Mixed PV-Wind Small Power Microgrid [624081]

Mixed PV-Wind Small Power Microgrid

Catalin Patrascu*, Cezara-Liliana Rat**, Dan Hulea*, Danut Vitan*, Nicolae Muntean*
Electrical Engineering Department*, Automation and Applied Informatics Department**
Politehnica University Timisoara
Timisoara, Romania
[anonimizat]

Abstract- The paper presents a microgrid structure, based on
wind and photovoltaic (PV) renewable energies inputs, implemented in a dedicated configurable microgrid laboratory.
The wind turbine is realized through a Hardware-in-the-Loop
(HIL) emulator. PV energy can be taken from a set of panels or from a programmable DC source, which can implement the I-V
characteristics of the real system. The microgrid configuration,
its components and functionality are presented also through experimental results.
Keywords—DC microgrid; PV systems; wind turbines;
hardware-in-the-loop.
I. INTRODUCTION
Harvesting energy on a large scale is undoubtedly one of
the main challenges of our time. Future energy sustainability depends heavily on how the renewable energy problem is addressed in the next few decades [1].
Renewable energy has begun to be seen as a long-term
sustainable energy source only due to the increase in energy demand worldwide and environmental concerns. However, two major limitations exist that prevent widespread adoption: availability and variability of the electricity generated and the cost of the equipment. Seasonal effects and limited predictability result in intermittent generation. Industry must overcome this in order to deliver renewable energy in significant quantities [1-5].
Alternative resources have different characteristics; it is
therefore essential to have a well-defined and standardized
framework/procedure for integrating them into a hybrid system. The methods, that integrate different alternative energy sources to form a hybrid system, can be generally classified into three categories: DC-coupled, AC-coupled, and hybrid-coupled. Direct current (DC) electrical systems are
gaining popularity due in part to high efficiency, high reliability and ease of interconnection of the renewable sources compared to alternating current (AC) systems.
In a DC-coupled configuration, different alternative energy
sources are connected to a DC bus through appropriate power electronic interfacing circuits. DC microgrids have been proposed to improve point-of-load energy availability and to integrate distributed renewable energy sources with energy storage. Various renewable energy sources such as photovoltaic (PV) systems have natural DC couplings; therefore it is more efficient to connect these sources directly to DC microgrid by using DC/DC converters. The system can supply power to AC loads, or be interfaced to a utility grid through a bi-directional converter. The greatest advantages of the DC coupling scheme is its simplicity and the fact that no
synchronization is needed [4, 6].
For a successful implementation of a mixed PV-Wind
microgrid in a real system, it must first be tested in various situations. This is possible by testing the system at laboratory level with real components.
This paper presents a mixed microgrid structure based on
real components implemented in a configurable microgrid laboratory.
In next chapters, the microgrid structure is first presented.
Then the implemented solution of the HIL wind emulators is analyzed extensively. Finally, some experimental results, performed on the microgrid, in different operation modes, are evaluated.
II. M
ICROGRD STRUCTURE
The microgrid presented here has relatively few (low)
power components, constituting a so called “nanogrid”. Nevertheless, in accordance with the literature in this domain, it is going to be referred to as a “microgrid”. This term will be used throughout this paper.
The microgrid has a DC-coupled configuration, which can
be seen in Fig. 1, and it contains two DC Busses: a high voltage DC Bus (HVDC – 350Vdc) and a low voltage DC Bus (LVDC – 50Vdc).
These renewable power sources are wind and PV. The
wind system is connected to the HVDC Bus through an inverter (INV). The INV controls the AC-DC conversion of the wind system’s induction generator [7, 8]. The PV module is connected directly to the HVDC Bus, because the voltage of the HVDC bus was chosen to correspond to the MPP of the PV module.
The energy storage is implemented through supercapacitors
(SC) and rechargeable battery bank (BAT). The BAT is composed of Valve Regulated Lead-Acid Gel Batteries suited for small photovoltaic installatio ns [9]. The SC are capable of
capturing and storing, but also of delivering quick bursts of energy necessary for the stability of the microgrid. They can save excess energy during peak power supply that is otherwise lost. They are self-contained energy storage devices that discharge and recharge quickly, providing a much needed energy buffer for the microgrid [10].
The bidirectional hybrid DC-DC (HBDC) converter is
responsible for charging/discharging the supercapacitors from/to the HVDC. HBDC works as a boost converter in one direction and a buck in the other [11, 12]. 978-1-5090-4489-4/17/$31.00 ©2017 IEEE 699

