Vous êtes sur la page 1sur 11

IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 25, NO.

9, SEPTEMBER 2010

2409

Issues on Solar-Generator Interfacing With Current-Fed MPP-Tracking Converters


Teuvo Suntio, Senior Member, IEEE, Jari Lepp aaho, Student Member, IEEE, Juha Huusari, Student Member, IEEE, and Lari Nousiainen, Student Member, IEEE

AbstractThe large-scale harvesting of solar energy is an important action to decelerate the observed climate changes. Reliably operating solar-energy systems composing of solar arrays and their interfacing converters are of prime importance to maximize solar-energy harvesting. The paper investigates the solar-generator interfacing in terms of current-fed (CF) maximum-power-point (MPP) tracking converter. The investigations show that the CF converter under input-voltage control can usually operate from the short-circuit to open-circuit conditions of the solar generator without stability problems. When the output voltage or current has to be controlled constant, the converter may become unstable at the MPP due to the negative incremental resistance appearing at its input terminals. In practice, this means that the operating range of the CF converter has to be limited to the voltages less than the MPP voltage, when the output-voltage or current control is active. Practical evidence is provided based on a CF superbuck converter derived from the corresponding voltage-fed converter applying duality-transformation methods and supplied by an actual solar panel. Index TermsCurrent-fed (CF) converter, current source, maximum power point (MPP), solar cell, voltage source.

I. INTRODUCTION

HE NEED for large-scale harvesting of renewable and clean energy, such as solar or photovoltaic energy, has recently been fully recognized for reducing greenhouse gas emissions and securing the availability of energy in the future [1][5]. Worldwide actions have been planned and initiated for developing technologies capable of fullling the recognized needs [2]. Solar energy can be utilized as heat power or as electrical power by means of photovoltaic conversion [4]. The simplied electrical equivalent model of a solar cell composes of a photocurrent source and a diode connected across the current source, as depicted in Fig. 1(a) [5][7]. The solar cell is a highly nonlinear and nonideal current source due to the behavior of the semiconductor junction, yielding limited output voltage and power because of the forward biasing of the parallel diode, as depicted in Fig. 1(b). In addition, with the nonlinear terminal behavior, its dynamic behavior can be represented by means of operatingpoint-dependent nonlinear resistance and capacitance [8][10]. The typical maximum voltage of a single-junction silicon cell

Fig. 1. Solar cell characteristics. (a) Simplied electrical model. (b) Static terminal characteristics as I /U and P /U curves.

Manuscript received December 7, 2009; revised February 23, 2010; accepted April 11, 2010. Date of current version September 17, 2010. Recommended for publication by Associate Editor D. Xu. The authors are with the Department of Electrical Energy Engineering, Tampere University of Technology, Tampere FI-33101, Finland (e-mail: teuvo.suntio@tut.; jari.leppaaho@tut.; lari.nousiainen@tut.). Digital Object Identier 10.1109/TPEL.2010.2048580

is in the order of 0.5 V at the ambient temperature of 25 C with negative temperature coefcient [9]. As a consequence, a large number of cells have to be usually connected in series to form a solar generator for fullling the practical solar-energy harvesting. In order to maximally utilize the energy provided by the solar generator, its operating point has to be kept at the maximum power point (MPP) [see Fig. 1(b), the point (1,1)], where the static and dynamic resistances of the solar generator coincide [11], [12]. A lot of methods have been developed for tracking the MPP, as discussed in [13][15]. It has been observed that the MPP tracking has a signicant contribution to the reliability problems in photovoltaic energy systems [16]. The output of the solar generator is very seldom suited as such for powering the load because of its highly varying terminal characteristics due to the changes in the environmental conditions, such as ambient temperature, level of solar irradiation, cloud passing, etc. The interfacing converter can be based either on the principle of shunt [17], [18] or series

0885-8993/$26.00 2010 IEEE

2410

IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 25, NO. 9, SEPTEMBER 2010

regulation [10], [19][28]. The terminal characteristics of the solar generator in Fig. 1(b) imply that the series regulation can be accomplished either by using voltage-fed (VF) or current-fed (CF) converters, as discussed in [19], because of the constantcurrent-type source behavior at the voltages less than the MPP and the constant-voltage-type source behavior at the voltages higher than the MPP, respectively [10]. The use of VF MPP-tracking converters is actually quite problematic because the tracking control has to be based on the input current of the converter, which has to be always less than the photocurrent of the solar generator for avoiding the saturation of the control [26]. The constraints of a VF converter are not usually fully recognized [21], [27], [28], but observed and proved to take place in the practical applications [10], [26]. In order to avoid the named difculties, the input voltage of an MPP-tracking converter is recommended to be controlled instead of the input current [10]. As a consequence, the VF converter will actually change to a CF converter having totally different static and dynamic properties compared to the original converter. A CF converter can be implemented by two different methods such as 1) adding an input capacitor at the input terminals of a VF converter and applying positive feedback from the input voltage [22], [23] as well as 2) by transforming a VF converter into a corresponding CF converter by applying duality-transformation methods [29][34]. The rst method actually changes the properties of the converter in such a way that a buck-type VF converter becomes a boost-type CF converter and vice versa, which is not recognized in the literature, as e.g., in [19], [20], and [22][25]. The validation of the MPP-tracking converters presented in the public literature is not usually carried out in accordance with sound scientic methods, because the real solar generator is often substituted with a voltage source, as in [25], or with a power-electronic-based solar array simulator, as in [19], which may not have the same dynamic characteristics as a solar generator has. In addition, the MPP-tracking converter can be loaded with a pure resistor, as in [21], during the measurements yielding results, which cannot be generalized to cover also the other type of loads, such as a storage battery or a large capacitor bank. Therefore, it is not a surprise that there exist reliability problems in practical installations, because the real behavior of the MPP-tracking converters is so poorly understood. The similar constraints and issues, as discussed earlier, are valid also in the solar-generator interfacing by means of grid-integrated inverters although not covered in this paper. This paper investigates the basic constraints of the solargenerator interfacing in terms of CF converters based on the general internal dynamics of such a converter and scientically sound stability-assessment method. The similar investigations of a VF converter can be found from [26]. The theoretic ndings are supported with experimental evidence from a CF fourth-order buck converter with cascaded input/output-voltage control. The theory and experiments clearly conrm that a CF converter has less constraints than the VF converter in the MPPtracking applications. In reality, many of the reported MPPtracking converters are CF converters [10], [20], [22], [23][25],

