You are on page 1of 8

Available online at www.sciencedirect.

com

ScienceDirect
Availableonline
Available onlineatatwww.sciencedirect.com
www.sciencedirect.com
Energy Procedia 00 (2017) 000–000

ScienceDirect
ScienceDirect
www.elsevier.com/locate/procedia

Energy
EnergyProcedia
Procedia117 (2017) 000–000
00 (2017) 458–465
www.elsevier.com/locate/procedia

1st International Conference on Power Engineering, Computing and CONtrol, PECCON-2017, 2-


4 March 2017, VIT University, Chennai Campus

Photovoltaic based
The 15th DVRSymposium
International with Power Quality
on District Detection
Heating and Cooling using
Wavelet Transform
Assessing the feasibility of using the heat demand-outdoor
temperature function for a long-term
D.Divyalakshmi a
district heat
, N.P.Subramaniam a
demand forecast
a
Department of Electrical and Electronics Engineering, Pondicherry Engineering College, Puducherry – 605010, India
I. Andrića,b,c*, A. Pinaa, P. Ferrãoa, J. Fournierb., B. Lacarrièrec, O. Le Correc
a
IN+ Center for Innovation, Technology and Policy Research - Instituto Superior Técnico, Av. Rovisco Pais 1, 1049-001 Lisbon, Portugal
Abstract
b
Veolia Recherche & Innovation, 291 Avenue Dreyfous Daniel, 78520 Limay, France
c
Département Systèmes Énergétiques et Environnement - IMT
of Atlantique, 4 rueand
Alfred Kastler, 44300 Nantes,
The increased usage of non-linear loads, abrupt switching of heavy loads and line faults have flared up the power
increased usage of non-linear loads, abrupt switching heavy loads line faults have flaredFrance
up the
quality issues to greater extent. This paper deals with a Photovoltaic based Dynamic Voltage Restorer to compensate
any voltage deviations or disturbances that occur in a system caused by such power quality issues. The DC link
capacitor
Abstract
of the DVR is supplied by the PV panel which supplies the load as well, thereby, considerably reducing
the power drawn from the grid. A charge controller and associated DC-DC converter takes care of the charging and
discharging of the batteries as well as supplying the remote DC and AC loads. A switching controller based on
District heating networks are commonly addressed in the literature as one of the most effective solutions for decreasing the
Wavelet Transform is developed to detect any power quality event that occurs and automatically switch the DVR to
greenhouse gas emissions from the building sector. These systems require high investments which are returned through the heat
inject the compensated voltage. With the effect of Wavelet Transform, the switching of DVR is done faster.
sales. Due to the changed climate conditions and building renovation policies, heat demand in the future could decrease,
2017 The Authors. Published by Elsevier Ltd.
©prolonging the investment
© 2017 The Authors. return
Published byperiod.
Elsevier Ltd.
Peer-review under responsibility ofthe scientific committee of the 1st International Conference on Power Engineering,
The main scope
Peer-review underofresponsibility
this paper is of
to assess the feasibility
the scientific of using
committee of thethe
1stheat demand –Conference
International outdoor temperature
on Power function for heat demand
Engineering,
Computing
Computing and CONtrol.
and district
CONtrol.
forecast. The of Alvalade, located in Lisbon (Portugal), was used as a case study. The district is consisted of 665
buildings that vary in both construction period and typology. Three weather scenarios (low, medium, high) and three district
Keywords:Power Quality, Dynamic Voltage Restorer (DVR), Harmonics, Wavelet Transform, Photovoltaic array.
renovation scenarios were developed (shallow, intermediate, deep). To estimate the error, obtained heat demand values were
compared with results from a dynamic heat demand model, previously developed and validated by the authors.
The results showed that when only weather change is considered, the margin of error could be acceptable for some applications
1.(the
Introduction
error in annual demand was lower than 20% for all weather scenarios considered). However, after introducing renovation
scenarios, the error value increased up to 59.5% (depending on the weather and renovation scenarios combination considered).
the present
TheInvalue of slopeday’s progressive
coefficient modern
increased world,within
on average the usage of non-linear
the range of 3.8% up devices
to 8%like
per computers,
decade, that mobile chargers,
corresponds to the
variable
decreasefrequency drives
in the number etc. have
of heating flared
hours up, thereby
of 22-139h duringpolluting
the heatingourseason
electrical systemonwith
(depending harmonics. ofAlso
the combination abrupt
weather and
switching
renovationof heavy considered).
scenarios loads and line faults
On the have
other further
hand, functionincreased
intercept the powerfor
increased quality disturbances
7.8-12.7% per decadelike voltage on
(depending sag,
the
swell,
coupledharmonics, interruption,
scenarios). The flicker,could
values suggested transient
be usedetc. [1]. Such
to modify power parameters
the function quality disturbances might
for the scenarios result and
considered, in
malfunction or damage
improve the accuracy of demand
of heat the sensitive devices or can even cause complete collapse of the system, when not
estimations.
properly rectified. The emerging applications of renewable energy resources like wind, solar and biomass have
© 2017inflated
further the problem
The Authors. ofby
Published harmonics and voltage fluctuations, due to the intermittent nature of such sources and
Elsevier Ltd.
power electronic
Peer-review underconverters associated
responsibility with them
of the Scientific [2]. of The 15th International Symposium on District Heating and
Committee
Cooling.

