MODELING AND SIMULATION OF GRID-CONNECTED PHOTOVOLTAIC DISTRIBUTED GENERATION SYSTEM

Transcription

1 Journal of Theoretical and Applied Information Technoloy th November. Vol. 45 No. 5 - JATIT & S. All rihts reserved. ISSN: E-ISSN: MODEING AND SIMUATION OF GRID-CONNECTED PHOTOVOTAIC DISTRIBUTED GENERATION SYSTEM M.MAKHOUF, F.MESSAI, H.BENAA Department of Electrical Enineerin, Faculty of Enineerin Sciences, Mentouri University Route d Ain El Bey, Constantine, Aleria, m makhlouf@hotmail.com ABSTRACT Compared to the traditional enery resources, photovoltaic (PV) system that uses the solar enery to produce electricity considered as one of renewable eneries has a reat potential and developin increasinly fast compared to its counterparts of renewable eneries. Such systems can be either stand-alone or connected to utility rid. However, the disadvantae is that PV eneration depended on weather conditions. Thus there is also a need for developin control techniques for three phase rid connected PV systems includin a method for DC link voltae control that stabilizes the voltae at the inverter input to insure a continuous flow of enery exchane between the rid and the PV system. An C filter also is necessary to filter the output current and voltae from the harmonics and protect the rid from their destructive effect. This paper presents detailed modelin of the rid-connected photovoltaic eneration system components, in Simulink / MATAB software. Simulation results presented here validate the component models and the chosen control schemes. Keywords: Grid Connected PV Systems, DC-DC Converter, PWM Inverter, MPPT, PI Controller.. INTRODUCTION The increasin of the world enery demand, due to the modern industrial society and population rowth, is motivatin a lot of investments in alternative enery solutions, in order to improve enery efficiency and power quality issues. The use of photovoltaic enery is considered to be a primary resource, because there are several countries located in tropical and temperate reions, where the direct solar density may reach up to W/m. At present, photovoltaic (PV) eneration is assumin increased importance as a renewable enery sources application because of distinctive advantaes such as simplicity of allocation, hih dependability, absence of fuel cost, low maintenance and lack of noise and wear due to the absence of movin parts. The cell conversion ranes vary from % of efficiency up to a maximum of 9% for very expensive units []. In spite of those facts, there has been a trend in price decreasin for modern power electronics systems and photovoltaic cells, indicatin ood promises for new installations. However, the disadvantae is that photovoltaic eneration is intermittent, dependin upon weather conditions. Thus, the MPPT makes the PV system providin its maximum power and that enery storae element is necessary to help et stable and reliable power from PV system for both loads and utility rid, and thus improve both steady and dynamic behaviors of the whole eneration system. In this paper we have studied a rid-connected photovoltaic eneration system which is composed of PV array, power electronic converters, filter, controllers, local loads and utility rid as shown in fiure. 78

2 Journal of Theoretical and Applied Information Technoloy th November. Vol. 45 No. 5 - JATIT & S. All rihts reserved. ISSN: E-ISSN: The paper discusses the detailed modelin of the whole system. PV array is connected to the utility rid by a boost converter to optimize the PV output and DC/AC inverter to convert the DC output voltae of the solar modules into the AC system. The DC input of the inverter must be constant and it is controlled by the use of a PI control circuit. An C filter has been introduced to insure a clean current injection to the rid.the proposed model of the entire components and control system are all simulated in Matlab/Simulink Software. Two different cases are simulated steady and transient states, and all simulation results have verified the validity of models and effectiveness of control methods. Fi.: Confiuration Of The Grid-Connected Photovoltaic Generation System. MODEING THE PV ARRAY The direct conversion of the solar enery into electrical power is obtained by solar cells. A PVG is composed by many strins of solar cells in series, connected in parallel, in order to provide the desired values of output voltae and current. Fi. shows the equivalent circuit of a PVG, from which non linear I V characteristic can be deduced. The liht current is related to irradiance and temperature and the liht current measured at some reference conditions: ( + ( ) G I I, REF ISC T C T C, REF G µ () REF Where I,REF iht current at reference conditions [A]. G, G REF Irradiance, actual and at reference condition [W/m²]. T, T C, REF Cell temperature, actual and at reference condition [ K]. µ ISC Manufacturer supplied temperature coefficient of short circuit current [A/ K]. The diode current is iven by Shockley equation: q( V + IR ) s ID I exp () γkt c Where: V terminal voltae [V], I reverse saturation current [Amps], γ shape factor. R s series resistance [Ω], q electron chare K Boltzmann constant.8. - J/K. The reverse saturation current is: qε G I DT c exp (4) AkT c Where: D diode diffusion factor, ε G material band ap enery (. ev for Si,.5 ev for GaGs) A completion factor The reverse saturation current is actually computed by takin the ratio of equation (4) at two different cell temperatures, thereby eliminatin D, similar to the determination of I, I is related to the temperature and the saturation current estimated at some reference conditions: Fi. :Solar-Cell Equivalent Circuit. The cells are connected in series and in parallel combinations in order to form an array of the desired voltae and power levels. Applyin Kirchoff s law of current, the terminal current of the cell is: I I I () D T C qε G I I REF exp (5), T C, REF ka T C, REF T C And thus the I-V characteristic is described by: ( V + IR ) q I I exp s I γkt c (6) The shape factor γ is a measure of cell temperature and is related to the completion factor 79

