Get documents about "Identifying Harmonics By Empirical Mode Decomposition For Effective Control Of Active Filters During Electric Vehicle Charging
Document Sample
(IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 9, September 2011
Identifying Harmonics by Empirical Mode
Decomposition for Effective Control of Active
Filters During Electric Vehicle Charging
B. V. Dhananjay and T. Senthil
B.V. Dhananjay, Dr.T.Senthilkumar,
Research Scholar, Professor, Automobile Engineering,
Vinayaka Mission’s University, Bharathidhasan University,
Salem,India Trichirapalli, India
Abstract-This paper provides Hilbert Huang affecting charging are temperature, state of charge,
Transform(HHT) method (an empirical mode plate area, impurities, gassing.
decomposition(EMD)) for identifying the presence of Electric vehicles (EV) will become an
harmonics during electric vehicle battery charging attractive alternative to internal combustion engine
when harmonics are generated into the electric line,
due to switching actions of the power electronics.
vehicles in the event that their range can be
Activation of the active filters based on the difference extended. One way to achieve this in the short term
between load current and fundamental current is to provide a fast charger infrastructure. Such a
measured from the line is done. By using active power structure would provide greater mobility for the EV
filter (APF) injection of the required current to user, since during short stops (<1 hour) the EV
minimize the harmonics is done. As part of batteries could be charged from typically 20 to 80
simulation, the accuracy of the HHT is above 95%. % of nominal charge. This would significantly
By correctly recognizing the harmonics using HHT extend the EV range.Fast charger infrastructure
and injecting the compensating current into the line, cost is high. Chargers adversely affect the grid
the charging time of the battery can be reduced. The
reduction in the charging time also depends on the
power quality due to presence of power electronic
battery condition. loads like diode rectifiers and thyristor bridge
converters in the distribution network that result in
Keywords-Hilbert Huang Transform; active power
voltage distortion and current harmonics, (Akagi
filter. 1996).
High increase of problems in the electric
power distribution networks due to the presence of
I INTRODUCTION harmonics. Loads that use switching control with
semiconductor devices are the main cause. One of
the most important tools for correcting the lack of
The battery is the primary source of electrical electric power quality are the active power filters
energy. It stores chemicals. Two different types of (APF), (Udom et al. 2008). The objective of this
lead in an acid mixture react to produce an work has been proving that back propagation
electrical pressure. This electrochemical reaction neural networks, previously trained with a certain
changes chemical energy into electrical energy. A number of distorted waveforms, are an alternative
battery can be of primary cell, secondary cell, wet to the rest of the techniques used and proposed at
charged, dry charged and low maintenance type. A the present time for controlling the APF's, as the
fully charged battery contains a negative plate of ones based on the use of the Fast Fourier
sponge lead(Pb), a positive plate of lead Transform (FFT). A large number of these control
dioxide(Pbo2) and an electrolyte of sulphuric acid techniques are based on ANN’s, (Pecharanin et al.
(H2So4) and water (H2o). During charging, sulphate 1994).
leaves the plates and combines with hydrogen(H2)
to become sulphuric acid (H2So4). Free oxygen
combines with lead on the positive plate to form II MATERIALS AND METHODS
lead dioxide. Gassing occurs as the battery nears
full charge and hydrogen bubbles out at the
negative plates, oxygen at the positive. Factors A Materials
109 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
(IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 9, September 2011
Figure 1shows a three-phase diagram of M=
(a + b) (1)
an HHT controlled shunt APF. A load current 2
signal iLis acquiredand used by the ANN to obtain
the distortion current waveform as reference signal Where a = Maximum_envelope and b =
for the control of the APF.The power converter Minimum_envelope.
injects the necessary compensation current iLin the 4. Obtain a new signal using the following
power circuit, achieving thus a sinusoidal source equation:
current. h 11 (t) = X(t) − M 11 (t) (2)
Where h11(t) is called first IMF. Subsequent
IMF’s had to be found if there are some
overshoots and undershoots in the IMF. Hence,
the envelope mean differs from the true local
mean and h11(t) becomes asymmetric.
