A Phase-Controlled 12-Pulse Rectifier with Unity Displacement Factor without Phase Shifting Transformer

Transcription

1 A Phase-Controlled 12-Pulse Rectifier with Unity Displacement Factor without Phase Shifting Transformer Yeddo B. Blauth Federal University of Rio Grande do Sul Electrical Engineering Department - DELET Rua Osvaldo Aranha, Port0 Alegre - RS - Brazil yeddo@inep.ufsc.br Ivo Barbi Federal University of Santa Catarina Power Electronics Institute - INEP P.O. BOX Florian6polis - SC - Brazil ivo@inep.ufsc.br Abstract - This paper presents a three-phase 12-pulse phasecontrolled rectifier with unity displacement factor, low line current harmonic content, two quadrant operation and absence of phase shifting transformer. The circuit is composed by two three-phase 6-pulse rectifiers parallel connected by four balancing reactors. One is a conventional thyristor rectifier with a lag firing angle (a). The other is an active rectifier composed by GTOs or IGBTs and diodes, operating with a lead (and symmetrical) firing angle (-a). The 12-pulse operation with line frequency modulation provides a 5-level line current with reduction or cancellation of certain harmonics. Circuit operation, theoretical analysis, key equations, along with experimental and simulation results are presented. Symbology firing angle interval limits of the firing angle Ti-6, Si-6 firing angles peak to peak value of Iuc(t) I~1-4(t) time constant = 271f line frequency auxiliary variable defined in Fig. 12 line current P 1, P2 input currents RMS line current alternating reactors current RMS alternating reactors current RMS (9th harmonic of the line current average current of the reactors current of the reactors average output current inductance of the reactors (equal values) inductances of the reactors = 1,2,3... output power proposed rectifier power factor resistive load equivalent resistances defined in Fig. 13 Rn-6, Rs1-6 RL1-4 s1-6 T1-6 THD VO VO(6n> VOR(G~) v1-3 v12, v13 Vi -4eq(t) VI -4eq Vi VL VM, Vm VPU Vth equivalent resistances of Ti-6, Si-6 switches intrinsic resistances of reactors L1-4 P2 thyristors P1 IGBTs total harmonic distortion of the line current average output voltage RMS (6n)th harmonic of the output voltage with a constant current source load RMS (6n)th harmonic of the output voltage with a resistive load phase voltages line voltages voltage sources defined in Fig. 13 average values of VI -4eq(t) input voltage RMS line voltage amplitudes defined in Fig. 9 and table I normalized output voltage ThCvenin voltage I. INTRODUCTION Mainly in the industrial environment, a great part of the consumed electric power is nowadays processed by front end three phase 6-pulse diode or thyristor rectifiers [l]. Although these 6-pulse rectifiers can still be considered the most important converters in the power electronics universe, they have low power factor and introduce excessive harmonic current in the AC source. They do not comply with the International Standards, such as IEEE-5 19 [2], unless heavy and expensive AC line filters are used. in order to reduce these harmonics and get a larger power factor, some techniques have been proposed and investigated in recent years. One of them is the multi-pulse converter technique, which uses phase shifting transformers or autotransformers to provide cancellation of certain harmonics [3,4]. This is a good solution when the rectifier is composed only by diodes. However, if output voltage control is needed and thyristors /98/$ IEEE. 370

2 are used, the rectifier presents a low displacement factor and a lower power factor. Anyway, this technique is expensive and bulky, due to the input power transformer. Another technique is the pulsewidth modulated controlled rectifier using high frequency modulation, which operates with a unity power factor and sinusoidal input currents [5,6]. However, rectifiers of this type are not well adapted to high power applications, due to high switching losses. In this paper, unity displacement factor, high power factor, low line current harmonic content and low switching losses are obtained with a parallel connection of two rectifiers. The proposed rectifier does not use a phase shifting transformer and uses current multi-levels and a simple line frequency modulation strategy, instead of high frequency modulation. 11. GENERATION OF THE PROPOSED CIRCUIT The proposed circuit has been generated from the generic current multi-level cell shown in Fig. 1 [SI. Fig. 3. Two-level cell bridge circuit identical to the circuit of Fig. 2, but. with a more convenient appearance. VI -@ v2,-@ V3 --@ Fig. 1. Generic current multi-level cell. A bridge circuit composed by two-level cells is shown in Fig. 2 and redrawn for convenience in Fig. 3. In these circuits switches and nodes are respectively numbered and lettered to make easier the topology change. The voltage and current sources have alternate symbols to indicate that the power flow can be in any direction, and the circuits have a 4 quadrant operation. Two 3-phase versions of the circuit shown in Fig. 3 are presented in Figs. 4 and 5. The circuit of Fig. 5 can be viewed as a simple parallel connection of two rectifiers (PI and P2). The nature of the switches will be defined in the next section, after the definition of the modulation strategy. Fig. 5. Three-phase symmetrical rectifier THE CIRCUIT OPERATION Fig. 2. Two-level cell bridge circuit. The adopted mcidulation strategy is shown in Fig. 6. The first harmonic of the input current of the P1 rectifier (IlP1) leads the phase voltage and the first harmonic of the input current of the P2 rectifier (IlP2) lags the phase voltage by the same angle (a). The phasor diagram clearly shows that the first harmonic of the total input current (11) and the phase voltage (Vl) are in phase and, therefore, the displacement factor is unitary for any firing angle (a). 97 1

