solar pv
Content
1
Implementing Frequency Regulation Capability in a Solar Photovoltaic Power Plant
Venkata Ajay Kumar Pappu, Student Member, IEEE, Badrul Chowdhury, Senior Member, IEEE and Ravi Bhatt, Student Member, IEEE
system is to maintain a balance between demand and supply of active power at any given point of time. Any difference between these two quantities directly impacts the system frequency, which is 60Hz or 50Hz in most countries. Traditional approach to achieve frequency regulation is to use fossil fuel and hydro reserves to generate electrical energy [3]. However the method depends on the geographical location and is not very efficient. In the special case when the microgrid is comprised of many other Distributed Energy Resources (DER) that are renewable in nature, load following becomes a critical function needed to maintain frequency and voltage of the microgrid [4]. This paper discusses a method to provide reserve power capability to the PV systems so that they can participate in frequency regulation in a microgrid environment. Energy storage systems are promising technologies which may work in compliance with PV systems to regulate frequency. II. OBJECTIVE In conventional power systems, the action of the prime mover allows the system to gain control over frequency [5]. Fig.1 shows the droop line control of a power plant. For a change in frequency from f1 to f2, there is a corresponding increase in prime mover power from P1 to P2 [6].
Abstract-- Photovoltaic power plants pose some challenges when integrated with the power grid. The PV plants always focus on extracting the maximum power from the arrays. This makes the PV system unavailable for helping in regulating the grid frequency as compared to conventional generators. The objective of this paper is to introduce a pseudo power point tracking which provides frequency regulation functionality to PV systems. This method is fast and helps in creating a reserve which can be utilized for frequency stability. Control technique discussed in this paper allows the control over the amount of active power injected into the grid. This paper also discusses the energy storage option for frequency regulation. Index Terms—Frequency regulation, reserve power, pseudo power point tracking, modified fractional open circuit voltage algorithm, online search algorithm.
I. INTRODUCTION n ever-increasing demand for energy and environmental concerns in the 21st century has led to a sustained effort to generate power from various alternative energy resources. Although coal and hydro power plants are still used to produce electric power on a large scale, renewable energy has been of primary focus in recent years due to their abundance. One of many possible scenarios indicates that the growth of the PV system by the year 2030 would reach 200GWp and provide 7% of the total U.S. electricity [1]. The rate at which photovoltaic (PV) power generation is projected to increase is promising in many aspects. Photovoltaic systems form an important part of a renewable energy generation portfolio since they are pollution free when operating, modular in nature which makes construction easier, and have relatively longer life. PV generation has increased on a large scale especially in the US and Europe, with largest plant rated at 60MW at Olmedilla in Spain [2]. If proper control technique is available such large plants can work hand in hand with the utility grid to increase the system reliability. The ultimate goal of any power
This work was supported in part by the Intelligent Systems Center of Missouri University of Science & Technology (Missouri S&T). V. A. K. Pappu is with the Department of Electrical & Computer Engineering, Missouri S&T, Rolla, MO 65409, USA (e-mail: [email protected]). B. H. Chowdhury is with the Department of Electrical & Computer Engineering, Missouri S&T, Rolla, MO 65409 USA (e-mail: [email protected]). R. Bhatt is with the Department of Electrical & Computer Engineering, Missouri S&T, Rolla, MO 65409 USA (e-mail: [email protected]).
A
Fig.1. Droop line control of a power plant
In this paper, control over the active power of the PV plant has been developed. With the suggested technique, PV systems can reserve variable amounts of power to respond to a drop in frequency. This enables load following capability along with frequency regulation. Thus, voltage and frequency management is possible even under islanded operating conditions.
2
III. DESCRIPTION OF THE MODEL A. Overview The system being modeled is shown in Fig 2. The full system is implemented in a two-stage power conversion, which is connected to the grid on the inverter side. It has two independent control loops for the dc-dc converter and the single-phase dc-ac inverter. The raw power generated from the solar array is tracked using two separate algorithms integrated into the Maximum Power Point (MPP) controller. The Online search/Perturb and Observe algorithm [7], [8], [9], is used for the true MPP and the modified fractional Voc method is used for the pseudo MPP. This controls the duty cycle of the boost converter to maintain a constant voltage at the dc bus link.
that has to be tracked by the system. The difference in the two powers that are drawn from the solar array will provide the reserve power P1-P2.
P P1 Online Search algorithm Fractional Voc algorithm
Power
P2
V Voltage Fig. 3. Combination of algorithms for tracking two different operating points on the P-V Curve
The on-line search algorithm to track the point P1 which represents maximum available power is shown in Fig. 4 [10]. The operating current and voltage of the array are constantly monitored and the difference in these variables is checked at every instant of time. The difference between the present and previous values of power is checked to determine the sign of the slope on the power characteristic curve.