HVDC
400VLOADS
AC
DC
HYBRID INVERTER
Server
Ethernet network
AC Bus
DC BusLVDC
50VSmart
MeterDC
DC
HBDC
CCDCDCWT
SC
PVDC
LOADSINV
HI
BATMAIN GRID
Fig. 1. Microgrid Structure
The HBDC has a rather new topology which uses a
nonlinear droop control method for charging and discharging the SC [12]. The hybrid topology has the advantage of higher conversion ratio and lower switching element stress, compared to conventional topologies. The control method used for charging and discharging the SC has a decentralized structure, therefore is immune to communication losses from within the microgrid and is easier to be implemented than a centralized structure.
The charge controller (CC) and the HBDC are responsible
for maintaining the voltage on the HVDC Bus at around 350Vdc based on the amount of energy available from the renewable sources. The CC links the HVDC Bus, where the renewable sources transfer their power, with the LVDC Bus, where BAT is connected [13]. If the renewable sources cannot provide enough power, a Hybrid Inverter (HI) can take power from the main grid in order to charge the batteries (if the system is grid connected). On the other hand, when too much energy is produced, the inverter can inject power into the main grid.
The HI is also responsible for supplying the AC microgrid
loads. The HI contains an in tegrated AC transfer switch,
capable of being grid-interactive or grid-independent [14]. The local AC loads are usually supplied by the renewable sources through the CC. If their energy output does not suffice and the microgrid is in on-grid mode, the grid will provide the power deficit through the HI. If this situation occurs while in off-grid mode, the loads are supplied from the BAT through the HI and/or from the SC through the CC and the HBDC.
The HVDC bus also contains DC loads, such as lighting,
plug-in hybrid electric vehicle (PHEV) chargers in combination with a DC-DC converter or other components
which can use 350Vdc.
The general data regarding the converters and storage
elements used in the microgrid are summarized in the table I.
This microgrid has a local network based on Ethernet in
order to ensure connectivity between converters, power meters and server for microgrid settings, data acquisitions and monitoring purposes.

III. HIL
WIND EMULATORS
Historically, wind turbine prototypes were tested in the
field, which was and continues to be a slow and expensive procedure [15]. As in many such situations, it is much more feasible to define and conduct experiments using meaningful simulations of renewable energy sources rather than the actual sources themselves.

TABLE I.
THE CONVERTER PARAMETERS
Component Parameters
Charge Controller
[CC] Xantrex XW MPPT 60/150 Charge Controller
Uin=400 (V), I in=16.6(A), U out=600 (V),
Iout=6(A), P out=4 (kW)
AC-DC Convertor
[INV]
Danfoss FC-302
Uin≈3×380(Vac), I in=7.4 (A), f in = 50/60 (Hz),
P=4 (kW), U out=400 (Vdc), I out=8.2 (A)
Hybrid DC-DC
Converter [HBDC ] VLowDC =50(V), V HighDC =400(V), P N = 5(kW)
Hybrid Inverter [HI] Schneider Xantrex Hybrid Inverter
Uin:=48(V), I in=96(A), U out≈240(V), I out=40 (A),
Pout=4.5 (kW)
Batteries [BAT] 8 x Moll 6V 6 OPzV.block.solar 420 6 V/480Ah
Supercapacitor [SC] Maxwell, 2×36 (F), Rated voltage=125(V) 700