Keywords: Heat demand; Forecast; Climate change


1876-6102© 2017 The Authors. Published by Elsevier Ltd.
Peer-review under responsibility ofthe scientific committee of the 1st International Conference on Power Engineering, Computing and CONtrol.

1876-6102 © 2017 The Authors. Published by Elsevier Ltd.


Peer-review under responsibility of the Scientific Committee of The 15th International Symposium on District Heating and Cooling.
1876-6102 © 2017 The Authors. Published by Elsevier Ltd.
Peer-review under responsibility of the scientific committee of the 1st International Conference on Power Engineering, Computing
and CONtrol.
10.1016/j.egypro.2017.05.171
D. Divyalakshmi et al. / Energy Procedia 117 (2017) 458–465 459
2 D.Divyalakshmi, N.P.Subramaniam / Energy Procedia 00 (2017) 000–000

So as to protect such critical devices from power quality disturbances, numerous topologies of custom power
devices have been developed as per IEEE 519 standards [3]. One such commonly used device is Dynamic Voltage
Restorer (DVR), to improve the quality of load voltages against any disturbances and isolating the load from any
power quality event occurring in the supply side or lines. It is a series compensation device and injects a series
compensating voltage which is synchronised with the utility gird with equal and opposite magnitude so as to protect
the sensitive loads from any power quality issue and supply clean and harmonic-free power to the consumers.
The PWM pulses for the DVR have been developed using various control schemes in [4]-[7]. Synchronous
Reference frame (SRF) theory is reported in [7]. The rating and design of series injection transformer of the DVR is
presented in [8]. Many research works have been carried out in the detection of voltage disturbances [9,10] This
paper proposes a photovoltaic (PV) array fed three phase Dynamic Voltage Restorer (DVR) for voltage
compensation. Besides voltage regulation, the proposed DVR reduces the energy consumption by distributing solar
power from PV array to local DC and AC loads through efficient converters.

2. Photovoltaic-Dynamic Voltage Restorer (PV-DVR)

The Voltage Source Converter (VSC), injecting transformer and the DC link capacitor or external energy source
constitute the foremost important components of DVR. The output voltage of the VSC is controlled by the gate
pulses produced by a controller based on various control theories. In this work, Synchronous Reference Frame
(SRF) theory along with hysteresis control, is used for production of PWM pulses. The system voltage, load voltage
and reference voltage values are utilised to accordingly produce gate pulses, which further controls the series
injected voltage for compensation. In general, DC link capacitor of DVR can be of two types: i. Battery-supported,
and ii. Self-supported (capacitor supported) [11]. In a self-supported or capacitor supported DVR, the compensating
voltage and line current are in quadrature with each other and does not draw active power from the lines. But the
voltage before the power quality event and during the event (i.e) compensated voltage and the supply voltage are not
in phase with each other, which is a drawback of this self-supported DVR. Hence, self-supported DVR is chosen
only when this phase difference is acceptable. This paper deals with a PV-DVR, in which a battery fed by a PV
source feeds the DC link capacitance of DVR.