Fig. 2. Topology of the two stage power conversion
For the above system a boost converter is used for stepping up the PV array voltage, to integrate it with a single-phase inverter and the grid. The inverter is controlled by a feedback loop, which injects the power into the grid at unity power factor. The two-stage conversion topology provides effective control over the power transfer and gives flexibility to implement storage devices at different stages. B. PV Array The open source PV array model developed by the University of Colorado has been used as a reference to develop this temperature dependent solar array in Matlab/Simulink [11]. The model is dependent on temperature and insolation to which the array is exposed. For the simulation purpose, a GE-PV 200W solar panel is used to build a 2kW PV system. To get a 2 kW output, two parallel strings with five panels in each string are used to generate a MPP voltage of (26.3*nS) 131.5 V and a MPP current of (7.6*nP) 15.2 A, where nS is the number of panels in series and nP is the number of strings in parallel. C. MPPT Controller Traditional MPP controllers draw the maximum available power from the array. However in this paper; a pseudo-MPP tracking is introduced. The objective is to have reserve power available, to be utilized for frequency regulation. To achieve this, a combination of two different MPPT algorithms is used as shown in Fig. 3. The point P1 in the figure represents the maximum possible power that can be drawn for a given insolation, whereas the point P2 represents the false or pseudo maximum power point
Fig. 4. On-Line search MPP algorithm
The slope thus calculated is fed as an error to the reference voltage. The PI controller uses this reference voltage to generate switching pulses for the boost converter. A new reference value indicates that the operating point has moved towards the MPP. The slope and its sign determine the direction that the search algorithm tracks the MPP. This process is repeated until the slope is zero, i.e., the MPP is found.
3
To get the reserve power, one has to trac a power point ck below the optimal power point for the a array. Hence, the modified fractional Voc method, shown in Fig. 5, has been circuit method the adopted. In the conventional fractional open c maximum power point voltage is represented as a fraction k of d the open circuit voltage of the array given by (1): (1) wer where k is the ratio of Vmpp (maximum pow point) and and Voc (open circuit voltage). The ratio k varies between 0.71 to s 0.78 [12]. At every instant, the open circuit voltage has to be measured and the maximum power point is tracked with the most current value of Voc. and Voc. The m methodology was modified for this application such that the fra action k resides in the range 0.8 to 0.95. The ratio k in the m modified algorithm represents the controlling variable for the a amount of reserve power from the PV plant.
control blocks are the change in the reference voltage value. The difference between Vref and Vpv is the error that is fed to v the PI block which decides the sw witching function for the converter, thus delivering the associa active power. ated D. DC-DC Boost Converter and Inv verter A two-stage converter topolog has been used. Singlegy phase boost converter is in series with an inverter has better w control over the amount of power be eing injected into the grid. Pulse width modulation (PWM) is used for switching the boost converter. The switching is do by using a carrier wave one with a frequency of 30 kHz that take the control voltage input es from the MPP controller. The inv verter has a closed loop control which takes the grid voltag current and the dc bus ge, reference voltage [13] as inputs and controls the gate signals d to transfer the power at unity pow factor. The system is wer designed using PLECS and is shown in Fig. 7. n
V mpp = k × V occ
Fig. 7. DC-AC power conv version system
The output voltage of the boost co onverter is 260Vdc and this is converted to 260Vac p-p using th single-phase inverter. A he transformer is used to step up the voltage to the grid voltage of tance, equal to 1000µF, is 340V peak. A high value of capacit used at the dc output to maintain a stiff voltage and to reduce the ripple to about 1.5% (3 to 4V) at the terminals. The simulink model of the inverter is designed by research partner Luke Watson. The inverter also makes use of the PWM hes technique to drive all the four switch of the inverter. E. Simulation and Results An analysis and study of the comp plete system is done under various scenarios and presented as different cases. Each case nk simulation is done in Matlab/Simulin R2008b. Case 1: Insolation is 800 W/m2 and temperature is 25oC. t The system is checked for pseud and true power modes do and the outputs at the grid are analyzed when the insolation and the temperature are constant. For the pseudo maximum F power mode, the ratio k in the modif fractional open circuit fied voltage algorithm is 0.95. This se election is arbitrary as at k=0.95, the controller will track the point on the right hand n side of the PV characteristics shown in Fig. 3. If the ratio k is smaller, the controller will track a point on the left part of the p characteristics. Any value below 0. (left of MPP) or above .7 0.8 (right of MPP) would help. Fig 8 shows the variation in g. the PV voltage and current corresponding to the mode change. The change in mode is triggered by an external frequency change signal at t=0.82 sec.
Fig. 5. Modified fractional open circuit voltag algorithm ge
Both algorithms are implemented as pro ogram modules in the Simulink block as shown in Fig. 6.
