Determine Battery Capacity Basing on the Distribution of Power Flows in Hybrid Photovoltaic and Wind Generations Operating in Half-Isolation Mode

Determine Battery Capacity Basing on the Distribution of Power Flows in Hybrid Photovoltaic and Wind Generations Operating in Half-Isolation Mode

Le Tien Phong

Electrical Faculty

Thai Nguyen University of Technology Thai Nguyen city, Viet Nam

Ngo Duc Minh

Electrical Faculty

Thai Nguyen University of Technology Thai Nguyen city, Viet Nam

AbstractHybrid models exploiting photovoltaic and wind generations are combined with energy storage to overcome disadvantages of each other. This paper mentions the distribution of power flows through power converters to make operating plans for each element. It can be done by using known data collected or predicted and applied in half-isolation mode for whole system. This paper also proposes a new program to determine the optimal capacity for DC storage using battery. The operation of whole system is simulated in Matlab software accounting for random characteristic of input data. Experimental

grid problem and the available value for lack of power Plp is not limited.

Wg

PVg

Wg

Electric grid

results represented the ability of controlling power flows in a photovoltaic system.

PVg

Energy

storage

KeywordsBattery, distribution of power flows, energy management system, optimal battery capacity, photovoltaic generation, power converter, smart grid, wind generation.

1. INTRODUCTION

   Photovoltaic generation (PVg) and wind generation (Wg) are potential energy types to replace traditional sources. Because of advance in producing, PVg has enhanced its efficiency and reduced cost price whereas Wg has some silent types to install at any place in residental zone and used as distributed source.

   In smart grid, power converters can help us operate hybrid generations flexibly in isolation or half-isolation mode to have some support from grid. One case of operating in half-isolation

   Fig. 1. Connect islands in the smart grid.

   This paper proposes a new algorithm to determine C basing on collected or predicted input data (solar irradiance, wind speed), load diagram in periodic time T and execute the distribution of power flows at each time for all elements in the system operated in half-isolation mode and limited Plp.

2. SYSTEM STRUCTURE AND DISTRIBUTE POWER FLOWS IN WHOLE SYSTEM

   1. System structure

      A system operates in half-isolation mode often has DC parallel structure as Fig. 2 [7-9].

      mode, Fort Collins – Colorado State (America) is executing a special energy policy consume energy less than generate by using the following ways: photovoltaic systems are located on roofs and gardens, wind turbine, thermal and electrical energy storage, etc, and saving energy programs. Separating the grid into many small islands represented in Fig. 1 will become a good growth trend to keep our existence on the earth [1]. In this system, battery storage is used to overcome random characteristic of input data (solar irradiance, wind speed, etc). Two criterions, enough capacity to absorb all energy from

      Sun

      PVg

      Battery

      DC/DC1h

      Converter

      DC/DC2h

      Converter

      EMS

      DC

      load

      DC/AC2h

      Converter

      APS

      ~

      AC

      load

      source or enough to provide for load power, can be often applied to determine battery capacity C for an individual generation but cant be applied for hybrid generations [2].

      Wind ~

      Wg

      AC/DC1h

      Converter

      DC/DC1h

      Converter DCbus

      ACbus

      References [3-6] are introduced an approximative method to calculate the distribution of power flows in hybrid systems for all periodic time T (day, month, year) and execute an economic problem to compare cost price between purchasing and selling electricity but it only can be applied in completely connected

      Fig. 2. System structure.

      Where:

      APS (Auxiliary Power Source), such as grid, diesel generators, etc, provides power in a limited range because of

      DC load

      DC load

      AC load

      AC load

      PVg

      DC/DC1h

      Battery

      the ability of grid, other systems or rated power of generators.

      DC/DC1h and AC/DC1h converters are used to control the operation point for PVg and Wg.

      AC/DC1h

      Wg DC/DC1h

      DC

      bus

      DC/AC2h

      AC

      bus

      APS

      DC/DC2h and DC/AC2h are are bidirectional converters used to control charging or discharging mode for battery and generate power to ACbus and connect to APS.

      EMS (Energy Management System) collects all

      PVg

      1. Scenarios 1

         DC load

         DC load

         AC load

         AC load

         Battery

         DC/DC2h

         information about load diagram, solar irradiance, wind speed and instant battery capacity to calculate power flows having to operate for all elements in periodic time T [10-12].

         DC/DC1h

         AC/DC1h

         Wg DC/DC1h

         DC

         bus

         DC/AC2h

         AC

         bus

         APS

         To limit power from APS, battery has to have enough large capacity to adapt technical requirements but not allow abundance because of high price and low working life.