but the presented discussions and dynamic analyses do not reveal their correct nature. The rest of the paper is organized as follows. An introduction to the general dynamic representations of a CF switched-mode converter including the effect of MPP-tracking control are given and discussed in Section II. The stability-analysis method is introduced in Section III. The experimental dynamic analyses of a solar generator and the input impedance of the MPP-tracking converter are presented and discussed in Section IV. The conclusions are drawn nally in Section V.

II. GENERAL DYNAMIC REPRESENTATIONS The dynamics of a switched-mode converter can be represented by using only certain two-port-network parameters, which are uniquely determined by the input source and the output variable kept constant by the feedback controller or by the applied internal open-loop controlling method (e.g., the directduty-ratio-controlled VF converter is a voltage-output converter and the corresponding current-mode-controlled converter is a current-output converter, respectively [35]). It is well known that the type of the converter output can be changed from voltage to current or vice versa by applying proper feedback control, as discussed in detail in [35]. The change of mode also requires changing the ideal load from constant current to constant voltage or vice versa for proper operation [35]. It is quite poorly understood that the use of positive feedback, adding an input capacitor in a VF converter, and changing the input source from voltage to current as well as the ideal load from constant current to constant voltage will change the voltage-tovoltage converter into a current-to-current converter. In order to protect the storage battery or large capacitor bank connected typically at the output of a CF converter from damage, the output-voltage-limiting control has to be applied, which changes the output mode of the converter accordingly. As a consequence, the set of two-port-network parameters uniquely characterizing the different converters can be given, according to [36] as follows: the voltage-to-voltage converter by G, the voltage-to-current converter by Y , the current-tocurrent converter by H , and the current-to-voltage converter by Z . The rest of the discussions in the paper are limited to the CF converters characterized by H and Z parameters. The source and load impedances are known to affect the dynamics of a VF converter [35] and will also affect a CF converter. The load effect in the voltage-output mode will be only treated in this paper because of its signicant role in the stability of the converter, when loaded by a storage battery or a large capacitor bank. A more comprehensive introduction of the loadimpedance interactions is outside the scope of the paper and will be treated in other publications. The theoretical formulations derived in the subsequent sections based on the named transfer functions as H and Z parameters are general, and therefore, equally valid for model-based analytical transfer functions as well as for the experimentally measured frequency responses that are used in this paper for validating the theoretical ndings.

SUNTIO et al.: ISSUES ON SOLAR-GENERATOR INTERFACING WITH CURRENT-FED MPP-TRACKING CONVERTERS

2411

Fig. 2. Linear two-port model of a current-to-current converter with the ideal input and output sources.

Fig. 3. Linear two-port model of a current-to-voltage converter with the ideal input and output sources.

A. Current-to-Current Converter The set of H parameters can be given in a matrix form both at open and closed loop, according to (1), where the meaning of the transfer functions can be deduced based on the correT T sponding input io vectors. The in o c and output u iin u variable c denotes the general control variable and the superscript c denotes the current-output mode, respectively. The set in (1) can be also represented by using the linear two-port model shown in Fig. 2, which explicitly illustrates the very nature of the input and output of the corresponding converter. The steady-state ideal input and output relations of a CF converter can be given by Io = M (D)Iin and Uin = M (D)Uo , which dictate that the input current will be reected into the output and the output voltage into the input modulated by the modulo M (D) regardless of the mode of the output. The name or type of the converter is naturally based on its inputto-output relation in respect to M (D), as in a VF converter, where M (D) is chosen in such a way that the corresponding output variable will increase along the increase in the duty ratio (D ) u in io
c Zin Gc io c Toi

Fig. 4. Linear two-port model of a current-to-voltage converter with the nonideal load.