Fig. 6. MPPT controller block with embedded Matlab code
As the system frequency changes, an external signal n triggers the change in the algorithm from the true MPP to of pseudo MPP and vice-versa. The outputs o both algorithm
4
Voltage (Volts)
Current (Amp)
Current (Amp) Time (sec) Fig. 10. Current injected into the grid at 1490 W (True MPP) Time (sec) Fig. 8. PV Voltage and Current Current (Amp)
Fig.9 shows the variation in the power level. At t=0.82 sec, the average output power of the PV array has decreased from 1.49 kW to 400 W as a result of a commanded change representing a drop in load.
Power (Watts)
Time (sec) Fig. 11. Current injected into the grid at 400W (Pseudo MPP)
Figs. 10 and 11 show the current injected into the grid for the true MPP and the pseudo MPP respectively. The current decreases from 8.5 A to 2.3 A with the change in the operating mode. One can conclude that if the grid frequency increases, the amount of active power being injected into the grid can be decreased instantaneously by changing the MPP mode. Figs. 12 and 13 show the power output at the inverter terminals for true and pseudo MPP respectively for Case 1. The power is injected into the grid at unity power factor. By controlling the phase angle of the current, the reactive power can be controlled; this is discussed later in the paper.
Power (Watts)
Time (sec) Fig. 9. Change in Power output from 1490 W to 400 W.
Time (sec) Fig. 12. Power (average: 1490W) injected into the grid for true MPP.
5
Power (Watts)
Figs. 15 and 16 show the variation in the power output when there is a non-uniform rise and fall of the insolation level. For a very short duration at the start of the simulation, the power level is zero, as the MPPT takes some time to start tracking the power and reach the operating point.
Time (sec) Fig. 13. Power (average: 400W) injected into the grid for pseudo MPP
Fig.14 shows the boost converter’s output voltage for the entire operating time of the system. Although there is a ripple of 4V, the average value of the output is 260V. There is a spike in the voltage waveform at t = 0.85 sec due to the change in the operation mode of the system; however it settles down in less than 0.3 sec to an average value of 260V.
Voltage (Volts)
Time (sec) Fig. 14. Boost converter output voltage
Case 2: Varying Insolation and temperature is 25oC. The insolation data for Case 2 is shown over a period of 1 sec in Table I.
TABLE I Varying Insolation Time(s) Insolation(W/m2) 0 200 0.1 450 0.2 650 0.3 600 0.4 400 0.5 700 0.6 900 0.7 550 0.8 850 0.9 620 1 450
One can conclude from Figs. 15 and 16 that the controller is able to create a reserve margin even for variable insolation levels. Frequency regulation can be achieved throughout the day limited only by the amount of power available from the PV plant. Case 3: Insolation is 1000 W/m2; temperature is 25oC and injecting power at 0.8 pf leading. Fig. 17 shows the variation in the active and reactive power magnitudes for the high and low power operations at 0.8 power factor leading. By changing the phase angle of the
Insolation (W/m2) Time (sec) Fig. 16. PV power output under varying insolation for the true MPP
Power (Watts)
Insolation (W/m2) Time (sec) Fig. 15. PV power output under varying insolation for pseudo MPP
Power (Watts)
6
reference current in the inverter block, the am mount of reactive power injected into the grid can be controlled d.
[2] [3] [4]
[5]
[6]
[7] [8] [9] [10]
Time (sec) Fig.17. Active and reactive power magnitudes at 0.8 p leading on the grid pf side
IV. CONCLUSION AND FUTURE WORK A new technique to implement freq quency regulation capability in solar photovoltaic power plant i discussed. This is method allows effective control of solar p power for various purposes without compromising the efficien of conversion. ncy A system comprising of a 2 kW PV ar rray, a two-stage converter and the grid was modeled, a and simulated in MATLAB. Results show that based on solar i insolation, the PV plant system can be given a reserve power m margin and active power injected into the grid can be effectively controlled. With pseudo power point tracking, the activ power flowing ve into the system can be balanced in order to regulate the r frequency of the system. The reserve power technique is very useful for fast response actions which last fo few seconds to or several minutes. m Work is going on in getting help from energy storage systems, such as batteries, to make the r renewable energy sources more dispatchable. Battery ener rgy storage, for example, is capable of providing spinnin reserves, load ng leveling, frequency control, VAR support, v voltage regulation, etc. [14]. However depth of discharge, life c cycle expectancy, self discharge rate and installation cost a some of the are challenges batteries are facing today and further study is necessary. A detailed and comparative study is also required y in areas such as the type of the battery t technology to be selected, sizing and location. V. ACKNOWLEDGMENT n The authors are thankful to Dr. Jonathan W. Kimball for his support and helpful discussions during t research. The the authors gratefully acknowledge the help of PhD candidate, f Luke D. Watson of the ECE department for his design of the control block for the inverter. VI. REFERENCE