3 Phase 1 current, of P1 (IlP1) Phase 1 voltage (Vl) Vab I,. Phase 1 current 1 \ ofp2 (IlP2) ;-m-., Phase 1 total k -- -* 4..U-. - # 21,VI Phasor IlP2 diagram Fig. 6. Proposed rectifier modulation strategy. P2 is the well known 6-pulse phase-controlled rectifier fired in the usual way with a lag firing angle (a), while P1 is an active rectifier fired in the same way, but with a lead (and symmetrical) firing angle (-a). Therefore, thyristors can be used in P2, and GTOs or IGBTs and diodes must be used in P 1, Fig. 7 shows the final proposed rectifier circuit. Fig. 8 shows P1 rectifier theoretical output voltage and input current, the gate strategy, and how P1 works with a lead firing angle. It can be noticed from Fig. 8 that any firing pulse lasts 120, and at each time only two switches are conducting (ON state). Fig. 8. P1 waveforms. From top to bottom: output voltage (Vab), phase 1 voltage (Vl), line voltages (V12 and V13), phase 1 input current (IlPl) and firing pulses (Vgl to Vg6). At the tx instant an arbitrary commutation happens. It can be concluded from the gate pulse pattern that immediately before tx, TI and T5 arc ON, and after tx, T5 is OFF and TI and T6 are ON. After tx the output voltage (V13) is smaller than before tx (VIZ), and it can not force any device to turn itself off. Therefore, natural commutation is impossible and self-extinction devices must be used in P 1. The modulation strategy has line frequency and, therefore, low switching losses. Fig. 6 also shows that the total input current has a complex waveform, and how current multilevels are achieved. Because of the multi-levels of the input current, low harmonic content is expected. IV. RELEVANT ANALYSIS RESULTS Fig. 7. Proposed rectifier circuit configuration. A. Output Voltage The average output voltage of the proposed rectifier is given by (1). 3Jz vo = -. v,. cos(a) (1) 7r It can be noticed that (1) is the same as that of the traditional 6-pulse rectifier and the term COS(CI) explains the two quadrant operation. Fig. 9 shows the generic 12 quadrant output voltage waveform when the load is a constant current source. It can be noticed that there are three different 30 intervals in which the output voltage waveform varies in a different way. 972

4 L Interval am am Vm VM Fig. 10. Thevenin equivalent circuit of the proposed rectifier. 24aM - a) I 2,(a - aln) B. Input Current Assuming that the current in reactors L1 to L4 are equal and perfectly smooth, the line current will have the waveform shown in Fig. 11. The three different current waveforms of each 10 quadrant 30 interval are presented in Fig. 11. To any 20 quadrant firing angle (a > 90 ) the current waveform is the same as that of its supplement (1 SOo - a). Validity interval : 60''-2~( 001~< a,+- + 2a 2a _- IC+ >I a Fig. 9. Output voltage waveform with a constant current load a 180O-2~~ Fig Different total line current waveforms according to the firing angle a. '1 3. V, CO(( 6n - 1). a] - cos[( 6n + 1). a] (2) V0(6n) = 7 6n-1 6n+l The line current harmonic content is given by (4), the line current total harmonic distortion is given by (5) and the RMS line current is given by (6). & (cos(ia)l.lo i = 1,5,7,11,... z IL(,) =

