Solar Energy 153 (2017) 262–275

Contents lists available at ScienceDirect

Solar Energy
journal homepage: www.elsevier.com/locate/solener

Modeling and analysis of a smart grid monitoring system for renewable

energy sources
Yasin Kabalci a,⇑, Ersan Kabalci b
a
Department of Electrical and Electronics Engineering, Faculty of Engineering, Omer Halisdemir University, 51240 Nigde, Turkey
b
Department of Electrical and Electronics Engineering, Faculty of Engineering and Architecture, Nevsehir HBV University, 50300 Nevsehir, Turkey

a r t i c l e i n f o a b s t r a c t

Article history: The observing and metering processes are necessary in renewable energy conversion systems as applied
Received 21 November 2016 in smart grid applications of conventional grid. In this study, the requirements of the renewable energy
Received in revised form 15 May 2017 sources examined with a solar microgrid model that is developed via Matlab/Simulink. The dc-ac conver-
Accepted 20 May 2017
sion system utilized in this paper contains three solar power plants with maximum power point tracking
(MPPT) system and a multilevel inverter to create three-phase ac line voltages. The transmission line is
modeled at the output of the inverter with a length of 25 km by using real line parameters. The infras-
Keywords:
tructure of power line communication (PLC) is managed by binary phase shift keying (BPSK) modems that
Renewable energy
Smart grid
are located in different places. The usage of the proposed energy monitoring technique eliminates addi-
Power line communication tional monitoring costs due to the fact that the power lines are not only exploited to carry the generated
Microgrid voltage, but also utilized to convey the drawn power rate of loads at the back-end of the microgrid.
Remote monitoring Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction 220 V/50 Hz (Mannah et al., 2011; Kosonen and Ahola, 2010; Son
et al., 2010; Tsuzuki, 2006; Lin et al., 2009).
The microgrid is defined as a small autonomous system that is The most studied area of the renewable energy sources is the
constituted with combining several distributed energy sources use of the PVs since generating energy via solar energy sources
such as solar, wind, fuel cells, biomass, and so on together. The dis- offers several advantages such as precluding air pollution, sound-
tributed generation (DG) systems based on renewable energy less operation in consequence of the motionless, and reduced
sources demonstrate a rapid emerging. Small scale power genera- maintenance costs. Although the PVs are frequently assumed as a
tors in the level of medium and low-voltage power systems can be costly process to generate electrical energy, the most appropriate
established by the DG systems. Therefore, smart grid (SG) systems solution to supply the required energy is the usage of stand-
are important parts of the DG systems. The SG applications include alone PV applications. In grid connected systems, voltage source
two main infrastructures that are smart transmission grid (STG) inverters (VSI) capable of pulse-width modulation (PWM) are gen-
and smart distribution grid (SDG). The emerging smart grid sys- erally employed to connect renewable energy sources and the grid.
tems require sensing of data from all the sensors located on the In addition, the current control is an important feature of this type
system within a few power cycles (Kabalci et al., 2012; Ginot converter to supply high quality power to the utility grid. When
et al., 2010; Kurohane et al., 2010; Zhang et al., 2010). The power the cost of the grid connected systems is considered, the cost can
line communication (PLC) term is exploited to describe transmit- be reduced by employing fewer power conversion stages and com-
ting and receiving process of any type data over the conventional ponents (Kabalci and Kabalci, 2010; Kabalci, 2013; Kabalci, 2015;
transmission lines. This notion also defines the main framework Ma et al., 2013).
of the smart grid. The usage of power lines as a transmission med- Several remote monitoring methods in terms of energy moni-
ium prevents the supplementary establishment costs to get the toring, weather monitoring and fault detection systems have been
systems communicated together. The PLC applications allow data proposed for both traditional and renewable energy sources
rates of transmission as much as 200 Mbps using single phase with (Wilkinson et al., 2014; Almas et al., 2014; Ahmed et al., 2016;
Padilla et al., 2014; Silvestre et al., 2013; Kamel et al., 2015;
Vanfretti et al., 2016; Senthilnathan and Annapoorani, 2016;
⇑ Corresponding author. Venkatraman et al., 2016; Silva et al., 2017; Tina and Grasso,
E-mail addresses: [email protected] (Y. Kabalci), [email protected] 2014; Shariff et al., 2013; Gaurav et al., 2014; Kabalci et al.,
(E. Kabalci).

http://dx.doi.org/10.1016/j.solener.2017.05.063
0038-092X/Ó 2017 Elsevier Ltd. All rights reserved.
 Y. Kabalci, E. Kabalci / Solar Energy 153 (2017) 262–275 263