Instead of relying solely on software simulation, a HIL
emulator can be developed to allow direct experimentation and measurements of how such a system might operate [16].Nowadays, HIL solutions can be used as an intermediary step before the implementation on the actual process [17].
The wind turbine HIL structure, used in this microgrid,
consists of:
1. A software system which implements the real-time
mathematical model of the wind turbine. This was made in LabVIEW and it includes a variable wind speed profile generator and the aerodynamic and mechanical models of the wind turbine. The code developed, which is a graphical description of the dynamic and electrical behavior of the turbine, was deployed (compiled and downloaded) to the real time target, a National Instruments CompactRIO 9068 platform (cRIO). The NI cRIO includes a dual core (Cortex A9) real-time processor and an integrated FPGA, along with 8 RIO (reconfigurable I/O) modules, which provide high-speed signal acquisition and generation, necessary in real time processing. In [18] a similar system was programmed using a LabVIEW toolkit “Control Design and Simulation Module”.
2. A physical system that provides the similar static
and dynamic characteristics as the real studied system. This contains a voltage source inverter (ABB ACS 800) and a 7.5 kW three phase squirrel cage induction motor (IM) with a gearbox (GB) (the wind turbine equivalent) and an induction generator (IG) [19, 20]. The gear box is needed in order to adjust the speed to the value necessary for the generator [18].
The wind turbine chosen for laboratory emulation has the
characteristics shown in table II [18].
Being in a close loop, the two systems communicate with
each other [21]. The software reads the torque estimated by ACS800 and calculates the new rotating speed ( ω
r) based on
the previous rotating speed and the torque calculated using the aerodynamic model of the turbine. The 1/6 gearbox is also accounted for in this step.
Another method would be to read the rotating speed and
calculate the torque, as can be seen in [19].
Several control schemes have been tested, each of them
providing certain benefits as far as microgrid testing and functionality is concerned.
a) PC – NI cRIO – ABB ACS800
This first method, Fig. 2a, is the fastest, but also the most
complex. The method consists of controlling the NI cRIO board through an external computer in order to monitor the entire process and to prescribe wind profiles, but also to offer the interface between the system and the human operator in charge of the overall simulation. In this case the external computer takes complete control of the process as long as the
interface between NI cRIO and it remains alive. This method requires the best reaction time possible, therefore the NI cRIO board communicates with the computer directly through alocal Ethernet connection and with the ACS800 inverter through analogue connection. TABLE II.
THE PARAMETERS OF THE WIND TURBINE SYSTEM
Component Parameters
Rated power P୬=5 . 5[ k W ]
Rated wind speed v଴=1 1[ m / s ]
Maximum turbine speed n=1 2 6[ r p m ]
Turbine inertia J୵୲=1 4 0[ k gmଶ]
Blade swept area A = 19.6 [mଶ]
Radius of the turbine blade R=2 . 5 [ m ]
Maximum coefficient of power conversion Cp = 0.42
Nominal tip-speed ratio λ଴=3

The main advantage of this method is that it provides full
control. The disadvantage of this variant is the existence of signal interference. The microgrid laboratory contains a large number of equipment that cause unwanted background noise. In order to suppress interfering signals and reduce background noise, several filters have been added. Another option would be to use an Ethernet connection between NI cRIO and the ACS800 inverter, but this configuration is significantly slower and, since the real-time processing capabilities of NI cRIO are not taken full advantage of, and the external computer practically controls the process, the NI
cRIO controller board becomes redundant. This conclusion leads to the next configuration.
b) PC – ABB ACS 800
In order to simplify the system while maintaining its
functionality, the NI cRIO was not used. A computer communicates directly to the ACS800 inverter through Ethernet cable.
This solution, Fig. 2b, is significantly slower, but it has the
benefit of being simple and immune to noise. This configuration uses a toolkit, namely NI OPC Servers to communicate between the computer and the ACS800 inverter through the RETA-01 Ethernet Adapter [22].
c) NI cRIO – ABB ACS800
The NI cRIO controller board, as an embedded system, is
usually dedicated to controlling the I/O system and does not provide any direct human interface. It has the advantage of running a fixed program rather than running any user application, therefore being much more capable of rigorous testing than a computer would be. Due to this and its restricted communication with other systems, an embedded system based on NI cRIO, Fig. 2c, provides much more accurate data. In this case, the controller board does not depend on the computer; it stores and uses the last value it received from a computer through the network. Another option here would be to include predefined wind variation on the controller board.
d) Results
The output of the IG is controlled programmatically by a
Danfoss Inverter (INV) using a dSpace Controller Board through a computer running Matlab/Simulink. The control method is based on the Maximum Power Point Tracking (MPPT) algorithm. The effect of this can be seen in Fig. 3. 701

