Volume 2, Issue 5, May 2017 International Journal of Innovative Science and Research Technology

ISSN No: - 2456 - 2165

Ultrasonic Non Destructive Technique an effective

tool for Characterization of Aluminium
A.R.Golhar N.K.Choudhari
Assistant Professor in Physics Principal
Priyadarshini Bhagwati College of Engineering, Nagpur Priyadarshini Bhagwati College of Engineering, Nagpur
India India
artigolhar@gmail.com drnitinchoudhari@gmail.com

Abstract Ultrasonic techniques are providing fast and non- material such as composite laminates, new methods is develop
destructive information for quality assurance of the composite for in situ structure, health monitoring of these materials[11].
and help to optimize process parameters. The Ultrasonic Ultrasonic measurements are useful for determining several
parameters are used to indicate the correlation between the important material properties [12]. In this present paper by
acoustic properties and the microstructure of the material. To using ultrasonic non destructive techniques and IDASM
characterize the aluminium metals by knowing the aluminum Neural Network a relationship is developed between
and iron percentage present in the Aluminum metals so that it aluminum and iron percentage present in the aluminium
can be classified into the types of aluminium metals which are
available. Grade of the aluminium samples help user in a position
sample and various observed NDT parameters.
to decide its applications. In this paper an attempt is made to
II. MATERIAL CHARACTERISTICS OBSERVATIONS
characterize the aluminium metals by ultrasonic non destructive
techniques and signal processing technique. To develop the The Various specimen used in the experiments has been
relationship between aluminum and iron percentage present in prepared from aluminium alloys of different grades with
the aluminum metals and the various observed NDT parameters different dimensions. For Ultrasonic testing the sample
such as density, ultrasonic velocity, attenuation, compositions surfaces are smooth to perform investigations. The hardness of
present in aluminium samples, peak amplitude of FFT, Time alloys has measured by Hardness tester. Digital vernier caliper
signal, Power Spectral Density etc IDASM Neural network is
have been used to measure the thickness and dimensions of
used. This Neural model calculates the percentage of aluminium
and iron present in the aluminium samples and it is compare the different samples with a greater accuracy. Density of
with the Experimental data. The impact of various variables on different samples has been calculated using conventional
aluminum and iron percentage present in the aluminum samples method by knowing the masses of the sample which has
is also discussed in this paper. measured in digital weighing machine. The chemical
composition of aluminium alloys have been observed by
Keywords Ultrasonic, Aluminium, , Characteristics, Neural OXFORD instrument, which produces x-rays when energized.
Network.
Ultrasonic NDT Techniques:
I. INTRODUCTION
A. Ultrasonic Velocity Measurement
Non-destructive testing techniques are most commonly Ultrasonic device Ultrasonic thickness gauge using 5 MHz
employed for detection and characterization of flaws in the transducer has been used for the measurement to be carried
component. Apart from flaw characteristics, another parameter out. A direct pulse echo method is used for the measurements.
which is equally important to assess the structural integrity of The ultrasonic device measures the velocity of the acoustic
engineering components is the material property. With the waves in the aluminium samples. By knowing the thickness or
development in electronics and digital technology, ultrasonic distance between the two parallel external surfaces of the
testing parameters, which are affected by changes in material samples in which acoustic wave travel with different
properties [1,2,3] can be measured with greater accuracy . The composition, Velocity is calculated in m/sec according to the
ultrasonic wave/microstructure interaction established new equation
methodologies for non-destructive assessment of various Velocity = Thickness/ Velocity
microstructures in 9% Chromium ferrites steels useful for
practical situations [4]. From non linear ultrasonic assessment B. Ultrasonic attenuation Measurement
the damage parameter can be obtained to quantify pitting
damage in 7075 Aluminium alloy [5] and by thermography The lab set up used for the NDT ultrasonic test is shown in fig
NDT technique [6]. By heat treatment and age hardening (1). The Aluminium samples are placed between the
treatments material characterization is done by ultrasonic non transducer, through BNC cable. The transducer is mounted on
destructive techniques. [7,8] The effective elastic constants of the two ends of a clamp as shown in the figure (1). Glycerin is
the metals composites are calculated by using the values of used as a couplant of ultrasonic vibration through transducer
velocities and the mass densities of composites [9,10].With
the development of new technology and use of light weight