Yoc

Gc ci Gc co

iin o . u c

(1)

compared to the current-output mode in (1) iin v v v Toi Gci Zin u in = io v v u o Gv Z G o co io c c c c c Gc G T T Toi co c c + io c oi oi G + Zin ci c c Y Y Y u in o o o = c c u o Gio Gco 1 c Yoc Yo Yoc

(2) iin io . c (3)

C. Current-to-Voltage Converter and Load Effect B. Current-to-Voltage Converter The CF converter is usually a current source internally. If its output voltage has to be kept constant, the mode of the converter will change from the current-output mode to voltage-output mode. As a consequence, the new unique set of parameters dening its dynamics is the Z parameters given in (2), where the superscript v denotes the voltage-output mode. The corresponding linear two-port model is given in Fig. 3. When the original converter is characterized by H parameters, the most convenient way is to express the Z parameters as a function of the H parameters. The transformation can be found by interchanging the input and output variables at the output port in (1) and substituting u o at the input port by its new denition. This yields (3), which indicates that the internal dynamics will change profoundly due to the change of the output mode The effect of the load impedance (ZL in Fig. 4) on the dynamics of the current-to-voltage converter can be found by computing io from Fig. 4 and substituting it in (2) with its new denition, similar to the corresponding VF converter [35]. This procedure yields the load-affected dynamic description, as dev Zo , then the gains ned in (4). If the load impedance ZL L of the load-affected control-to-output transfer function [Gv co , the right-most bottom-row entry in (4)] and the output-voltage feedback loop are very small. This yields also very short control bandwidth in the output-voltage feedback loop. As a consequence, the load-affected input impedance would approach the v short-circuit (CS) input impedance Zin sc dened in (5), which equals the internal input impedance of the current-to-current converter. Such a situation may be the consequence of connecting a storage battery or large capacitor bank at the output of the

2412

IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 25, NO. 9, SEPTEMBER 2010

Fig. 5.

Solar-generator-interfacing principle with a CF converter.

converter. u in u o =
v Gv io Toi v +Z Zo L v v Gco Toi v Zo + ZL Gv co

v Zin +

v Toi v Zo v /Z ) 1 + (Zo L

Gv ci +

Gv io v /Z ) 1 + (Zo L iin jo c

v /Z ) 1 + (Zo L

(4)

v v Zin sc = Zin +

v Gv io Toi c = Zin . v Zo

(5)

Fig. 6. Control system arrangements. (a) Cascaded controllers. (b) Separate controllers.

D. Feedback-Controlled Converter A typical MPP-tracking-converter connection to a solar generator and load is shown in Fig. 5, where the MPP-tracker (MPPT) locates the MPP and provides its coordinates in terms of voltage to the input-voltage feedback controller. The outputvoltage-limiting controller is normally inactive, but will be activated automatically when the output voltage reaches the predened level for protecting the load from damage. The input- and output-voltage controllers can be arranged to work in cascade [see Fig. 6(a)] or separately [see Fig. 6(b)], where the controller requiring the minimum pulsewidth to be applied is only active. If positive feedback has to be used, the polarities of the feedback and reference voltage signals are to be interchanged. Gcc i and Gse i denote the corresponding voltage-loop-controller transfer functions and the scaling factors applied in the corresponding loop, respectively. The controller arrangements affect naturally the behavior of the closed-loop transfer functions. u in io
c Zin o 1 + Lin

The closed-loop dynamics of the input-voltage-controlled converter can be given in (6), at the bottom of this page, where Lin = Gse 1 Gcc 1 Gc ci o is the input-voltage loop gain accordc ing to Fig. 6, as well as Gc io and Yo , as dened in (7), governing the low-frequency behavior of the dened closed1. The superscript loop output variables in (6) when Lin c and the subscript extension o indicate that the corresponding transfer functions are the transfer functions of the current-to-current converter according to (1) at open loop. The closed-loop input impedance under the input-voltage control in c Zin c = Zin o /(1 + Lin ) usually does not have characteristics resembling a negative incremental resistance
c Gc io = Gio o c c Zin o Gco o Gc ci o

Yoc = Yoco +
c Toi o 1 + Lin

c c Toi o Gco o . Gc ci o

(7)

1 Gse 1

= c Gio o Lin + Gc 1 + Lin 1 + Lin io

Lin 1 + Lin

Yoco Lin + Yc 1 + Lin 1 + Lin o

Gc Lin co o 1 + Lin Gse 1 Gc ci o

iin o u u r in

(6)

SUNTIO et al.: ISSUES ON SOLAR-GENERATOR INTERFACING WITH CURRENT-FED MPP-TRACKING CONVERTERS

2413

The closed-loop dynamics of the output-voltage-controlled converter can be given in (8), at the bottom of this page, where Lout = Gse 2 Gcc 2 Gv co o is the output-voltage loop gain and the transfer functions labeled with the subscript extension -o are the corresponding closed-loop (6) (i.e., cascaded control) or open-loop (1) (i.e., separate control) transfer functions of the current-output converter, which are transformed into the v voltage-output mode using (3) as well as Zin (i.e., ideal input v impedance) and Toi (i.e., ideal output-to-input transfer function), as dened in (9), respectively. The ideal transfer functions are specic for a certain topology and independent of state of feedback and load, similar to the same transfer functions in a VF converter [35]
v Zin

Fig. 7.