In order to find the additional IMF’s, h11(t)
is taken as the new signal. After nth iteration, we
have:
h1n (t) = h1(n −1) (t) − M1n (t) (3)
Where M1n(t) is the mean envelop after the nth
iteration and h1(n-1)(t) is the difference between
the signal and the mean envelope at the (k-1)th
iteration.
5. Calculate C2F as follows:
C2F1 = IMFn (4)
Figure 1 APF control using HHT
Where IMFn = final IMF obtained
B Methods C2F2 = IMFn + IMF(n −1) (5)
Empirical Mode Decomposition (Huang) and Similarly,
Hilbert Transform
A signal can be analyzed in details for its
frequency, amplitude and phase contents by using C2Fn = IMFn + IMF(n −1) + ....... + IMF1
(6)
EMD followed by HT (Jayasree et al. 2010 and
Stuti et al. 2009), The EMD produces the mono Where C2Fn is the original signal.
components called IMFs from the original signal.
In a given frame of signal, there can be many
IMFs. Each IMF will contain a wave form of 6. Calculate F2C as follows:
different amplitude. Hilbert Transform is applied F2C1 = IMF1 (7)
on an IMF to obtain, IF and IA. It is mandatory that
a signal be symmetric regarding the local zero
mean, and should contain same number of extreme F2C 2 = IMF1 + IMF2 (8)
and zero crossings.
The steps involved in EMD of a signal X(t) with F2C n = IMF1 + IMF2 + ....... + IMFn (9)
harmonics into a set of IMFs are as follows.
1. Identify all local maxima of X(t). Connect the Where F2Cn is the original signal.
points using a cubic spline. The interpolated 7. Hilbert transform is applied for each
curve obtained. The upper line is called the IMF and analytical signal is obtained.
upper envelope (Maximum_envelope). A complex signal is obtained from each
2. Identify all local minima of X(t) connect the
point using a cubic spline.. The lower line is
IMF:
called the lower envelope Analytic(I MF) = real(IMF) + imag(IMF) (10)
(Minimum_envelope) obtained by cubic spline.
3. Compute the average by:
110 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
(IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 9, September 2011
8. Instantaneous frequencies are obtained a simulation are presented where a step load change
from analytical signal using occurs at time 60 ms. One additional resistance is
connected in parallel with the load, increasing the
IF =
0.5 × (angle( − X(t + 1) × conj(X(t − 1))) + π) (11) total load current.
2× π
Figure 4 shows the EMD process. In the
sample harmonics signal considered, only one
9. Instantaneous amplitudes are obtained instantaneous mode function is present. A flat
from the analytical signal using the residue signal is also presented. This plot is only
following for 1000 samples. This will be repeated for the
remaining length of the signal. Figure 5 shows the
IA = real(IMF) 2 + imag(IMF) 2 (12)
extraction of different signals present from the fine
level to coarse level. Similarly, Figure 5 shows the
extraction of different signals present from the
coarse level to fine level. Figure 7 presents the
III EXPERIMENTAL SIMULATION instantaneous frequency present in every sample of
the signal. Figure 8 presents the instantaneous
amplitude present in every sample of the signal.
Figure 9 and Figure 10 presents statistical values of
instantaneous frequencies and instantaneous
amplitudes. Based on the statistical values, the
amount of harmonics will be estimated and
appropriately, required compensating current will
be injected into the line.
V CONCLUSION
A Hilbert Huang Transform method has
been used at the control of a shunt active power
filter. Based on the amount of harmonics
recognition, the APF is activated. By correctly
injecting the compensating current into the line, the
charging time of the battery can be reduced. The
circuit has to be verified with the implementation
of HHT in real time for improved charging of the
EV battery. The reduction in the charging time also
depends on the battery condition.