3 Journal of Theoretical and Applied Information Technoloy th November. Vol. 45 No. 5 - JATIT & S. All rihts reserved. ISSN: E-ISSN: as γ A NCS NS (7) NCS is the number of cells connected in series per module. A module is defined as an array of cells, usually encapsulated for protection, as it is supplied by manufacturer; NS is the number of modules connected in series of the entire array. While R s and γ are assumed to be constant, I is a function of irradiance and cell temperature and I is a function of temperature only. The cell temperature can be determined from the ambient temperature and with the help of some standard test information. In [] the way to Evaluate these parameters based on the four parameters model proposed by Townsend (989) [], Eckstein (99) [4] and Fry, Bryan (998) [5] to be the most précised model that ood produced the I-V characteristics. Now only the four parameters I, I o, Rs and γ need to be evaluated, a method to calculate these parameters has been developed by Eckstein [4]. Since there are four unknown parameters, four conditions of the current I and the voltae V are needed. Generally, available manufacturer s information are set at three points at the reference conditions,(gw/m²,t5 C),the voltae at open circuit V oc,ref, the current at short circuit I sc, ref and the voltae and current at maximum power V mp, ref and I mp, ref. The 4th condition comes from the knowlede of the temperature coefficients at short circuit µ ISC and at open circuit µ VC. I MP, REF I SC, REF I O, REF γ REF.( +. I ) q V MP, REF K. T C, REF R s. ln MP, REF I MP I SC, REF, REF V OC, REF () The indices OC, SC, MP and REF refer to the open circuit, the short circuit, the maximum power and the reference condition respectively. The cell s parameters chane with the solar radiation G (W/m²) and ambient temperature T ( K) and they can be estimated by the followin relations: I ( G )( I + ) G REF, REF Isc ( T C I O I O, REF / µ () T REF ( ) N s E q T C, REF T / T REf exp. A T C ( / ) γ REF TC TC, REF (4) γ (5) Fi. : Mathematic Model Of A PV Array The parameters evaluated in this section are based on data input for a sinle module at some reference condition. To describe the I V characteristic for the entire array that contains series and parallel modules as shown in fi., parameters need be scaled up in the followin way: I tot N p I, () I (4), tot N p I (.( V I. ) expv t MPREF + MP, REF R s (8) (. ) I SC, REF I O, REF.exp V t V OC, REF (9) I REF I SC, REF, () Substitutin equation (7) into equation (9) and solvin forγ and I, REF ives I, REF I SC, REF. exp (V t. V OC, REF ) () γ tot NS γ (5) N S R s tot R s N P, (6) Where, N S and N P are cell numbers of the series and parallel cells respectively. Connectin cells in series will increase the output voltae, and connectin them in parallel will increase the output current, correspondin to the expression I tot N p I (7) 8