2.1. Principle and Operation of DVR

Fig. 1. (a) Single line diagram of DVR; (b) Phasor Diagram of DVR; (c) Equivalent circuit of DVR.

Fig 1(a) and (c) shows the circuit diagram and equivalent circuit of the DVR. The DVR injects a series voltage Vc
with the source voltage VS to obtain a fluctuation and disturbance free load voltage VL. Thus, the load voltage of
DVR can be written as,
460 D. Divyalakshmi et al. / Energy Procedia 117 (2017) 458–465
D.Divyalakshmi, N.P.Subramaniam / Energy Procedia 00 (2017) 000–000 3

VL  VS  VC (1)
Here, VC  VDVR  V presag  Vsag (2)
Where, Vpresag is the voltage prior to sag, and Vsag is the voltage during the occurrence of sag event. The angle of
the compensated voltage can be given as,
VC   C   S (3)
Where, ӨS is the angle of the supply voltage.
Fig.1 (b) shows the phasor diagram of different compensating voltages of the DVR[11]. The compensated
voltage can be analysed in four different ways depending on the phase angle of the load voltage. Here, Vload1, Vload2,
Vload3 and Vload4 represents the load voltages injected by the DVR for different compensating voltages Vc1,Vc2, Vc3,
Vc4respectively. Vc1 denotes the compensating voltage for a load Vload1and is injected in phase with the supply
voltage. Vc2 represents the compensating voltage for the same load voltage but is injected with a small leading phase
angle with the supply voltage. Similarly, Vc3 represents the compensating voltage for the same load voltage but is
injected with another phase angle with the supply voltage. However, under this condition, the phase of the load
voltage remains the same as the load voltage during pre-sag condition. This phase angle of compensating voltage Vc3
is considered to be the optimum phase angle for a DVR with an energy source or storage device at the DC side. The
compensating voltage Vc4 for the same load voltage is in quadrature with the line current IL. Under this condition, no
active power is injected, and hence, this condition is highly suitable for self-supported (capacitor supported) DVR.
But however, small losses occur during this condition. So as to minimise such losses and involve active power
injection, DC link with an energy source is preferred. In this paper, a PV array is chosen as the DC link energy
source. However, when the compensating voltage is injected in phase with supply voltage, the voltage rating of
VSC is minimum.

2.2. Photovoltaic- Dynamic Voltage Restorer(PV-DVR)

Fig. 2. Block diagram of the PV-DVR


D. Divyalakshmi et al. / Energy Procedia 117 (2017) 458–465 461
4 D.Divyalakshmi, N.P.Subramaniam / Energy Procedia 00 (2017) 000–000

In PV-DVR, the DC link capacitor is charged and discharged by PV panel through DC-DC converter, which
further feeds local AC and DC loads. The battery bank operated during back-up, is controlled by a charge controller.
the semiconductor switch controller is operated based on Wavelet Transform to detect the beginning and end of any
power quality event and the switches are operated accordingly. The bye pass switch across the injecting transformer
is closed by default unless a power quality disturbance is detected by the switching controller and opens the switch,
bringing in the DVR into connection for voltage compensation.

3. Wavelet based Switching Controller for Power Quality Detection

Unlike Fourier Transform, which analyses a signal only in frequency domain, Wavelet Transform analyses a
signal in time-frequency domain, which greatly helps in time localization of any fault or any power quality event
that occurs in the system. In low frequency part, wavelet transform has low time resolution and high frequency
resolution. Whereas, in the high frequency part, it has high time resolution and low frequency resolution. Wavelet
Transform is characterized by its three main advantages:
 Simultaneous localization of any power quality event both in time and frequency domain.
 It is computationally very fast.
 Capable of separating even the fine details in a signal.