[1] P. Denholm, R. M. Margolis, K. Zweibel, “Tackling Climate Change in n 2007 [Online] the U.S.,” ases.org, page 98, Jan
[11]
[12] [13] [14]
Available:http://www.ases.org/images/s stories/file/ASES/climate_change .pdf, [Accessed: Jan 12,2010]. D. Lenardic, “Large Scale Photovolta Power plants”, May 31,2010 aic [Online] Available: http://www.p pvresources.com/en/top50pv.php, [Accessed: June 06,2010]. M.L. Lazarewicz, A. Rojas, "Grid fre equency regulation by recycling electrical energy in flywheels," Proc. IEEE Power Engineering Society I General Meeting, pp.2038-2042, vol.2, June 2004. K.D. Brabandere, K. Vanthournout, D.J. Deconinck, G.R. Belmans, D 2007, “Control of Microgrids,” Proc. IEEE Power Engineering Society I General Meeting, pp. 1 – 7. 2007. I Serban, C.P Ion, C Marinescu, M Geo orgescu, “Frequency Control and Unbalances Compensation in Autonom mous Micro-Grids Supplied by RES,” IEEE International Electric Mac chines & Drives Conference, vol. 1, pp. 459 – 464, May 2007. J S. Wijnbergen, S.W.H. de Haan, J.G. Slootweg, “A System for Dispersed Generator Participation in Voltage Control and Primary Frequency Control of the grid,” Proc 36th IEEE Power Electronics c. Specialists Conference, pp. 2918 – 2924 June 2005. 4, R. Ito, Y. Matsuzaki, T. Tani,; T. Yach “Evaluation of performance of hi, MPPT equipment in photovoltaic sys stem,” Proc. 25th International Telecommunications Energy Conference pp. 256 – 260, Oct 2003. e, T. Esram, P.L. Chapman, “Comparison of Photovoltaic Array Maximum power Point Tracking Techniques,” IEEE Transactions on Energy e Conversion, Vol 22, pp. 439 – 449, June 2007. K.A. EL-Serafi, A.E. Kalas, M. H. El lfar, “On-Line Maximum Power Tracking In a Stand Alone Photo ovoltaic System,” Suez Canal University, Faculty of Engineering, 2003. Y-E. Wu, C-L. Shen, C-Y. Wu, “R Research and Improvement of Maximum Power Point Tracking for Photovoltaic Systems,” Proc. 8th P IEEE Int. Conference on Power Electr ronics and Drive Systems (IEEE PEDS, Taipei, Taiwan, R.O.C.), pp.1308 8-1312, 2009. ECEN 2060 Renewable Sources an Efficient electrical Energy nd Systems, University of Colorado at Boulder [Online] Available: http://ecee.colorado.edu/~ecen2060/energyprogram.html, [Accessed: DEC 14,2008]. E K.A. EL-Serafi, A.E. Kalas and M. H. Elfar, “On-Line Maximum Power Tracking in a Stand Alone Photovoltaic System,” Suez Canal University, Faculty of Engineering, 2003 N.A Ninad, L.A.C. Lopes, “Operation of Single Phase Grid Connected Inverters with Large DC Bus Voltage Ripple,” Proc. IEEE Canada Electrical Power Conference, EPC, pp. 172 – 176. Advancement of energy storage S.C. Smith, P.K. Sen, B. Kroposki, "A devices and applications in electrical po ower system," Proc. IEEE Power and Energy Society General Meeting - Conversion and Delivery of g Electrical Energy in the 21st Century, vol., no., pp.1-8, 20-24 July 2008. v
Power
VII. BIOGRAP PHIES
Venkata A. Pappu was bo in Ongole, India in 1985. He orn received his degree in Bach helor of Technology in Electrical and Electronics Engineerin from VIT University, Vellore, ng India in 2008. He was a graduate research assistant at Missouri S&T, Rolla, MO His research interests include O. Renewable Energy Generat tion and Power System Stability. Badrul H. Chowdhury (M M’83, SM’93) obtained his Ph.D. degree in Electrical Eng gineering from Virginia Tech, Blacksburg, VA in 1987. He is currently a Professor in the Electrical & Computer Engineering department of the E Missouri University of Sc cience and Technology, formerly known as the Univers sity of Missouri-Rolla. Dr. Chowdhury’s research in nterests are in power system modeling, analysis and control; renewable energy and distributed generation. He is a Senior Member of IEEE. Ravi Bhatt was born in Mumbai, India on June 18, n 1986. He received his Bachelor's degree in Electrical B Engineering from Mum mbai University, India. He is currently pursuing his Master's degree in Electrical ri Engineering at Missour University of Science & Technology, Rolla. Hi research interests are in is renewable energy sources s.
