Analysis of Shunt Active Filter With ANN Based Con
Analysis of Shunt Active Filter With ANN Based Con
net/publication/329716659
CITATIONS READS
0 139
2 authors, including:
Munazama Ali
Islamic University of Science and Technology
1 PUBLICATION 0 CITATIONS
SEE PROFILE
All content following this page was uploaded by Munazama Ali on 24 August 2021.
27
International Journal of Computer Applications (0975 – 8887)
Volume 182 – No. 32, December 2018
28
International Journal of Computer Applications (0975 – 8887)
Volume 182 – No. 32, December 2018
(t) = (t) / (t) = cos sin t = sin t(3.2.6) performance to eliminate harmonic components of current or
voltage waveform. Generally, Conventional PI Controller is
Where = cos used to regulate this voltage but due to its limitations, as the
There are also some switching losses in the PWM converter, detection of PI parameters are difficult and time taking
and hence the utility must supply a small capacitor leakage process also other disadvantages like inability to improve the
and converter switching losses in addition to the real power of transient response of the system therefore ANN controller has
the load. The total peak current supplied by the source is been introduced in this research work because of its learning
therefore ability, high speed recognition and able to themselves to any
system. An artificial neural network(ANN), is a mathematical
= + (3.2.7) model inspired by biological neural network. Artificial neural
network consists of an interlinked collection of artificial
If the active filter provide total reactive and harmonic power. neurons and it develops information using a connectionist
At this time, the active filter must provide following method to calculation. it resembles the brain in two factors:
compensation current:
(i) The data is accumulated by the network
= (t) - (t) through the learning process.
Hence, for accurate and instantaneous compensation of (ii) Interneuron connection strengths are used to
reactive and harmonic power it is necessary to estimate, store the data.
i.e.fundamental component of the load current as the reference The large data of the DC-link voltage for ‘n’ and ‘n-1’
current. intervals from the conventional method are gathered and are
stored in the MATLAB workspace. This data is then retrieved
3.3 Control Strategy and used for training the ANN Controller. The neurons in the
In anyactive power filter, control algorithm has major role in
input and output layer is almost a fixed quantity to obtain the
deciding the performance of harmonic compensation. Better
provided input. At each training session, seven hidden layers
control strategy leads to better dynamic response of the
are taken with 1000 iterations are done and six such a
system. In this research work, control strategy is provided by
validation checks are taken out in order to minimise the scope
hysteresis-band controlled technique.Hysteresis-band is
of error occurrence.
basically an instantaneous feedback current control technique
of PWM. The hysteresis band current controller for active The rapid spotting of the disturbing signals with high
power filter is used to generate the switching pattern of accuracy, and fast processing of the reference signals, are the
inverter. The control circuit generates the sine reference prerequisites for desired compensation in caseof ShAPF.
current wave of desired magnitude and frequency and it is
then compared with actual current wave. 5. PERFORMANCE EVALUATION
29
International Journal of Computer Applications (0975 – 8887)
Volume 182 – No. 32, December 2018
(b)
(c)
The shunt active power filter is switched to system at 0.02 Fig6: Three phase balanced (a) source current (b)SAPF
sec. The response of switching before and after can be easily Injected current (c) DC link voltage with PI Controller.
distinguished from the wave form and THD are given in
figures below.
(a)
(b)
(a)
(c)
Fig8: Three phase balanced (a)Source current (b)SAPF
Injected current(c)DC link voltage with Ann Controller.
30
International Journal of Computer Applications (0975 – 8887)
Volume 182 – No. 32, December 2018
[3] MunazamaAli,arminderkaur and Abdul [17] A. E. Emanuel, “Summary of IEEE Standard 1459:
hamidbhat“comparative Analysis of shunt active power definitions for the measurement of electric power
filter with pi and Ann controller”, International journal of quantities under sinusoidal, nonsinusoidal, balanced, or
computer applications(1-6, August) 2016. unbalanced conditions,” IEEE Trans. Ind. Appl., vol. 40,
no. 3, pp. 869–876, May/Jun. 2004.
[4] H. Rudnick, J.Dixon and L.Moran, “Active power
filters as a solution to power quality problems in [18] MetinKesler and EnginOzdemir, “Synchronous-
distribution networks,” IEEE Power and Energy Reference Frame-Based Control Method for SAPFUnder
Magazine. Unbalanced and Distorted Load Conditios”, IEEE
transactions on industrial electronics, VOL. 58, NO. 9,
[5] Elsa Susan Daniel and G.Abirami “Selective Harmonic SEPTEMBER 2011.
Elimination Using Shunt Hybrid Active Power Filters
Operating At Different Switching Frequencies” [19] R Rajalakshmi, Dr V Rajasekaran, “Improvement of
International Journal of Innovative Research In Electrical Energy Efficiency through power quality by
& Electronics, Instrumentation And Control Engineering compensation of harmonics with shunt active power
Vol. 1, Issue 1, April 2013 filter”2011 International Conference advancements in
electrical, electronics and control engineering IEEE.
[6] Chennai Salim and Benachaiba Mohamed Toufik,“
Three-phase three-level (NPC) Shunt Active power Filter [20] JR Vázquez, PatricioSalmeron, F Javieralcantra,
JainePrieto. A new active power line conditioner for
31
International Journal of Computer Applications (0975 – 8887)
Volume 182 – No. 32, December 2018
compensation in unbalanced /distorted electrical power [23] M. Aredes, H. Akagi, E. H. Watanabe, E. V. Salgado, L.
system. 14th PSCC, Sevilla. 2002. F. Encarnação, “Comparisons Between the p–q and p–
q–r Theories in Three- Phase Four- WireSystems,” IEEE
[21] Agelidis, V.G., "Real-Time Power System Phasors and Transactionson Power Electronics, , vol. 24, no.
Harmonics EstimationUsing a New Decoupled 4,924933, April 2009.
Recursive-Least-Squares Technique for
DSPImplementation," IEEE Transactions on Industrial [24] B.Singh,V.Verma, and J.Solanki, “Neural network-based
Electronics, vol.60, no.6,pp.2295,2308, June 2013. selective compensation of current quality problems in
distribution system,” IEEE Trans. Ind. E
[22] Elsa Susan Daniel and G.Abirami “Selective Harmonic
Elimination Using Shunt Hybrid Active Power Filters [25] W. Mack grady, S. Santoso, Understanding power
Operating At Different Switching Frequencies” system harmonics, IEEE power Eng.Rev. 21(November
International Journal of Innovative Research In Electrical (11)) (2001) 8-11. lectron., vol. 54, no. 1, pp. 53–60, Feb.
& Electronics, Instrumentation And Control Engineering 2007.
Vol. 1, Issue 1, April 2013.
IJCATM : www.ijcaonline.org 32