2016; Fateh et al., 2013; Madueño et al., 2016; Cabanas et al., 2007; The conducted literature review is obviously showed that the
Mohamed et al., 2014; Shariff et al., 2015; Tung et al., 2014; Chi remote monitoring systems are generally based on either wired
et al., 2016; Collotta and Pau, 2015; Le et al., 2016; Fabrizio communication systems utilizing fiber and Ethernet or wireless
et al., 2017; Harid et al., 2016; Atalik et al., 2014; Zaker et al., communication systems using GSM, ZigBee and Bluetooth. Even
2014) so far. When the studies related to the remote monitoring though energy monitoring or remote monitoring systems can be
are examined, it is clearly seen that these systems can be classified accomplished by employing these mentioned different communi-
as wired, wireless and hybrid remote monitoring systems. Remote cation protocols, these energy-monitoring methods come up with
monitoring systems utilizing wired communication infrastructure a significant amount of additional cost. On the other hand, there
cover Supervisory Control and Data Acquisition (SCADA) is a significant gap in the literature about usage of power lines as
(Wilkinson et al., 2014; Almas et al., 2014) and Ethernet-based sys- a communication medium both for remote monitoring and for
tems (Ahmed et al., 2016; Padilla et al., 2014; Silvestre et al., 2013; AMI applications. In this study, a new method to eliminate this
Kamel et al., 2015; Vanfretti et al., 2016; Senthilnathan and expense is proposed in which the conventional electrical power
Annapoorani, 2016; Venkatraman et al., 2016; Silva et al., 2017; lines are not only exploited to transmit the generated electricity,
Tina and Grasso, 2014). Wireless remote monitoring systems but also exploited to carry the measurements of solar plants
widely utilize GSM (GPRS, 3G or LTE), ZigBee or Bluetooth systems located in different places. Therefore, the main aim of this study
(Shariff et al., 2013; Gaurav et al., 2014; Kabalci et al., 2016; Fateh is to contribute at this point to the literature by considering power
et al., 2013; Madueño et al., 2016; Cabanas et al., 2007; Mohamed lines as an alternative communication medium for remote moni-
et al., 2014; Shariff et al., 2015; Tung et al., 2014; Chi et al., 2016; toring applications of the smart grids. In addition, this proposed
Collotta and Pau, 2015; Le et al., 2016; Fabrizio et al., 2017; Harid method provides several significant advantages such as eliminat-
et al., 2016). In addition, few systems associated with wired sys- ing channel installation cost, ability to use existing power lines
tems and wireless communication systems, for instance combining more efficiently and the possibility of different channel environ-
fiber or Ethernet systems with wireless methods exist ments that do not require obtaining license and possibility to cre-
(Venkatraman et al., 2016; Atalik et al., 2014; Zaker et al., 2014). ate smart grid infrastructure.
While a remote monitoring system based on SCADA for observing The outline of this paper is as follows: Section 2 describes the
status of the various parameters of wind turbines are investigated designed distributed generation and energy conversion stage, and
in Wilkinson et al. (2014), a phase measurement unit (PMU) for a the designed PLC modems are presented in Section 3. The measure-
laboratory is considered by open source SCADA system in Almas ment and analysis results of the proposed system are given in Sec-
et al. (2014). Ahmed et al. (2016) investigated an Ethernet based tion 4, and finally, conclusions are drawn in Section 5.
optical communication system for small scale wind turbine farms.
In another study, in Padilla et al. (2014), authors aimed to provide a
2. Generation and energy conversion stages
remote monitoring, control and protection system for electrical
power systems. In addition, several remote monitoring and fault
The modeled microgrid structure comprised by generation,
detection systems based on Ethernet communication protocol are
energy conversion, transmission and monitoring stages is illus-
considered in Kamel et al. (2015), Vanfretti et al. (2016), Tina
trated in Fig. 1. It is clear that the structure of the microgrid is asso-
and Grasso (2014).
ciated with microgrid loads at the end of the transmission line.
On the other hand, a large part of wireless communication sys-
Three separate solar farms placed in different locations are taken
tems exploited for remote monitoring are concentrated on using of
into account in the electricity generation part of this study. The
the GSM protocols (Shariff et al., 2013; Gaurav et al., 2014; Kabalci
modeled solar plants cover solar arrays and strings of 150 solar
et al., 2016; Fateh et al., 2013; Madueño et al., 2016; Cabanas et al.,
panels at each solar plant that are controlled with regular Perturb
2007; Mohamed et al., 2014). A GSM based remote monitoring sys-
and Observe (P&O) maximum power point tracking (MPPT) algo-
tem by employing several sensor structures is proposed for photo-
rithm. The output voltage of each solar plant are regulated by buck
voltaic energy generation systems in Shariff et al. (2013). A
converters and supplied to five-level diode clamped multilevel
different GSM system is examined for both solar and wind energy
inverter (MLI) that is controlled with PI regulator and phase dispo-
generation systems in Gaurav et al. (2014). Kabalci et al. (2016)
sition sinusoidal pulse width modulation (PD-SPWM) algorithm.
proposed a remote monitoring system for solar irrigation systems
The sections of generation and energy conversion stages are intro-
thanks to GSM and internet based communication. Other GSM
duced in the following subsections.
based works (Fateh et al., 2013; Madueño et al., 2016; Cabanas
et al., 2007; Mohamed et al., 2014) are focused on for monitoring
of transmission lines and electrical power systems in terms of leak- 2.1. Solar plants and DC conversion
age, fault detection and partial discharge. The most widely utilized
wireless communication method rather than the GSM based sys- The voltage and current parameters of the modeled solar panels
tems is ZigBee communication systems. A wireless remote moni- are adjusted regarding to NE-170UC1 Sharp PV panel that provides
toring system based on ZigBee communication protocol to maximum power out around 170 W (NE-170UC1 Multipurpose
observe grid connected PV systems is reported in Shariff et al. Module, 2008). The voltage-current (V-I) and voltage-power (V-P)
(2015). While Tung et al. (2014) proposed an AMI system for analysis of the modeled PV module is shown in Fig. 2 where irradi-
high-rise buildings, Chi et al. (2016) reported another AMI system ation is changed up to 1000 W/m2 by 200 W/m2 steps while tem-
for high-traffics smart metering by exploiting ZigBee communica- perature is stable at 25 °C.
tion system. In (Collotta and Pau (2015), authors studied for a mea- The obtained curves clearly verify that the modeled PV module
surement and energy management system that is originated from is perfect agreement with the PV module of the Sharp. Moreover,
Bluetooth low energy. A monitoring and fault detection system for the modeled solar panel infrastructure can be also used to set sev-
photovoltaics in which Bluetooth data transmission method is uti- eral parameters such as open circuit voltage (VOC), maximum
lized is reported in Le et al. (2016). Apart from these studies, there power voltage (VMP), short circuit current (ISC), and maximum
are some studies related to wireless sensor networks (WSN) power current (IMP) for any type of PV module. The designed solar
(Fabrizio et al., 2017) and wireless local area network (WLAN) plant models produce rated power at 25.5 kW.
(Harid et al., 2016) for energy demand issue and leakage current The solar plants are tested by applying various irradiation val-
detection, respectively. ues assuming all are located at geographically spanned areas.
264 Y. Kabalci, E. Kabalci / Solar Energy 153 (2017) 262–275

