Wavelet packet transforms were used to extract features from acoustic emission signals for tool wear monitoring. The acoustic emission signals were decomposed into different frequency bands using wavelet packet transforms. The root mean square values of the decomposed signals in each frequency band were extracted as features. Seven features were found to be most sensitive to tool wear based on analysis of experimental data. By dividing the features by cutting speed, the sensitivity of the features to changes in cutting conditions was reduced, providing effective monitoring of tool wear under different conditions using wavelet packet analysis of acoustic emission signals.
Ensemble Empirical Mode Decomposition: An adaptive method for noise reduction IOSR Journals
Empirical mode decomposition (EMD), a data analysis technique, is used to denoise non-stationary
and non-linear processes. The method does not require any pre & post processing of signal and use of any
specified basis functions. But EMD suffers from a problem called mode mixing. So to overcome this problem a
new method known as Ensemble Empirical mode decomposition (EEMD) has been introduced. The presented
paper gives the detail of EEMD and its application in various fields. EEMD is a time–space analysis method, in
which the added white noise is averaged out with sufficient number of trials; and the averaging process results
in only the component of the signal (original data). EEMD is a truly noise-assisted data analysis (NADA)
method and represents a substantial improvement over the original EMD.
Describes Pulse Compression in Radar Systems.
For comments please contact me at solo.hermelin@gmail.com.
For more presentations on different subjects visit my website at http://solohermelin.com.
Since some figures were not downloaded, I recommend to see this presentation on my website under RADAR Folder, Signal Processing subfolder.
Ensemble Empirical Mode Decomposition: An adaptive method for noise reduction IOSR Journals
Empirical mode decomposition (EMD), a data analysis technique, is used to denoise non-stationary
and non-linear processes. The method does not require any pre & post processing of signal and use of any
specified basis functions. But EMD suffers from a problem called mode mixing. So to overcome this problem a
new method known as Ensemble Empirical mode decomposition (EEMD) has been introduced. The presented
paper gives the detail of EEMD and its application in various fields. EEMD is a time–space analysis method, in
which the added white noise is averaged out with sufficient number of trials; and the averaging process results
in only the component of the signal (original data). EEMD is a truly noise-assisted data analysis (NADA)
method and represents a substantial improvement over the original EMD.
Describes Pulse Compression in Radar Systems.
For comments please contact me at solo.hermelin@gmail.com.
For more presentations on different subjects visit my website at http://solohermelin.com.
Since some figures were not downloaded, I recommend to see this presentation on my website under RADAR Folder, Signal Processing subfolder.
IJERA (International journal of Engineering Research and Applications) is International online, ... peer reviewed journal. For more detail or submit your article, please visit www.ijera.com
IJERA (International journal of Engineering Research and Applications) is International online, ... peer reviewed journal. For more detail or submit your article, please visit www.ijera.com
IJERA (International journal of Engineering Research and Applications) is International online, ... peer reviewed journal. For more detail or submit your article, please visit www.ijera.com
On The Fundamental Aspects of DemodulationCSCJournals
When the instantaneous amplitude, phase and frequency of a carrier wave are modulated with the information signal for transmission, it is known that the receiver works on the basis of the received signal and a knowledge of the carrier frequency. The question is: If the receiver does not have the a priori information about the carrier frequency, is it possible to carry out the demodulation process? This tutorial lecture answers this question by looking into the very fundamental process by which the modulated wave is generated. It critically looks into the energy separation algorithm for signal analysis and suggests modification for distortionless demodulation of an FM signal, and recovery of sub-carrier signals
International Journal of Engineering Research and Development (IJERD)IJERD Editor
journal publishing, how to publish research paper, Call For research paper, international journal, publishing a paper, IJERD, journal of science and technology, how to get a research paper published, publishing a paper, publishing of journal, publishing of research paper, reserach and review articles, IJERD Journal, How to publish your research paper, publish research paper, open access engineering journal, Engineering journal, Mathemetics journal, Physics journal, Chemistry journal, Computer Engineering, Computer Science journal, how to submit your paper, peer reviw journal, indexed journal, reserach and review articles, engineering journal, www.ijerd.com, research journals,
yahoo journals, bing journals, International Journal of Engineering Research and Development, google journals, hard copy of journal
An Optimized Transform for ECG Signal CompressionIDES Editor
A significant feature of the coming digital era is the
exponential increase in digital data, obtained from various
signals specially the biomedical signals such as
electrocardiogram (ECG), electroencephalogram (EEG),
electromyogram (EMG) etc. How to transmit or store these
signals efficiently becomes the most important issue. A digital
compression technique is often used to solve this problem.
