IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 23, NO.

5, SEPTEMBER 2008

2301

Three-Phase Boost-Type Grid-Connected Inverter

Yang Chen and Keyue Smedley, Fellow, IEEE
AbstractAlternative energy sources, such as solar energy and fuel cells, are desirable due to their pollution-free property. In order to utilize the present infrastructure of the utility grid for power transmission and distribution, grid-connected dc-to-ac inverters are required for alternative energy source power generation. For many of these applications, the input dc voltage is usually below peak voltage of the output and may vary in a wide range. Thus single-stage buck-type inverters may not be adequate, since they have very limited input voltage range and require the input dc voltage to be higher than the peak of the output voltage. For this reason, two-power-stage topologies, cascaded topologies and multilevel topologies are reported for applications where the input voltage is lower than the peak of the output voltage. Typically, one dcdc power stage is required to boost the dc voltage in addition to an inverter for dcac conversion, which yields increased circuitry complexity. One-stage inverters for low dc voltage to high ac voltage conversion have been reported for non-grid-connected inverters based on the topology of a current source inverter. In this paper, the one-cycle control (OCC) method and the pulse width modulation (PWM) method have been proposed for a three-phase boost-type grid-connected inverter. The inverter features a single power stage that converts dc power to grid-connected ac power by injecting three in phase sinusoidal currents into grids, which may reduce power losses and circuit complexity. The input dc voltage is lower than the peak grid voltage and can vary in a wide range, which greatly suits the power conversion from photovoltaic or fuel cells to grid lines. The dc inductance may be kept low because the average dc current is maintained constant in a switching cycle. With the OCC method, the inverter preserves the advantages of simple circuitry, good stability and fast dynamic response and maximum power point tracking (MPPT) function can be conveniently integrated into the control core. Experiments have been performed with a 1.5-kW laboratory prototype that demonstrated the good performance of the inverter and MPPT function. Index TermsBoost type, grid connected, maximum power point tracking (MPPT), one cycle control (OCC), photovoltaic (PV), three-phase inverter.

I. INTRODUCTION N RECENT years, serious concerns have been raised over the fossil fuel electricity generation, because it pollutes our environment and depletes the energy supply. As a result, alternative energy sources, such as solar energy and fuel cells, have gained great attention because they are friendly to the environment and exible for installation. However, these types

Manuscript received September 07, 2006; revised May 13, 2007. Current version published November 21, 2008. Recommended by Associate Editor J. Espinoza. This manuscript is revised from the original papers, Three-Phase Boost-Type Grid-Connected Inverters, presented at APEC 2006, March 610, 2006, Dallas, TX, and A Cost-Effective Three-Phase Grid-Connected Inverter with Maximum Power Point Tracking, presented at the Industry Applications Conference 2006, 41st IAS Annual Meeting, Oct. 812, 2006, Tampa, FL. The authors are with the Power Electronics Laboratory, University of California, Irvine, CA 92697 USA (e-mail: smedley@uci.edu). Color versions of one or more of the gures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identier 10.1109/TPEL.2008.2003025