Fig. 1. The block diagram of the modeled microgrid.

Fig. 2. I-V and P-V characteristics of the modeled PV module (irradiation increases from 200 W/m2 to 1000 W/m2).

Therefore, the output voltage of each plant are regulated by a buck tral point of MLI. Furthermore, the voltage at each capacitor is at
converter that is controlled with a regular P&O MPPT algorithm Vdc/4 level that generates the five-level line voltage. The voltage
allowing to obtain output voltage around 480–520 V at the dc stress on each switching device is limited to Vdc through the
bus outputs where it is intended to be set at 500 V. The performed clamping diodes that have been named as D1. . .3 and D11. . .3 in this
analysis results are depicted in Fig. 3 where each solar plant volt- topology. The clamping diodes provide the staircase output level
ages are shown respectively. The dc busbar voltage where the solar against to conventional full-bridge or two-level inverter. The
plant outputs are coupled and applied to inverter is shown in the switching orders of any phase leg to generate staircase voltage syn-
4th axis around 500 V. The maximum output power available at thetize that have been analyzed for R-phase voltage VRN are listed
the dc busbar output is measured around 75 kW as seen in the last in Table 1 (Çolak et al., 2011). The four switches (Sa1–Sa4) of eight in
curve of the figure. a phase leg should be switched on at any interval to generate Vdc/2
voltage level. In the next switching interval, the switches from Sa2
2.2. Energy conversion section to Sa11 are switched to generate Vdc/4 voltage level. The remaining
switching states that constitute zero and negative outputs are
The coupled dc bus voltage is applied to MLI that is imple- repeated in the order given in Table 1.
mented in diode clamped five-level topology. The phase legs and The generated line voltages are compared by a PI controller in
controllers are illustrated in Fig. 4 where R-phase is shown with the modulator block seen on the right hand-side of Fig. 4. This con-
circuit and block diagrams. The advantage of diode clamped MLI ventional controller where the parameters are detected referring to
topology is eliminating the separate dc source requirement at each Ziegler-Nichols method is used to limit the output voltage and
phase leg or at each cascaded bridge cells. The dc input voltage is power by adjusting the duty cycle of switching signals to ensure
shared to five voltage levels by the dc capacitors on the common the output phase voltages at 230Vrms yielding line voltages around
dc bus. The connection of C2 and C3 capacitors comprises the neu- 380 Vrms.
 Y. Kabalci, E. Kabalci / Solar Energy 153 (2017) 262–275 265

Fig. 3. DC output voltages of the solar plants, bus bar voltage, and DC power graph.

Fig. 4. Multilevel inverter and controllers.

266 Y. Kabalci, E. Kabalci / Solar Energy 153 (2017) 262–275

Table 1
Voltage levels of five-level DC-MLI and switching states.

Voltage VRN Switching state

Sa1 Sa2 Sa3 Sa4 Sa11 Sa21 Sa31 Sa41
V4 = Vdc/2 1 1 1 1 0 0 0 0
V3 = Vdc/4 0 1 1 1 1 0 0 0
V2 = 0 0 0 1 1 1 1 0 0
V1 = Vdc/4 0 0 0 1 1 1 1 0
V0 = Vdc/2 0 0 0 0 1 1 1 1

The generated staircase and sinusoidal line voltages of each Table 2

phase are shown in Fig. 4 that are converted from dc bus voltage Values of transmission line parameters.
seen in Fig. 5. The inverter analyses are performed regarding to Unit Value
varying dc input voltage that is generated by the solar plants oper- Length 25 km
ated under various irradiation magnitudes. The voltage levels are Frequency 50 Hz
kept stable owing to control capacity of modulator, and total har- Resistance 0.2568 O/km
monic distortion (THD) rates that are observed against various Inductance 4
107 H/km
Capacitance 8.6
109 F/km
conditions verified the robust structure of controller. The THD
analysis of line voltages and load currents are analyzed in the last
section.
3.1. Transmission section of the system

3. Design of transmission and monitoring stages in the Three-phase PI section transmission line model is utilized in the
advanced metering infrastructure (AMI) model designed remote monitoring system. The length of the transmis-
sion line is selected as 25 km at the output of the multilevel inver-
The dc-dc conversion realized at the buck converter ensures the ter. In order to determine channel characteristic of the
stability of the generated energy in each solar plant separately. The transmission line that will be utilized as a communication medium
acquired dc voltages at the output of the solar farms are collected as well as power delivery, a detailed simulation study is firstly per-
at the dc bus bar part of the designed system. The output voltage of formed in the Advanced Design System (ADS) simulation software.
dc bus bar is fed to diode clamped MLI that is controlled with PD- The schematic diagram of the modeled transmission line in ADS
SPWM to generate three-phase ac line voltages. The length of the software is shown in Fig. 6. When the transmission line modeled,
transmission line that is introduced in the following subsection is the equivalent of the three-phase PI section line is utilized as can
selected as 25 km at the output of the inverter with parameters be seen from the figure. The values of the transmission line param-
seen in Table 2. eters are determined by using following equations and values