Interconnected (cascaded) system.

the resulting mappings are given in (11) and (12), respectively u in1 i u io 2 = = S11 S21 L11 L21 S12 S22 L12 L22 iin1 u i u o 2

c Zin

Gc Gc io c ci , Gco

v Toi v (Zin )

c Toi

Y c Gc + o c ci . Gco

(9)

(10)

resembles the negative The ideal input impedance incremental resistance (rcon Uin /Iin ) at least at the low frequencies, similar to the inverse of the ideal input admittance (Yin ) in a VF converter [35]. According to (8), the closedv v loop input impedance (Zin c ) equals (Zin o (1 + Lout )) + v (Lout /(1 + Lout ))Zin . At the low frequencies, the outputvoltage loop gain (Lout ) is typically high. Consequently, v v Zin c Zin . A storage battery or large capacitor bank as a load may, however, also change the behavior of the input impedance by reducing the gain of the output-voltage loop, as discussed earlier in Section II-C. As a consequence, the closed-loop input impedance may not have anymore characteristics resembling the negative incremental resistance (i.e., v v Zin c Zin o ; Lout < 1). III. INTERNAL AND INPUTOUTPUT STABILITY The similar methods, as applied in the interconnected VF systems for assessing the stability [37], can also be utilized in the CF systems. The internal and input-to-output stabilities of an interconnected system can be determined at an arbitrary interface by constructing the mappings from the system input variables to the intermediate and output variables. For stability to exist, all the resulting transfer functions have to be stable i.e., their poles have to be located in the closed left half plane of the complex plane. A cascaded or interconnected system consisting of two subsystems S and L is shown in Fig. 7 with the corresponding input, intermediate, and output variables. The subsystems are assumed to be both CF systems without the loss of generality and characterized by four transfer functions, because the control variable c = 0 [see (1) and (2)]. o 2 ), and the output variThe system input variables are ( iin1 ,u io 2 ), respectively. The intermediate variables are ables are (u in1 , given by ( i,u ) and denote two physical variables by denition. The H parameter sets of the subsystems are given in (10) and
v Zin Lout o v 1 + Lout + 1 + Lout Zin = Gv io o 1 + Lout

L12 S21 L11 1 S22 L11 1 S22 L11 u i in1 = u o 2 i S21 S22 L12 1 S22 L11 1 S22 L11 S12 L12 S12 S21 L11 S11 + 1 S22 L11 1 S22 L11 u in1 = io 2 S21 L21 S22 L12 L21 L22 + 1 S22 L11 1 S22 L11 iin1 . u o 2

(11)

(12)

If the subsystems are stable as stand-alone systems, then the stability of the interconnected system is dependent on the stability of 1/(1 S22 L11 ), similar to an interconnected VF system [37], which can be inferred by means of S22 L11 applying Nyquist stability criterion. According to (1), S22 L11 = c YocS Zin L . The similar product in a VF system yields Zo S Yin L [37], which is commonly known as minor-loop gain according to [38]. The main difference between the minor loop gains of the CF and VF systems is that they are the inverse of each other. The inverse minor-loop gain provides an excellent method for analyzing the stability of the solar-generator interfacing, when the output impedance of the solar-generator system and the input impedance of the tracking converter are known. According to Section III, the input impedance of a CF converter under output voltage or current feedback control can behave as a negative incremental resistor at least at the low frequencies, similar as in a VF converter. Suntio et al. [26] showed that the VF MPP-tracking converter becomes unstable at the MPP because of the negative-incremental-resistance behavior at its input impedance. The experimental measurements introduced Gv Lout ci o u in 1 + Lout Gse 2 Gv co o io Lout 1 u r o 1 + Lout Gse 2

iin u o

v Toi Lout o + Tv 1 + Lout 1 + Lout oi v Zo o 1 + Lout

(8)

2414

IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 25, NO. 9, SEPTEMBER 2010

TABLE I MAIN CHARACTERISTICS OF RALOSS SR30-36 SOLAR PANEL AT THE IRRADIATION OF 1 KW/M2 AND THE PANEL TEMPERATURE OF 25 C

in Section IV show that the output-voltage-feedback-controlled CF converter becomes unstable at the MPP for the same reason. IV. EXPERIMENTAL EVIDENCE The parameters of the Raloss SR30-36 solar panel used for powering the CF converter are dened in Table I. The articial sun forming the light source can produce irradiation, corresponding only to 500 W/m2 , yielding short-circuit (SC) current of 1 A. The interfacing converter is a direct-duty-ratiocontrolled CF superbuck converter (see Fig. 12) derived from the corresponding VF converter [35] by applying the dualitytransformation technique described in detail in [29][34]. The dynamic details and design of such a converter are outside the scope of the paper. The methods to extract the dynamic models can be found, e.g., from [34]. The solar power system used in the measurement is dened earlier in Fig. 5, where the storage battery is a 12-cell lead-acid battery of 7 Ah yielding output voltage of 27 V. The frequency responses shown in the subsequent sections are obtained by Venable Industries frequency-response analyzer Model 3120 with an impedance measurement kit and imported into MATLAB for efcient gure handling. A. Solar Generator The measured solar-generator static terminal characteristics at the SC currents of 0.5 and 1 A at the panel temperature of 44 are shown in Fig. 8 with the corresponding MPP voltage/current, SC current, and open-circuit (OC) voltage denitions. The small-signal output impedance of the solar panel including the cabling impedance was measured at several panel output voltage/current operating points starting from the SC and ending to the OC condition. Fig. 9 shows the measured frequency responses at the panel SC current of 1 A indicating that the dynamic output resistance of the panel is highest at SC and lowest at OC, and its phase starts from 0 not 180 , as claimed, e.g., in [10]. The high-frequency series-resonant behavior is a consequence of the cabling inductance (2 H) and the nonlinear output capacitance of the solar generator. Fig. 10 shows the behavior of the low-frequency dynamic output resistance and capacitance of the solar generator at the SC current of 1 A, when the operating point moves from the SC to the OC based on the frequency responses shown in Fig. 9. The resistance and capacitance values are proportional in such a way that the authentic values are divided by the maximum resistance
Fig. 9. Set of measured small-signal output impedances of the solar panel, including the cabling impedance. Fig. 8. Static terminal characteristics of the Raloss solar panel.