Implementing Frequency Regulation Capability in a Solar Photovoltaic Power Plant
Venkata Ajay Kumar Pappu, Student Member, IEEE, Badrul Chowdhury, Senior Member, IEEE and Ravi Bhatt, Student Member, IEEE
system is to maintain a balance between demand and supply of active power at any given point of time. Any difference between these two quantities directly impacts the system frequency, which is 60Hz or 50Hz in most countries. Traditional approach to achieve frequency regulation is to use fossil fuel and hydro reserves to generate electrical energy [3]. However the method depends on the geographical location and is not very efficient. In the special case when the microgrid is comprised of many other Distributed Energy Resources (DER) that are renewable in nature, load following becomes a critical function needed to maintain frequency and voltage of the microgrid [4]. This paper discusses a method to provide reserve power capability to the PV systems so that they can participate in frequency regulation in a microgrid environment. Energy storage systems are promising technologies which may work in compliance with PV systems to regulate frequency. II. OBJECTIVE In conventional power systems, the action of the prime mover allows the system to gain control over frequency [5]. Fig.1 shows the droop line control of a power plant. For a change in frequency from f1 to f2, there is a corresponding increase in prime mover power from P1 to P2 [6].
Abstract-- Photovoltaic power plants pose some challenges when integrated with the power grid. The PV plants always focus on extracting the maximum power from the arrays. This makes the PV system unavailable for helping in regulating the grid frequency as compared to conventional generators. The objective of this paper is to introduce a pseudo power point tracking which provides frequency regulation functionality to PV systems. This method is fast and helps in creating a reserve which can be utilized for frequency stability. Control technique discussed in this paper allows the control over the amount of active power injected into the grid. This paper also discusses the energy storage option for frequency regulation. Index Terms—Frequency regulation, reserve power, pseudo power point tracking, modified fractional open circuit voltage algorithm, online search algorithm.
I. INTRODUCTION n ever-increasing demand for energy and environmental concerns in the 21st century has led to a sustained effort to generate power from various alternative energy resources. Although coal and hydro power plants are still used to produce electric power on a large scale, renewable energy has been of primary focus in recent years due to their abundance. One of many possible scenarios indicates that the growth of the PV system by the year 2030 would reach 200GWp and provide 7% of the total U.S. electricity [1]. The rate at which photovoltaic (PV) power generation is projected to increase is promising in many aspects. Photovoltaic systems form an important part of a renewable energy generation portfolio since they are pollution free when operating, modular in nature which makes construction easier, and have relatively longer life. PV generation has increased on a large scale especially in the US and Europe, with largest plant rated at 60MW at Olmedilla in Spain [2]. If proper control technique is available such large plants can work hand in hand with the utility grid to increase the system reliability. The ultimate goal of any power
This work was supported in part by the Intelligent Systems Center of Missouri University of Science & Technology (Missouri S&T). V. A. K. Pappu is with the Department of Electrical & Computer Engineering, Missouri S&T, Rolla, MO 65409, USA (e-mail: [email protected]). B. H. Chowdhury is with the Department of Electrical & Computer Engineering, Missouri S&T, Rolla, MO 65409 USA (e-mail: [email protected]). R. Bhatt is with the Department of Electrical & Computer Engineering, Missouri S&T, Rolla, MO 65409 USA (e-mail: [email protected]).
A
Fig.1. Droop line control of a power plant
In this paper, control over the active power of the PV plant has been developed. With the suggested technique, PV systems can reserve variable amounts of power to respond to a drop in frequency. This enables load following capability along with frequency regulation. Thus, voltage and frequency management is possible even under islanded operating conditions.
2
III. DESCRIPTION OF THE MODEL A. Overview The system being modeled is shown in Fig 2. The full system is implemented in a two-stage power conversion, which is connected to the grid on the inverter side. It has two independent control loops for the dc-dc converter and the single-phase dc-ac inverter. The raw power generated from the solar array is tracked using two separate algorithms integrated into the Maximum Power Point (MPP) controller. The Online search/Perturb and Observe algorithm [7], [8], [9], is used for the true MPP and the modified fractional Voc method is used for the pseudo MPP. This controls the duty cycle of the boost converter to maintain a constant voltage at the dc bus link.
that has to be tracked by the system. The difference in the two powers that are drawn from the solar array will provide the reserve power P1-P2.
P P1 Online Search algorithm Fractional Voc algorithm
Power
P2
V Voltage Fig. 3. Combination of algorithms for tracking two different operating points on the P-V Curve
The on-line search algorithm to track the point P1 which represents maximum available power is shown in Fig. 4 [10]. The operating current and voltage of the array are constantly monitored and the difference in these variables is checked at every instant of time. The difference between the present and previous values of power is checked to determine the sign of the slope on the power characteristic curve.