Fig. 5. Multilevel inverter; (a) multilevel and sinusoidal line voltages, (b) line voltages at load node.
 Y. Kabalci, E. Kabalci / Solar Energy 153 (2017) 262–275 267

Fig. 6. Modeling of three-phase PI section transmission line in ADS software.

listed in Table 2 that the parameters are calculated regarding to

following equations,
ð2R1 þ R0 Þ ðR0  R1 Þ
RS ¼ ; RM ¼ ð1Þ
3 3

ð2L1 þ L0 Þ ðL0  L1 Þ
LS ¼ ; LM ¼ ð2Þ
3 3

3C 1 C 0
CP ¼ C1; CG ¼ ð3Þ
ðC 1  C 0 Þ
where while the RS and RM denote self-resistance and mutual resis-
tance, LS and LM are self-inductance and mutual inductance, respec-
tively. CP and CG show phase capacitance and ground capacitance,
respectively. The R0, R1, L0, L1, C0 and C1, parameters shown in equa-
tions are defined as follows.
R0 ¼ r 0  llength  kr0 ; R1 ¼ r 1  llength  kr1 ð4Þ
Fig. 7. Characteristic behavior of the transmission line with increasing frequency.
L0 ¼ l0  llength  kl0 ; L1 ¼ l1  llength  kl1 ð5Þ
the other is established with 1500 W rated load. Each of the load
C 0 ¼ c0  llength  kc0 ; C 1 ¼ c1  llength  kc1 ð6Þ
plants is monitored by designed modems that are separately
where r1 and r0 are resistances, l1 and l0 are inductances and c1 and located and utilized at dissimilar frequencies to convey measured
c0 are capacitances in terms of per unit length and positive- and power rate. Different carrier frequencies are selected for mapping
zero-sequence, respectively. multi-channel input data thanks to designed modems at the
The llength parameter is length of the transmission line whose energy generation part of the system.
value is selected as 25 km as mentioned before. After the transmis- In the proposed system, 6 kHz and 8 kHz carrier frequencies are
sion line is modeled in the ADS software, simulation studies are preferred due to two important reasons. One of the reasons is to
performed to obtain characteristics of the transmission line up to cope with adjacent channel interference problem while the second
100 kHz frequency. The behavior of the transmission line in case reason is to decrease disruptive effects of the transmission medium
of exploiting as a communication channel medium is acquired as as mentioned before. Moreover, the designed modems are con-
shown in Fig. 7. structed to recover carrier frequencies accurately. The designed
When the characteristic is analyzed, it is shown that the trans- modems can be reconfigured when the numbers of load plants
mission line has an increasing attenuation feature with rising fre- are increased. The monitoring system of the proposed design will
quency value. Therefore, when the communication system is be explained in the next section in detail.
designed, below of the 10 kHz frequency is selected to ensure that
the destructive effects of the channel less affect the communica- 3.2. Monitoring section of the system
tion signals.
Two different rated loads are considered in the designed sys- Communication infrastructure of the system is constituted
tem. While one of the load plants is formed as a 2500 W rated load, thanks to BPSK modems that are seen in the Fig. 1. These modems
268 Y. Kabalci, E. Kabalci / Solar Energy 153 (2017) 262–275

8 qffiffiffiffi
are very important parts of the proposed monitoring system for >
< 2E
cosð2pf c tÞ; for a data 1
T0
conveying and receiving values of measurement. The BPSK modu- sBPSK ðtÞ ¼ qffiffiffiffi ð8Þ
lation scheme is one of the most robust digital scheme against to >
:  2E cosð2pf c tÞ; for a data 0
T0
noise effects. Two different phases that are generally separated
by 180 or p radians from each other are employed in this mapping
method. Mathematical expression of this mapping technique can where the symbol energy ðEÞ is equivalent to E ¼ ð12ÞA2 T 0 ðcðtÞÞ2
be given as (Xiong, 2000; Glover and Grant, 2000; Harada and Prasad, 2002).
Fig. 8a shows the block diagram of the designed modem for one


cðtÞðA cosð2pf c tÞÞ; for a data 1 channel case. The modulator scheme of the system is shown at
sBPSK ðtÞ ¼ ð7Þ
cðtÞðA cosð2pf c tÞÞ; for a data 0 the top of the figure. The power measurement process from the
PV panels in real-time is realized by measured power block. Follow-
where A shows amplitude of the carrier signal, cðtÞ is a random bin- ing the measurement process, acquired information is primarily
ary pulse with period of T 0 and its level changes between -1 and 1, transformed to the digital data. Message signal and carrier signal
and f C is frequency parameter of the carrier signal. When the value are usually multiplied to obtain the BPSK waveform. However, this
of the data is 1 then the phase of the sBPSK ðtÞ signal will be obtained method is not an proper process to generate modulated signal
as 0, otherwise the phase degree will be p. Symbol energy ðEÞ of the because of the bandwidth issue of real systems. Probability of error
BPSK signal can be expressed as follows: at the receiver side is increased due to the limited bandwidth.