5 C. Currents of the Reactors Fig. 12 shows L2 and L1 voltage and current waveforms assuming Oo 5 a 5 60, constant output current and perfect current sharing among reactors. The ripple current of the reactors shown in Fig. 12 is caused by instantaneous differences between P1 and P2 output voltages, and its analysis allows the design of the reactors. # v 1L I Fig. 12. L2 and L1 voltages and currents assuming Oo< a 2 60, constant output current and perfect current sharing among reactors. It is convenient to separate the current of the reactors in two parts: a direct current and an alternating current, like shown in (7). 10 IL1-4(t) = -- + ILac(t) (7) 2 The average current of the reactors is given by (8). IO IL1-4 = 7 L The constant K defined in Fig. 12 is given by (9). K= 3. J2.VL.[sen(a)- a.cos(a)] (9) 2.n.w. L The voltage across the reactors reaches its maximum value at a = 60 and, therefore, this is the worst case (a = 60 ). The worst case peak to peak value ofih,.# is given by (10). M=- &.V, (10) 4.w. L The worst case RMS value of l ~~~(t) is given by (1 1) VL ILac W.L To guarantee continuous conduction even in the worst case, the inductance of the reactors is given by (12) V~ D. Displacement Factor and Power Factor Because of the adopted modulation strategy, the proposed rectifier has a unity displacement factor over the entire operating band. The power factor is presented in Fig. 11 and is given by (1 3). PF= &.cos(a) Io (13) IL IV. CURRENT SHARING ANALYSIS To guarantee a correct operation of the proposed rectifier, the average current of the reactors must be equal. If the firing angle absolute value of all switches of each rectifier are equal, if the amplitudes and the relative phases of the line voltages are equal, and if all switches, wires and balancing reactors resistances and inductances are precisely equal, then the proposed rectifier is composed by two identical rectifiers with symmetrical firing angles, and it can be affirmed that the average current of the reactors will be exactly the same. In this case, there will be a perfect current sharing between the two rectifiers (P1 and P2). Unfortunately, this is not the case in practical circuits. Small variations in any of the above mentioned quantities may cause unbalanced currents. To quantify these unbalanced currents it can be noticed from Fig. 8 and Fig. 13 that the proposed rectifier actually is composed by four different rectifiers: the upper (Tl, T2, T3; Vleq) and the bottom (T4, T5, T6; V4eq) P1 rectifiers, and the upper (Sl, S2, S3; V2eq) and the bottom (S4, S5, S6; V3eq) P2 rectifiers. I m, Fig. 13. Proposed rectifier equivalent circuit Based on Fig. 13, the following equation can be written: From (14) it can be concluded that when the output current is constant, the time evolution of the L1 reactor current Z~l(t) is governed by the time constant presented in (15). It can be 974

6 also concluded that in a steady state, the average current of the L1 reactor is given by (16). Neglecting wire resistances and assuming equal line voltages, the component values of the circuit of Fig. 13 are approximately given by (1 7, 18). 3.JT.vL COS(OTI) + co~(0j-2) + COS(OT~) (17) yecl 2*n - 3 The equations to the other voltage sources and to the other equivalent resistances (R2-4) are analogous to (17, 18). To confirm the above analysis, Fig. 14 shows simulation results where two reactors, one firing angle (an) and one switch equivalent resistance (RT6) are made purposively different. It can be concluded from Fig. 14 that the circuit has only two different time constants and all average currents of the reactors are different. 1 OA CURRENIS of the REACTORS V. COMPARATIVE ANALYSIS The following graphics (figs. 15 and 16) are useful for a comparison between the proposed rectifier (normal lines) and the well-known 6-pulse rectifier (broken lines). All of them have in the horizontal axes the normalized output voltage defined in (1 9). v,, =-- v0 -costa) (19) V0M.X It can be concluded from Fig. 15 that for any output voltage, the RMS input harmonic current and, therefore, the RMS input current of the proposed rectifier are smaller than those of the 6-pulse rectifier. From Fig. 16 lit can be concluded that the proposed rectifier displacement and power factors are larger than those of the 6-pulse rectifier over the entire operating band. The power factor is PF == when a = 15O and is always larger than 0.95 in the usu,al operation band (Yo 2 VoMAX 12). From figs. 15 and 16 it can be concluded that the overall performance of the proposed rectifier is much better than that of the 6-pulse one. Input current Harmonic current 03 I 8A 6A 4A Fig. 15. RMS normalized input current and RMS normalized input harmonic current of the proposed rectifier (normal lines) and the 6-pulse rectifier (broken lines). Power factor Displacement factor 2A TIME 0.5s I.os 1.5s 2.0s Fig. 14. Simulation results of a forced unbalanced situation. VL = 155V, IO = 10A, TI = 58.30, a~2.6, = as14 = 64,80, RL1-4 = OQ, R~1-5 = RS1-6 = 0.8R, RT6 = 2.3R, L1,2 = 0.15H, L3,4 = 0.3H. Table 11 shows that simulation and theoretical results are quite similar. simul. theor. Table IT. Simulation and theoretical results r 1 IL1 IL2 IL3 IL4 21,2 23,4 7.19A 2.81A 6.14A 3.86A 173ms 276ms 7.17A 2.83A 6.19A 3.81A 187ms 286ms [,,, 1 I,-- I U 0 5 I vpu (1 05 I vpu Fig. 16. Power and displacement factors of the proposed rectifier (normal lines) and the 6-pulse rectifier (broken lines).. VI. EXPERIMENTAL RESULTS In order to verity the theoretical analysis and the circuit operation, a laboratory prototype with the following specifications has been built: Po=3kW VL = 380 V 975