Fig. 10. Low-frequency magnitudes of the dynamic output resistance and capacitance of the Raloss solar panel.

SUNTIO et al.: ISSUES ON SOLAR-GENERATOR INTERFACING WITH CURRENT-FED MPP-TRACKING CONVERTERS

2415

Fig. 11. Dynamic resistances of the Raloss solar generator (rsg ) and a general interfacing converter (rc o n ).

of 1 k and the maximum capacitance of 22 F, respectively. The output resistance and capacitance at the MPP (16.0 V) are approximately 17.4 and 2.0 F, respectively. The reason for the piecewise linear shape of the resistance is the low number of samples at the voltages much lower than the MPP. Fig. 11 shows the extended view of the behavior of the solargenerator dynamic resistance in the vicinity of the MPP at the SC current of 1 A (solid line) as well as the behavior of v the low-frequency ideal input impedances (|Zin | = rcon Uin /Iin = Usg /Isg ) of a MPP-tracking converter at the same operating points (dash-dot line) with the corresponding MPP marked with a dot. Fig. 11 indicates that the dynamic output (rsg ) and input (rcon ) resistances of the solar generator and the interfacing converter, respectively, at the MPP equals each other in accordance to the maximum power theorem [11]. Therefore, according to Section III, the inverse minor-loop gain would indicate that the CF converter may become unstable when its operating point enters into the input voltages equal to or higher than the MPP under output-voltage or current control. The stability problems may arise already earlier due to the possible source interactions, i.e., due to the series resonant behavior in the output impedance of the solar supply. This argumentation supports the classication of the output characteristics of the solar generator into the constant-current and constant-voltage regions in respect to the MPP, but the existence of the power regions proposed in [10] cannot be justied. If the output-voltage control loop is inactive, the CF converter can usually operate without stability problems from the SC to OC conditions in the solar panel. Its current-output nature suits also very well for charging the storage batteries and it may be an optimal converter type for interfacing solar generators especially in Telecom uninterrupted power supply (UPS) systems and space power systems. B. CF Superbuck Converter The power stage of the experimental CF superbuck converter is shown in Fig. 12. The converter works in such a way that

Fig. 12. control.

CF superbuck converter with cascaded input- and output-voltage

the diode conducts during the ON time and the MOSFET during the OFF time, respectively. The ideal steady-state input-tooutput relations in the converter are as follows: Io = DIin and Uin = DUo , which justies the use of buck converter as its name. When using a conventional PWM modulator, its gatecontrol output has to be inverted for correct operation. The source uB and resistance rB denote the storage battery load. The switching-frequency ripple components of the voltages across the capacitor C1 and the series connection of the capacitors C1 and C2 are shown in Fig. 13, including the gate-control signal of the MOSFET also. The capacitor ripple voltages in a CF converter are typically triangle shaped due to the charging of constant currents similar to the ripple currents of inductors in a VF converter. The detailed dynamic analysis and modeling of the converter is outside the scope of the paper. The measured closed-loop input impedances under inputvoltage control and the output impedances of the solar generator when the operating point in terms of the input voltage varies from 5 to 18 V are shown in Fig. 14. Even if there are impedance overlaps, the inverse minor-loop gain Zcon /Zsg indicates stable operation, as discussed earlier. The converter operates well over the operating range dened earlier, as shown in Fig. 15, where the time-domain behavior of the solar-panel output voltage and current, as well as the output current of the converter are shown. The duty ratio (D) of the converter can be directly computed from the ratio icon o /isg . Fig. 16 shows the measured closed-loop input impedances (Zcon , dashed line) of the CF converter, when the output-voltage control is active and the output impedances (Zsg , solid lines) of the solar generator from 5 V to the MPP voltage of 16 V, when the

2416

IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 25, NO. 9, SEPTEMBER 2010

Fig. 13. Ripple voltages of the capacitor C 1 and the series connection of C 1 and C 2 , as well as the gate-control signal of the MOSFET.

Fig. 16. Measured closed-loop input impedances (Z c o n , dashed lines) of the CF converter under output-voltage control and the output impedances (Z sg , solid lines) of the Raloss solar generator from the panel output voltage of 5 V to the MPP of 16 V.

Fig. 14. Measured closed-loop input impedances (Z c o n , dashed lines) of the CF converter under input-voltage control and the output impedances of the solar generator (Z sg , solid lines) at the converter input voltages from 5 to 18 V. Fig. 17. Time-domain behavior of the converter input voltage (u c o n , bottom curve) and input current (ic o n , upper curve) due to the instability taken place at the input voltage of 17 V.