Acos(2πfct)

Band Pass
sBPSK(t)
Measured Pulse Shaping Digital/Analog
Power Filter Converter Filter

Modulator

Power Line Channel

Band Pass Band Pass Pulse Shaping Pulse Decision Received

Filter Filter Filter Compensator Circuit Data

Acos(2πfc't) Demodulator

(a)

(b)

(c)
Fig. 8. Communication infrastructure of designed AMI system; (a) Block diagram of the designed PLC modem, (b) developed modulator structure in Matlab/Simulink, (c)
demodulator structures designed in Matlab/Simulink.
 Y. Kabalci, E. Kabalci / Solar Energy 153 (2017) 262–275 269

Pulse-shaping methods are generally exploited at the modulator channel conditions are took into account in the study. The received
side before the mapping process to handle this problem. After the signal is primarily fed to the BPF filter to separate noises from the
pulse-shaping process, data is transformed to the analog signal type received message signal at the receiver part of the system. Follow-
by using a digital to analog converter (DAC) block. In the next step, ing the filtering process, obtained signal is again multiplied with
analog signal is multiplied with a sinusoidal carrier signal for map- carrier signal for moving it to baseband. Later, different processes
ping process and the modulated signal is applied to a band pass fil- such as pulse-shaping, filtering, and pulse compensation techniques
ter (BPF) to limit bandwidth of the signal. After these processes, the are carried out to eliminate inter symbol interference (ISI) effect. In
signal is fed to the transmission channel in which a power line the final step, decision circuit block detects whether the received

Fig. 9. Solar plant generation analyses, (a) irradiation magnitudes at each solar plant, (b) solar plant output voltage, dc bus voltage, and total output power of the plants.
270 Y. Kabalci, E. Kabalci / Solar Energy 153 (2017) 262–275

signal is a zero data or a one data (Xiong, 2000; Glover and Grant, modulator and demodulator systems will be presented by the fol-
2000; Harada and Prasad, 2002). lowing section.
The Simulink implementation of modulator and demodulator
structures for one channel case is shown in Fig. 8b and Fig. 8c,
respectively. As can be seen from the input of the modulator struc- 4. Comprehensive performance analysis of the proposed system
ture, the measured signal is firstly filtered and transformed digital
signal by Filtering & A/D conversion block. The designed monitoring system is examined in terms of gener-
After the mapping process is performed, the modulated signal ation and energy conversion section, transmission/monitoring sec-
are filtered and the final message signal is fed to the communica- tion and microgrid load section as mentioned previously. The dc
tion medium. The message signal received by the demodulator sys- voltages generated at solar plants of energy generation part are
tem is firstly filtered to reduce destructive effects of the channel delivered to the bus bar over the interphase transformers in order
and then carrier recover procedure is performed. In addition, to keep the total dc voltage at a fixed value. The solar plants are
demodulation process is also realized at the Filtering & Carrier tested under various irradiation conditions as shown in Fig. 9a
Recovery block. After these processes are conducted, pulse shaping where the irradiations are arbitrarily changed to analyze the
and decision processes are carried out at the demodulator struc- success of dc converter and coupled operation on dc busbar. The
ture, respectively. At the demodulator structure, a fourth order fil- irradiations applied to 1st solar plant is get varied as an index of
ter is utilized to obtain appropriate results. The results of the [880 940 1020 940] W/m2 at [0 0.2 0.5 0.8] second intervals as

Fig. 10. Energy conversion section analyses, (a) line voltages at MLI output, (b) staircase and sinusoidal MLI voltages.
 Y. Kabalci, E. Kabalci / Solar Energy 153 (2017) 262–275 271

Fig. 11. Energy conversion section THD analyses, (a) line voltages THD at unloaded operation, (b) line voltages THD at fully loaded operation, (c) line current THD at loaded
operation.