Fig. 2. Topology of the two stage power conversion
For the above system a boost converter is used for stepping up the PV array voltage, to integrate it with a single-phase inverter and the grid. The inverter is controlled by a feedback loop, which injects the power into the grid at unity power factor. The two-stage conversion topology provides effective control over the power transfer and gives flexibility to implement storage devices at different stages. B. PV Array The open source PV array model developed by the University of Colorado has been used as a reference to develop this temperature dependent solar array in Matlab/Simulink [11]. The model is dependent on temperature and insolation to which the array is exposed. For the simulation purpose, a GE-PV 200W solar panel is used to build a 2kW PV system. To get a 2 kW output, two parallel strings with five panels in each string are used to generate a MPP voltage of (26.3*nS) 131.5 V and a MPP current of (7.6*nP) 15.2 A, where nS is the number of panels in series and nP is the number of strings in parallel. C. MPPT Controller Traditional MPP controllers draw the maximum available power from the array. However in this paper; a pseudo-MPP tracking is introduced. The objective is to have reserve power available, to be utilized for frequency regulation. To achieve this, a combination of two different MPPT algorithms is used as shown in Fig. 3. The point P1 in the figure represents the maximum possible power that can be drawn for a given insolation, whereas the point P2 represents the false or pseudo maximum power point
Fig. 4. On-Line search MPP algorithm
The slope thus calculated is fed as an error to the reference voltage. The PI controller uses this reference voltage to generate switching pulses for the boost converter. A new reference value indicates that the operating point has moved towards the MPP. The slope and its sign determine the direction that the search algorithm tracks the MPP. This process is repeated until the slope is zero, i.e., the MPP is found.
3
To get the reserve power, one has to trac a power point ck below the optimal power point for the a array. Hence, the modified fractional Voc method, shown in Fig. 5, has been circuit method the adopted. In the conventional fractional open c maximum power point voltage is represented as a fraction k of d the open circuit voltage of the array given by (1): (1) wer where k is the ratio of Vmpp (maximum pow point) and and Voc (open circuit voltage). The ratio k varies between 0.71 to s 0.78 [12]. At every instant, the open circuit voltage has to be measured and the maximum power point is tracked with the most current value of Voc. and Voc. The m methodology was modified for this application such that the fra action k resides in the range 0.8 to 0.95. The ratio k in the m modified algorithm represents the controlling variable for the a amount of reserve power from the PV plant.
control blocks are the change in the reference voltage value. The difference between Vref and Vpv is the error that is fed to v the PI block which decides the sw witching function for the converter, thus delivering the associa active power. ated D. DC-DC Boost Converter and Inv verter A two-stage converter topolog has been used. Singlegy phase boost converter is in series with an inverter has better w control over the amount of power be eing injected into the grid. Pulse width modulation (PWM) is used for switching the boost converter. The switching is do by using a carrier wave one with a frequency of 30 kHz that take the control voltage input es from the MPP controller. The inv verter has a closed loop control which takes the grid voltag current and the dc bus ge, reference voltage [13] as inputs and controls the gate signals d to transfer the power at unity pow factor. The system is wer designed using PLECS and is shown in Fig. 7. n
V mpp = k × V occ
Fig. 7. DC-AC power conv version system
The output voltage of the boost co onverter is 260Vdc and this is converted to 260Vac p-p using th single-phase inverter. A he transformer is used to step up the voltage to the grid voltage of tance, equal to 1000µF, is 340V peak. A high value of capacit used at the dc output to maintain a stiff voltage and to reduce the ripple to about 1.5% (3 to 4V) at the terminals. The simulink model of the inverter is designed by research partner Luke Watson. The inverter also makes use of the PWM hes technique to drive all the four switch of the inverter. E. Simulation and Results An analysis and study of the comp plete system is done under various scenarios and presented as different cases. Each case nk simulation is done in Matlab/Simulin R2008b. Case 1: Insolation is 800 W/m2 and temperature is 25oC. t The system is checked for pseud and true power modes do and the outputs at the grid are analyzed when the insolation and the temperature are constant. For the pseudo maximum F power mode, the ratio k in the modif fractional open circuit fied voltage algorithm is 0.95. This se election is arbitrary as at k=0.95, the controller will track the point on the right hand n side of the PV characteristics shown in Fig. 3. If the ratio k is smaller, the controller will track a point on the left part of the p characteristics. Any value below 0. (left of MPP) or above .7 0.8 (right of MPP) would help. Fig 8 shows the variation in g. the PV voltage and current corresponding to the mode change. The change in mode is triggered by an external frequency change signal at t=0.82 sec.
Fig. 5. Modified fractional open circuit voltag algorithm ge
Both algorithms are implemented as pro ogram modules in the Simulink block as shown in Fig. 6.