of sources produce dc power while the present power grid accepts ac power; therefore, grid-connected inverters are necessary for dcac conversion. To avoid introducing additional distortions to the power grid, the generated currents from these inverters are required to have low harmonics and high power factor. When the output currents are in phase with the grid voltages, the maximum real output power is achieved. The inverters are desired to have high power quality, high efciency, high reliability, low cost, and simple circuitry. Compared to single-phase inverters, three-phase grid-connected inverters have distinctive advantages. First, the power ow is constant, which results in a reduced dc capacitor number and value; and second, fewer switches are used for three-phase dcac conversion. As far as the alternative source is concerned, the photovoltaic (PV) module or fuel cell stack usually supplies a dc voltage lower than the peak grid voltage, and their output voltage varies in a wide range according to various operation conditions [1], [2]. Series connection of several such modules can be a simple way to increase their output voltage so as to employ a buck-type grid-connected inverter for power conversion [1], [3]. However, this method may reduce the input power collection range because the inverter stops conversion when its input dc voltage drops below its output peak grid voltage. Additionally, series connection of PV modules may reduce the overall efciency of the PV array (a PV array consists of several PV modules) when individual modules are running under different conditions [4]. For boosting PV output voltage in order to accommodate the buck-type grid-connected inverter, a two-stage topology that boosts the PV voltage by a dcdc converter in the rst stage and then inverts it into ac voltages by the second stage was reported in previous literature [5], [6]. In [4], a cascaded structure is used in which the individual PV voltage is stepped up to a higher module output voltage and such module output voltages are series connected to build up an array output voltage as the front-end of a buck-type grid-connected inverter. In [7], a multi-level topology is employed to increase the input voltage of the buck-type inverter. In [4][7], either two power stages or extra power switches are used for increasing the dc side voltage, which increases circuitry complexity and may reduce system overall efciency. An alternative way to achieve dc-to-ac inversion without directly boosting input dc voltage is to use the power stage topology of a current source inverter (CSI) as reported in [8] and [9]. The input side inductor in a CSI topology is shorted to the input dc source to build up magnetic energy in certain sections of a switching period and then the energy is released to the ac side during the rest of the switching period, just as in a basic boost converter. With CSI, the input dc voltage is lower than the output ac peak voltage while there is only one power stage in the system. Two different control methods for non-grid-connected applications are reported in [8] and [9].

0885-8993/$25.00 2008 IEEE

2302

IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 23, NO. 5, SEPTEMBER 2008

Fig. 1. Power stage of the boost-type inverter.

In this paper, the one-cycle control (OCC) method and the conventional pulse width modulation (PWM) method are proposed based on the CSI topology for a grid-connected application. The property of single power stage is preserved and the input dc voltage of the inverter is lower than the output peak grid voltage, which perfectly suits the property of wide output voltage range in PV or fuel cells. The dc side inductance can be kept small in a balanced three-phase system, so the size, weight, and power dissipation of the dc inductor are reduced and system dynamic response is improved. The proposed OCC method features a simple control circuit, fast transient response, and good stability. Multipliers or microprocessors are not necessary for a three-phase OCC control algorithm, while the maximum power point tracking (MPPT) capability can be conveniently integrated into the control core with an acceptable precision. The power stage and switching strategy are described in Section II. In Section III, the control principle is explained. Section IV presents some practical application issues, and the performance of the boost-type inverter is validated by experiments as shown in Section V. In Section VI, conclusions are given. II. POWER STAGE AND SWITCHING STRATEGY Fig. 1 shows the power stage of the three-phase boost-type is a dc voltage source, , , grid-connected inverter, where are three-phase grid voltages, and , , and and , , form an output lter. Each switch of the bridge is realized by an insulated gate bipolar transistor (IGBT) in series with a diode as shown in the dashed line box. According to the zero-crossing points of grid voltages, each line cycle can be divided into six regions as shown in Fig. 2. In each region, two line-to-neutral voltages have the opposite ), polarity of the third one. For example, in Region I ( , , and . The line-to-line voltages are (1) is the rms value of the line-to-neutral voltage. where Similarly, in the other regions, two voltages can always be selected so as to obtain the two line-to-line voltages referred to . the third voltage with the amplitude greater than Therefore, a boost converter operation is guaranteed as long as satises . the dc voltage At any given instant, one of the upper switches ( , , ) and one of the lower switches ( , , ) are kept

Fig. 2. Six regions in a line cycle.

Fig. 3. Three stages for different switching patterns in Region I. (a) Stage I. (b) Stage II. (c) Stage III.

on to form the inductor current ow passage. For example, in and are kept off and is on for the enRegion I, tire region, while switches , , and are operated at the switching frequency. There are three stages with different switching strategies in Region I: 1) In Stage I [Fig. 3(a)], is turned on and , are off. The inductor current increases, and output currents are supplied by , , . 2) In is turned on and , are off. The Stage II [Fig. 3(b)], discharges through , and the grid , inductor current . Current is supplied by , . 3) In Stage III [Fig. 3(c)], is turned on and , are off. decreases through , and , . is supplied by , .