   2. Distribute power flows in whole system

   PVg

   Battery

   1. Scenarios 2

      DC load

      AC load

      DC load

      AC load

      DC/DC2h

      Power flows in whole system are executed by the following criterions [3-6]:

      • Power from generations is prior to provide load.

        DC/DC1h

        AC/DC1h

        Wg DC/DC1h

        DC

        bus

        DC/AC2h

        AC

        bus

        APS

      • Battery operates in discharging mode when generations dont provide enough power for load or in charging mode when load cant absorb power from generations

   2. Scenarios 3

      DC load

      DC load

      AC load

      AC load

      Battery

      whereas instant capacity Cins is at admittable limitation (Cmin=0.2C Cins C).

      PVg

      DC/DC1h

      DC/DC2h

      • APS isnt used to charge battery or aborb power from battery. The value of lack of power from APS is always less than admittable value Plpcp (PlpPlpcp).

        AC/DC1h

        Wg DC/DC1h

        DC

        bus

        DC/AC2h

        AC

        bus

        APS

        Because of above criterions, power flows in whole system can be clasified into the following scenarios:

        Scenarios 1: generations provide enough power for load by themselves. In this case, load doesnt use power from battery and APS

        PVg

        Battery

        DC/DC1h

   3. Scenarios 4

      DC load

      AC load

      DC load

      AC load

      Scenarios 2: Abundunt power from generations after providing for load power is used to charge battery

      AC/DC1h

      Wg DC/DC1h

      DC

      bus

      DC/AC2h

      AC

      bus

      APS

      Scenarios 3: Generations dont provide enough power for load. Battery operates in discharging mode.

      Scenarios 4 and 5: Generations dont provide enough power for load. Load consumes power from both battery and APS.

      AC load

      AC load

      PVg

      Battery

      DC/DC1h

      AC/DC1h

   4. Scenarios 5

   DC load

   DC load

   AC

   APS

   Scenarios 6 and 7: Generations dont provide enough power for load. Load consumes power from only APS.

   Wg DC/DC1h

   DC

   bus

   DC/AC2h

   bus

   Power flows in whole system are represented clearly in Fig.

   1. Where, DC/DC1h, DC/DC2h, AC/DC1h, DC/AC2h are efficiencies for converters.

      PVg

      DC/DC1h

      e. Scenarios 6

      DC load

      AC load

      AC load

      Battery

      AC/DC1h

      Wg DC/DC1h

      DC

      bus

      DC/AC2h

      AC

      bus

      APS

      g. Scenarios 7

      Fig. 3. Power flows in whole system.

3. DETERMINE BATTERY CAPACITY

   1. Proposed algorithm

      Value of lack of power in scenarios 6 and 7 can be determined by Kirhoff law 1:

      In preliminary calculation, the value of power from generations can be received by a simple model basing on solar irradiance S, start-up wind speed vcut-in, cut-out speed vcut-out,

      Plp (i) Pload(i) PG (i)

      rated speed vr, rated power PrPVg and PrW at standard conditions.

      • For PVg

        Using half-fold diagram to calculate exact power at each instant time [13] or estimate power approximately [14], [15]:

        Basing on preliminary calculation, battery capacity C can

        be determined by a proposed algorithm in Fig. 4. New ideals for this algorithm is that only applies in half-isolation mode limiting received power from the grid and battery capacity step

        C will be increased immediately if current capacity doesnt adapt to requirements.

        P P . S

        PVg rPVg 1000

        For Wg [3], [4]

        Collect data about irradiance, wind, load

        Collect data about irradiance, wind, load

        0 when v vcutin or v vcutout

        Start

        Calculate PPVg(i), PWg(i), PG(i)

        Calculate PPVg(i), PWg(i), PG(i)

        3

        P P . v vcutin when v v v

        Wg rW v v

        cutin r

        r cutin

        Choose intial value C

        Choose intial value C

        PrWg

        when vr v vcutout

        At ith time, total power PG(i) generates from generations:

        PG (i) PPVg (i) PWg(i)

        C=C+C N

        i=1

        i=1

        i=i+1

        i=i+1

        PG(i)=Pload(i) Y

        N

        Basing on data about DC and AC load, total equivalent load at ith time:

        Plp(i)Plpcp

        Y

        Y

        Scenarios 2:

        PG(i)>Pload(i)

        Scenarios 1

        Scenarios 1

        P (i) P

        (i) PloadAC(i)

        Battery can N

        Scenarios 3: Battery can discharge

        to Cp(i)

        Scenarios 3: Battery can discharge

        to Cp(i)

        load

        loadDC

        DC / AC2h

        charge to Cn(i)