This paper proposed a comparative study of transform based
approach for ECG signal compression. Adaptive threshold is
used on the transformed coefficients. The algorithm is tested
for 10 different records from MIT-BIH arrhythmia database
and obtained percentage root mean difference as around
0.528 to 0.584% for compression ratio of 18.963:1 to 23.011:1
for DWT. Among DFT, DCT and DWT techniques, DWT has
been proven to be very efficient for ECG signal coding.
Further improvement in the CR is possible by efficient
entropy coding.
IJERA (International journal of Engineering Research and Applications) is International online, ... peer reviewed journal. For more detail or submit your article, please visit www.ijera.com
IJERA (International journal of Engineering Research and Applications) is International online, ... peer reviewed journal. For more detail or submit your article, please visit www.ijera.com
IJERA (International journal of Engineering Research and Applications) is International online, ... peer reviewed journal. For more detail or submit your article, please visit www.ijera.com
On The Fundamental Aspects of DemodulationCSCJournals
When the instantaneous amplitude, phase and frequency of a carrier wave are modulated with the information signal for transmission, it is known that the receiver works on the basis of the received signal and a knowledge of the carrier frequency. The question is: If the receiver does not have the a priori information about the carrier frequency, is it possible to carry out the demodulation process? This tutorial lecture answers this question by looking into the very fundamental process by which the modulated wave is generated. It critically looks into the energy separation algorithm for signal analysis and suggests modification for distortionless demodulation of an FM signal, and recovery of sub-carrier signals
International Journal of Engineering Research and Development (IJERD)IJERD Editor
journal publishing, how to publish research paper, Call For research paper, international journal, publishing a paper, IJERD, journal of science and technology, how to get a research paper published, publishing a paper, publishing of journal, publishing of research paper, reserach and review articles, IJERD Journal, How to publish your research paper, publish research paper, open access engineering journal, Engineering journal, Mathemetics journal, Physics journal, Chemistry journal, Computer Engineering, Computer Science journal, how to submit your paper, peer reviw journal, indexed journal, reserach and review articles, engineering journal, www.ijerd.com, research journals,
yahoo journals, bing journals, International Journal of Engineering Research and Development, google journals, hard copy of journal
An Optimized Transform for ECG Signal CompressionIDES Editor
A significant feature of the coming digital era is the
exponential increase in digital data, obtained from various
signals specially the biomedical signals such as
electrocardiogram (ECG), electroencephalogram (EEG),
electromyogram (EMG) etc. How to transmit or store these
signals efficiently becomes the most important issue. A digital
compression technique is often used to solve this problem.
This paper proposed a comparative study of transform based
approach for ECG signal compression. Adaptive threshold is
used on the transformed coefficients. The algorithm is tested
for 10 different records from MIT-BIH arrhythmia database
and obtained percentage root mean difference as around
0.528 to 0.584% for compression ratio of 18.963:1 to 23.011:1
for DWT. Among DFT, DCT and DWT techniques, DWT has
been proven to be very efficient for ECG signal coding.
Further improvement in the CR is possible by efficient
entropy coding.
Studies on jet penetration and kerf width at various operating pressure in ma...eSAT Journals
Abstract Abrasive Water Jet (AWJ) machining is one of the non-traditional machining method popular method for machining of hard, heat sensitive and brittle materials. The present work attempts to investigate the effect of operating pressure on depth of penetration and kerf characteristics generated while machining of D2 heat treated steel. It is found that increase in operating pressure increases the depth of penetration and decreases the surface taper on the work-piece. Keywords: Kerf width, Depth of cut, Operating pressure, AWJ Machining
This slide contains theoritical and analytical study about "about abrasive water jet machining process" and it has also discription about " OMAX 60120 abrasive water jet machine.Here analytical stydy is done with mainly ss-304 material.