Fig. 15. Time-domain behavior of the solar-generator output voltage and current as well as the converter output current when a microprocessor-based MPPtracking controller sweeps the input-voltage reference from 5 to 18 V and back.

battery impedance rB = 7.7 (i.e., an additional resistors was connected in series with the battery, Fig. 12). Fig. 16 indicates that the input impedance of the converter approaches the output impedance of the solar generator along approaching the MPP with opposite phases. As a consequence, the converter would become unstable at the MPP. The operating point of the converter was placed at the input voltage of 17 V, when only the inputvoltage feedback loop was active. As a consequence of the activation of the output-voltage feedback loop, the operating point moved instantaneously from the unstable operating point to 15.5 V, as shown in Fig. 17, which is a stable operating point for the converter. The converter was observed to be stable at the aforementioned-treated operating point (i.e., 17 V), when the additional battery resistor was removed, lowering rB from 7.7 to 0.3 . Fig. 18(a) shows the measured input impedances of

SUNTIO et al.: ISSUES ON SOLAR-GENERATOR INTERFACING WITH CURRENT-FED MPP-TRACKING CONVERTERS

2417

Section II as well as shows that the input-voltage-controlled CF converter can operate within the whole practical operation range of the solar generator. This observation cannot be, however, generalized to cover all the converter topologies and applied internal control methods, because the behavior of the input impedance may have properties making the converter prone to instability, as in the VF converters (see [35], e.g., current-mode controls, etc.) The observed behavior of the superbuck converter means that different MPP-tracking methods, such as the hill-climbing method can be readily used, where the operation has to be extended over the MPP in order to locate it. When the output-voltage control is active, the converter cannot usually operate in the voltages higher than or equal to the MPP voltage. If the MPPT moves the operation point to the MPP, the instability phenomenon will move the operating point back to the lower voltage, which leads to the phenomenon known as limit cycle oscillation [39]. When the constraints of the MPPtracking converter are known, the microcomputer-based control and monitoring system can be implemented to avoid the known problems, and thereby, improve substantially the reliability of the photovoltaic energy systems. V. CONCLUSION This paper has investigated the operational constraints and stability of a CF converter in interfacing a solar generator and introduced the impedance-ratio-based inverse minor-loop gain to assess the stability of the interconnected CF systems. The investigation shows that the CF interfacing does not contain the same constraints as the VF interfacing reported in [26], but the operating point has to be limited at the voltages less than the MPP, when the output-voltage or current control is active. The validity of the theoretical ndings was proved by time- and frequency-domain experiments. The paper reveals that the most usually applied interfacing technique has been the CF interfacing for a long time, but not recognized in reality. Many of the dynamic and even steady-state analyzes of the used interfacing converters are quite erroneous. REFERENCES
[1] A. J ager-Waldau, Photovoltaic and renewable energies in Europe, Renewable Sustainable Energy Rev., vol. 11, no. 7, pp. 14141437, Sep. 2007. [2] A Strategic Research Agenda for Photo Solar Energy Technology (2007, Mar.). The report repaired by Working Group 3 Science, Technology and Applications of the EU PV Technology Platform [Online]. Available: www.eupvplatform.org [3] S. R. Bull, Renewable energy today and tomorrow, in Proc. IEEE, Aug., 2001, vol. 89, no. 8, pp. 12161226. [4] S. Rahman, Green power: What is it and where can we nd it? IEEE Power Energy Mag., vol. 1, no. 1, pp. 3037, Jan./Feb. 2003. [5] S. Lyi and R. A. Dougal, Dynamic multiphysics model for solar array, IEEE Trans. Energy Convers., vol. 17, no. 2, pp. 285294, Jun. 2002. [6] Y. T. Tan, D. S. Kirschen, and N. Jenkins, A model of PV generation suitable for stability analysis, IEEE Trans. Energy Convers., vol. 19, no. 4, pp. 748755, Dec. 2004. [7] M. G. Villalva, J. R. Gazoli, and E. R. Filho, Comprehensive approach to modeling and simulating photovoltaic arrays, IEEE Trans. Power Electron., vol. 24, no. 5, pp. 11981208, May 2009. [8] R. A. Kumar, M. S. Suresh, and J. Nagaraju, Effect of solar array capacitance on the performance of switching shunt voltage regulator, IEEE Trans. Power Electron., vol. 21, no. 2, pp. 543548, Mar. 2006.

Fig. 18. Measured input impedances of the converter when the battery in denotes the inimpedance (a) rB = 0.3 and (b) rB = 7.7 . (Note: Z c on in o u t denotes the input put impedance under pure input-voltage control and Z c on impedance when both of the loops are active.)

the converter, when only the input-voltage control is active in , dashed line) and when both of the controllers are active (Zcon in out ) without the additional resistance in the battery sys(Zcon tem. The same impedances are also shown in Fig. 18(b) when the extra resistance is added. Fig. 18 clearly validates the theoretical conclusions presented in Section II concerning the load interactions in a output-voltage-controlled converter. The low impedance of the battery has reduced the output-voltage loop gain, the magnitude, and phase of the input impedance in such a way that it has approached the input impedance under the input-voltage control. The stable operation at the input voltage of 17 V is the consequence of the reduction of the phase from 180 to approximately 166 . C. Discussions The frequency- and time-domain experimental evidence clearly validates the general theoretical analysis presented in