IJISRT17MY09 www.ijisrt.com 15
Volume 2, Issue 5, May 2017 International Journal of Innovative Science and Research Technology
ISSN No: - 2456 - 2165

aluminum & Iron present in it. There may be different

and Aluminium surfaces. The DPR 300 Pulser /receiver of percentage of aluminum & Iron present in samples. It may
JSR Ultrasonic (USA) have been used to generate high voltage not affect velocity, but may impact other ultrasonic parameters
pulse. Ultrasonic transducer is connected to the pulser via like attenuation etc.Results obtained using attenuation,
cable which converts electrical energy to ultrasonic pulse that density, MOE, densities were not sufficient and hence we
is propagated into a test sample. The receiving transducer is introduced frequency domain analysis that has produced very
used to detect acoustic pulses that have propagated through encouraging results. The variation of magnitude of the
test sample. The receiving transducer is connected to the spectrum can be used as a tool for predicting the percentage of
TDS2024 200 MHz Testronix Digital Storage Oscilloscope. A aluminum & Iron.
pair of MODSONIC transducer of 4MHz has been used as a
To calculate the estimated values of percentage of aluminum
transmitting and receiving transducer. Attenuation coefficient
& Iron in aluminium or the observed NDT parameters
, is calculated in dB/mm accordance to equation
Integrated Data Analysis and Stimulation Model (IDASM)
Neural Networks model has used. There are large numbers of
= (20/w) log(Vi/Vo) ----------------- (1)
variables for predicting the percentage of aluminum & Iron of
the aluminium Metals which is the dependent variable. The
where,
dependency analysis is a technique which allows us to build a
Vi is the input Voltage
mathematical description of the relationship between the
Vo is the output Voltage
independent and dependent variable. The network report is
W is the thickness of the sample
generated by IDASM. It shows the results of trained file . The
result is displayed after the file has been trained to the
Fast Fourier Transform (FFT) and power Spectral Density
expected levels and accuracy, and the number of iterative
(PSD) using MATLAB is used to analyzed the received time
cycle is reached. The report contains the impact of
signal . The observed values of peak amplitude Time signal,
independent variables NDT observed parameters on the
FFT, and PSD have recorded. The Modulus of Elasticity is
dependent variables Percentage of aluminium in the sample.
calculated by following mathematical relation
Table (1) shows the impact on Aluminium percentage at
Modulus of Elasticity MOE = (velocity)2 x (density) in N/m2
minimum and maximum values of the Aluminium percentage
(dependent variable) by changing the requisite observed NDT
III. RESULTS & DISCUSSIONS parameters (Independent variable) values by 1%. Table (1)
shows the summary results of behavior of various NDT
observed parameters around minimum and maximum
To establish the relation between these observed NDT
Aluminium percentage.
parameters to characterize the Aluminium metals by
Table I
calculating the percentage of aluminum & Iron of the sample,
the graphs have been plotted for the measurement of Summary Report
percentage of aluminum & Iron with respect to various Behavior around Minimum AL
observed NDT parameters like density, ultrasonic velocity,
attenuation, MOE, Peak amplitude of Time signal, FFT, PSD AL = ( 0.01 )HARDNESS + ( 0.01 )DENSITY + ( -0.01 )VELOCITY
etc. + (0.00) ATTEN + ( 0.00 )MOE + ( 0.01 )TS Y + ( 0.01 )FFT Y
+ ( 0.00 )FFT X + (0.00) PSD Y + ( 0.01 )PSD X
Percentage of percentage of aluminum & Iron present in
aluminium sample by nondestructive ultrasonic method has
been investigated with a variety of parameters. Most of this Behavior around Maximum AL
work has been carried out using ultrasonic waveform AL = ( 0.00 )HARDNESS + ( 0.01 )DENSITY + ( 0.00 )VELOCITY
parameters such as velocity measurement, attenuation, etc. + ( 0.00) ATTEN + ( 0.00 )MOE + ( 0.01 )TS Y + ( 0.01 )FFT Y
The basis of these studies is that as the percentage of + ( 0.01 )FFT X + (0.00) PSD Y + ( 0.01 )PSD X
aluminum & Iron present in aluminium samples change and Table (1) Summary of Network report generated
hence changes in ultrasonic signal propagation is observed.
However all the said parameters may not be sufficient to
characterize the samples and to predict the parentage of