The application wavelet transform algorithm in testing adc effective number o...ijcsit
In evaluating Analog to Digital Convertors, many parameters are checked for performance and error rate.
One of these parameters is the device Effective Number of Bits. In classical testing of Effective Number of
Bits, testing is based on signal to noise components ratio (SNR), whose coefficients are driven via
frequency domain (Fourier Transform) of ADC’s output signal. Such a technique is extremely sensitive to
noise and require large number of data samples. That is, longer and more complex testing process as the
device under test increases in resolutions. Meanwhile, a new time – frequency domain approach (known as
Wavelet transform) is proposed to measure and analyze Analog-to-Digital Converters parameter of
Effective Number of Bits with less complexity and fewer data samples.
Correlation Analysis of Tool Wear and Cutting Sound SignalIJRES Journal
With the classic signal analysis and processing method, the cutting of the audio signal in time
domain and frequency domain analysis. We reached the following conclusions: in the time domain analysis,
cutting audio signals mean and the variance associated with tool wear state change occurred did not change
significantly, and tool wear is not high degree of correlation, and the mean-square value of the audio signal
changes in the size and tool wear the state has a good relationship.
ECG signal denoising using a novel approach of adaptive filters for real-time...IJECEIAES
Electrocardiogram (ECG) is considered as the main signal that can be used to diagnose different kinds of diseases related to human heart. During the recording process, it is usually contaminated with different kinds of noise which includes power-line interference, baseline wandering and muscle contraction. In order to clean the ECG signal, several noise removal techniques have been used such as adaptive filters, empirical mode decomposition, Hilbert-Huang transform, wavelet-based algorithm, discrete wavelet transforms, modulus maxima of wavelet transform, patch based method, and many more. Unfortunately, all the presented methods cannot be used for online processing since it takes long time to clean the ECG signal. The current research presents a unique method for ECG denoising using a novel approach of adaptive filters. The suggested method was tested by using a simulated signal using MATLAB software under different scenarios. Instead of using a reference signal for ECG signal denoising, the presented model uses a unite delay and the primary ECG signal itself. Least mean square (LMS), normalized least mean square (NLMS), and Leaky LMS were used as adaptation algorithms in this paper.
Rolling Element Bearing Condition Monitoring using Filtered Acoustic Emission IJECEIAES
The defect present in the bearing of a rolling element may affect the performance of the rotating machinery and may reduce its efficiency. For this reason the condition monitoring of a rolling element bearing is very essential. So many measuring parameters are there to diagnose the fault in a rolling element bearing. Acoustic signature monitoring is one of them. Every rolling element bearing has its own acoustic signature when it is in healthy condition and when the bearing get defected then there is a change in its original acoustic signature. This change in acoustic signature can be monitored and analyzed to detect the fault present in the bearing. But the noise present in the acquired acoustic signal may affect the analysis. So the noisy acoustic signal must be filtered before the analysis. In this work the experiment is performed in two stages. In first stage the filtration of the acquired acoustic signal is done by employing the active noise cancellation (ANC) filtering techniques. In second stage the filtered signal is used for the further analysis. For the analysis initially the static analysis is done and then the frequency and the time-frequency analysis is done to diagnose the defect in the bearing. From all the three analysis the information about the defect present in the bearing is well detected.
Evaluation of phase-frequency instability when processing complex radar signals IJECEIAES
A new radar system for digital signal processing before detection is proposed. These are the guidelines for selecting an intermediate frequency for signal processing. The features of signal processing in the case of echo-signal selection by the features of the correlation properties of their complex bypass are described. This paper presents the study of ambiguity function (AF) when processing complex radar signals. In this work, the AF synthesis was performed considering non-determined components and the presence of phase-frequency instability. The received result enhances the potentials for distinguishing an incoherent radar signal. The numerical simulation results of received AF are presented. Considering fluctuation components in the complex AF, depending on the laws of the distribution of amplitude and frequency fluctuations and their parameters, allowed to get the gain in the width of the main lobe from the units to tens of times. Paper represents original analytical expressions for AF of radio-signals modulated by narrow band random processes with various distribution laws.