a) PC – NI cRIO – ABB ACS800
b) PC – ABB ACS 800

c) NI cRIO – ABB ACS800
Fig. 2. The emulator configurations

Fig. 3. Emulator response to variation in wind speed

IV. PV SYSTEM
The PV system is composed of 12 Polycrystalline Silicon
PV panels (table III) which provides close to 3 kW. The panels are connected in series to increase voltage and thus a converter is unnecessary for microgrid integration. The PV systems can be replaced using a programmable dc source, which can simulate various I-V curves for different solar cell materials [23]. TABLE III.
THE PARAMETERS OF THE PV SYSTEM
Component Parameters
PV Panels
12 x Renesola JC250M
Umax=30.1(V), Imax=8.31(A)
Programmable DC Power
Supply Chroma 62100H 600S
Uout =0-600(Vdc), Iout =0-17(A),
P=10(kW)

The PV panels are connected directly to the HVDC bus in
order to eliminate the cost of an additional interface converter. Therefore the voltage of the HVDC bus is chosen as close as possible to the maximum power point voltage of the panels. Small adjustments of the bus voltage can be made also for this scope. As literature shows [24] this method does not add significant power losses.

V. M
ICROGRID FUNCTIONALITY
The microgrid presented in this paper uses decentralized
control, based on the voltages of the DC buses. This type of control requires no communication between subsystems because all control decisions are made with local information only. Practically, each subsystem has its own controller [5].
The microgrid was designed to operate in two modes: on-
grid and off-grid. A normal operation mode is defined as follows: the renewable energy sources produce enough power to supply the DC and AC loads and maintain the BAT charged. This means that the HVDC bus has a voltage of 350V. The CC sends a level of power to the LVDC bus that is enough to keep the BAT charged and the AC loads supplied.
In on-grid mode, in a case of excess power, meaning more
power than is necessary for the loads and BAT, the HVDC bus voltage increases above 350V. The HBDC starts charging the supercapacitors. If this is not a sudden spike of power and the excess power appears to persist for a longer duration, the CC begins pumping power into the LVDC bus. When the BAT comes to be overcharged (the LVDC bus increases the voltage level) the HI sends the supplementary power to the main grid.
If the power supplied by the renewable energy sources
decreases bellows the point in which the loads can be sustained, the microgrid attempts to compensate the shortage using the storage system: the HBDC turns on discharging the SC in an attempt to maintain the voltage in the HVDC bus and normal operation in the LVDC circuit. If the lack of power persists, the DC load is disconnected. If this does not suffice, the CC will shut down and the AC loads will be supplied through the HI form the BAT first and then from the main grid.
In off-grid mode, when a sudden surge of power appears
the SC start charging. When the SC is fully charged, the CC sends excess energy to the LVDC bus. There are certain limitations to how much power can be transferred to the LVDC bus, limitations included in the CC`s control algorithm.
702