Fig. 3. Schematic diagram of Wavelet Transform Fig.4. Flowchart of Wavelet Based Switching Controller

This paper deals with a Switching Controller based on Wavelet Transform for detecting the beginning and ending
of any power quality event. The signal subjected to Wavelet Transform provides two types of coefficients as output
namely, Approximate coefficients (Ca) and Detail Coefficients (Cd). The approximate coefficients are the high-
scale, low-frequency components of the signal and the detail coefficients are the low-scale, high-frequency
components, obtained by passing the signal through Low pass and High pass filters respectively.
462 D. Divyalakshmi et al. / Energy Procedia 117 (2017) 458–465
D.Divyalakshmi, N.P.Subramaniam / Energy Procedia 00 (2017) 000–000 5

The discrete wavelet decomposition of a digitized time signal Co(n)is decomposed into approximate Ca(n) and
detail coefficients Cd(n) as expressed by,
Ca (n)   h(k  2n)Co (k ) (4)
k

Cd (n)   g (k  2n)Co (k ) (5)


k
where, h(n) denotes the response of a low-pass filter and g(n) represents the response of a high-pass filter.
The wavelet decomposition of the next level is dependent on the approximate coefficients of the previous level
and not on Co(n) (i.e.) the approximate coefficients of the present level acts as the base signal for next level
decomposition. The number of approximate and detail coefficients obtained is half of the decomposed signal and
hence the coefficients must be up-sampled at each level to maintain the number of samples constant.
Fig.4 explains the operational flow of the wavelet based switching controller. Energy, mean and Standard
Deviation of the signal coefficients are obtained using the following formulae:
N
Energy, E   (C d (n) 2 ) (6)
n 1

1 N
Mean,  
N
 (C
n 1
d ( n) 2 ) (7)

1 N
S .D, 
N
 (C
n 1
d ( n)   ) 2 (8)

Where, N is the number of samples and Cd(n) is the detail coefficient value. When the values of Energy and
Standard Deviation is greater than the threshold value, a power quality disturbance is detected and the switching
controller takes necessary steps to operate the DVR from the PV or battery source based on the availability. The
above mentioned Switching Controller uses these Detail coefficients to localize various power quality events like
voltage sag, voltage swell, flicker, transient, interruption etc.

4. Simulation Results

The dynamic performance of PV-DVR under various power quality events and disturbances is simulated using
MATLAB Simulink and the results are discussed below.The simulated system parameters are shown in Table 1.

Table 1.System Parameters.


System Parameters Values
System voltage 415V(VL-L), 50/Hz
KVA Rating of DVR 14.28kVA
DC capacitor voltage 150V
DC bus capacitance 3000µF
Interfacing inductor of DVR 0.946mH
Injecting transformer ratio 1:1
Ripple filter 10 µF
Power of PV panel 20kW
Battery bank 2000Ah
D. Divyalakshmi et al. / Energy Procedia 117 (2017) 458–465 463
6 D.Divyalakshmi, N.P.Subramaniam / Energy Procedia 00 (2017) 000–000

4.1. Case-1: Voltage Sag and Voltage swell:

A voltage sag is introduced in the system as a fault occurs between 0.1 and 0.2 sec and a voltage swell is caused
between 0.3 and 0.4 sec by sudden switching off of a heavy load. The Fig. 5 shows the System voltage (Vs),
Wavelet Coefficients (Cd), Load voltage (V L), compensated voltage (Vc) for voltage sag occurring between 0.1 and
0.2 sec and voltage swell occurring between the time 0.3 and 0.4 sec.

