Scribd
Upload
39 views11 pages

Yu-Cheng Chiu (邱昱誠) 2019/01/02: 力學與機電能源實驗室 Mechanics and Electromechanical energy La

Uploaded by

邱昱誠

This document discusses an algorithm for signal superposition. It involves applying FFT to an input signal to obtain the frequency spectrum, then determining the phase speed and delay for each frequency using dispersion curves. The phase of each frequency component is delayed accordingly, and an inverse FFT transforms the modified spectrum back to the time domain to produce the output signal. Experimental results using this method on a hanning window input show the output is 4 cycles instead of the expected 5 cycles. Future work is discussed to better understand if the output can be viewed as the superposition of the input signal with different spectral contents and group velocities, or the superposition of individual cosine/sine waves.

Copyright:

© All Rights Reserved
Download as PPTX, PDF, TXT or read online from Scribd
Download as pptx, pdf, or txt

Yu-Cheng Chiu (邱昱誠) 2019/01/02: 力學與機電能源實驗室 Mechanics and Electromechanical energy La

Uploaded by

邱昱誠
39 views11 pages
This document discusses an algorithm for signal superposition. It involves applying FFT to an input signal to obtain the frequency spectrum, then determining the phase speed and delay for each frequency using dispersion curves. The phase of each frequency component is delayed accordingly, and an inverse FFT transforms the modified spectrum back to the time domain to produce the output signal. Experimental results using this method on a hanning window input show the output is 4 cycles instead of the expected 5 cycles. Future work is discussed to better understand if the output can be viewed as the superposition of the input signal with different spectral contents and group velocities, or the superposition of individual cosine/sine waves.

Original Description:

22

Original Title

19.01.02

Copyright

© © All Rights Reserved

Available Formats

PPTX, PDF, TXT or read online from Scribd

Did you find this document useful?

Is this content inappropriate?

This document discusses an algorithm for signal superposition. It involves applying FFT to an input signal to obtain the frequency spectrum, then determining the phase speed and delay for each frequency using dispersion curves. The phase of each frequency component is delayed accordingly, and an inverse FFT transforms the modified spectrum back to the time domain to produce the output signal. Experimental results using this method on a hanning window input show the output is 4 cycles instead of the expected 5 cycles. Future work is discussed to better understand if the output can be viewed as the superposition of the input signal with different spectral contents and group velocities, or the superposition of individual cosine/sine waves.

Copyright:

© All Rights Reserved
Download as PPTX, PDF, TXT or read online from Scribd
Download as pptx, pdf, or txt
39 views11 pages

Yu-Cheng Chiu (邱昱誠) 2019/01/02: 力學與機電能源實驗室 Mechanics and Electromechanical energy La

Uploaded by

邱昱誠
This document discusses an algorithm for signal superposition. It involves applying FFT to an input signal to obtain the frequency spectrum, then determining the phase speed and delay for each frequency using dispersion curves. The phase of each frequency component is delayed accordingly, and an inverse FFT transforms the modified spectrum back to the time domain to produce the output signal. Experimental results using this method on a hanning window input show the output is 4 cycles instead of the expected 5 cycles. Future work is discussed to better understand if the output can be viewed as the superposition of the input signal with different spectral contents and group velocities, or the superposition of individual cosine/sine waves.

Copyright:

© All Rights Reserved
Download as PPTX, PDF, TXT or read online from Scribd
Download as pptx, pdf, or txt
You are on page 1/ 11

Yu-Cheng Chiu(邱昱誠)

2019/01/02

力學與機電能源實驗室
Mechanics and Electromechanical energy Lab
Review from last discussion
𝟏 𝝎𝒕
 Hanning window function: 𝑨 = (𝟏 − 𝒄𝒐𝒔 )𝒔𝒊𝒏 𝝎𝒕 , where
𝟐 𝟓
w = central(excited) frequency
 We can identify the frequency content by applying FFT
 The received signal can be viewed as the superposition of same
input signal with different group velocity and amplitude

力學與機電能源實驗室
Mechanics and Electromechanical energy Lab
1
Algorithm of signal superposition
 Step 1: Apply FFT on input signal (Hanning window,
f(t))
 Step 2: Find the two frequencies (f1, f2) with 20 dB
lower energy than central frequency
 Step 3: Find the energy content (f(w)) and group
velocity (Vg) of all frequency within f1 & f2
 Step 4: Shift the input signal then be multiplied by
spectral content
 Step 5: Superpose all signal

力學與機電能源實驗室
Mechanics and Electromechanical energy Lab
2
Algorithm of signal superposition
• The distribution of group velocity below 350 kHz is denser than
that beyond 350 kHz
• This sometimes will cause an-symmetric signal in time domain

力學與機電能源實驗室
Mechanics and Electromechanical energy Lab
3
Algorithm of signal superposition

• The top figure show

……
displacement results in
different frequency due
central frequency

+
to phase shift and

……
spectral content
+

• The bottom figure


show displacement
results by superposing
all signal
=

力學與機電能源實驗室
Mechanics and Electromechanical energy Lab
4
Result & discussion
• With the input of 5 cycle hanning window, the simulated
result is 4 cycle signal while experimental one is over 5
cycle

力學與機電能源實驗室
Mechanics and Electromechanical energy Lab
5
Result & discussion
 Assumption from paper (M. Lowe):
1. The start of received signal – start of input signal propagate at
Vmax
2. The end of received signal – end of input signal propagate at Vmin
 This assumption seems to have relationship with superposition
theory initially

力學與機電能源實驗室
Mechanics and Electromechanical energy Lab
6
Future work
 Discussion:
Whether the received signal can be viewed as
1. superposition of input signal(hanning window) with
different spectral content and group velocity when giving
a hanning window function?
2. superposition of pure cosine(sine) wave with different
spectral content and phase velocity?(Inverse FFT)
time wave number frequency

propagation distance

FFT of input signal

力學與機電能源實驗室
Mechanics and Electromechanical energy Lab
7
Algorithm of signal superposition(suggested by M. Lowe)
 The calculation is done by 4 steps below:

Step1. The initial signal is transformed by FFT to the frequency domain.


This is a frequency spectrum of complex values (amplitude and phase).
Step2. For each frequency in the spectrum, the phase speed is found
from the dispersion curve
Step3. For each frequency, the phase delay from the source to the
receiver is calculated, according to the phase speed. The phase of the
spectrum at this frequency is delayed by this value of phase delay.
Step4. The modified spectrum is transformed back by inverse FFT to
the time domain.
力學與機電能源實驗室
Mechanics and Electromechanical energy Lab
8
Thanks for your listening

力學與機電能源實驗室
Mechanics and Electromechanical energy Lab
Gant chart
Date 107 108
Work Oct Jan Feb Mar Apr May Jun
完成10件A構型實驗與討論
剩下10件A構型之實驗（含EDM）
B構型實驗
108年第1季報告
108年期中報告
POD model 之研究與應用
論文撰寫與修訂

 The tool of POD model is R language


 The number of Model A with EDM is 6 piece

力學與機電能源實驗室
Mechanics and Electromechanical energy Lab
10

You might also like