Ensemble Empirical Mode Decomposition: An adaptive method for noise reductionIOSR Journals
Abstract:Empirical mode decomposition (EMD), a data analysis technique, is used to denoise non-stationary and non-linear processes. The method does not require any pre & post processing of signal and use of any specified basis functions. But EMD suffers from a problem called mode mixing. So to overcome this problem a new method known as Ensemble Empirical mode decomposition (EEMD) has been introduced. The presented paper gives the detail of EEMD and its application in various fields. EEMD is a time–space analysis method, in which the added white noise is averaged out with sufficient number of trials; and the averaging process results in only the component of the signal (original data). EEMD is a truly noise-assisted data analysis (NADA) method and represents a substantial improvement over the original EMD. Keywords –Data analysis, Empirical mode decomposition, intrinsic mode function, mode mixing, NADA,
In many situations, the Electrocardiogram (ECG) is
recorded during ambulatory or strenuous conditions such that the
signal is corrupted by different types of noise, sometimes
originating from another physiological process of the body. Hence,
noise removal is an important aspect of signal processing. Here five
different filters i.e. median, Low Pass Butter worth, FIR, Weighted
Moving Average and Stationary Wavelet Transform (SWT) with
their filtering effect on noisy ECG are presented. Comparative
analyses among these filtering techniques are described and
statically results are evaluated.
DESPECKLING OF SAR IMAGES BY OPTIMIZING AVERAGED POWER SPECTRAL VALUE IN CURV...ijistjournal
Synthetic Aperture Radar (SAR) images are inherently affected by multiplicative speckle noise, due to the coherent nature of scattering phenomena. In this paper, a novel algorithm capable of suppressing speckle noise using Particle Swarm Optimization (PSO) technique is presented. The algorithm initially identifies homogenous region from the corrupted image and uses PSO to optimize the Thresholding of curvelet coefficients to recover the original image. Average Power Spectrum Value (APSV) has been used as objective function of PSO. The Proposed algorithm removes Speckle noise effectively and the performance of the algorithm is tested and compared with Mean filter, Median filter, Lee filter, Statistic Lee filter, Kuan filter, frost filter and gamma filter., outperforming conventional filtering methods.
DESPECKLING OF SAR IMAGES BY OPTIMIZING AVERAGED POWER SPECTRAL VALUE IN CURV...ijistjournal
Synthetic Aperture Radar (SAR) images are inherently affected by multiplicative speckle noise, due to the coherent nature of scattering phenomena. In this paper, a novel algorithm capable of suppressing speckle noise using Particle Swarm Optimization (PSO) technique is presented. The algorithm initially identifies homogenous region from the corrupted image and uses PSO to optimize the Thresholding of curvelet coefficients to recover the original image. Average Power Spectrum Value (APSV) has been used as objective function of PSO. The Proposed algorithm removes Speckle noise effectively and the performance of the algorithm is tested and compared with Mean filter, Median filter, Lee filter, Statistic Lee filter, Kuan filter, frost filter and gamma filter., outperforming conventional filtering methods.
Design and Implementation of Low Ripple Low Power Digital Phase-Locked LoopCSCJournals
We propose a phase-locked loop (PLL) architecture, which reduces the double frequency ripple without increasing the order of loop filter. Proposed architecture uses quadrature numerically–controlled oscillator (NCO) to provide two output signals with phase difference of π/2. One of them is subtracted from the input signal before multiplying with the other output of NCO. The system also provides stability in case the input signal has noise in amplitude or phase. The proposed structure is implemented using field programmable gate array (FPGA), which dissipates 15.44mw and works at clock frequency of 155.8 MHz.