2418

IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 25, NO. 9, SEPTEMBER 2010

[9] J. Thongpron, K. Kirtikara, and C. Jivicate, A method for determination of dynamic resistance of photovoltaic modules under illumination, Solar Energy Mater. Solar Cells, vol. 90, no. 18/19, pp. 30783084, Nov. 2006. [10] W. Xiao, W. G. Dunford, P. R. Palmer, and A. Capel, Regulation of photovoltaic voltage, IEEE Ind. Electron., vol. 54, no. 3, pp. 13651374, Jun. 2007. [11] J. C. McLaughlin and K. L. Kaiser, Deglorifying the maximum power transfer theorem and factors in impedance selection, IEEE Trans. Educ., vol. 50, no. 3, pp. 251255, Aug. 2007. [12] J. Ding and R. Radharkrishnan, A new method to determine the optimum load of a real solar cell using the Lambert W-function, Solar Energy Mater. Solar Cells, vol. 92, no. 12, pp. 15661569, Dec. 2008. [13] V. Salas, E. Olias, A Barrado, and A. Lazaro, Review of the maximum power point tracking algorithms for stand-alone photovoltaic systems, Solar Energy Mater. Solar Cells, vol. 90, no. 12, pp. 15551578, Jul. 2006. [14] T. Esram and P. L. Chapman, Comparison of photovoltaic array maximum power point tracking techniques, IEEE Trans. Energy Convers., vol. 22, no. 2, pp. 439449, Jun. 2007. [15] S. Jain and V. Agarval, Comparison of the performance of maximum power point tracking schemes applied to single-stage grid connected photovoltaic systems, IET Electr. Power Appl., vol. 1, no. 5, pp. 753762, Sep. 2007. [16] G. Petrone, G. Spagnuolo, R. Teodorescu, M. Veerachary, and M. Vitelli, Reliability issues in photovoltaic power processing systems, IEEE Trans. Ind. Electron., vol. 55, no. 7, pp. 25692580, Jul. 2008. [17] R. D. Middlebrook and S. G. Kimble, Design of the HEAO main bus shunt regulator, IEEE Trans. Aerosp. Electron. Syst., vol. AES-12, no. 2, pp. 162172, Mar. 1976. [18] A. Garrigos, J. M. Blanes, J. A. Carrasco, A. H. Weinberg, E. Maset, E. Sanchis-Kilders, J. B. Ejea, and A. Ferreres, The sequential switching shunt maximum power regulator and its application in the electric propulsion system of spacecraft, in Proc. IEEE PESC, 2007, pp. 1374 1379. [19] A. Capel, J. C. Marpinard, J. Jalade, and M. Valentin, Current fed and voltage fed switching DC/DC convertersSteady state and dynamic models, their applications in space technology, in Proc. IEEE INTELEC, 1983, pp. 421430. [20] W. Xiao, N. Ozog, and W. G. Dunford, Topology study of photovoltaic interface for maximum power point tracking, IEEE Trans. Ind. Electron., vol. 54, no. 3, pp. 16961704, Jun. 2007. [21] J. M. Enrique, E. Duran, M. Sidrach-de-Cardona, and J. M. Andujar, Theoretical assessment of the maximum power point tracking efciency of photovoltaic facilities with different converter topologies, Solar Energy, vol. 81, no. 1, pp. 3138, Jan. 2007. [22] N. Femia, G. Lisi, G. Petrone, and M. Vitelli, Distributed maximum power point tracking of photovoltaic arrays: Novel approach and system analysis, IEEE Trans. Ind. Electron., vol. 55, no. 7, pp. 26102621, Jul. 2008. [23] K. Siri, Study of system instability in solar-array-based power systems, IEEE Trans. Aerosp. Electron. Syst., vol. 36, no. 3, pp. 957964, Jul. 2000. [24] M. G. Villalva and E. Rupert, Input controlled buck converter for photovoltaic applications: Modeling and design, in Proc. IET PEMD, 2008, pp. 505509. [25] J. H. Yang, H. S. Bae, J. H. Lee, and B. H. Cho, A simplied seriesparallel structure for maximum power point tracking (RPPT) system, in Proc. IEEE APEC, 2008, pp. 160166. [26] T. Suntio, J. Lepp aaho, and J. Huusari, Issues on solar-generatorinterfacing with voltage-fed converter, in Proc. IEEE IECON, 2009, pp. 600605. [27] H. Dehbonei, S. R. Lee, and H. Nehrir, Direct energy transfer for high efciency photovoltaic energy systems. Part I: Concepts and hypothesis, IEEE Trans. Aerosp. Electron. Syst., vol. 45, no. 1, pp. 3145, Jan. 2009. [28] H. Dehbonei, S. R. Lee, and H. Nehrir, Direct energy transfer for high efciency photovoltaic energy systems. Part II: Experimental evaluation, IEEE Trans. Aerosp. Electron. Syst., vol. 45, no. 1, pp. 4657, Jan. 2009. [29] C. A. Desoer and E. S. Kuh, Basic Circuit Theory. Tokyo, Japan: McGraw-Hill Kogausha, Ltd, 1969, pp. 444449. [30] S. Cuk, General topological properties of switching structures, in Proc. IEEE PESC, 1979, pp. 109130.