CHEN AND SMEDLEY: THREE-PHASE BOOST-TYPE GRID-CONNECTED INVERTER

2303

Fig. 4. Diagram of the controller. Fig. 5. Diagram of the OCC core for a boost-type inverter. TABLE I VECTOR OPERATION MECHANISM OF THE OCC CONTROLLER

Fig. 6. Time sequence of control diagram.

A. One-Cycle Control Method 1) Control Signal Generation: Fig. 5 shows the diagram of the OCC core for the boost-type inverter. It comprises an integrator with reset, two comparators, two ip-ops, and other is a constant and related to linear and logic components. output power. and are integration components and . represents the period of a switching cycle. The time sequence of signals CLK1, CLK2, and Reset is shown in Fig. 6. The operation procedures are as follows (e.g., in Region I): at the beginning of each switching cycle, CLK1 in the upper ip-op. Switch is turned on and sets the power stage is operated in Stage I. Meanwhile, the integrator starts to work. Dene (2) where is the dc component of the inductor current on the dc is the current sensing resistance. side, and When the falling ramp meets , the ip-op is and , which turns off and turns reset. Then, . The power stage switches to Stage II. At the time of on , the lower ip-op is set by CLK2. and the power stage turns back to Stage I again. Similarly, when the falling

III. CONTROL PRINCIPLE Fig. 4 shows the diagram of the general controller, where is the voltage sensing ratio and is the current sensing resistance. The Voltage/Region Selection Circuit selects correand in each region according to Table I sponding and also generates the region signals for the Drive Signal Distributor. Since the switching frequency is much higher than the and can be considered as constant line frequency, , in each switching cycle. The Control Core takes signals , and and generates PWM signals , , and , which are distributed to the corresponding switches by the Drive Signal Distributor according to Table I. The control core can be achieved by OCC method or the conventional PWM method. These are described respectively in the following two subsections.

2304

IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 23, NO. 5, SEPTEMBER 2008

ramp intersects , and and the power stage switches to Stage III until the next switching cycle begins. 2) DC Side Current: According to the above control method, the following equations can be obtained (in Region I):
Fig. 7. Single-phase equivalent circuit in Region I.

(3) Studying the inductor current on the dc side, has a dc component and a high-frequency ac component. In each switching cycle

(4)
Fig. 8. Conventional PWM control core.

where is the instantaneous inductor current at the end of a is the instantaneous inductor current switching cycle, and at the beginning of a switching cycle. Since the switching freand can quency is much higher than the line frequency, be treated as constant. Substitute (3) into (4): (5)

. Similarly, the same a large range as long as relationship as in (9) can be obtained in other regions. 3) AC Side Current: In order to simplify the analysis of ac side current, a single-phase equivalent circuit, e.g., phase A in Region I, is presented in Fig. 7. From (3) (10)

For a balanced three-phase system, the three line-to-neutral voltage can be expressed as (6) Derived from (6): (7) Substituting (6) and (7) into (5), the following equation can be obtained: (8) is small so is In the start-up process, inductor current large, which causes . Then drops when increases. At a certain value of , the steady state of , i.e., , is reached when the right side of (8) is equal to zero. Then, the steady-state value of can be obtained from (8) by plugging in (2) (9) From (9), it can be observed that the dc side average current is constant provided that , , , , and are all is not xed and the three-phase system is balanced. Since related to directly, it is not necessary to keep very large constant, which greatly reduces the losses, size, to maintain and weight of the inductor. The dc input voltage can vary in