Fig. 5. System voltage (Vs), Wavelet Coefficients (Cd),Load voltage (V L), compensated voltage (Vc) for Voltage sag between 0.1 and 0.2 sec and
voltage swell between 0.3 and 0.4 sec
The load voltage VL is seen with constant voltage magnitude after compensation of the DVR against voltage sag
and voltage swell. It can be seen that the magnitude of compensating voltage is more for sag and is injected in
phase with supply voltage. Whereas, for swell its magnitude is less and is injected with a phase angle so as to reduce
the swell magnitude. The wavelet coefficients exactly detect the beginning and end of the event: voltage sag
between 0.1 and 0.2 sec and voltage swell between 0.3 and 0.4 sec. The detection of power quality event by this
controller exactly switches the DVR instantly for compensation. The DVR voltage is zero other than compensation
time as the switch controller maintains the DVR in bye pass condition. The THD of load voltage is observed to be
0.84% in this case.

4.2. Case-2: Interruption:

The Fig. 6 shows the System voltage (Vs), Wavelet Coefficients (Cd), Load voltage (V L), compensated voltage
(Vc) for interruption occurring between 0.2 and 0.35 sec.

Fig. 6. System voltage (Vs), Wavelet Coefficients (Cd),Load voltage (V L), compensated voltage (Vc) for interruption between 0.2 and 0.35 sec
464 D.Divyalakshmi, N.P.Subramaniam
D. Divyalakshmi et al.//Energy
Energy Procedia
Procedia 00
117(2017)
(2017)000–000
458–465 7

Just, as in the previous case, the interruption was exactly detected by the wavelet based controller between 0.2
and 0.35 sec and was able to operate the DVR appropriately. The THD of the Load voltage is observed to be 0.81%
in this case.

4.3. Case-3: Sag with harmonics and Swell with harmonics:

A non-linear harmonic load is introduced in the system and analysed for voltage sag and swell in this case. The
Fig. 7 shows the System voltage (Vs), Wavelet Coefficients (Cd), Load voltage (V L), compensated voltage (Vc) for a
harmonic voltage with voltage sag between 0.1 and 0.2 sec and voltage swell between 0.3 and 0.4sec. The wavelet
coefficients exactly detect the beginning and end of the event.

Fig. 7. System voltage (Vs), Wavelet Coefficients (Cd),Load voltage (VL), compensated voltage (Vc) for non-linear harmonic load with voltage
sag between 0.1 and 0.2 sec and voltage swell between 0.3 and 0.4 sec

The load voltage VL is seen with constant harmonic free voltage after compensation of the DVR against voltage
sag and voltage swell. The DVR was able to detect and compensate both sag or swell and harmonics
simultaneously. The wavelet coefficients exactly detect the harmonics at 0 and 0.5 sec and operates the DVR for the
entire period. In this case, the THD of uncompensated supply voltage is observed to be 32.76%, and the THD of
load voltage after compensation is observed to be 3.53%.

Table 2. Comparison of wavelet detection and actual occurrence of the PQ event

Simulation result for switching of Actual occurrence of the Delay in switching of


WT based DVR (ms) PQ event (ms) DVR (ms)
Start End Start End Start End
Sag 102.1 201.6 100 200 2.1 1.6
Swell 300.9 401.2 300 400 0. 9 1.2
Interruption 201.8 351.3 200 350 1.8 1.3
Harmonics 0.800 495.5 0 495 0.8 0.5
Sag with harmonics 102.8 202.0 100 200 2.8 2.0
Swell with harmonics 301.6 401.3 300 400 1.6 1.3
D. Divyalakshmi et al. / Energy Procedia 117 (2017) 458–465 465
8 D.Divyalakshmi, N.P.Subramaniam / Energy Procedia 00 (2017) 000–000

Unlike, conventional Overcurrent relays which take typically 0.1-1 sec for operating [12], this Wavelet based
switching controller operates in a faster rate as shown in the following table 2. Also, the conventional overcurrent
relays depend on the magnitude pick up current for their faster and effective operation, which is not the case in
wavelet based controller.