[31] D. Shmilovitz and S. Singer, A switched mode converter suitable for superconductive magnetic energy storage (SMES) systems, in Proc. IEEE APEC, 2002, pp. 630634. [32] D. Shmilovitz, Application of duality for derivation of current converter topologies, HIT J. Sci. Eng. B, vol. 2, no. 3/4, pp. 529557, Dec. 2005. [33] D. Shmilovitz, Gyrator realization based on a capacitive switched cell, IEEE Trans. Circuits Syst.-II: Exp. Briefs, vol. 53, no. 12, pp. 14181422, Dec. 2006. [34] J. Lepp aaho, M. Karppanen, and T. Suntio (2008). Current-sourced buck converter. presented at Nordic Workshop Power Ind. Electron. (NORPIE) [CD-ROM], pp. 17. [35] T. Suntio, Dynamic Prole of Switched-Mode ConverterModeling, Analysis and Control. Weinheim, Germany: Wiley-VCH, 2009. [36] C. K. Tse, Linear Circuit Theory. Harlow, England: Addison Wesley Longman, 1998, pp. 226240. [37] K. Zenger, A. Altowati, and T. Suntio, Dynamic properties of interconnected power systemsa system theoretic approach, in Proc. IEEE ICIEA, 2006, pp. 835840. [38] R. D. Middlebrook, Input lter considerations in design and applications of switching regulators, in Proc. IEEE IAS, 1976, pp. 91107. [39] R. P. Venturini, V. V. R. Scarpa, G. Spiazzi, and S. Buso, Analysis of limit cycle oscillations in maximum power point tracking algorithms, in Proc. IEEE PESC, 2008, pp. 378384.

Teuvo Suntio (M98SM08) received the M.Sc. (Tech.) and D.Sc. (Tech.) degrees in electrical engineering from Helsinki University of Technology, Espoo, Finland, in 1981 and 1992, respectively. From 1977 to 1991, he was as a Design Engineer and the Research and Development Manager at Fiskars Power Systems. From 1991 to 1992, he was the Research and Development Manager at Ascom Energy Systems Oy, Espoo. From 1992 to 1994, he was an Entrepreneur in Power Electronics Design Consultancy. From 1994 to1998, he was a Consultant and a Project Manager at Efore Oyj, Espoo. In 1998, he joined the Electronics Laboratory, University of Oulu, Oulu, Finland, as a Professor specializing in switched-mode power converter technologies. Since August 2004, he has been a Professor in the Department of Electrical Energy Engineering, Tampere University of Technology, Tampere, Finland. He is the author of the book Dynamic Prole of Switched-Mode ConverterModeling, Analysis, and Control (Weinhein, Germany: Wiley-VCH, 2009). He has authored or coauthored about 160 international scientic journal and conference papers. He holds several international patents. His current research interests include dynamic modeling, control design, optimal electromagnetic interference design of switched-mode power converters, as well as interfacing of renewable energy sources. Prof. Suntio is a member of the IEEE Power Electronics, the IEEE Industrial Electronics, and the IEEE Circuits and Systems Societies, as well as a member of European Power Electronics Association. Since 2010, he has been an Associate Editor of the IEEE TRANSACTION OF POWER ELECTRONICS.

Jari Lepp aaho (S09) received the M.Sc. (Tech.) degree in electrical engineering from Tampere University of Technology, Tampere, Finland, in 2008, where he is currently working toward the Ph.D. degree from the Department of Electrical Energy Engineering. Since 2008, he has been a Researcher in the Department of Electrical Energy Engineering, Tampere University of Technology. He has authored or coauthored six conference and two journal papers. His current research interests include dynamics of currentfed maximum-power-point-tracking converters and the interfacing of solar generators. Mr. Lepp aaho is a member of the IEEE Power Electronics and the IEEE Industrial Electronics Societies.

SUNTIO et al.: ISSUES ON SOLAR-GENERATOR INTERFACING WITH CURRENT-FED MPP-TRACKING CONVERTERS

2419

Juha Huusari (S09) received the M.Sc. (Tech.) degree in electrical engineering in 2009 from Tampere University of Technology, Tampere, Finland, where he is currently working toward the Ph.D. degree from the Department of Electrical Energy Engineering. Since 2009, he has been a Researcher in the Department of Electrical Energy Engineering, Tampere University of Technology. He has authored and coauthored three conference publications and two journal publications. His current research interests include analysis and design of transformer-isolated currentfed maximum-power-point-tracking converters, on the interfacing of solar generators, as well as practical switching-power-supply design issues. Mr. Huusari is a member of the IEEE Power Electronics, the IEEE Industrial Electronics, and the IEEE Power Engineering Societies.

Lari Nousiainen (S09) received the M.Sc. (Tech.) degree in electrical engineering in 2009 from Tampere University of Technology, Tampere, Finland, where he is currently working toward the Ph.D. degree from the Department of Electrical Energy Engineering. Since 2009, he has been a Researcher in the Department of Electrical Energy Engineering, Tampere University of Technology. His current research interests are focused on the analysis and design of grid-tied current-fed photovoltaic inverters. Mr. Nousiainen is a member of the IEEE Power Electronics, the IEEE Industrial Electronics, and IEEE Power and Energy Societies.

Vous aimerez peut-être aussi