In steady state, is constant, and so is . can also be treated as a constant in each switching cycle since the switching frequency is much higher than the line frequency. From Fig. 6 is and (10), in a switching cycle, the average value of (11) Since the LC lter blocks most of the high-frequency component of from going into grids, for the average value of the . Thus, phase A curoutput line current , there is and its value is rent has the same shape of the grid voltage determined by (11). Similarly, phase C current can also be derived that has the same shape of . Since is satised in such a topology, three sinusoidal currents will be injected to the grids. The displacement factor caused by output capacitors, e.g., , can be reduced if the switching frequency is increased and is reduced. In high-power applications, the control method in [10] can be used to compensate the displacement factor in order to let , , and be in phase with grid voltages. B. Conventional PWM Method 1) Control Signal Generation: The boost-type inverter can also be controlled by the conventional PWM control method. The whole system remains the same except that the control core is replaced by the conventional PWM control core. Fig. 8 shows the conventional PWM control core which comprises two multipliers, sawtooth generation circuits, two comparators, and other linear or logic components. The time sequence of the conventional PWM method is shown in Fig. 9. The operation procedures are as follows (e.g.,

CHEN AND SMEDLEY: THREE-PHASE BOOST-TYPE GRID-CONNECTED INVERTER

2305

Since the LC lter blocks most of the high-frequency compofrom going into grids, . Thus, phase A nent of current has the same shape of the grid voltage and its value is determined by (15). Similarly, phase C current can also be derived to have the same shape of . is determined by the equa, which is satised in such a topology. tion of Consequently, three sinusoidal currents can be provided to the grids. C. Summary of the Proposed Control Methods The control methods in this paper have many advantages and share some similarities to previously published control methods, such as the space vector modulation (SVM) method and the sinusoidal PWM (SPWM) method. A summary of the comparison follows. 1) Circuit complexity: Compared to the SVM method, the proposed control methods in this paper greatly simplify the implementation, because transformation from abc-coordinates to dq-coordinates is not needed, and computing turn-on time for corresponding switches is not needed. 2) Total harmonic distortion (THD): However, some shortcomings of the proposed methods were observed. Firstly, the output ac currents are actually under an open-loop control state, thus, their THDs, which are related to the distortions in ac grid voltages and dc current as well, are not controllable. In addition, when the switching frequency and output LC resonant frequency are close to one another, oscillations were observed on the ac output currents. This appears as a common issue for inverters with a LC lter at the same side and special consideration is required in the selection of the circuit parameters to minimize this effect. 3) Current ripple: In this paper, the power stage is operated as two correlated boost converters in each switching period and the average dc inductor current is automatically maintained constant by the inherent balance property of a balanced three-phase system. Therefore, the dc inductance may be kept small. The switching ripple of the dc inductor current is related to the input line-to-line voltage, the duty cycle, the switching frequency, and the inductance, and remains the same for the two proposed methods provided that the above factors are identical. IV. PRACTICAL APPLICATION ISSUES A. Fluctuation of DC Side Current In Section III, the balanced three-phase grid voltages were considered. However, there will always be some degree of unbalance in the grid lines. The most common case is that the three grid voltages have unequal amplitudes. This will cause undesired uctuation of the power ow, and in turn causes the dc side current to uctuate. For simplicity of analysis, it is suphas a small deviation on its amplitude from posed that only , and the rms values of and do not the nominal value change in (6). Then the variation in caused by the change of its rms value can be derived as (16)

Fig. 9. Time sequence of the conventional PWM method.