Design and Implementation of Low Ripple Low Power Digital Phase-Locked Loop
12
1. 1
Wavelet packet transforms of acoustic emission signals
for tool wear monitoring
Xiaoli Li*
Institute of Precision Engineering, Harbin Institute of Technology, Harbin, 150001, China
Patri, K. Venuvinod
Department of Manufacturing Engineering, City University of Hong Kong, HK
*Corresponding author and address: Xiaoli Li, Department of Manufacturing Engineering, City
University of Hong Kong, Hong Kong
E-mail: me150001@cityu.edu.hk
Abstract
In the tool wear monitoring systems, one of
the most important issues is to extract signal
features from signal detected under given
cutting conditions. This paper uses wavelet
packet transforms method to extract features
from acoustic emission (AE) signal. Wavelet
packet transforms can decompose AE signal
into different frequency bands in time
domain, the root means square (RMS)
values computed from the decomposed
signal of each frequency band are taken as
features. Analyzing above features, the
features are direct relation to tool wear. The
experimental results indicate that it is an
effective method to extract the features of
tool wear monitoring using the wavelet
packet transforms of AE signals.
Keywords: Wavelet packet transforms,
acoustic emission (AE), tool wear
monitoring
1. Introduction
In flexible manufacturing systems (FMS),
tool wear monitoring plays a critical role in
dictating the dimensional accuracy of the
workpiece and guaranteeing automatic
cutting process. It is therefore essential to
develop simple, reliable and cost-effective
tool wear condition monitoring strategies in
this vitally important area. Various methods
for tool wear monitoring have been
proposed in the past, these methods have
been classified into direct (optical,
radioactive and electrical resistance, etc.)
and indirect (acoustic emission (AE),
spindle motor current, cutting force,
vibration, etc.) sensing methods according to
used sensors [1-2]. Recent attempts have
been concentrated on development of the
2. 2
indirect methods. Among indirect methods,
AE is the most effective mean of sensing
tool wear. The major advantage of using AE
to monitor tool condition is that frequency
range of the AE signal is much higher than
of the machine vibrations and environmental
noises and not interfere with the cutting
operation. But AE signals often have to be
treated with additional signal processing
schemes to extract the most useful features
from signals [3-5]. If AE signal can
effectively be analyzed, tool wear may be
detected using AE signals. Among various
approaches have been taken to analyze AE
signals, spectral analysis has been found to
be the most informative for monitoring tool
wear [6-7]. Spectral analysis such as fast
Fourier transform (FFT) is the most
commonly used signal processing
techniques in tool wear monitoring. A
disadvantage of FFT method is that it has a
good solution only in frequency domain and
a very bad solution in time domain.
Recently, wavelet packet transform
proposed is a significant new tool in signal
analysis and processing. Wavelet transform
has a good solution in frequency and time
domain synchronously can extract more
information in time domain at different
frequency bands. It has been to analyze tool
failure monitoring signal [8-10]. The
wavelet packet transform has been used for
on-line monitoring of machining process. It
can capture important features of the sensor
signal that are sensitive to the change of
process condition (such as tool wear) but is
in sensitive to the variation of process
working condition and various noises [11].
The wavelet packet transform can
decompose sensor signal into different
components in different time windows and
frequency bands, the components, hence,
can be considered as the features of the
original signal.
The objective of this paper is to extract
features from acoustic emission (AE) signal
using wavelet packet transform method. A
wavelet packet transform can decompose
AE signal into different frequency bands in
time domain, the root means square (RMS)
values computed from the decomposed
signal for each frequency band are used as
features. Analyzing above features, the
features that are direct relation to tool wear
are used as final monitoring features. The
experimental results indicate that the
monitoring features had a low sensitivity to
changes of the cutting conditions so that
wavelet packer transform is shown to be an
effective method to extract the features of
the AE signals for tool wear monitoring.
2. Wavelet packet transform
3. 3
Given a time varying signal f (t); wavelet
transforms (WT) consist of computing
coefficient that inner products of the signal
and a family of wavelets, namely
( ) ( )w a b f t t dtf a b( , ) ,
*
= ò ψ
(1)
where
( )ψ ψa b a
t b
at a b R a, ( ) , ,= ∈ ≠−1
0; a
and b are the dilation and translation
parameters, respectively; “*”denotes the
complex conjugation.