The CC will never pump more power into the LVDC bus
than is necessary for the BAT and the AC loads, therefore a part of the excess power might not pass through the CC into the LVDC bus. If such a case occurs the only solution is to control the output of the wind system through the INV’s resistive dissipative circuit [8].
In the case of a fall in power supply, during off-grid mode,
the storage system starts discharging in order to compensate the loss. If this action is not sufficient, the DC load is disconnected. If the renewable power supply does not rebound, the entire system fails and shuts down.
Microgrid testing was achieved by using the structure from
figure 1, implemented on a SCADA system, based on LabVIEW software. Notations present in the waveforms are explained in table IV.
The experimental results, shown in Fig. 4, present the main
current and voltage variations throughout the microgrid in the case of a major variation of load consumption. In this case all renewable sources generate energy. At the beginning the DC load absorbs all the available power maintaining the voltage of the HVDC bus at 350V. The CC is not active because it only starts at 355V. When the DC loads receives the command to decrease power consumption, the voltage on the HVDC bus is raised. Than the HBDC converter goes in charging mode and raises the HBDC current and voltage to indicate that the SC is being charged. As a reaction to the increase in the voltage on the HVDC bus and due to the fact that they function at a constant power, the currents from the PV and wind generator decrease. After a period of about 5-10 seconds, the CC starts providing energy to the LVDC bus, and then this energy is injected into the main grid and the AC loads by the HI, which can be observed from its entry current, IHybInv.
In Fig. 5, the microgrid is supplied by both energy sources
until the PV is stopped at t=5s. The DC load is set to a constant value during the entire test, consuming constant power. The wind generator provides a constant power of about 800W. The voltage supplied by the wind generator is enough to maintain the HVDC bus at 350V, hence the HBDC does not activate and the voltage on the SC is kept constant. The HBDC activates only when the voltage in the HVDC is above 365V.
TABLE IV.
THE NOTATION EXPLANATIONS
Notation Name Explanation
VHVbus HV Bus Voltage
VSC Supercapacitor Voltage
VLVbus LV Bus Voltage
IWIND Wind Current
IHBDC In/Out HBDC Current
IPV PV Current
IDCload DC Load Current
IINchrg In Charge Controller Current
IOUTchrg Out Charge Controller Current
IBAT In/Out Battery Current
IHybInv In/Out Hybrid Current Fig. 4. Microgrid current and voltage variations at different power
consumption
Fig. 5. Microgrid current and voltage variations at constant power
consumption and without PV
703

The fluctuations present in the HBDC current are very
small and of no consequence. The actual reaction of stopping the PV system can be seen at the current of the CC, which decreases rapidly. Due to this, the HI stops pumping power into the main grid and attempts to take energy from the BAT in order to compensate the lack of energy in the LVDC bus. Because of BAT state of charge, the VLVbus has a variation which can be seen also in VHVbus through the CC current.

VI. C
ONCLUSIONS
Working in a microgrid (Smart Grid) environment
facilitates mixing the renewable and conventional energy sources with loads and storage elements in an intelligent energy system. Though, before it can be fully implemented to the real system, the microgrid solutions must be studied and tested in various conditions, using a laboratory configurable infrastructure. The paper has presented a mixed microgrid, including configuration description, emulation facilities for input sources, various energy conversion modules and their control, sustained by experimental results.
The microgrid uses two DC busses and the corresponding
power interfaces between inputs, storage elements, DC or AC loads and the main grid
The solution adopted for the wind turbine emulator and the
corresponding experimental results in transients were also presented. The proposed structure was implemented in a configurable laboratory facility, which can ensure experimental conditions for various configurations and control strategies.
.
R
EFERENCES

