BATTERY CHARGER BASED ON DOUBLE-BUCK AND BOOST CONVERTER Miro Milanovič, Andrej Roškarič and Milan Auda Faculty of Electrical and Computer Engineering, Maribor, Slovenia Key words: electrical accumulators, battery charger, electric converters, voltage converters, buck converters, boost converters, nonlinear current controllers, experimental results Abstract: The paper describes a double-buck and boost converter structure suitable for battery charger in electrical vehicle application. It is very convenient to use inverter elements for battery charging task. Different DC to DC converter structures can be organized from the inverter semiconductor elements. In the case of the induction motor electrical drive application, the required inductors could be established from the motor windings. Such combined converter should satisfy the unity power factor operation as well. By using the double-buck circuit the current high harmonic distortion will be either avoided or significantly reduced. The boost converter established from the inverter elements enables the energy transfer from the mains to the load during the whole mains voltage half period. Such combined converter does not need any additional semiconductor or inductor elements except the diode bridge. Polnilnik baterij zasnovan na dvojnem pretvorniku navzdol in pretvorniku navzgor Ključne besede: akumulatorji električni, polnilniki akumulatorjev, pretvorniki električni, pretvorniki napetosti, pretvorniki navzdol, pretvorniki navzgor, regulatorji toka nelinearni, rezultati eksperimentalni Povzetek: V prispevku je predstavljena kombinacija vezij osnovnih DC/DC pretvornikov z namenom uporabe v polnilniku akumulatorskih baterij. Novo vezje je ocenjeno iz zahtev bremena (napetostna regulacija) in vpliva na obliko vhodnega toka (tokovna regulacija). Z ozirom na delne rezultate je podrobneje razčlenjeno sinhronizirano delovanje vzporedne vezave dveh pretvornikov navzdol in pretvornika navzgor. Na osnovi dinamične karakteristike sestavljenega pretvornika, ki je bila ocenjena z avtoregresivno metodo za identifikacijo sistemov, je bil določen regulator. Na sestavljenem pretvorniku so bile izvedene meritve vhodnih in izhodnih veličin. Rezultati so komentirani glede zahtev bremena in zahtev standarda lEG 1000-3-2. 1 Introduction Power conversion system is sometimes constructed by paralleling converters in order to improve performance or reliability, or attain a high system rating. The parallel converters operation is a usual operation mode in the telephone-exchange power supply systems. Some authors put their atention in the study of the current sharing techniques depending on the load requirements /1/, /2/, /3/. All of them considered the buck converter parallel operation as an autonomous system. The current reference values have been provided to the buck converters from the central control unit and the DC to DC converters do not operate under the synchronous mode /1/. The parallel operation of boost converter is an attractive operation mode as well. In /5/ authors discuss current ripple cancellation by simultaneous operation of two boost converters. The inverters in electric vehicle consist of six transistors and diodes. It is simple to establish the battery charger circuit from this set of elements. It is possible to establish the different battery charger structures by using the inverter elements. Same converter circuits are going to be describe here. From the set of the converter circuits, the double buck and single boost have been chosen. By using this structure, it is possible to avoid the high current distortion when the single buck converter operates. This property could be reached if the double-buck structure is operated in synchronous mode. The voltage control and current inner loop control will be also presented. The objectives for such operation mode is to reach the unity power factor operation with less current distortion than that of ordinary buck converter /4/. Such battery charger structure is also appropriate for the telephone-exchange power supply system. 2 Circuit proposals In electrical vehicle applications and in telephone-ex-change power supply systems, the battery charger circuits require an output voltage which is usualy less than the input voltage. In Fig. 1 (a) the inverter circuit is shown. This is a voltage source inverter and it is supplied from the battery (Uo). With dashed line the diode bridge is shown as well. The diode bridge will operate only when the battery charging function will be required. To transform the inverter circuit function into converters behaviour, suitable for battery charging tasks, some of the connections in inverter circuits should be disconected. The first circuit of the stage is the simple buck circuit which is shown in Fig. 1 (b). By using this configuration, the battery task function has been solved on an appropriate way, but regarding the mains this circuit have a poor power factor. As could be concluded from the input current waveform in Fig. 2(a), the power factor is poor because of two reasons: - the input current is discontinuous; - there is no energy transfer during the whole half period (only when Ud > Uo). ........ JHU i D,i T, D-li T,;' T,i' ID,< •......• Ln :? -----------; . Hk 1 A D^i 1 - • - i "iL t T 1 1 ■ i -------------^ : L, (a) L;, IJ-, i,,. ........i ->...... ^ ■■'"■-r^-lt f- Ds --1 [-1 T T,. • L T\ • -i-N-i 21 ■* A » i D •...... (b) (C) (d) Fig. 1: Inverter to DC-DC converters transformation: (a) Electrical vehicle inverter, (b) Buck converter, (c) Boost and buck converter, (d) Double-buck and boost converter. - Tt (b) (c) The circuit shown in Fig. 1 (c) has an advantage regarding the converter shown in Fig. 1 (b). This circuit enables the energy transfer from the nnains, even if the instantaneous input voltage is lower than the DC output voltage. This advantage has been provided by two boost converters. These two converters could work in simultaneous mode or in synchronous mode as was proposed in /5/. Regarding the current waveform shown in Fig. 2(b), the power factor of this circuit is better than in the previous case but there is still discontinuous current when buck converter operates. To avoid this disadvantage from the inverter circuit, the converter shown in Fig. 1(d) also proposed. This is a double-buck and boost converter circuit. By using the two buck converters, it is possible to avoid the discontinuous current, which appears in the previous proposed converters, as shown in Fig. 2(c). The boost converter provides the energy to the load when the instantaneous input voltage is lower than the output voltage. By using this structure, the previous mentioned disadvantages disappear. The circuit advantages can be summarized as follows: Fig. 2: Input current shapes: (a) Input current for single buck converter only; (b) Input current for single buck and boost converter; (c) Input current for double-buck and boost converter. - the input current is continuous; - there is energy transfer during the whole half period. For the further investigation the double-buck and boost converter has been chosen. .... i / V. >1^ / \ Un / ; \ / ' \ ' \ i.....\ w be Fig. 3: The double-buck and boost converter complete control scheme. jrivertei: co!ive::rei Fig. 5: Intervals where the converter changes Its sturucture depending of Input-output voltage conditions. reference will be established by multiplication of the current reference produced by the voltage controller with the "sinusoidal" signal (ux) of the input voltage normalized to a unity magnitude amplitude as follows: ....... 1 U J -J -w- 1 i , - flir- ...................' I current controller Fig. 4: The current control scheme. T I, 3 Double-buck and boost converter control For the double-buck and boost rectifier, the control requirements are defined regarding the load. There are two operating modes required by the bateries. The rectifier should operate as: - a current source; ^ a voltage source. Regarding the battery condition, the converter will work as a current source or as a voltage source. The unity power factor requirements cause that the rectifier input current on the main side should follow the sinusoidal reference waveshape. This means, that when the rectifier works as a voltage source, a current loop control as an inner loop should be included in the control algorithm. In this case, the current reference will be produced by a voltage controller. The sinusoidal current U. sin(cDt) (1) where Us represents the magnitude ofthe input voltage. In Fig. 3 the complete control scheme is shown. 3.1 Current controller and triggering unit The combined operation of two different converter structures requires a non-linear approach to control the system. In Fig. 4, a current control loop is shown. During the mains voltage half period the converter changes the structure as depicted in Fig. 5. The boost converter operates only when the input voltage is lower than the DC output voltage (Ud < Uo) and the double-buck converter operates when the input voltage is higher than the DC output voltage (Ud > Uo). Each of those two modes requires a special atention regarding inner current references. Regardless what happens with the sinusoidal reference value (if it changes the magnitude), the input current average value in boost converter and the input current average value in the double-buck converter must follow the sinusoidal current reference. 