7 The output current can be estimated by: % 1, = ~ =! o w (1.35)( 380) The inductance of the reactors was evaluated from (12): (0.462)( 380) L2 L = 8OmH (2n.60)(5.85) The firing pulses of the P2 rectifier were generated with the scheme shown in Fig. 17. The PI lag firing pulse also fires an upldown counter in the up position. When the line voltage changes its signal, the counter begins to count down. PERCENTUAL AMPLITUDE Fig. 19. Harmonic analysis of the prototype input current. The harmonic analysis of the prototype input current shown in Fig. 18 is presented in Fig. 19. It can be seen that the prototype has very low harmonic content. VII. CONCLUSION Fig. 17. P2 firing pulses generation scheme. The firing pulse of the P2 rectifier is generated when the counter reaches zero. Because of the intrinsic symmetry of the line voltage, this scheme guarantees symmetric fring pulses. Fig. 18 shows experimental results obtained with the prototype that confirms the theoretical analysis. It can be concluded from Fig. 18 that the input current has 5 different levels and a unity displacement factor. The measured power factor is PF = 0,974 and the total harmonic distortion of the input current is THD = 20%, while the theoretical values are PF = 0,983 and THD = 18,7%. In this paper, a three-phase controlled rectifier based on a parallel connection of two 6-pulse rectifiers has been presented and analyzed. It was shown that the proposed rectifier has two quadrant operation, low line current harmonic content, high power factor and a much better overall performance than that of the 6-pulse rectifier. Simulation results have been presented to confirm the analysis of the current sharing of the reactors and experimental results have been included to demonstrate the feasibility of the proposed circuit. It is the authors opinion that this new rectifier will find many commercial and industrial applications mainly in the medium and high power range owing to its low frequency modulation and its superior performance. REFERENCES Ref4 300V Z.5Oms Fig Experimental results of a 3kW prototype feeding a 8OQ resistive load with a firing angle a = 200. From top to bottom: output voltage and phase voltage (300V/div.), line current, conventional 6-pulse rectifier (P2) input current and active rectifier (Pl) input current (5A/div.). Time in 2.5mddiv. [ B.R.Pelly, Thyristor Phase-Controlled Converters and Cycloconverters New York: Wiley-Interscience, IEEE Std , IEEE Recommended Practices and Requirements for Harmonic Control in Electrical Power Systems. J.Schaeffer, Rectifier Circuits: Theory and Design New York: Wiley-Interscience, D.A.Paice, Power Electronic Converter Harmonics IEEE Press, T.Kataoka, K.Kawakami, J.Kitano, A Pulsewidth Modulated AC to DC Converter Using Gate Turn-off Thyristors, IEEE IAS Annual Meeting, pp , Y.Sato et al., A New Control Strategy to Improve AC Input Current Waveform of High-Power Parallel Connected PWM Rectifiers, IEEE PCC - Yokohama, 1993, pp M.Chandorkar, D.Divan, R.Lasseter, Control Techniques for Dual Current Source GTO Inverters, Proc. of IEEE Power Conversion Conf - Yokohama, 1993, pp H.Braga, I.Barbi, A New Technique for Parallel Connection of Commutation Cells: Analysis, Design and Experimentation, Proceedings of the IEEE PESC, 1995, pp