in Region I): Sawtooth1 and Sawtooth2 are out of phase, and each of them accounts for half the switching period. The peak , which is a constant, voltage of Sawtooth1 or Sawtooth2 is . In the rst half cycle, when is and greater than , and . The circuit falls is in Stage I. As soon as the ramp signal , the upper comparator ips and , below . Then the circuit then switches to Stage II. Similarly, in the second half cycle when is greater than , and . The circuit switches back to is less than , Stage I again. As soon as , . The circuit the lower comparator ips and then switches to Stage III. 2) DC Side Current: From the above control method, the following equations can be obtained (in Region I): (12) Through the same derivation as (4)(8), the following equation can be obtained: (13) It can be observed from (13) that the dc side current is constant if , , , , and are all xed. Therefore, is not necessarily very large to maintain a constant current, which greatly reduces the losses, size, and weight of the inductor. The can vary in a large range as long as dc input voltage . 3) AC Side Current: From (12), there is (14) Referring to the single-phase equivalent circuit in Fig. 6 (for is constant. Voltage phase A in Region I), in steady state, can also be treated as a constant in each switching cycle if the switching frequency is much greater than the line frequency. From Fig. 6 and (14), in a switching cycle, the average value of is (15)

2306

IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 23, NO. 5, SEPTEMBER 2008

Fig. 10. Zero current switching.

can also be switched off with ZCS, and thus, off. Thus, can be softly switched on and off while in Region I is in the hard switch state. By the same means, can also be operated in ZCS and is still hard switched when the inductor current commutes between phases B and C in Region I. As a result, two of the three switches that are operated at the switching frequency can be switched on and off in ZCS without adding any auxiliary components in Region I. Thus, the switching losses can be reduced. Similar results can be obtained in other regions. C. Stability Analysis The boost-type inverter is a large-signal nonlinear system. With the help of the Poincare mapping technique [11], the system stability can be analyzed as follows. With the OCC method, during a short period of time when the at the beginning dc side inductor current is varying from at the end of the th cycle, the following of the rst cycle to equation can be obtained from Figs. 6 and (8): (19)

where is the rms voltage of phase A. From (2) and (5), it can be derived that

(17) Neglecting the second-order small increment, the maximum disturbance on caused by the deviation of is

The sufcient and necessary conditions for steady-state stability are (20) In steady state, (20) is automatically satised. Therefore, the proposed control method shows a robust stability. Similarly, in the conventional PWM method, at the end of the th cycle can be obtained from Fig. 9 as (21) The term in the second parentheses on the right side equals zero, which is also satised by the steady-state condition. Thus, the control method is stable. D. Output Power Control Form (9), the input power of the three-phase boost-type inverter can be derived as (22) Thus, it is convenient to adjust the total input power of the inverter by varying the value of . The output power, which is the product of the input power and efciency, can then be regulated accordingly. Practically, when the voltage of the dc source, such as a battery, is relatively stable within a certain period of time, the inverter converts the amount of power requested by the demand from the supervision. When the voltage of the dc source, such as PV or fuel cells, is varying in a wide range due to its inherent properties, it is desirable that the inverter should be operated at the maximum power point in order to achieve a high efciency. In this case, the power converted is determined by the source, as explained in the following subsection.

(18) Equation (18) shows that the maximum dc side current uctuation per unit is proportional to the nominal rms value of grid voltage and inversely proportional to the input dc voltage , if amplitude variation is considered here. Other factors, such as multi-amplitude variations or phase shift, may be analyzed similarly. From power ow point of view, (18) also reveals that any output power uctuation will induce input dc is constant, while the dc current uctuation providing that side inductance does not suppress such uctuation in the line frequency range. For heavily distorted grid voltages, extra steps, such as ltering the sensed grid voltage for gating signal comparison or adding a closed-loop portion to enhance output current quality, should be considered. B. Zero Current Switching (ZCS) From the input power point of view, the boost inverter can be viewed as a current-source inverter because its dc side current is maintained almost constant by the proposed control methods. In order to avoid high-voltage spikes across the switches during current commutating between two stages (as depicted in Section II), the switches in the consecutive stages should be turned on before the switches in the current stage are turned off. Gate signal overlapping is usually adopted and thus some switches can be switched at zero current. For example, in Region I as shown in Fig. 10, when the circuit starts to switch is turned on before is turned from Stage I to Stage II, can be switched on with ZCS because off by . This way, is reversely biased and will not shift to until is fully turned off. Similarly, when the circuit is changing from is turned on before is turned Stage II back to Stage I,

CHEN AND SMEDLEY: THREE-PHASE BOOST-TYPE GRID-CONNECTED INVERTER

2307

Fig. 11. Simulation results for P and P

versus V .