3.1.1 The boost converter operation Regarding Fig. 5 the boost mode of the converter operation is mandatory when the instantaneous input voltage is lower than output voltage value.The hysteresis-controlled principle of operation in the boost mode has been used. To avoid the discontinuous current mode operation from the current reference module, two current references are needed as shown in Fig. 6. From the timing sequences in Fig 6, the hysteresis controller is designed as shown Fig 7. 3.1.2 The double-buck converter operation For double-buck converter operations, the intervals when the instantaneous input voltage is higher than the output voltage are obligatory. Otherwise there will be no energy transfer from source to load. In Fig. 8, the two buck converters synchronous operation is indicated. i_]TLi_m_ From Figs. 4 and 8, the double-buck converter operation can be described. When the current from the first converter, composed of transistor T3 and inductance L2 reaches the sinusoidal current reference value, then this transistor switches OFF and the transistor Ti from the second converter switches ON. The current through T3 is instantly zero and the current through inductance L2 and diode D4 is decreasing. At the same time, the current through transistor Ti is increasing and when the current reaches the sinusoidal current reference, transistor Ti switches OFF and the whole procedure is repeated. The above description can be summarized in the timing diagrams, of Fig. 8. This time sequences help us to design the double-buck hysteresis controller (Fig. 9). Fig. 6: The boost mode hysteresis control. Fig. 7: The boost mode hysteresis controiler. Fig. 9: The double-bucl< hysteresis controiier. rn 1—1 i~L, r ::i u u u n u (1 u n u ...... 1 n : n h n 1 n n 1 n J n J n 1 n 1 n 1 1 1 ► > > > Fig. 8: The double~buci< mode hysteresis control. As derived from the above description, when the converter is operated in double-buck mode, the input current iin does not reach a zero value. This means that in the double-buck converter structure the continuous input current operation mode has been achieved. Actually there are some commutation problems when current goes from transistor Ti to transistor T3 but, appropriate snubber circuit design will make this problems negligeable. 3.1.3 The current reference module As follows from the above description, the current controller needs three current reference values. Two of them are needed by the boost part of the converter and the third one is required by the double-buck converter part. The current reference module provides to the current controllers the apropriate current references. At the output Ud the circuit provides the information to the comparator K3 which will decide which converter will operate. Regarding the circuit in Figs. 7 and 9, the references U(lmax) and U(lmin) will be active when the boost structure has been chosen and the reference U(lref) will be active when the double-buck structure has been chosen. ;;; :-: 1 •-i U, u. ! i,-£ - .......... P, i Fig. 10: The current reference module. + 5V +15V iJ,, R cHIIIr U(ireC o j. a L1 o n Fig. 11: The PWM multiplier. T 3.1.4 The multiplier As it has been shown in Fig. 10, for generating appropriate "sinusoidal" current references the multiplier function should be organized. At the voltage controller output, the PWM signal of iref is available in its voltage form U(iref)- This signal should be multiplied by a voltage "wave-shape" signal Ux. The multiplierfunction can be realized by using the circuit shown in Fig. 11. The circuit consist of an analog switch which is controlled by a PWIVl signal from the digital voltage controller, and a lowpas filter. The signal UxU(iref) = U(Iref) shown in Fig. 12 can be described by Fourier series, as follows: U(lref) = ^ + X®K coskü)t-t-^bkSinkcot (2) k=-.1 k==1 The coefficients ao, ak and bk, can be evaluated in the usual way: t-, t. sSli.- UvU(i.ef) t Fig. 12:The PWM multiplier timing diagrams. U(lre,)dt a^ =■ U(lref)cos kfflodt (3) ' J J U(l|.ef)sin kcoot dt 0 Bn = — T J U^ sin o)t dt 2 an = — T J t. sin üot dt (5) where coo = 1/T. The low-pass filter in Fig. 11 causes that the spectrum higher harmonics will be rejected because its cut-off frequency is much higher than the frequency of Ux which can be described as follows: Bq^^ = -rp J U^ sin cot dt. For coefficients aon, the next equation will be taken into consideration: Ux = Ux sin cot, (4) where co = 23xf = 314.159 rad/s. Because ofthat, only coefficients ao will be identified. Regardig the time diagrams in Fig. 12 for ao coefficients yields: an = — T t Uv sin cot dt = — Ux sin cot dt. -yy':--* I 1 - t jntroliei - _ I - Fig. 13: The voltage control loop. Fig. 14: Dynamic response to a resistive load change: (a) Voltage response; (b) Current response. After soloving (6), it follows: 2U ao =—^ Asin(cot2n +