        Calculate Plp(i)

        Calculate Plp(i)

        For battery, battery capacity can charge to Cn(i) or discharge to Cp(i) at ith time can be determined by time step t [3-6], [16]:

        Scenarios (4 ÷7)

        Scenarios (4 ÷7)

        Cins(i+1)

        =Cins(i)

        Cins(i+1)

        =Cins(i)

        Y

        Cn(i)>C

        C (i)

        Pload PG (i).t

        N Y

        Cp(i)<Cmin

        p Cins(i

        1).(1

        )

        DCDC2h

        VDCbus

        Cins(i+1)=Cn(i) N

        C (i)

        Pload PG (i).t

        C (i+1)=C

        Cins(i+1)=Cp(i)

        n

        Where:

        Cins(i

        1).(1

        )

        VDCbus

        . DCDC2h

        ins

        min

        Y

        i<=24

        S

        Cins(i-1) is instant battery capacity at (i-1)th time.

        Value of lack of power consuming from APS in scenarios 4 and 5 can be determined by Kirhoff law 1:

        Stop

        Fig. 4. Proposed algorithm to determine battery capacity.

   2. Simulate the problem determining battery capacity and

   Plp (i) Pload(i) PG (i) Pb (i)

   distributing power flows

   Periodic time T=24 [h], time step t=1 [h], battery capacity

   Discharging capacity from battery to DCbus can be determined by:

   step C=10 [Ah], admittable value of power Plpcp = 1600 [W].

   Generations: Rated power of Wg is 5000W and PVg is 1360W. Generating diagrams of PVg and Wg are represented

   Pb (i) (Cins(i) Cins(i 1)).VDCbus

   in Fig. 5.

   5000

   4500

   4000

   3500

   Power [W]

   Power [W]

   3000

   2500

   2000

   1500

   1000

   500

   0

   Fig. 8 represents time ranges for charging and discharging battery, where positive power presents charging power and negative power presents discharging power.

   PVg

   PVg

   Wg

   Wg

   Charging capacity

   Discharging capacity

   Charging capacity

   Discharging capacity

   200

   150

   Capacity of Battery [Ah]

   Capacity of Battery [Ah]

   100

   50

   0

   ]

   ]

   0 4 8 12Time [h 16 20 24

   Fig. 5. Generating diagrams of PVg and Wg.

   DC and AC load diagrams are represented in Fig. 6.

   4000

   3500

   -50

   -100

   -150

   0 4 8 12 16 20 24

   Time [h]

   Fig. 8. Charging and discharging battery.

   3000

   Power [W]

   Power [W]

   2500

   2000

   DC load

   AC load

   DC load

   AC load

   1500

   1000

   500

   0

   0 4 8 12 16 20 24

   Time [h]

   Fig. 6. DC and AC load diagrams.

   Apply proposed algorithm in Fig. 4 for above data, suitable battery capacity resulting from program is 200 [Ah]. Use some local values of above suitable battery capacity to recalculate maximum value of Plp and abundant power Pabundant of generations. Results in TABLE. I show that 200 Ah is the most value for above input data.

   Connect to computer, display

   Connect to computer, display

   TABLE I. RECALCULATE PLPMAX AND PABUNDANT FOR SOME LOCAL VALUES OF SUITABLE BATTERY CAPACITY

   In Fig. 8, the battery operates in discharging mode in time range (0÷1), (2÷3), (6÷7), (9÷12), (17÷18)h and charging mode

   in time range (1÷2), (3÷6), (9÷10), (14÷15), (16÷17), (23÷24).

4. EXPERIMENTAL DISTRIBUTION OF POWER FLOWS

   1. Experimental model

      Modules of PVg located in Thai Nguyen university of technology by Phoenix Solar Pte, Singapore are parameters shown in TABLE II. Parameters of battery are shown in TABLE III.

      TABLE II. PARAMETERS OF PVg

      Type	Kyocera KC85/Japan
      Rated power [Wp]	1360 Wp

      TABLE III. PARAMETERS OF BATTERY

      Type	Lead-acid, Dong Nai N200
      Nominal voltage	12 V
      Capacity	200Ah

      Structure model for PVg is represented in Fig. 9, where DC/DC converter is flyback type.

      C [Ah] 180 190 200 210

      Plpmax[W] 1582.6

      Sun

      Iref

      I + e

      Pulse

      EMS

      Pabundant [W] 117,1 67,5 0 0

      out –

      generator Control

      signal

      Power generating from generations and battery, load power, lack of power Plp diagrams corresponding C=200Ah are represented in Fig. 7.