(a)

E. PV Connection and MPPT [12] When the dc source in Fig. 1 is a PV array, the operation conand output current from PV, are determined ditions, such as by the actual input power of the boost inverter as well as the solar irradiance and cell temperature at any instant. It is desired to automatically extract the maximum power from the PV array. A simple and cost-effective method was proposed and is depicted as follows. is sensed and introduced to the control circuit in Fig. 5 (23)
(b)

is the voltage sensing ratio and remains constant. where Substituting (23) into (22), the following can be obtained: (24) When , , and are constant and is xed in a specic case, the input power of the inverter is a function of its input is drawn in Fig. 11 with the help of Matlab. In the voltage. can same picture, generic curves of the PV output power are at three different solar also be drawn. The solid lines of are the irradiance levels at 25 C, and the dashed lines of corresponding curves at 50 C. At any time, when equals , a temporary steady state of the system is achieved. These steady-state operating conditions are represented by a series of intersection points, which move up and down with the change of and in Fig. 11. solar irradiance or temperature, between If these operating points are close to the real maximum power points (MPPs), a cost-effective MPPT method can be achieved with in Fig. 5. By this means, the by simply replacing inverter can adjust its input power automatically according to the variation of . For any particular application case in which the output properties of the PV source are known, parameters , , and in (24) are tuned to make the operating points approach MPPs. Thus, the MPPT function is achieved with an acceptable precision. V. EXPERIMENTAL AND SIMULATION VERIFICATION A 1.5-kW laboratory prototype of the OCC three-phase boost-type inverter was built. The major parameters of the prototype are as follows: CLK1 and CLK2 frequencies are 5.4 F, mH 40 kHz; the ac side lter

Fig. 12. (a) Experimental output: grid voltage (210 V/div) and output currents (5 A/div) with one-cycle control method. (b) Experimental input: voltage V (50 V/div) and current I (5 A/div) with one-cycle control method.

6.85 kHz); the line-to-neutral gird (resonant frequency V ; input voltage V. Ripple voltage current of dc side current is set to be 5%, and considering ) case, the minimum is the maximum duty cycle ( calculated to be 0.55 mH. So a dc side inductor is selected to be and . 0.6 mH. Other parameters are Fig. 12(a) shows the measured ac side voltage and currents V. The proposed OCC method with the input voltage regulates the average inductor current during each switching cycle, and effectively controls the three-phase line currents , , and to be sinusoidal and in phase with corresponding gird voltages. The THD of each line current is below 5%. Fig. 12(b) presents the dc side voltage and current. It can be seen that the in the dc side inductor is kept almost conaverage current stant, even though the inductor value is not large ( mH). can be observed around 3 A, In Fig. 12(b), the ripple of which is in the same range of the theoretical analysis in (18) (1.8A uctuation due to 10% amplitude variation on Phase A). However, practically, several other factors, such as the non-ideal input voltage or the amplitude variation from Phase B and/or C, may add to the uctuation in . The overall efciency of the whole system measures about 92.7% in the tested condition. The efciency may vary when power devices, switching frequency, and the output LC lter are different. Fig. 13(a) and (b) shows the transient response when changes between 50 and 100 V. It can be observed that the

2308

IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 23, NO. 5, SEPTEMBER 2008

Fig. 15. Extracted power time. (a)

, the maximum

, and relative error versus

(b) Fig. 13. (a) Experimental step up response V (50 V/div), i (5 A/div), I (10 A/div) with one-cycle control method. (b) Experimental step down response V (50 V/div), i (5 A/div), I (10 A/div) with one-cycle control.