depicted with the blue line. The corresponding dc output voltage of curve is easily regulated by the dc converters. The following varia-
solar plant is shown in the second curve of Fig. 9a by blue line as in tions are not significantly effective on regulated voltage and power
irradiation curve. The irradiation index and intervals of 2nd plant since they were higher than 900 W/m2.
are as [1000 950 900 950] W/m2 and [0 0.3 0.6 0.9] s depicted with The MLI output voltages are analyzed under various irradiations
red line, and the corresponding output voltage variation is shown of solar plants that are presented above. The measurement results
by red line on second curve of Fig. 9a. The 3rd solar plant where of staircase and sinusoidal VRS, VST, and VTR line voltage are shown
the irradiation and output voltages are shown with green lines is in Fig. 10a. Since regular PI controller independently manages each
tested by [875 975 925 920] W/m2 irradiation index changing at phase voltages as shown in Fig. 4, effects of the dc voltage fluctua-
[0 0.35 0.7 0.95] s. The MPPT algorithm limits the effect of the irra- tions seen on Fig. 9b are compensated at the phase outputs of MLI.
diation change that causes around 5 V fluctuation at the output Although there some increases seen at 0.45 s and 0.7 s, the most
voltage of solar arrays. On the other hand, the dc converters regu- significant fluctuations of dc voltages occurred up to 0.3 s are elim-
late the output voltages of each solar plant to ensure the dc busbar inated at the output of MLI. The detailed view of staircase and sinu-
to provide stable voltage to MLI. soidal voltages are presented in Fig. 10b where the line voltages are
The regulated output voltage of each solar plant and dc busbar at 380 VRMS and frequencies at 50 Hz. The quality of line voltages at
voltage where all the solar plants are coupled are illustrated in unloaded and fully load conditions, and line current at loaded
Fig. 9b where the supplied dc power to the full loaded analyze operations are analyzed in Fig. 11a–c respectively.
are shown in the last curve. The MPPT algorithms used in the con- Fig. 11a presents the line voltage THD analysis while MLI is
trol of dc buck converters are responsible to provide available max- operated without any load. The measured THD ratio at this analysis
imum output power to the energy conversion section comprised by is 0.04% where almost all the baseband and carrier harmonics are
the MLI inverter. Furthermore, the regular MPPT algorithm based eliminated. While MLI is fully loaded with 75 kW resistive and
on Perturb and Observe approach that many surveys can be found inductive load, the line voltage is decreased from 386.8 V to
in the literature limits the output voltage in a hysteresis band of 383.7 V and the THD of line voltage is increased to 2.12%
480–520 V. The first voltage curve seen on Fig. 9b belong to the (Fig. 11b) where 5th, 7th, 9th, 11th ordered harmonics that are
output voltage of Solar Plant #1 that the irradiation and solar array assumed as the most significant ones are lower than 1%. In another
output voltages are shown with blue lines in Fig. 9b. measurement, the MLI is loaded with 13 kW resistive and induc-
The second curve belongs to Solar Plant #2 that the irradiation tive load and current THD analysis is performed where the resul-
and solar array output voltages are red lines in Fig. 9a while the tant THD is at 2.37% as seen in Fig. 11c.
third curve is the resultant voltage of Solar Plant #3 where the The transmission line is employed to deliver power load plants
array voltage and irradiations are shown with green line in that are different and with dissimilar power consumption features.
Fig. 9a. The solar plant output voltages and dc bus voltage fluctu- The first plant presented in the lower part of the microgrid load
ates around 490–500 V up to 0.3 s interval since all the plant irra- side is constructed for consuming approximately 2500 W while
diations are independently varied where the first and third plant the second one expends about 1500 W. The power consumption
irradiations were real lower than 1000 W/m2 of 2nd plant. The dc values of the plants are acquired by the designed wattmeter.
bus voltage ripple was around ±5 V on 490 V during these circum- Figs. 12 and 13 depict obtained analysis results for the examined
stances. Once the irradiation of 2nd plant increased from 875 W/m2 loads, respectively. Each one of the load plants contains a modem
to 975 W/m2 at 0.35 s, the rough fluctuation seen on 3rd curve of that can quantize the detected power consumption and mapping
Fig. 9b is decreased and the power fluctuation seen in the last the data which will be conveyed to transmission line as a transmis-
272 Y. Kabalci, E. Kabalci / Solar Energy 153 (2017) 262–275

Fig. 12. Transmitted, demodulated and calibrated signals of first load plant, (a) filtering at 400 Hz 2nd order, (b) filtering at 200 Hz 2nd order, (c) filtering at 50 Hz 4th order.

sion medium. The measured power consumptions of each load sured data are transformed to mapping data in modulator system
plant are presented in the first axis of the Figs. 12 and 13, respec- by attenuating at a rate of 1:1000. While the mapping data of
tively. As can be seen in the second axis of the figures, the mea- the first plant is mapped by a carrier signal at 8 kHz frequency,
 Y. Kabalci, E. Kabalci / Solar Energy 153 (2017) 262–275 273

Fig. 13. Transmitted, demodulated and calibrated signals of second load plant, (a) filtering at 400 Hz 2nd order, (b) filtering at 200 Hz 2nd order, (c) filtering at 50 Hz 4th order.

the other one is mapped by a carrier signal at 6 kHz frequency. The 50 Hz 4th order in Fig. 12a–c respectively for the first load plant
filtering analyses are performed by adjusting the filter order and while second plant analyses are shown in Fig. 13a–c under same
cut-off frequency at 400 Hz 2nd order, 200 Hz 2nd order, and circumstances.
274 Y. Kabalci, E. Kabalci / Solar Energy 153 (2017) 262–275