      Charge battery Battery

      PVg

      Battery

      DC/DC1h

      converter

      DC

      load

      DCbus

      5000

      Power [W]

      Power [W]

      4000

      3000

      Load

      PVg+Wg Plp

      1. Control structure

         Control signal

         Control signal

         Iout

         2000

         1000

         *

         *

         C1

         C1

         D

         C2

         D

         C2

         PVg Battery,

         *

         *

         load

         0

         0 4 8 12 16 20 24

         Time [h]

         Fig. 7. Power generating from generations and battery, load power, lack

         of power diagrams.

      2. DC/DC converter type flyback Fig. 9. Structure model for PVg.

         Experimental model for PVg is represented in Fig. 10. Where, load block has 4 compact lights (4x33W) and each light can be switch on or off invidually.

         er

         er

         Collect data and controller

         Results of TRACKING PV SYSTEM PROGRAM in 27

         minutes are represented in Fig. 12. Where, the red curve represents input power and the green curve represents output power of the converter.

         PVg

         DC

         load

         Battery

         Control cabinet

         Comp

         Control cabinet

         Comp

         ut

         Fig. 10. Experimental mode.

   2. Experimental results

      Reference current Iref is installed to controller by buttons on board or computer. Maximum power point tracking in control program is only actived when output current of converter Iout is less than Iref. Operation parameters can be observed online on display panel of control board and program TRACKING PV SYSTEM PROGRAM in computer using Visual Studio software version 10. At the end of observation process, the program will create a file collected data about the operation process.

      Experimental time: from 8p5 to 8h43 on 11 September 2014, battery voltage is 12,8V. Operation scenarios are shown in TABLE IV.

      Operation time	Operation scenarios
      From 8p5 to 8p1	Iref =16A, PloadDC=0
      From 8p1 to 8p7	Iref =10A, PloadDC=0
      From 8p7 to 8p3	Iref =10A, PloadDC=33W
      From 8p3 to 8p7	Iref =10A, PloadDC=0
      From 8p7 to 8h43	Iref =17A, PloadDC=0

      Operation time	Operation scenarios
      From 8p5 to 8p1	Iref =16A, PloadDC=0
      From 8p1 to 8p7	Iref =10A, PloadDC=0
      From 8p7 to 8p3	Iref =10A, PloadDC=33W
      From 8p3 to 8p7	Iref =10A, PloadDC=0
      From 8p7 to 8h43	Iref =17A, PloadDC=0

      TABLE IV. EXPERIMENTAL SCENARIOS

      Fig. 12. Results of observation program.

      Results from control program are shown:

      • Converter only exploited power from PVg in time ranges (8p5÷8p1), (8p1÷8p7), (8p3÷8p7) and (8p7÷8h43) to charge battery. After changing Iref, input and output power curves of the converter changed pulse to track energy requiments.

      • From 8p7 to 8p3, Iref was hold in constant by 10A whereas added a 33W light load at 8p7. Immediately after adding load, EMS hadnt changed pulse so input and output power curve of PVg was built-up corresponding 33W. After transient time, EMS changed pulse control to take output current converter to Iref and power charging battery curve dropped corresponding 33W. It showed that it has amount of power from PVg shared partial power into light, partial power into battery.

      Power flows controlled in all operation modes of experimental model are represented in Fig. 13.

      Iref, current and voltage output converter diagrams are repr ted in Fig. 11.

      Sun

      PVg

      Power flow from PVg DC/DC

      Converter

      DC

      load

      esen

      17

      Battery

      DCbus

      Iref [A]

      Output current of converter [A]

      Iref [A]

      Output current of converter [A]

      Output voltage of converter [V]

      Output voltage of converter [V]

      16

      15

      14

      Add DC load

      13

      12

      Sun

      1. Charging battery and not adding load

         Power flow from PVg DC/DC

         converter

         PVg

         DC

         load

         11 Battery

         10

         DC bus

         9

         0 3 6 9 12 15 18 21 24 27

         Time [minute]

      2. Both charging battery and adding load

   Fig. 13. Distribution of power flows in experimental model corresponding to

   scenarios.

   Fig. 11. Iref, current and voltage output converter diagrams corresponding

   scenarios.

5. CONCLUSION

Operating hybrid generations in half-isolation mode to receive support from grid but not depending much on the grid can provide power for load in site and take some benefits such as reducing transmition power loss and pressure on the national power system. This is also the future model for all power systems.

Experimental model showed that converters can help us control power flows between all elements flexibly and adapt to many different requirements.