IJISRT17MY09 www.ijisrt.com 16
Volume 2, Issue 5, May 2017 International Journal of Innovative Science and Research Technology
ISSN No: - 2456 - 2165
Table (2) gives the average effect of Independent measured Table (4) gives the average effect of Independent measured
NDT parameters on Aluminium percentage.
Table II Summary Report
Average effect of independent attributes:- Behavior around Minimum FE
Independent Variables Average Effect on AL Rank FE = ( 0.56 )HARDNESS + ( 0.63 )DENSITY + ( 1.83 )VELOCITY + (
DENSITY 0.010000 1 0.23 )ATTEN + ( -2.48 )MOE + ( -0.23 )TS Y + ( -0.06 )FFT Y + ( 1.23
)FFT X + ( 0.16 )PSD Y + ( 0.05 )PSD X
TS Y 0.010000 1
Behavior around Maximum FE
FFT Y 0.010000 1
PSD X 0.010000 1
FE = ( 0.03 )HARDNESS + ( 0.09 )DENSITY + ( 0.19 )VELOCITY + (
HARDNESS 0.005000 2 0.05 )ATTEN + ( 0.00 )MOE + ( -0.05 )TS Y + ( -0.02 )FFT Y + ( 0.25
FFT X 0.005000 2 )FFT X + ( 0.03 )PSD Y + ( -0.34 )PSD X
ATTEN 0.000000 3
MOE 0.000000 3
NDT parameters on iron percentage.
PSD Y 0.000000 3 Table III
VELOCITY -0.005000 4 Table (3) Summary of Network report generated actual and estimated values
Table (2) Average effect of Independent variables on Aluminium percentage. for the iron percentage used to build the Neural Networking Model.

Actual and Estimated values for the Aluminium percentage Table IV

used to build the Neural Networking Model. The graph was
plotted between Actual Aluminium percentage measured Average effect of independent attributes:-
experimentally and the estimated Aluminium percentage by Independent Average Effect on
Rank
Variables FE
ISDAM Neural network model as shown in fig (3). The value
VELOCITY 1.010000 1
of coefficient of determination R2 is close to 1, it shows the
FFT X 0.740000 2
extremely good fit of data. The ISDAM Neural network model
build for this study shows more than 99% accuracy and error DENSITY 0.360000 3
is less than 1%. HARDNESS 0.295000 4
ATTEN 0.140000 5
PSD Y 0.095000 6
FFT Y -0.040000 7
TS Y -0.140000 8
PSD X -0.145000 9
MOE -1.240000 10

Table (4) Average effect of Independent variables on iron percentage

The graph was plotted between Actual iron percentage

measured experimentally and the estimated iron percentage by
IDASM Neural network model as shown in fig (4). The value
of coefficient of determination R2 is close to 1, it shows the
extremely good fit of data. The IDASM Neural network model
build for this study shows more than 99% accuracy and error
is less than 1%.

Figure (3) plots between Actual aluminium percentage measured

experimentally by Estimated aluminium percentage by ISDAM Neural
Network Model of all samples of aluminium.