Fig. 8. Percentage of error for various PQ events

Fig. 8 shows the percentage of error for various PQ events. It is observed that the DVR responds faster to swell
than sag, as the rate of change of voltage is high for swells. Similarly, the DVR responds to interruption faster than
sag as well. This error can be further decreased by optimised choice of dynamic threshold values according to the
type of PQ event occurring, rather than a fixed value.

5. Conclusion

A Dynamic Voltage Restorer based on Photovoltaic source with automatic detection of power quality event
using Wavelet Transform is proposed. The PV panel supplies the load when solar power is available and charges the
battery in absence of solar power. Also the PV source plays the twin role of supplying the loads as well as the DVR,
in times of operation. The wavelet transform based control algorithm exactly detects the beginning and ending of
any power quality event and based on this detection, the switching controller automatically switches the DVR. The
charge controller is used to appropriately charge and discharge the battery. With the effect of Wavelet Transform,
the switching of breakers is done at a faster rate and by appropriate switching of DVR, energy can be considerably
saved, as the injecting transformers are not energised all the time. Compared to conventional overcurrent relays, this
wavelet based switching controller is found to more faster in operation and has effective control over compensation.
References
[1]. N.G Hingorani, “Introducing Custom Power”, IEEE Spectrum, June, pp: 41-48, 1995
[2]. L. Hadjidemetriou, E. Kyriakides and F. Blaabjerg, "A Robust Synchronization to Enhance the Power Quality of Renewable Energy
Systems," in IEEE Transactions on Industrial Electronics, vol. 62, no. 8, pp. 4858-4868, Aug. 2015.
[3] A.M.Sharaf, S Abu-Azab “Power Quality Enhancement of Time Dependent Interharmonic Loads “Proceedings of the Ninth International
IEEE Conference on Harmonics and Quality of Power ICHPS‟2000, Orlando, FL, October 2000.
[4]. P. Jayaprakash, B. Singh, D. P. Kothari, A. Chandra, and K. Al-Haddad, “Control of reduced rating dynamic voltage restorer with battery
energy storage system,” in Proc. Power Syst. Technol. IEEE POWERCON, Oct. 12–15, 2008, pp. 1–8.
[5] B. Singh, P. Jayaprakash, and D. P. Kothari, “Adaline based control of capacitor supported DVR for distribution systems,” J. Power
Electron., vol. 9, no. 3, pp. 386–395, May 2009.
[6] S. Jothibasu and M. K. Mishra, "A Control Scheme for Storageless DVR Based on Characterization of Voltage Sags," in IEEE Transactions
on Power Delivery, vol. 29, no. 5, pp. 2261-2269, Oct. 2014.
[7] ].„Power Quality problems and mitigation techniques‟ by Bhim Singh, Ambrish Chandra and Kamal-Al-Haddad, Wiley Publications.
[8] N. Mohan, T. M. Undeland, andW. P. Robbins, : “Power Electronics: Converter Applications and Design”. New York: Wiley, 1999.
[9] B. Bae, J. Jeong, J. Lee and B. Han, "Novel Sag Detection Method for Line-Interactive Dynamic Voltage Restorer," in IEEE Transactions on
Power Delivery, vol. 25, no. 2, pp. 1210-1211, April 2010.
[10] A. Teke, K. Bayindir and M. Tumay, "Fast sag/swell detection method for fuzzy logic controlled dynamic voltage restorer," in IET
Generation, Transmission & Distribution, vol. 4, no. 1, pp. 1-12, January 2010.
[11] Fitzer C, Barnes M, Green P. Voltage sag detection technique for a dynamic voltage restorer. IEEE Trans Ind Appl 2004;40(1):203–12.
[12] D. S. Alkaran, M. R. Vatani, M. J. Sanjari, G. B. Gharehpetian and A. H. Yatim, "Overcurrent Relays Coordination in Interconnected
Networks Using Accurate Analytical Method and Based on Determination of Fault Critical Point,"in IEEE Transactions on Power Delivery,
vol.30,no.2,pp. 870-877, April 2015.

You might also like