[1] E. F. Camacho, T. Samad, M. Garcia-Sanz, I. Hiskens, “Control for
renewable energy and smart grids,” The Impact of Control Technology, IEEE Control Systems Society, pp. 69-88, 2011.
[2] Hui J., “An Adaptive Control Algorithm for Maximum Power Point
Tracking for Wind Energy Conversion Systems”, Ph. D. Dissertation, Queen’s University Kingston, Canada, December 2008.
[3] L. Mihet-Popa, “Wind Turbines Using Induction Generators Connected
to the Grid”, Ed. Politehnica, 2007.
[4] S. B. Mohammad, “Model Predictive Control techniques with
application to photovoltaic, DC Microgrid, and a multi-sourced hybrid energy system,” Ph. D. Dissertation, Texas A&M University, 2015.
[5] C. M. Colson, “Towards real-time power management of microgrids
for power system integration: a decentralized multi-agent based approach”, Ph.D. Dissertation, Montana State University, 2012.
[6] C. Wang, H. Nehrir, F. Lin, J. Zhao, “From hybrid energy systems to microgrids: Hybridization techniques, configuration, and control,”
IEEE Power and Energy Society General Meeting, pp. 1-4, 25- 29 July 2010.
[7] C. Patrascu, N. Muntean, O. Cornea, A. Hedes, “Microgrid Laboratory
for Educational and Research Purposes,” IEEE 16th International Conference on Environment and Electrical Engineering (EEEIC), pp. 1-6, 7-10 June 2016.
[8] Danfoss datasheet, available online at: http://drives.danfoss.com
/products/vlt/low-voltage-drives/vlt-automation-drive-fc-301-302/#/
[9] Moll Batteries datasheet, available online at: http://www.moll-
batterien.de/dateien/pdfs/MOLL_TechSpez_OPzV_block_solar_en.pdf.
[10] Maxwell Technology UltraCapacitors datasheet, available at:
http://www.maxwell.com/images/documents/MANUAL_HTM125_20BMOD0063_1014343_8.pdf
[11] H. Nomura, K. Fujiwara, M. Yoshida, “A New DC-DC Converter
Circuit with Larger Step-up/down Ratio,” IEEE 37th Power Electronics Specialists Conference, pp. 1-7, 18-22 June 2006
[12] D. Hulea, O. Cornea, N. Muntean, “Nonlinear droop charging control
of a supercapacitor with a bi-directional hybrid DC-DC converter,” IEEE 16th International Conference on Environment and Electrical Engineering (EEEIC), pp. 1-6, 7-10 June 2016.
[13] Conext MPPT 60/150 datasheet, available at: https://www .altestore.
com/static/datafiles/Others/conext-mppt-60-150-datasheet.pdf
[14] Hybrid Invertor datasheet, available at: http://cdn.solar.schneider-
electric.com/wp-content/uploads/2014/12/xw_hybrid_inverter-charger operation_guide_975-0240-01-01_rev-d_eng.pdf
[15] R. Schkoda,C. Fox, R. Hadidi, V. Gevorgian, R. Wallen, S. Lambert,
“Hardware-in-the-Loop Testing of Utility-Scale Wind Turbine Generators,” National Renewable Energy Laboratory Technical Report NREL/TP-5000-64787, Jan. 2016.
[16] T. Hardy, W. Jewell, “Hardware-in-the-loop wind turbine simulation
platform for a laboratory feeder model,” IEEE Transactions on Sustainable Energy, vol. 5, no. 3, pp. 1003-1009, July 2014..
[17] A. Bouscayrol, X. Guillaud, R. Teodorescu, P. Delarue, W. Lhomme,
“Hardware-in-the-loop simulation of different wind turbines using Energetic Macroscopic Representation,” IECON 2006 – 32nd Annual Conference on IEEE Industrial Electronics, pp. 5338-5343, : 6-10 Nov. 2006.
[18] M. Topor, N. Muntean, C. Sorandaru, “Wind Turbine Emulator
Development using NI cRIO 9068 and ABB ACS 800 Drive”, Journal of Electrical Engineering.
[19] N. Muntean, L. Tutelea, D. Petrila, O. Pelan, “Hardware in the loop
wind turbine emulator,” International Aegean Conference on Electrical Machines and Power Electronics and Electromotion, pp. 53-58, 8-10 Sept. 2011.
[20] ABB ACS800 Hardware Manual, available at: https://library.e.abb.
com/public/f17eb3f499af3aa7c1257b97005603ef/EN_ACS800_04_HW_F_A4.pdf
[21] G. Caraiman, C. Nichita, V. Mînzu, B. Dakyo, “Marine current Turbine
Emulator Design Based on Hardware in the Loop Simulator Structure”, 14th International Power Electronics and Motion Control Conference EPE-PEMC, pp.101-107, 6-8 Sept. 2010.
[22] ABB RETA-01 User’s Manual, available at: https://library.e.abb.com
/public/b3b3c13dhttps://library.e.abb.com/public/b3b3c13dcead1c3ec12572aa002977cc/en_RETA_01_um_d.pdfcead1c3ec12572aa002977cc/en_RETA_01_um_d.pdf
[23] Chroma User’s Manual, available at: http://www.chromausa.com/pdf /62150H-600S-E.pdf [24] K. Shenai, A. Jhunjhunwala, P. Kaur, “Electrifying India: Using solar
dc microgrids,” IEEE Power Electronics Magazine, vol. 3, no. 4, pp. 42-48, Dec. 2016.
704
Powered by TCPDF (www.tcpdf.org)