Fig. 15 presents the experimental observations of the boosttype inverter with PV connection. PV array consists of 8 modules (Shell SP75) and is installed on the rooftop of the Engineering Laboratory Facility at the University of California, Irvine with a 60 angle of incidence and facing roughly 45 east of due south. The upper two curves in Fig. 15 show the of the PV output power and the actual maximum power during the daytime. (All the data of Fig. 15 are based on the experiments conducted on Aug. 26, 2005, in Irvine, CA. Weather conditions were sunny, 6591 F, humidity of 50%, and winds generally SSW at 6 mph.). The bottom bar chart shows the reland . The achieved power curve ative error between closely matches maximum power throughout the period. When hits its peak around 13:00 the relative error is only 3.8%. However, before 10:00 and after 16:00 when the temperature drops signicantly from that of midday, the error increases up to almost 20%, because the proposed MPPT method does not account for temperature variations. Nonetheless, the proposed the MPPT method closely tracks MPPs (especially during the period of peak output) with an acceptable precision. The experimental results well match those predicted by the Matlab simulation. VI. CONCLUSION In this paper, the one-cycle control (OCC) method and the conventional PWM method have been proposed for a three-phase boost-type grid-connected inverter, which employs a single power stage for power conversion from a low dc ) with sinuvoltage source to the ac grid system (208 soidal and in-phase three-phase current injection. Compared to a buck-type inverter which needs an additional dcdc converter to boost the dc side voltage over 294 V, the boost-type inverter can accommodate a low dc voltage with a much wider range V). The average dc current is maintained constant ( in each switching cycle in a balanced three-phase system, so the dc side inductance can be kept small. With the proposed OCC method, the control circuit is simple and stable, and the overall system has a fast dynamic response. A cost-effective MPPT function can be integrated into the OCC core with acceptable precision. Experiments have been performed with a 1.5-kW prototype inverter to verify the performance of the proposed OCC methods. Compared to the buck-type inverter, the boost-type inverter has a slightly higher output current THD, but it can be designed to satisfy the UL1741 standards.

Fig. 14. Simulated phase voltage and output currents with the conventional PWM method.

circuit has a wide input voltage range and a fast transient response. Additionally, there is no overshoot on dc or ac currents to go into steady because the stability criteria (20) forces state immediately when has a step change. Therefore, the dynamic process is quite smooth. Fig. 14 shows the simulation results of the conventional PWM control method under the same input and output condition as CLK2 kHz. It can be seen that the above with CLK1 three phase currents , , and are also sinusoidal (THD is close to 5%) and a near unity power factor is achieved.

CHEN AND SMEDLEY: THREE-PHASE BOOST-TYPE GRID-CONNECTED INVERTER

2309

The demonstrated features are very suitable for PV, fuel cell, and many renewable and alternative energy sources, as well as battery power conversion.

[12] Y. Chen and K. M. Smedley, A cost-effective three-phase grid-connected inverter with maximum power point tracking, presented at the Industry Applications Conf. 2006, 41th IAS Annu. Meeting, Tampa, FL, Oct. 812, 2006. Yang Chen received the B.S. and M.S. degrees, both in electrical engineering, from Tsinghua University, Beijing, China, in 1998 and 2001, respectively, and the Ph.D. degree in power electronics from the University of California, Irvine (UCI), in 2005. His research interests include single/three-phase inverter, solar energy usage with MPPT, parallel control of three-phase PFC rectiers/APFs/inverters, grid-connected boost-type inverter and power quality control. Since January 2006, he has been with International Rectier (IR) and worked in the DC-DC power management area.