When a=2j
, b=k2j
, j, k 0Z, wavelet are in
this case
( )ψ ψj k
j
j
t k, = −
− −
2 22
(2)
The discrete wavelet transform (DWT) is
defined
( ) ( )c f t tj k j k, ,
*
= ò ψ
(3)
where cj,k is defined as wavelet coefficient, it
may be though of as a time frequency map
of the original signal f(t). Here, a multi-
resolution analysis approach is used in
which a discrete scaling function
( )φ φj k
t k
j
j
j, =
− −
2 2 2
2
(4)
set
( ) ( )d f t t dtj k j k, ,
*
= ò φ
(5)
where dj,k is called as scaling coefficients, it
is the sampled version of original signal,
when j =0, it is the sampled version of the
original. Wavelet coefficients cj,k ( j=1,þ,J)
and scaling coefficients dj,k given by
c x n h n kj k
n
j
j
, [ ] [ ]= −å 2
(6)
and
d x n g n kj k
n
j
j
, [ ] [ ]= −å 2
(7)
where x[n] are discrete-time signals, hj[n-
2j
k] is the analysis discrete wavelets, the
discrete equivalents to 2-j/2
symbol 121 f
"Symbol" s 10 (2-j
(t-2j
k)), gj[n-2j
k] are
called scaling sequence. At each resolution
j>0, the scaling coefficients and the wavelet
coefficients
c g n k dj k j k
n
+ = −å1 2, ,[ ]
(8)
d h n k dj k j k
n
+ = −å1 2, ,[ ]
(9)
In fact, it is well known that the structure of
computations in a DWT is exactly an octave
- band filter band[12]. The terms g and h are
high-pass and low-pass filters derived from
the analysis wavelet ψ(t) and the scaling
function φ (t).
Wavelet packets are particular linear
combinations of wavelets. They form bases
4. 4
that retain many of the orthogorality,
smoothness and location properties of their
parent wavelets [13]. The coefficients in the
linear combinations are computed by
recursive algorithm, with the results that
expansions in wavelet packet bases have low
computational complexity.
The discrete wavelet transform can be
rewritten as follows:
[ ] ( ) ( )[ ]
( )[ ] ( ) [ ][ ]
( )[ ] ( )
c f t h t c f t
d f t g t c f t
c f t f t
j j
j j
( ) = ∗
= ∗
=
−
−
1
1
0
(10)
Set
{} ( )
{} ( )
H h k t
G g k t
k
k
⋅ = −
⋅ = −
å
å
2
2
(11)
then equation can been written below
[ ] ( )[ ]{ }
( )[ ] ( )[ ]{ }
c f t H c f t
d f t G c f t
j j
j j
( ) =
=
−
−
1
1
(12)
Clearly, DWT only is the approximation cj-
1[f(t)] but not the detail signal dj-1[f(t)],
Wavelet packet transform do not omit the
detail signal, therefore, wavelet packet
transform is
[ ] ( )[ ]{ } ( )[ ]{ }
( )[ ] ( )[ ]{ } ( )[ ]{ }
c f t H c f t G d f t
d f t G c f t H d f t
j j j
j j j
( ) = +
= +
− −
− −
1 1
1 1
(13)
let Qj
i
(t) is the ith packet on jth resolution,
then, the recursive algorithm can also
compute the wavelet packet transform, and
it is below:
( ) ( )
( ) ( )
( ) ( )
Q t f t
Q t HQ t
Q t GQ t
j
i
j
i
j
i
j
i
0
1
2 1
1
2
1
=
=
=
−
−
−
(14)
Where t=1,2, ..., 2J-i
, i=1,2,...,2j
, j=1,2,..., J,
J=log2N, N is data length.