Figure 2 Power circuit with HHT controlling
active power filter
v=0.1pu,cy=50
The model of Figure2 has been created v=0.1pu,cy=50
using Matlab 10. Different sets of parameters have 400 v=0.1pu,cy=50
v=0.1pu,cy=50
been employed at the power circuit and APF. In
200 v=0.1pu,cy=50
most cases the reference current obtained by the
V o l ta g e (V )
HHT controller was accurate enough to enable the
0
APF to compensate harmonic distortion. If an
elevated content of high order harmonics were
-200
present in the load current, the HHT controller
helps in obtaining reference signal. -400
6
System parameter used: 4 1
Power circuit:Phase voltage = 220 VRMS , 2 0.5
Frequency = 50 Hz , Source resistance = 0.1 Ω, 0 0
Signal patterns
Load resistance = 20 Ω, Load inductance = 30 mH Time(Sec)
APF:Vdc = 500 V, R = 10 Ω, L = 30 mH,
Switching frequency = 40 KHz. Figure 3 Sample plot of signals for harmonics
IV RESULTS AND DISCUSSION
More than 500 different harmonics
waveforms (Figure 3) have been used in HHT
analysis with different load changes.The results of
111 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
(IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 9, September 2011
Empirical Mode Decomposition 50.1
s ig n a l
50.05
F re q u e n c y s p e c tru m
In s ta n ta n e u o s
50
49.95
200
im f1
0 49.9
-200
49.85
0.015 0.02 0.025 0.03 0.035 0.04
Time (sec)
200 Figure 7 Instantaneous frequency
re s .
0 231.5
-200
0.01 0.02 0.03 0.04 0.05 0.06 0.07
Figure 4 Empirical mode decomposition
Am plitude
231
f2c
s ig n a l
230.5
0.015 0.02 0.025 0.03 0.035 0.04
Time (sec)
200 Figure 8 Instantaneous amplitude
f2 c 1
0 150 10
S td o f F re q u e n c y
F re q u e n c y
-200 100
M ean of
5
50
200
s ig n a l
0 0
0 0 1 2 3 0 1 2 3
Imf Imf
-200
150
0.01 0.02 0.03 0.04 0.05 0.06 0.07 Maximum Frequency
1500
Minimum Frequency
F re q u e n c y
F re q u e n c y
100
N o rm o f
Figure 5 Fine to coarse signals 1000
50
500
c2f 0 0
0 1 2 3 0 1 2 3
200 Imf Imf
s ig n a l
0 Figure 9 Statistical values of the instantaneous
frequencies
-200
0.01 0.02 0.03 0.04 0.05 0.06 0.07
10
300
1
A m p litu d e
A m p litu d e
M ean of
200
S td o f
c 2 f1
0 5
100
-1
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
0 0
0 1 2 3 0 1 2 3
500 Imf
Imf
s ig n a l
0
300
4000 Maximum Amplitude
-500
A m p litu d e
A m p litu d e
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 Minimum Amplitude
N o rm o f
200
Figure 6 Coarse to fine signals 2000
100
0 0
0 1 2 3 0 1 2 3
Imf Imf
Figure 10 Statistical values of the instantaneous
amplitudes
112 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
(IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 9, September 2011
REFERENCES
[1] HAkagi., ”New Trends in Active Filters for Power
Conditioning”, IEEE Trans. on Industry Applications, vol.
32, No 6, Dec. 1996, pp 1312-1322.
[2] TJayasree., DDevaraj., RSukanesh.,”Power quality
disturbance classification using Hilbert transform and RBF
networks”, Neurocomputing , Vol 73 Issue 7-9, March,
2010, pp. 1451-1456.
[3] NPecharanin., MSone., HMitsui., “An application of
neural network for harmonic detection in active
filter”,IEEE World Congress on Computational
Intelligence.,IEEE International Conference on Neural
Networks, Vol.6, pp. 3756-3760, 1994.
[4] StutiShukla, S. Mishra, and BhimSingh,”Empirical-Mode
Decomposition With Hilbert Transform for Power-Quality
Assessment”, IEEE transactions on power delivery, Vol.
24, No. 4, October 2009.
[5] Udom. Khruathep, SuittichaiPremrudeepreechacharn,
YuttanaKumsuwan,“Implementation of shunt active power
filter using source voltage and source current detection”,
IEEE, pp.2364-2351, 2008.
113 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
Related docs
Other docs by ijcsiseditor
Digital Images Encryption in Spatial Domain Based on Singular Value Decomposition and Cellular Automata
Views: 0 | Downloads: 0
Agent Behavior in Multiagent Systems: Issues and Challenges in Design, Development and Implementation
Views: 1 | Downloads: 0
Optimizing Cost, Delay, Packet Loss and Network Load in AODV Routing Protocols
Views: 2 | Downloads: 0