REFERENCES
[1] Y. Chen and K. M. Smedley, A cost-effective single-stage inverter with maximum power point tracking, IEEE Trans. Power Electron., vol. 19, no. 5, pp. 12891294, Sep. 2004. [2] G. K. Andersen, C. Klumpner, S. B. Kjaer, and F. Blaabjerg, A new green power inverter for fuel cells, in Proc. Power Electron. Specialists Conf. (PESC02), Cairns, Queensland, Australia, Jun. 2327, 2002, vol. 2, pp. 727733. [3] C. Qiao and K. M. Smedley, Three-phase grid-connected inverters interface for alternative energy sources with unied constant-frequency integration control, in Proc. Ind. Applicat. Conf. 2001, Chicago, IL, Sep.-Oct. 304, 2001, vol. 4, pp. 26752682. [4] G. R. Walker and P. C. Sernia, Cascaded DC-DC converter connection of photovoltaic modules, IEEE Trans. Power Electron., vol. 19, no. 4, pp. 11301139, Jul. 2004. [5] F. Antunes and A. M. Torres, A three-phase grid-connected PV system, in Proc. Ind. Electron. Soc. (IECON 2000), Nagoya, Aichi, Japan, Oct. 2228, 2000, vol. 1, pp. 723728. [6] J. C. Lima, J. M. Corleta, A. Medeiros, V. M. Canalli, F. Antunes, F. B. Libano, and F. S. Dos Reis, A PIC controller for grid connected PV system using a FPGA based inverter, in Proc. Ind. Electron. (ISIE 2000), Cholula, Puebla, Mexico, Dec. 48, 2000, vol. 1, pp. 169173. [7] M. Calais, V. G. Agelidis, L. J. Borle, and M. S. Dymond, A transformerless ve level cascaded inverter based single phase photovoltaic system, in Proc. Power Electron. Specialists Conf. (PESC00), Ireland, Jun. 1823, 2000, vol. 3, pp. 11731178. [8] M. Kazerani, Z. C. Zhang, and B. T. Ooi, Linearly controllable boost voltages from Tri-Level PWM current-source inverter, IEEE Trans. Ind. Electron., vol. 42, no. 1, pp. 7277, Feb. 1995. [9] K. D. T. Ngo and J. Chen, Integrator-based linearizing pulsewidth modulator for three-phase inverters, IEEE Trans. Power Electron., vol. 18, no. 2, pp. 612618, Mar. 2003. [10] S. Hiti, V. Vlatkovic, D. Borojevic, and F. C. Lee, A new control algorithm for three-phase PWM buck rectier with input displacement factor compensation, IEEE Trans. Power Electron., vol. 9, no. 2, pp. 173180, Mar. 1994. [11] K. M. Smedley, Poincare stability analysis of switching converters with nonlinear control, IEEE Power Electron. Soc. Newsletter, vol. 14, no. 1, pp. 34, Jan. 2002.

Keyue Ma Smedley (S87M90SM97F08) received the B.S. and M.S. degrees in electrical engineering from Zhejiang University, Hangzhou, China, in 1982 and 1985, respectively, and the M.S. and Ph.D. degrees in electrical engineering from the California Institute of Technology, Pasadena, in 1987 and 1991, respectively. She was employed at the Superconducting Super Collider from 1990 to 1992, where she was responsible for the design and specication of acdc converters for all accelerator rings. She is currently a Professor and Associate Chair in the Department of Electrical Engineering and Computer Science, University of California, Irvine (UCI). She is also the Director of the UCI Power Electronics Laboratory. Her research interests include topologies, control, and integration of high-efciency dcdc converters, high-delity class-D power ampliers, active and passive soft switching techniques, single-phase and three-phase power factor corrected rectiers, active power lters, and grid-connected inverters for alternative energy sources. She has published numerous technical articles and holds nine U.S. patents. Dr. Smedley is an Associate Editor of IEEE TRANSACTIONS ON POWER ELECTRONICS, a Co-chair of the Industry/Education Committee of the Power Sources Manufacturers Association, and the General Chair of IASTED and IEEE Power Electronics Society International Conference on Power and Energy Systems in 2003, and a General Co-chair of the Industrial Conference on Power Electronics for Distributed and Cogenerations in 2004. She received the UCI Innovation Award in 2005.