block diagram of the developed card
Analysis menu has different tools and features as
summarized follow. The most important analysis tool is the
“graph” section and it visualizes the captured ultrasonic
echoes on smartphone screen. Horizontal axis gives the
number of samples (up to 16.000) and vertical axis gives
signal strength (between 0-255 steps). Recorded UT data
can also be recalled for visualization on the graph screen.
Figure 3 shows menus on the screen. The sampling speed
can be set with ADC FREQ menu. Amplification of the
card can be managed with GAIN menu. A gate for
measuring the thickness can be defined with GATE
CONFIG menu. Different gates are shown in Figure 3 (a, b,
c) as red lines. Thickness of a material can be displayed on
graph screen in mm or inch. Thickness gives distance of
back wall of a material or possible flaws as it shown in
Figure 1. UT devices are tested with calibration blocks.
CALIBRE menu is also used to test our ultrasonic card with
known materials in terms of composition and thickness. We
able define velocity of ultrasonic signal according to certain
thickness of a calibration block.
III. EXPERIMENTAL RESULTS
An ultrasonic evaluation was applied to a 63 mm tick
aluminum reference block, shown in Figure 4. There are
two artificially created defects on this aluminum test block
at different depths. The aim of this evaluation is to
demonstrate how successfully our system detects flaws in
the reference block. The sampling frequency was set to 100
MHz. This means time interval between two successive
samples is 10 ns. When evaluation screens (Figure 3) are
examined, there are three meaningful sections in returning
echoes. These include initial pulse between 800 and 1000
samples, flaw echoes returning from a defect in material
between 1700 and 2000, and the back wall echo reflected
from the bottom of the material between 2600 and 3000
samples. Some ultrasonic waves generated from transducers
penetrate into the material while some are reflected back
from the surface of the material. Signals reflected back
from the surface create initial pulse. When penetrating
waves encounter with a discontinuity some of them are
reflected back from that point and the remaining waves
continues to their path. Reflected signals from the
discontinuity regions cause flaw echoes. Back wall echo
refers to the ultrasonic waves that are reflected back from
the bottom of the material.
Figure 3: (a) Measurement of the thickness of the material, (b)
Position of the first defect in the material, (c) Position of the
second defect in the material
4
All these three kinds of reflections are caused of
acoustic impedance mismatches. When ultrasonic waves
transferring from one material to another, some of them are
reflected back and remaining continue to penetrate the
materials. The reflection and penetration rates of ultrasonic
waves vary according to materials’ acoustic impedances.
Reflection rate increases while penetration rate decreases
with increased difference between materials’ acoustic
impedances.
Figure 4: Aluminum reference block
The distance of a reflected echo is calculated as
Equation (1). Here d is the distance between the surface of
the material and the discontinuity region in an inspected
material, v is the speed of sound in the material and t is the
time of flight that sound needs to travel the particular
distance. The reason of dividing t into two is that the
ultrasonic evaluation is performed with pulse-echo method
in this study. In PE method only one ultrasonic transducer is
used to generate pulses and capture back echoes, therefore
obtained time is composed of arrival and departure times.
Since arrival and departure times are equal t is divided into
two [6].
)2/(tvd
(1)
Speed of sound through different materials is different
[7]. Thickness of a material and the depth of the defect in it
can easily be calculated by Eq. 1. Since the speed of sound
in aluminum and the duration are known the depth of defect
can easily be calculated. In figure 3 (a), the back wall echo
of the aluminum block was evaluated. In order to find the
duration of the back wall echo the number of samples
between the highest peak of the initial pulse and the highest
peak in the back wall echo are multiplied by 10 ns. The
speed of sound in aluminum is constant and available at
sources. The thickness of the material was found to be
approximately 63 mm. In figure 3 (b), the location of the
first defect in the aluminum block was detected at a depth
of about 32 mm. In figure 3 (c), second defect was detected
at a depth of about 26 mm in the aluminum block. As a
result, an accurate inspection was performed using
smartphone.
IV. CONCLUSION
A different point of view has been achieved by
integrating smartphones with conventional ultrasonic NDT
devices. The designed system can receive the inspection
data from custom designed ultrasonic card wirelessly. Smart
device can evaluate the data and control the ultrasonic card.
An android program developed for smart devices that can
record and evaluate the received ultrasonic signals.
Smartphone can share the saved inspection data thanks to
the mobile communication features. Integrating the
advanced features of the smartphone brings innovative
features to conventional ultrasonic NDT. For example, user
can save the coordinates of the location where the
inspection is made or take photos of the inspected
materials. Smartphone CPUs are powerful enough to
perform DSP calculations. With the development of mobile
technology, the numbers of academic study is increasing
[8]. For example, if DSP operations are needed and there is
no enough processor power to do this on the hardware, it
can be handled by a smartphone. Android NDK platform
gives ability to programmer use C/C++ coding skills in
android applications. For example, the grain structure of
materials is scattering the ultrasonic waves and this causes
noise in the received signal [9]. Removing these noises
from the signal is possible and this filter can be built on
smartphone instead of a DSP chip. These advantages
provide a development environment for different
applications for enhancing the features of the conventional
NDT application.
REFERENCES
[1] M. V. Brook, Ultrasonic Inspection Technology Development and
Search Unit Design: Examples of Practical Applications. IEEE
press, 2012.
[2] A.T. Özdemir, “A low-cost and highly reliable amplitude modulator
design for PC-based PCI plug ultrasonic pulser receiver boards”,
Electronics World, 120 (1937), pp.30-33, 2014.
[3] A.T. Özdemir, A. Orman, “Developing an iPhone smartphone based
fall detection algorithm”, Proc. 23rd Signal Processing and
Communications Applications Conference (SIU), pp. 2456-9,
Malatya, Turkey, 16-19 May, 2015.
[4] A.T. Özdemir, A. Atcı, “A TMS320C6416 DSP-based high-speed
data acquisition system”, Proc. 23rd Signal Processing and
Communications Applications Conference (SIU), pp. 1315-8,
Malatya, Turkey, 16-19 May, 2015
[5] GSMArena. (2016, February). LG G5 H850. [Online]. Available:
https://www.gsmarena.com/lg_g5-7815.php
[6] NDT Resource Center. (2012) Normal Beam Inspection. [Online].
Available: https://www.nde-ed.org/EducationResources/Communi
tyCollege/Ultrasonics/MeasurementTech/beaminspection.htm
[7] Olympus. Material Sound Velocities. [Online]. Available:
https://www.olympus-ims.com/en/ndt-tutorials/thickness-gage/
appendices-velocities/
[8] W.J. Yi, S. Gilliland, J. Saniie, “Mobile ultrasonic signal processing
system using android smartphone” IEEE 56th International Midwest
Symposium on Circuits and Systems (MWSCAS), pp. 1271-1274,
2013.
[9] J. Saniie, D.T. Nagle, K.D. Donohue, “Analysis of order statistic
filters applied to ultrasonic flaw detection using split-spectrum
processing,” IEEE Transactions on Ultrasonics, Ferroelectrics and
Frequency Control, vol. 38, no. 2, pp. 133-140, Mar. 1991.