3. Signal analysis and Features
extraction
3.1. Signal analysis
In monitoring of tool wear, monitoring
signals acoustic emission contains
complicated information on the cutting
processing. To ensure the accuracy and
reliability of monitoring, it is important to
extract the features of the signals that
describe the relationship between tool
condition. From a mathematical point of
view, the features extraction can be
considered as signal compression. Wavelet
packet transform is represented as a
compressed signal method. Therefore, it is
ideal to use the wavelet packets as the
extracted features [14,15]. According to
above pointed, each wavelet packet
transform represents certain information on
5. 5
the signal is a specific time-frequency
window.
Fig. 1 shows a typical cutting process
experiment in boring. The AE signal in time
domain is presented. At the beginning of the
cutting process, signal affected by tool wear
is smaller because the tool is fresh, the
magnitude of the AE is small, and cutting
process is stable. As the tool wear increases
progressing, the magnitudes of the AE have
increased.
a)
b)
c)
Fig.1 The AE signal in a typical tool wear cutting
process, cutting speed: 30m/min, feed rate:
0.2mm/rev, the depth of cut: 0.5mm; work material:
40Cr steel, tool material: high-speed-steel, without
coolant. (a) VB=0.06 mm; (b) VB=0.26 mm; (c)
VB=0.62 mm.
Fig.2 shows the decomposing results of
AE signal for the experiment shown in Fig.1
through the wavelet packet decomposition.
Fig.2 represent the constituent parts of the
AE signal at frequency band [0, 62.5], [62.5,
125], ..., [937.5, 1000] kHz, respectively.
Obviously, these decomposing results of AE
signal not only keep the same features which
are discussed above, but also provide more
information such as the time domains
constituent part of the AE signal at the
frequency band. The mean values of the
constituent parts of the AE at very frequency
band can represented the energy level of the
AE in the frequency band.
a)
b)
6. 6
c)
Fig.2 The composing results of AE by wavelet packet
transformation
3.2 Feature extraction
In tool wear monitoring, the feature
selection and feature number are very
important. The selected features must be
independent and the number of features
must be large enough. For the tool wear
monitoring, the cutting conditions (cutting
speed, feed rate and cutting depth) are also
the features related to wear, when signal
features extracted from AE signal
corresponding to different cutting
conditions, these cutting condition were also
represented by the features. In practice, the
cutting condition was not dependent on
features. So we hope that the selected
features should show a low sensitivity to
change of the cutting conditions, namely,
tool wear monitoring system could be
suitable for a wide range of machining
conditions.
According to discuss above, the RMS of in
each frequency band was used to describe
the features of different tool condition. As
wavelet packet transform processing, the
distribution of the wavelet packet transform
is in disorder [16], the distributions of above
wavelet packet transform results is as
follows in order:
Decomposed
order
n1 n2 n3 n4 n5 n6 n7 n8
Frequency
order
1 2 4 3 7 7 5 6
Decomposed
order
n9 n10 n11 n12 n13 n14 n15 n16
Frequency
order
16 15 13 14 9 10 12 9
But above features all are not sensitive to
tool wear. According to a large mounts of
data analysis, we found that n4, n3 , n7 , n8,
n6, n5 , n13 are sensitive to tool wear. Fig.3
and Fig.4 show two typical examples, above
features are replaced by q1, q2, q3, q4, q5, q6,
q7, respectively, those will be used to
classify tool wear satiates.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.05 0.1 0.3 0.5
q1
q2
q3
q4
q5
q6
q7
wear (mm)
feature value
Fig.3 The relationship between features extracted and
tool wear, cutting speed: 30m/min, feed rate:
0.2mm/rev, the depth of cut: 0.5mm; work material:
40Cr steel, tool material: high-speed-steel, without
coolant
7. 7
0
0.1
0.2
0.3
0.4
0.5
0.05 0.1 0.3 0.6
q1
q2
q3
q4
q5
q6
q7
wear (mm)
feature value
Fig.4 The relationship between features extracted and
tool wear, cutting speed: 40m/min, feed rate:
0.3mm/rev, the depth of cut: 1mm; work material:
40Cr steel, tool material: high-speed-steel, without
coolant
The selected features were summarized as
follows:
q1=RMS of wavelet coefficient in the
frequency band [125, 1875.]KHz
q2=RMS of wavelet coefficient in the
frequency band [187.5, 250]KHz
M M
M
q7=RMS of wavelet coefficient in the
frequency band [500, 562.5]KHz
It is known that RMS of continuous AE is
proportional to vc ap, tool flank wear VB,
but it is independent on feed rate. For the
purpose of elimination of the effects of
cutting conditions on features, divided vcap
into qi ( i=1, 2, ... , 7) and get new qi value,
the new qi value are final monitoring
features.