4 Journal of Theoretical and Applied Information Technoloy th November. Vol. 45 No. 5 - JATIT & S. All rihts reserved. ISSN: E-ISSN: V tot N S V (8) All parameters of the model use the data in table. Table : Parameters For PV Model Parameters Values Nombre des cellules en série 54 Ns Referenced solar irradiance G W/m² G ref Referenced cell temperature T ref T 5 C, Imp 7,695A Vmp 6,857V Pmp 4,8574W Voc,978V Isc 8,887A KV,V/K KI,A/K With different temperatures and solar radiations, output characteristics of PV array are presented in fi.4 and fi.5. environment conditions. Maximum Power Point Trackin (MPPT) aims at usin some control alorithms to ensure the PV array to operate at the maximum power point. At the maximum power point we have: V P (9) I V V + I With the current described by equation (6), the partial derivative of I with respect to V is: I q( V + IR S ) q I () V I exp kt C, REF γ kt C, REF An explicit expression for I V + R S V is obtained simply by rearranin equation (). Back substitution of this explicit expression () and usin I MP for I and V MP for V ives: I + I exp q ( V MP + I R S ) MP kt C, REF γ qv MP kt C, REF γ + qr S I q( V + ) + exp MP I MP R S kt C, REF γ kt C, REF γ () Fi. 4. Effect Of Irradiance On I V Characteristics. Fi. 5. Effect Of Temperature On I V Characteristics. As shown in fi.4 and 5, PV cell represents nonlinear voltae-current characteristics, and there is only one point that makes the PV enerator eneratin its maximum power under different environmental conditions... Maximum Power Point Trackin In order to improve the efficiency of the PV eneration system, PV array should be controlled to enerate the maximum power at the particular To eliminate V MP in equation (), the eneral I-V equation (6) is used, with I MP substituted for I and V MP substituted for V. Rearranin to solve for V MP ives V MP, γ I MP R S () kt C REF I I MP ln + q I An explicit expression for I MP is obtained by substitutin equation () into equation (): I ( ) ln + I MP I MP R S q I MP I I I kt C, REFγ I + () MP R + ( + ) S q I I I MP kt C, REFγ Newton Raphson is applied to solve for I MP usin an initial uess iven by ( + ( ) I G MP, GUESS NP I MP, REF ISC T C T C REF G, REF µ (4) Once I MP is found, V MP may be calculated usin equation (4) and thus the current and voltae at the maximum power point is determined as a consequence the maximum power. 8

5 Journal of Theoretical and Applied Information Technoloy th November. Vol. 45 No. 5 - JATIT & S. All rihts reserved. ISSN: E-ISSN: Boost Circuit And Its Control Since the output voltae of PV cell is low, the use of boost circuit will enable low-voltae PV array to be used, as a result, the total cost will be reduced. A capacitor is enerally connected between PV array and the boost circuit, which is used to reduce hih frequency harmonics. Fiure 6 is the confiuration of the boost circuit and its control system. Fi.6. Boost Circuit And Its Control The modelin of this converter depends on the analysis of the various sequences of operation which we will suppose of durations fixed by a control element U. There is two sequences of operation dependin on the state of the switch S, which we can represent each one by a differential equation [7, 8, 9]. - When the switch is closed: pv di pv V (5) dv C dc + I dc (6) - When the switch is open dv I dc pv C + I dc (7) By supposin U, when the switch is closed and U for S opened, we can represent the converter by a sinle system of equations, that we describe as instantaneous model. We consider the perfect switches here. V pv di pv + V dc ( U ) (8) dv U ) I C dc pv + I dc ( (9) The instantaneous model reveals a nonlinear behavior of the converter by the existence of products between the control element U and the state variables Vdc and Ipv. In addition it is advisable to announce that by replacin the variable U by its mean value over one switchin period of Td / fd, i.e. the duty cycle D (D Ton/Td) we can obtain the model with the mean values: di pv V ( D ) dc () dv dc I pv ( D ) C V pv + V dc RC () PV array can be controlled to operate at the maximum power point by reulatin the duty cycle D. The control scheme includes two levels. The control circuit aims to reulate the voltae usin the reference voltae V ref calculated by MPPT alorithm and enerate the control sinals for the boost ate as illustrated in Fi.6.. CONTRO OF THE GRID-CONNECTED INVERTER PV array is connected to the ac rid via a common DC/AC inverter. The inverter is used in current control method with PWM switchin mechanism to make the inductance current track the sinusoidal reference current command closely and obtain a low THD injected current... Uncoupled Watt-Var Method In order to understand the principle of this method in the eneral case, we consider an inverter connected to the network, via a resistor R and the inductance (which represent the simplified model of a transformer), as indicated on fi.7. We have the followin equation: i d i i R R R i i + i V e V e r V e () Where: V and I represent the voltae and current of the rid. 8

