FIELD OF INVENTION
The invention relates to a methodology for (defect detection in raw material semi- finished and finished products using ultrasonic testing equipment. More
particularly, the invention relates to a method for identifying weld defects by
classifying reflected signals in an ultrasonic testing system.
BACKGROUND OF THE INVENTION
Ultrasonic testing is a well established non destructive testing method used in
the industry for the detection of the defects at the different stages of
manufacture namely raw material stage, semi finished stage and final, stage after
manufacture. The method consists of in injecting an ultrasonic wave inside the
test specimen and observing, a reflected wave from the defect, if any. The
reflection follows the law of reflection of sound. Thus the presence of a defect is
essentially ascertained by seeing the presence of the reflected sound which is
indicated in a screen after required amplification by the instrument. The nature
of the reflecting surface is checked by the nature of the echo produced in the
screen and this gives the impression on the nature of the defect.
The signal pattern contains information about the nature of the discontinuity.
The present practice is to judge the echo dynamics and drawing conclusions
about the defect and is not standardized in predicting the nature of the defect.
RU2325637 describes a method, of exciting a pipe using acoustic damped
oscillations and collecting the signals at different iocations along the axis of the
pipeline at equal intervals using laser based systems. Patent US2008i83466
describes a procedure for removing transient signal from speech signals.
US2008179521 describes about a method to remove the noise from an image
produced by a coherent source.
US2008107311 describes a method for the recognition of face in the in a n
image.
WO2008081459 describes a method for compressing the data and wavelet
transformation is employed for this.
OBJECTS OF THE INVENTION
It is therefore an object of the invention to propose a method for identifying
weld defects by classifying reflected signals in an ultrasonic testing system.
Another object of the invention is to propose a method for identifying weld
defects by classifying reflected signals in an ultrasonic testing system which will
be independent of the operator.
An yet another object of the invention is to propose a method for identifying
weld defects by classifying reflected signals in an ultrasonic testing system which
is not influenced by the effectiveness of contact between the transducer and the
test specimen.
A still another object of the invention is to propose a method for identifying weld
defects by classifying reflected signals in an ultrasonic testing system which will
be independent of the surface roughness of the specimen.
A still further object of the invention is to propose a method for identifying weld
defects by classifying reflected signals in an ultrasonic testing system
independent of the strength of the reflected signal.
SUMMARY OF THE INVENTION
Accordingly, there is provided a method for identifying weld defects by
classifying reflected signals in an ultrasonic testing (UT) system, comprising the
steps of providing an ultrasonic testing (UT) equipment having an ultrasonic
probe, the probe is enabled to generate, receive and convert sound signals into
electrical signals, the U. T. equipment is capable of receiving electrical signals
from the probe and amplifying and displaying the received electrical signals;
injecting ultrasound wave signals via the probe at a predetermined location of a
weld in a job; the sound signals when encountering a defect in the weld get
reflected, the intensity of the reflection of the sound waves depending on the
orientation of the encountered defect in the weld; receiving the reflected sound
signals in the probe when the defect in the weld is so oriented that the reflected
signals are capable of reaching the probe; converting the received sound signals
in the probe to electrical signals and feeding the converted electrical signals into
the UT - equipment in which the electrical signals being amplified and displayed;
feeding the electrical signals into a computer apparatus embedded with a
software for carrying out wavelet analysis of the received signals; detecting the
plurality of features responsible for said reflection intensity of the sound waves
from the weld defect, depending on the parameters of wavelet transformation
produced through the wavelet analysis; and extracting the more important
features out of said plurality of detected features for identifying the nature of the
weld defects.
According to the present invention, a beam of ultrasound is directed towards a
defect and refection signals from a single defect is collected and analysed. The
sound is ultrasound and is directed towards a specific location. Wavelet analysis
has been associated with noise removal from signals and for image
enhancement.
The present invention encompasses an ultrasonic testing and for finding weld
defects.
BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWING
Fig. 1: Shows an arrangement for the measurement of the signal and analysis.
DETAILED DESCRIPTION OF A PREFERRED EMBODIMENT OF
THE INVENTION
The method includes injecting ultrasound at a specific location in a job (1)
towards a weld (2), by using an ultrasonic probe (3) which is capable of
generating and receiving and converting the received sound signal into electrical
signals. Sound signal when encounters a defect, is reflected back (7). If the
defect is oriented such that the reflected sound signal reaches the transducer
(3), the transducer (3) converts the sound signal to electrical signals which is fed
into an UT equipment (5). The ultrasonic testing equipment (5) displays the
received signal pattern (6) depending on the nature and the size of the defect.
The signal so received by the equipment (5) is again fed into a computer (8),
where it is analyzed by a software. The wavelet analysis is applied to the
received signal. The parameters of the wavelet transformation clearly indicate
the features of the defects responsible for the reflection. The important features
extracted by the wavelet analysis, namely kurtosis function, mean instantaneous
frequency are clearly indicative of the nature of the defect.
WE CLAIM
1. A method for identifying weld defects by classifying the reflected signals in an
ultrasonic testing system comprising the steps of:
providing an ultrasonic testing (UT) equipment (5) having an ultrasonic probe (3);
- enabling the probe (3) to generate, receive and convert sound signals into
electrical signals;
amplifying and displaying the electrical signals which are received;
injecting the ultrasound wave signals via the probe (3) at a predetermined
location of a weld (2) in a job (1).
enabling the sound signals to encounter a defect in the weld (2) to get reflected
where the intensity of reflection of the sound waves depending on the
orientation of the encountered defect in the weld.
receiving the reflected sound signal (7) in the probe (3) when the defect in the
weld is so oriented such that the reflected signals are capable of reaching the
probe.
2. The method as Claimed in Claim 1, wherein the important features extracted by
the wavelet analysis comprises Kurtosis function, Skewness function and mean
instantaneous frequency, which are indicative of the weld defects present as
well as nature of weld defects.

ABSTRACT

TITLE: A METHODOLOGY FOR CHARACTERIZATION OF WELD DEFECTS
BY ULTRASONIC TESTING
A method for identifying weld defects by classifying the reflected signals in
an ultrasonic testing system comprising the steps of: providing an ultrasonic
testing (UT) equipment (5) having an ultrasonic probe (3); enabling the
probe (3) to generate, receive and convert sound signals into electrical
signals; amplifying and displaying the electrical signals which are received;
injecting the ultrasound wave signals via the probe (3) at a predetermined
location of a weld (2) in a job (1), enabling the sound signals to encounter a
defect in the weld (2) to get reflected where the intensity of reflection of
the sound waves depending on the orientation of the encountered defect
in the weld , receiving the reflected sound signal (7) in the probe (3) when
the defect in the weld is so oriented such that the reflected signals are
capable of reaching the probe.