Similarly the report contains the impact of independent

variables NDT observed parameters on the dependent
variables iron percentage in the sample are also obtained.
Table (3) shows the impact on iron percentage at minimum
and maximum values of the iron percentage (dependent
variable) by changing the requisite observed NDT parameters
(Independent variable) values by 1%. Table (3) shows the
summary results of behavior of various NDT observed Figure (4) plot between Actual iron percentages measured experimentally by
estimated iron percentage by IDASM Neural Network Model of all samples of
parameters around minimum and maximum iron percentage. aluminium

IJISRT17MY09 www.ijisrt.com 17
Volume 2, Issue 5, May 2017 International Journal of Innovative Science and Research Technology
ISSN No: - 2456 - 2165
IV. CONCLUSIONS technique,Journal of composite Science and Technology,
61,(2001), pp1457-1463.
The result of this study demonstrates Digital signal processing [11] Macro Alfano, Leonardo pagnotta A non-destructive
used for ultrasonic signals associated with the IDASM Neural technique for the elastic Characterization of thin isotropic
Network having the potential for estimating the percentage of plates NDT&E International, 40 ,(2007), pp112-120.
aluminium and iron in aluminium sample which may help to [12] Meftaf Hrairi, Mirghani Ahmed, Yassin Nimir
identify the type of aluminium metals, process control, quality Compaction of fly ash-Aluminium alloy composites and
assurance and predicting the applications of existing evaluation of their mechanical and acoustic properties
aluminium metal. Due to this user may be in a position to Advance power Technology ,20 , (2009) , pp548-553.
decide the application of aluminium sample. However, it is to
be noted that the system needs further validation before it
made as commercial product. This will require a large data
base to be collected and documentation from various sources.
REFERENCES
[1] P.P. Nanekar and B. K. Shah, characterization of
material properties by ultrasonics, BARC Newsletter,
Issue No. 249, pp. 25-38.

[2] Dong Fei, David K. Hsu, and Mark Warchol,

Simultaneous Velocity, Thickness and Profile maging by
Ultrasonic Scan Journal of Nondestructive Evaluation,
Vol. 20, No. 3, September (2001).

[3] A.R.Golhar,N.K.Coudhari, Characteristics of alumnium

metals by ultrasonic Non-Destructive Techniques using
Digital signal processing International Journal for
Research and development in Technology, Volume 5,
Issue 56, May 2016, P.P. No.292-295

[4] Anish Kumar, B.K. Choudhary, K. Laha, T. Jayakumar,

K. Bhanu Sankara Rao and Baldev Raj, characterisation
of microstructure in 9 % chromium ferritic steels using
ultrasonic Measurements, Trans. Indian Inst.
Met.Vol.56, No. 5, (October 2003), pp. 483-497.

[5] Santanu De, S. Palit Sagar, S. Dey, Amit Prakash, I.

Chattoraj, Quantification of pitting in two tempers of
7075 Aluminium alloy by non-destructive evaluation,
Corrosion Science, 52, (2010), pp.1818-1823.
[6] M.P.Luong, Infrared thermographic scanning of Fatigue
in Metals Elsevier Science Publishers, 1993
[7] Fawad Tariq Nausheen Naz Rasheed Ahmed Baloch
Faisal, Characterization of Material Properties of 2xxx
Series Al-Alloys by Non Destructive Testing
Techniques,Journal of Nondestructive Eval, Springer
Science, (2011).
[8] M. Rosen, L.Ives, S.Ridder and F.Biancaniello, R.
Mehrabian, Correlation between Ultrasonic and Hardness
Measurements in Aged Aluminium Alloy 2024 ,
Materrial Science and Engineering, 74, (1985),pp. 1-10.
[9] M.O.Si-Chaib,S.Menad.H.Djelouah, M. Bocquet An
ultrasound method for the acoustoelastic Evaluation of
simple bending stresses NDT&E International 34 ,
(2001), 521-529.
[10] Zi-quan Li, Xiao-rong zhang, Shu-Yi Zhang, Zhonga-
hua Shen Determination of the elastic constants of metal-
matrix composites by a laser ultrasound

IJISRT17MY09 www.ijisrt.com 18