The monitoring modem that is placed at the output of the diode Almas, M.S., Vanfretti, L., Løvlund S., Gjerde, J.O., Open source SCADA
implementation and PMU integration for power system monitoring and
clamped multilevel inverter is also associated with a coupling cir-
control applications, In: 2014 IEEE PES General Meeting|Conference &
cuit similar to utilized at the modulator side. The decreased line Exposition, 2014, pp. 1–5.
voltage with mapped data is applied to the demodulator system Atalik, T. et al., 2014. Multipurpose platform for power system monitoring and
and is passed a filter to clear harmonic contents at the output of analysis with sample grid applications. IEEE Trans. Instrum. Meas. 63 (3), 566–
582.
the demodulator. The demodulator part of the designed modem Cabanas, M.F. et al., 2007. A new online method based on leakage flux analysis for
placed in the monitoring side is constituted to acquire carrier the early detection and location of insulating failures in power transformers:
signals at 6 kHz and 8 kHz frequencies to choose plant data sever- application to remote condition monitoring. IEEE Trans. Power Deliv. 22 (3),
1591–1602.
ally. The demodulated message signal are presented in the third Chi, H.R., Tsang, K.F., Chui, K.T., Chung, H.S.H., Ling, B.W.K., Lai, L.L., 2016.
axis of the Figs. 12 and 13 that are obtained by filtering a fourth Interference-mitigated ZigBee-based advanced metering infrastructure. IEEE
order low-pass Butterworth filters. In addition, carrier signal recov- Trans. Ind. Inf. 12 (2), 672–684.
Çolak, I., Kabalcı, E., Bayındır, R., 2011. Review of multilevel voltage source inverter
eries of the demodulators contain band-pass filters with ±100 Hz topologies and control schemes. Energy Convers. Manage. 52, 1114–1128.
lower and upper band edge frequencies. Besides, output filters of Collotta, M., Pau, G., 2015. A novel energy management approach for smart homes
the demodulator systems are adjusted to 50 Hz pass band edge fre- using bluetooth low energy. IEEE J. Sel. Areas Commun. 33 (12),
2988–2996.
quency to suppress the line voltages. Following the filtering pro- Fabrizio, E. et al., 2017. Monitoring and managing of a micro-smart grid for
cess, calibrating process is applied to determine real-time power renewable sources exploitation in an agro-industrial site. Sustain. Cities Soc. 28,
consumption values in terms of watt. While the obtained receiver 88–100.
Fateh, B., Govindarasu, M., Ajjarapu, V., 2013. Wireless network design for
outputs after filtering process are illustrated in the fourth axis of
transmission line monitoring in smart grid. IEEE Trans. Smart Grid 4 (2),
the Figs. 12 and 13, the calibrated outputs are depicted in the last 1076–1086.
axis of figures. The measured and monitored power signals of the Gaurav, D., Mittal, D., Vaidya, B., Mathew, J. A GSM based low cost weather
examined microgrid are considered, it is clear that designed mon- monitoring system for solar and wind energy generation, In: The Fifth
International Conference on the Applications of Digital Information and Web
itoring system can be efficiently utilized to monitor renewable Technologies (ICADIWT 2014), 2014, pp. 1–7.
energy sources placed in different locations. Ginot, N., Mannah, M.A., Batard, C., Machmoum, M., 2010. Application of power line
communication for data transmission over PWM network. IEEE Trans. Smart
Grid 1 (2), 178–185.
5. Conclusion Glover, I., Grant, P., 2000. Digital Communications. Prentice Hall, Great Britain.
Harada, H., Prasad, R. Simulation and software radio for mobile communication,
Artech House Publishers, 2002.
The observing and metering processes are necessary in renew- Harid, N., Bogias, A.C., Griffiths, H., Robson, S., Haddad, A., 2016. A wireless system
able energy conversion systems as applied in smart grid applica- for monitoring leakage current in electrical substation equipment. IEEE Access
4, 2965–2975.
tions of conventional grid. A microgrid model including separate Kabalci, E., 2013. Design and analysis of a hybrid renewable energy plant with solar
solar and load plants is established and is comprehensively exam- and wind power. Energy Convers. Manage. 72, 51–59.
ined in this study. The electricity generation part of the modeled Kabalci, E., 2015. A smart monitoring infrastructure design for distributed
renewable energy systems. Energy Convers. Manage. 90, 336–346.
microgrid contains three solar plants that are supposed to be placed
Kabalci, E., Kabalci, Y., BPSK modem based power line communication system for
in different locations. The solar plants are configured with 150 PV observing photovoltaic panels, In: Presented at the 2nd International
modules each at 170 Wp rated power, and generated solar array Conference on Nuclear and Renewable Energy Resources (NURER 2010),
voltages are regulated by buck converters to supply a diode Ankara, Turkey, July 4–7, 2010, 2010.
Kabalci, E., Kabalci, Y., Develi, I., 2012. Modelling and analysis of a power line
clamped MLI around 500 V. The buck converters are controlled by communication system with QPSK modem for renewable smart grids. Int. J.
using regular P&O MPPT algorithm that ensures to acquire the max- Electr. Power Energy Syst. 34 (1), 19–28.
imum available power and regulated dc output voltage at the Kabalci, Y., Kabalci, E., Canbaz, R., Calpbinici, A., 2016. Design and implementation
of a solar plant and irrigation system with remote monitoring and remote
desired magnitude. On the other hand, the diode clamped MLI, control infrastructures. Sol. Energy 139, 506–517.
which is a convenient alternative to conventional full bridge inver- Kamel, T., Biletskiy, Y., Chang, L., 2015. Fault diagnosis and on-line monitoring for
ter topology in terms of requiring a single dc voltage source, and grid-connected single-phase inverters. Electr. Power Syst. Res. 126,
68–77.
robustly decreasing the current and voltage THD ratios, is operated Kosonen, A., Ahola, J., 2010. Comparison of signal coupling methods for power line
by PD-SPWM modulation scheme with independent phase control communication between a motor and an inverter. IET Electr. Power Appl. 4 (6),
that enforces the inverter against unbalanced loads. This topic is 431–440.
Kurohane, K., Senjyu, T., Yona, A., Urasaki, N., Goya, T., Funabashi, T., 2010. A hybrid
planned to be researched in the future studies. The transmission
smart AC/DC power system. IEEE Trans. Smart Grid 1 (2), 199–204.
and distribution model of presented microgrid system is designed Le, P.T., Tsai, H.-L., Lam, T.H., 2016. A wireless visualization monitoring, evaluation
with realistic impedance values of a 25 km transmission line, and system for commercial photovoltaic modules solely in MATLAB/simulink
environment. Sol. Energy 140, 1–11.
the configured line is utilized as the communication medium in
Lin, C.K., Lu, M.T., Yeh, S.C., Chen, H.H., 2009. Video streaming over in-home power
addition to carrying power. The distribution line section is consti- line networks. IEEE Trans. Multimedia 11 (3), 523–534.
tuted with two load plants where the power consumptions of the Ma, T., Yang, H., Lu, L., 2013. Performance evaluation of a stand-alone photovoltaic
load plants are measured and are mapped by the designed modem system on an isolated island in Hong Kong. Appl. Energy 112, 663–672.
Madueño, G.C., Nielsen, J.J., Kim, D.M., Pratas, N.K., Popovski, P., 2016. Assessment of
to realize the PLC infrastructure. The carrier frequencies are set to LTE wireless access for monitoring of energy distribution in the smart grid. IEEE
6 kHz and 8 kHz for any load plant, and thus the communication J. Sel. Areas Commun. 34 (3), 675–688. S. Č.
channels on the power line are obtained. The acquired results Mannah, M.A., Batard, C., Ginot, N., Machmoum, M., 2011. A PLC-based method for
data transmission over a pulsewidth-modulated network. IEEE Trans. Power
clearly confirm that the installation costs of remote monitoring sys- Deliv. 26 (4), 2259–2266.
tems, which are necessary in the SCADA, wireless communication Mohamed, F.P., Siew, W.H., Soraghan, J.J., Strachan, S.M., Mcwilliam, J., 2014.
systems or in any Ethernet-based systems, are removed thanks to Remote monitoring of partial discharge data from insulated power cables. IET
Sci. Meas. Technol. 8 (5), 319–326.
the proposed monitoring system. In addition, the number of the Padilla, E., Agbossou, K., Cardenas, A., 2014. Towards smart integration of
load plants can be easily augmented by employing the proposed distributed energy resources using distributed network protocol over
monitoring system in the possible future works. ethernet. IEEE Trans. Smart Grid 5 (4), 1686–1695.
Senthilnathan, K., Annapoorani, I., 2016. Implementation of unified power quality
conditioner (UPQC) based on current source converters for distribution grid and
References performance monitoring through LabVIEW Simulation Interface Toolkit server:
a cyber physical model. IET Gen. Transm. Distrib. 10 (11), 2622–2630.
Shariff, F., Rahim, N.A., Ping, H.W., 2013. Photovoltaic remote monitoring system
Ahmed, M.A., Pan, J.-K., Song, M., Kim, Y.-C., 2016. Communication network
based on GSM. IEEE Conference on Clean Energy and Technology (CEAT) 2013,
architectures based on ethernet passive optical network for offshore wind
379–383.
power farms. Appl. Sci. 6 (3), 81.
 Y. Kabalci, E. Kabalci / Solar Energy 153 (2017) 262–275 275