4. Conclusions
One of the most complex problems for tool
wear condition monitoring system is that of
extracting the signal features and describing
the relationship between the tool wear
condition and the signal features under a
given cutting condition as accurately as
possible. In this paper, a method has been
developed for monitoring tool wear in
boring operations using acoustic emission
information. Several features were derived
from wavelet packet transform, and the
optimal features sensitive to tool flank wear
selected. Moreover, the feature extraction
with wavelet packet transform can be
implemented real time since wavelet packet
transform requires only a small amount of
computation.
References:
1. G.Byrne, D.Dornfeld, I.Inasaki,
G.Ketteler, W.Konig, and R.Teti. 1995
Annals of the CIRP 44, 541-567. Tool
condition monitoring (TCM) -the status
of research and industrial application.
2. Li Dan and J.Mathew. 1990 Int. J. Mach.
Tools Manufact. 30, 579-598. Tool wear
and failure monitoring techniques for
turning-A review.
3. Souquet, P., Gsib, N., Deschamps, M.,
Roget, J., and Tanguy, J. C. 1987 Annals
of the CIRP 36, 57-60. Tool monitoring
8. 8
with acoustic emission -industrial results
and future prospects. .
4. Liang, S. and Dornfeld, D. A.. 1989
Trans. ASME, J. Eng. Ind. 111, 199-204.
Tool wear detection using time series
analysis of acoustic emission.
5. Iwata, K. and Moriwaki, T. 1977 Annals
of the CIRP 25, 21-26. An application of
acoustic emission measurement to in-
process sensing of tool wear.
6. E.Emel and E..Kannatey-Asibu Jr . 1988
ASME, J.Engng Ind. 110,137-145. Tool
failure monitoring in turning by pattern
recognition analysis of AE signal.
7. Li Xiaoli, Yao Yingxue and Yuan
Zhejun. 1997 High Technology Letters.
10, 12-16. On-line tool monitoring in
drilling using fuzzy neural network.
8. IBRAHIM NUR TANSEL ect al. 1995
Int. J. Mach. Tools Manufact. 35,1137-
1147. Detection of tool failure in end
milling with wavelet transformations and
neural networks(WT-NN).
9. N.Kasashima, ect al. 1995 Annals of the
CIRP 44, 483-487. Online failure
detection in face milling using discrete
wavelet transform.
10.I. N. Tansel, C. Mekdeci, O. Rodriguez
and B. Uragun. 1993 Int. J. Mach. Tools
Manufact. 33, 559-575. Monitoring drill
conditions with wavelet based encoding
and neural network.
11.Ya Wu and R.Du. 1996 Mechanical
systems and signal processing 10, 29-53.
Feature extraction and assessment using
wavelet packets for monitoring of
machining process..
12.I. Daubechies. 1990 IEEE trans. on
information theory 36, 961-1005. The
wavelet transform, time-frequency
localization and signal analysis.
13.M. A. Cody. 1992 Dr.Dobb's Journal
April 16-28. The fast wavelet transform.
14.Li Xiaoli, Yao Yingxue and Yuan
Zhejun. 1997 Journal of Harbin Institute
of Technology 20,14-19. Study on tool
condition monitoring using fuzzy neural
network. .
15.Li Xiaoli, Yao Yingxue and Yuan
Zhejun.. 1997 Journal of Intelligent
Manufacturing (printed) On-line tool
condition monitoring using wavelet fuzzy
neural network.
16.M.V.Wickerhauser. !991 Lectures on
wavelet packet algorithms. Department of
mathematics, Washington University.