6 Journal of Theoretical and Applied Information Technoloy th November. Vol. 45 No. 5 - JATIT & S. All rihts reserved. ISSN: E-ISSN: R, resistance and inductance of the ird and e is the inverter voltae. Applyin abc to dq transformation of Park, the equation () is written in the followin way: Watt-VAR method is represented on fi.8. R ω i i d d d + i R i q q ω V d V q e d e q () Fi.8: Diaram Block Of The Uncoupled Watt-VAR Method The total diaram of the identification of the references and reulation of the current for this method is shown on fi.9 where the reulators are those of fi.8. Fi.7: Diaram Of An Inverter Connected To The Grid To know the advantaes of the control method, the traditional uncoupled Watt-VAR alorithm is briefly presented. The two new variables presented in (4), are the output variables of the control system which contains two PI controllers: X X ( V e ) d ( V e ) q d q (4) The values of id, ref. and iq, ref. are the references of the active and reactive currents []: K I X ( K P + S K I X ( K P + S ( I i ) d, ref ( I i ) ω i d q, ref d q ω i q (5) et s apply the aplace transformation to the equation (). Thus we transform the equations (4) and (5). We obtain the transfer functions (6): i d F ( S ) i d, ref i q i q, ref K + I SK P R K I + S + K + S P (6) The control is optimal when both K P and K I are maximum []. The diaram of control by the method uncoupled Fi.9: Total Diaram Of Control By The Uncoupled Watt-VAR Method The oal of this control method is to impose the values of the active and reactive powers injected into electrical rid. The powers and the voltae at the connection point are calculated in the dq reference frame []: P ref Q ref ( v i + v i ) dr ( v i + v i ) dr dr qr qr qr qr dr (7) Where P ref and Q ref are the reference powers. We can deduce the current in the dq frame as follows: i d, ref i q, ref ( + ) P V d V d Q V q + V q ( + v ) P V q Q d V d + V q (8) V d and V q are the direct and quadrature components of the voltae at the connection point in the dq reference frame. 8

7 t(s) THD THD t(s) Id,ref Iq,ref Journal of Theoretical and Applied Information Technoloy th November. Vol. 45 No. 5 - JATIT & S. All rihts reserved. ISSN: E-ISSN: I d,ref. and I q,ref. are the direct and quadrature current components injected into the rid. These currents depend on the power requested and the voltaes measured at the point of connection. This voltae is transformed in the dq frame before the calculation of the currents as explained by fi. 9. The role of the phase locked loop is to provide the rotation frequency, direct and quadrature voltae components at the point of common couplin (PCC) by resolvin the rid voltae abc components. Multiple control blocks of the PV system rely on this information to reulate their output command sinals. As stated earlier, the P computes the rotation frequency of the rid voltae vector by first transformin it to the dq frame, and then force the quadrature component of the voltae to zero to eliminate cross couplin in the active and reactive power terms []. A proportional-interal controller is used to perform this task. The proportional (Kp) and interal (Ki) ains of the controller were set throuh an iterative process to achieve a fast settlin time. The components of the current are compared with its references. The differences between them passed throuh reulators, which ive the components of the reference voltae in the dq reference. While passin by the reverse transformation dq to abc, we obtain the references of the PWM sinals for the inverter... DC BUS VOTAGE CONTROER The reulation of this voltae is carried out when absorbin or providin the active power to the rid. The correction of this voltae must be done by the addition of an active fundamental current into the reference currents. Based on a difference between (U dcref) and (U dc), the power (Pref) on the reulator output side is added to the fluctuatin active power and ives place to an active fundamental current thus reulatin the dc bus voltae. In order to obtain the (Pref) sinal, we have the choice between a proportional reulator and a proportional interal reulator. This last one is often used and ives better results in preventin the static errors. The schematic diaram of calculatin and of reulatin of the Dc bus voltae standard is iven by the fiure. Fi.: Schematic Diaram Of The DC ink Controller. 4. SIMUATION RESUTS Based on the above models and control methods, two simulation cases are studied: a. steady operation, when there is no chane in atmospheric conditions; b. chanes of solar irradiance and the dc bus control will stabilize the inverter input voltae; 4.. Steady Operation When the system is in steady state, solar irradiance is W/m, and temperature is 98K. Voltae (V) Voltae (V) (A): PV Generator Voltae Voltae (A) Voltae (V) (B): DC Bus Voltae (C) : Phase To Phase Inverter s (D): Phase To Phase Inverter s Voltae Before Filterin Voltae After Filterin (E): Thd In Voltae Before Filterin Current (A) (G): Inverter Output Currents Current (A) (F) : Thd In Voltae After Filterin (H): Id,Ref And Iq,Ref Current 84