Fig. 6. MPPT controller block with embedded Matlab code
As the system frequency changes, an external signal n triggers the change in the algorithm from the true MPP to of pseudo MPP and vice-versa. The outputs o both algorithm
4
Voltage (Volts)
Current (Amp)
Current (Amp) Time (sec) Fig. 10. Current injected into the grid at 1490 W (True MPP) Time (sec) Fig. 8. PV Voltage and Current Current (Amp)
Fig.9 shows the variation in the power level. At t=0.82 sec, the average output power of the PV array has decreased from 1.49 kW to 400 W as a result of a commanded change representing a drop in load.
Power (Watts)
Time (sec) Fig. 11. Current injected into the grid at 400W (Pseudo MPP)
Figs. 10 and 11 show the current injected into the grid for the true MPP and the pseudo MPP respectively. The current decreases from 8.5 A to 2.3 A with the change in the operating mode. One can conclude that if the grid frequency increases, the amount of active power being injected into the grid can be decreased instantaneously by changing the MPP mode. Figs. 12 and 13 show the power output at the inverter terminals for true and pseudo MPP respectively for Case 1. The power is injected into the grid at unity power factor. By controlling the phase angle of the current, the reactive power can be controlled; this is discussed later in the paper.
Power (Watts)
Time (sec) Fig. 9. Change in Power output from 1490 W to 400 W.
Time (sec) Fig. 12. Power (average: 1490W) injected into the grid for true MPP.
5
Power (Watts)
Figs. 15 and 16 show the variation in the power output when there is a non-uniform rise and fall of the insolation level. For a very short duration at the start of the simulation, the power level is zero, as the MPPT takes some time to start tracking the power and reach the operating point.
Time (sec) Fig. 13. Power (average: 400W) injected into the grid for pseudo MPP
Fig.14 shows the boost converter’s output voltage for the entire operating time of the system. Although there is a ripple of 4V, the average value of the output is 260V. There is a spike in the voltage waveform at t = 0.85 sec due to the change in the operation mode of the system; however it settles down in less than 0.3 sec to an average value of 260V.
Voltage (Volts)
Time (sec) Fig. 14. Boost converter output voltage
Case 2: Varying Insolation and temperature is 25oC. The insolation data for Case 2 is shown over a period of 1 sec in Table I.
TABLE I Varying Insolation Time(s) Insolation(W/m2) 0 200 0.1 450 0.2 650 0.3 600 0.4 400 0.5 700 0.6 900 0.7 550 0.8 850 0.9 620 1 450
One can conclude from Figs. 15 and 16 that the controller is able to create a reserve margin even for variable insolation levels. Frequency regulation can be achieved throughout the day limited only by the amount of power available from the PV plant. Case 3: Insolation is 1000 W/m2; temperature is 25oC and injecting power at 0.8 pf leading. Fig. 17 shows the variation in the active and reactive power magnitudes for the high and low power operations at 0.8 power factor leading. By changing the phase angle of the
Insolation (W/m2) Time (sec) Fig. 16. PV power output under varying insolation for the true MPP
Power (Watts)
Insolation (W/m2) Time (sec) Fig. 15. PV power output under varying insolation for pseudo MPP
Power (Watts)
6
reference current in the inverter block, the am mount of reactive power injected into the grid can be controlled d.
[2] [3] [4]
[5]
[6]
[7] [8] [9] [10]
Time (sec) Fig.17. Active and reactive power magnitudes at 0.8 p leading on the grid pf side
IV. CONCLUSION AND FUTURE WORK A new technique to implement freq quency regulation capability in solar photovoltaic power plant i discussed. This is method allows effective control of solar p power for various purposes without compromising the efficien of conversion. ncy A system comprising of a 2 kW PV ar rray, a two-stage converter and the grid was modeled, a and simulated in MATLAB. Results show that based on solar i insolation, the PV plant system can be given a reserve power m margin and active power injected into the grid can be effectively controlled. With pseudo power point tracking, the activ power flowing ve into the system can be balanced in order to regulate the r frequency of the system. The reserve power technique is very useful for fast response actions which last fo few seconds to or several minutes. m Work is going on in getting help from energy storage systems, such as batteries, to make the r renewable energy sources more dispatchable. Battery ener rgy storage, for example, is capable of providing spinnin reserves, load ng leveling, frequency control, VAR support, v voltage regulation, etc. [14]. However depth of discharge, life c cycle expectancy, self discharge rate and installation cost a some of the are challenges batteries are facing today and further study is necessary. A detailed and comparative study is also required y in areas such as the type of the battery t technology to be selected, sizing and location. V. ACKNOWLEDGMENT n The authors are thankful to Dr. Jonathan W. Kimball for his support and helpful discussions during t research. The the authors gratefully acknowledge the help of PhD candidate, f Luke D. Watson of the ECE department for his design of the control block for the inverter. VI. REFERENCE