Proposed algorithm determining battery capacity basing on distribution of power flows in hybrid generations in half- isolation mode can help system adapt to load more actively. The program basing on proposed algorithm collects multiple sets of data to evaluate the best value for battery or optimal battery capacity.

When input data (wind speed, solar irradiance) can be predicted, above program will help operator calculate distribution of power flows for each element or make operation plan for whole system.

REFERENCES

1. P. Nguyen, Distribution grid, Jounal of Infomation and Technology (STINFO), no. 9, 2014.

2. Gilbert M. Masters, Renewable and Efcient Electric Power Systems, John Wiley & Sons, Inc., Hoboken, New Jersey, ISBN 0-471-28060-7, 2004.

3. Jong Hwan Lim, Optimal Combination and Sizing of a New and Renewable Hybrid Generation System, International Journal of Future Generation Communication and Networking, Vol. 5, No. 2, June, 2012.

4. O. Gergaud, G. Robin, H. Ben Ahmed, M. Multon, Economic Formalusm for Optimizing the design and Energy Management of a Hybrid Wind/Photovoltaic System, International Conference on Renewable Energies and Power Quality, Vigo (Spain), 2003.

5. Said. Diaf, Djamila Diaf, Mayouf Belhamel, Mourad Haddadi, Alain Louche, A methodology for optimal sizing of autonomous hybrid PV/wind system, Energy Policy, Elsevier, 2007.

6. Satish Kumar Ramoji, B.Jagadish Kumar, Optimal Economical sizing of a PV-Wind Hybrid Energy System using Genetic Algorithm and Teaching Learning Based Optimization, International Journal of Advanced Research in Electrical, Electronics and Instrumentation Engineering, 2014.

7. Tushar Saxena, J.S.Jadon, Different Power Supply Planning Options Available for A BTS Site, International Journal of Engineering, Business and Enterprise Applications (IJEBEA), 2013.

8. Ziyad Salameh, The Impact of Renewable Energy Sources on Utilities, International Journal of Power and Renewable Energy Systems (IJPRES), Volume 2 Issue 2, 2015.

9. Huynh Quang Minh, Ngo Cao Cuong, Tran NguyenChau, A fuzzy-logic based MPPT method for stand-alone wind turbine system, American Journal of Engineering Research (AJER), e-ISSN : 2320-0847, p-ISSN : 2320-0936, Volume-3, Issue-9, 2014.

10. Caisheng Wang, M. Hashem Nehrir, Power Management of a Stand- Alone Wind/Photovoltaic/Fuel Cell Energy System, IEEE Transactions on Energy Conversion, Vol. 23, No. 3, 2008.

11. M. F. Almi, M. Arrouf, H.Belmili, S. Bouloumaa, B. Bendiba, Energy management of Wind/PV and Battery Hybrid System, International Journal of New Computer Architectures and their Applications (IJNCAA), 2014.

12. Nabeel Ahmad, Hybrid power system With Smart EnergyManagement System, Proceedings of the 2nd International Conference on Engineering & Emerging Technologies (ICEET), 2015.

13. Le Tien Phong, Ngo Duc Minh, Nguyen Van Lien, A New Method to Identify Maximum Power Point for Photovoltaic Generation, IEEE International Conference on Communations, Management and Telecommunations, 2015.

14. Mohammad Reza Javadi , Kazem Mazlumi, and Abolfazl Jalilvand, Application of GA, PSO and ABC in Optimal Design of a Stand-Alone Hybrid System for North-West of Iran, ELECO 2011 7th International Conference on Electrical and Electronics Engineering, 1-4 December, Bursa, Turkey, 2011.

15. Yann Riffonneau, Seddik Bacha, Franck Barruel, Stephane Ploix, Optimal Power Flow Management for Grid Connected PV Systems With Batteries. IEEE Transactions on Sustainable Energy, Vol. 2, No. 3, 2011.

16. O. Gergaud, G. Robin, H. Ben Ahmed, Energy Modeling of A Lead- Acid Battery within Hybrid Wind/Photovoltaic Systems, European Power Electronic Conference, 2003.

Le Tien Phong, received the M.Sc. degree in 2010 in Electrical Engineering from HaNoi University of Technology and Science and working in Thai Nguyen University of Technolgy now. His research interests include renewable energy in power system, control electrical energy conversions, control of power flow in power system.

Ngo Duc Minh, received the PhD. degree in Automation from Ha Noi University of Technology and Science in 2010 and working in Thai Nguyen University of Technolgy now. His research interests include active rectifier, active filter, FACTS BESS, D-STATCOM, Control of power system, distribution grid, renewable energy in power system.