Shariff, F., Rahim, N.A., Hew, W.P., 2015. Zigbee-based data acquisition system for Tung, H.Y. et al., 2014. The generic design of a high-traffic advanced metering
online monitoring of grid-connected photovoltaic system. Expert Syst. Appl. 42 infrastructure using zigbee. IEEE Trans. Ind. Inf. 10 (1), 836–844.
(3), 1730–1742. Vanfretti, L., Baudette, M., Domínguez-García, J.-L., Almas, M.S., White, A., Gjerde, J.
Sharp NE-170UC1 Multipurpose Module, Sharp, USA, 2008. Available at: http://files. O., 2016. A phasor measurement unit based fast real-time oscillation detection
sharpusa.com/Downloads/Solar/Products/sol_dow_NE170U1.pdf. application for monitoring wind-farm-to-grid sub–synchronous dynamics.
Silva, A.S., dos Santos, R.C., Bottura, F.B., Oleskovicz, M., 2017. Development and Electr. Power Comp. Syst. 44 (2), 123–134.
evaluation of a prototype for remote voltage monitoring based on artificial Venkatraman, K., Reddy, B.D., Selvan, M.P., Moorthi, S., Kumaresan, N., Gounden, N.
neural networks. Eng. Appl. Artif. Intell. 57, 50–60. A., 2016. Online condition monitoring and power management system for
Silvestre, S., Chouder, A., Karatepe, E., 2013. Automatic fault detection in grid standalone micro-grid using FPGAs. IET Gen. Transm. Distrib. 10 (15), 3875–
connected PV systems. Sol. Energy 94, 119–127. 3884.
Son, Y., Pulkkinen, T., Moon, K.d., Kim, C., 2010. Home energy management system Wilkinson, M., Darnell, B., Delft, T.V., Harman, K., 2014. Comparison of methods for
based on power line communication. IEEE Trans. Consum. Electron. 56 (3), wind turbine condition monitoring with SCADA data. IET Renew. Power Gen. 8
1380–1386. (4), 390–397.
Tina, G.M., Grasso, A.D., 2014. Remote monitoring system for stand-alone Xiong, F., 2000. Digital Modulation Techniques. Artech House, USA.
photovoltaic power plants: the case study of a PV-powered outdoor Zaker, N., Kantarci, B., Kantarci, M.E., Mouftah, H.T., 2014. Smart grid monitoring
refrigerator. Energy Convers. Manage. 78, 862–871. with service differentiation via EPON and wireless sensor network convergence.
Tsuzuki, S., 2006. High-speed power-line communication and its application to a Optical Switching and Networking 14 (1), 53–68.
localization system. IEICE Trans. Fundament. Electron. Commun. Comput. Sci. Zhang, P., Li, F., Bhatt, N., 2010. Next-generation monitoring, analysis, and control
E89A (11), 3006–3012. for the future smart control center. IEEE Trans. Smart Grid 1 (2), 186–192.