[1] P. Denholm, R. M. Margolis, K. Zweibel, “Tackling Climate Change in n 2007 [Online] the U.S.,” ases.org, page 98, Jan
[11]
[12] [13] [14]
Available:http://www.ases.org/images/s stories/file/ASES/climate_change .pdf, [Accessed: Jan 12,2010]. D. Lenardic, “Large Scale Photovolta Power plants”, May 31,2010 aic [Online] Available: http://www.p pvresources.com/en/top50pv.php, [Accessed: June 06,2010]. M.L. Lazarewicz, A. Rojas, "Grid fre equency regulation by recycling electrical energy in flywheels," Proc. IEEE Power Engineering Society I General Meeting, pp.2038-2042, vol.2, June 2004. K.D. Brabandere, K. Vanthournout, D.J. Deconinck, G.R. Belmans, D 2007, “Control of Microgrids,” Proc. IEEE Power Engineering Society I General Meeting, pp. 1 – 7. 2007. I Serban, C.P Ion, C Marinescu, M Geo orgescu, “Frequency Control and Unbalances Compensation in Autonom mous Micro-Grids Supplied by RES,” IEEE International Electric Mac chines & Drives Conference, vol. 1, pp. 459 – 464, May 2007. J S. Wijnbergen, S.W.H. de Haan, J.G. Slootweg, “A System for Dispersed Generator Participation in Voltage Control and Primary Frequency Control of the grid,” Proc 36th IEEE Power Electronics c. Specialists Conference, pp. 2918 – 2924 June 2005. 4, R. Ito, Y. Matsuzaki, T. Tani,; T. Yach “Evaluation of performance of hi, MPPT equipment in photovoltaic sys stem,” Proc. 25th International Telecommunications Energy Conference pp. 256 – 260, Oct 2003. e, T. Esram, P.L. Chapman, “Comparison of Photovoltaic Array Maximum power Point Tracking Techniques,” IEEE Transactions on Energy e Conversion, Vol 22, pp. 439 – 449, June 2007. K.A. EL-Serafi, A.E. Kalas, M. H. El lfar, “On-Line Maximum Power Tracking In a Stand Alone Photo ovoltaic System,” Suez Canal University, Faculty of Engineering, 2003. Y-E. Wu, C-L. Shen, C-Y. Wu, “R Research and Improvement of Maximum Power Point Tracking for Photovoltaic Systems,” Proc. 8th P IEEE Int. Conference on Power Electr ronics and Drive Systems (IEEE PEDS, Taipei, Taiwan, R.O.C.), pp.1308 8-1312, 2009. ECEN 2060 Renewable Sources an Efficient electrical Energy nd Systems, University of Colorado at Boulder [Online] Available: http://ecee.colorado.edu/~ecen2060/energyprogram.html, [Accessed: DEC 14,2008]. E K.A. EL-Serafi, A.E. Kalas and M. H. Elfar, “On-Line Maximum Power Tracking in a Stand Alone Photovoltaic System,” Suez Canal University, Faculty of Engineering, 2003 N.A Ninad, L.A.C. Lopes, “Operation of Single Phase Grid Connected Inverters with Large DC Bus Voltage Ripple,” Proc. IEEE Canada Electrical Power Conference, EPC, pp. 172 – 176. Advancement of energy storage S.C. Smith, P.K. Sen, B. Kroposki, "A devices and applications in electrical po ower system," Proc. IEEE Power and Energy Society General Meeting - Conversion and Delivery of g Electrical Energy in the 21st Century, vol., no., pp.1-8, 20-24 July 2008. v
Power
VII. BIOGRAP PHIES
Venkata A. Pappu was bo in Ongole, India in 1985. He orn received his degree in Bach helor of Technology in Electrical and Electronics Engineerin from VIT University, Vellore, ng India in 2008. He was a graduate research assistant at Missouri S&T, Rolla, MO His research interests include O. Renewable Energy Generat tion and Power System Stability. Badrul H. Chowdhury (M M’83, SM’93) obtained his Ph.D. degree in Electrical Eng gineering from Virginia Tech, Blacksburg, VA in 1987. He is currently a Professor in the Electrical & Computer Engineering department of the E Missouri University of Sc cience and Technology, formerly known as the Univers sity of Missouri-Rolla. Dr. Chowdhury’s research in nterests are in power system modeling, analysis and control; renewable energy and distributed generation. He is a Senior Member of IEEE. Ravi Bhatt was born in Mumbai, India on June 18, n 1986. He received his Bachelor's degree in Electrical B Engineering from Mum mbai University, India. He is currently pursuing his Master's degree in Electrical ri Engineering at Missour University of Science & Technology, Rolla. Hi research interests are in is renewable energy sources s.