8 x x Active Power Reactive Power Active Power Reactive Power Active POwer Reactive Power Journal of Theoretical and Applied Information Technoloy th November. Vol. 45 No. 5 - JATIT & S. All rihts reserved. ISSN: E-ISSN: Active and Reactive Power (W, VAR) (I): Active And Reactive Power Of The Inverter Active and Reactive Power (W, VAR) (J): Active And Reactive Power Of The oad When solar irradiance is 6 W/m, DC bus voltae is less than 6 V so the dc bus reulator must act to maintain the DC bus voltae constant (at 6 V). In fiure the DC bus voltae controller has insure a constant voltae that make PV system able to feed the inverter and the load, Active and Reactive Power (W, VAR) (K): Active And Reactive Of The Grid Fi.: Simulation Results Of The System In Steady State From fiures (a)-(d), at this operation situation,. In steady state, the operation point of PV array is just its maximum power point. The power requested by the load is KW and KVAR. Since the PV array enerate only an active power of KW and no reactive power (because the reulation system as shown in fi (l) has two inputs Id,ref and Iq,ref that is the imae of reactive power and obviously its value is set to Zero) the rest of the requested power comes from the rid as illustred in fiures ((i),(j),(k)., The voltae at dc bus is reulated to be 6V the currents of inverter have sinusoidal forms and its maximum is A. Fiures ((c), (d), (e), (f)) represents the voltae of the inverter before and after filterin it seems that the filter has improve the THD from about 9 % into less than.%. Tension (V) Fi. :The Effect Of DC Controller On The DC Bus Voltae 5. CONCUSIONS In this paper, a rid-connected photovoltaic eneration system is studied. In order to convert the solar enery efficiently, the maximum power point of the PV array should be tracked to ensure the PV array provide most power to both rid and the load. When solar irradiance or temperature fluctuates, PV eneration will chane as a result. The controller must act to maintain the DC bus voltae constant as possible and improve the stability of the whole system. the simulation results presented in this paper validate the component models and the chosen control scheme. REFERENCES 4.. Chanes Of Solar Irradiance Assumin solar irradiance chanes: durin to.s, solar irradiance is W/m ; durin.s to.s, solar irradiance is 5 W/m ; durin.s to.s, solar irradiance returns to W/m. Due to this chane we have droop in both pv voltae and DC bus voltae as shown in fiure. Current (A) (): Photovoltaic Current (M): DC Bus Voltae Fi.: Simulation Results Of The PV System When Solar Irradiance Chanes From W/M² To 6 W/M² And Then To W/M² Aain Voltae (V) [] M.G. SIMOES,. FRANCE SCHETTI,N.N.«Arisc-Microcontroller Based Photovoltaic System for Illumination Applications, Proceedin of IEEE Applied Power Electronics Conference and Exposition 5 (). [] CHENNI, R. MAKHOUF, M. et al: «Detailed Modellin Method for Photovoltaic Cell, Enery (7), [] Townsend, Timothy U. «A Method for Estimatin the on-term Performance of Direct- Coupled Photovoltaic Systems». M. S. Thesis. Solar Enery aboratory, University of Wisconsin, Madison:

9 Journal of Theoretical and Applied Information Technoloy th November. Vol. 45 No. 5 - JATIT & S. All rihts reserved. ISSN: E-ISSN: [4] Eckstein, Juren Helmut, «Detailed Modelin of Photovoltaic Components», M. S. Thesis Solar Enery aboratory, University of Wisconsin, Madison: 99. [5] F.Bryan, «Simulation of rid-tied buildin interated photovoltaic systems» M. S. Thesis Solar Enery aboratory, University of Wisconsin, Madison: 998. [7] M.Fadel, «ois de commande pour une alimentation AC/DC à absorption de courant sinusoïdal» EI. [8] J.P. Ferrieux, F. Forest, «Alimentations à découpae Convertisseurs à résonance. Principes-composants, modélisation», Dunod e édition ISBN [9] J.achaize, M.Fadel, S.Caux, P.Shott,.Nicod, «Modellin and Control of a fuel cell system for electrical rail transport», EPE Toulouse, France Sept -4. [] I. Papic, P. Zunko, D. Povh and M. Weinhold, «Basic Control of Unified Power Flow Controller», IEEE Trans. On Power Systems, vol., No. 4, pp , November 997. [] G.RAMI, «Contrôle de tension auto adaptatif pour des productions décentralisées d éneries Connectées au réseau électrique de distribution», Thèse de doctorat DE INP Grenoble [] Kroutikova, N.; Hernandez-Aramburo, C.A.; Green, T.C, «State-space model of rid-connected inverters under current control mode», Electric Power Applications, IET, vol., no., pp.9-8, May 7, , 8. 86