DC/AC Converters – Inverters
DC/AC converters named inverters are employed to convert a dc supply to an ac level of a definite frequency and value. Controlled semiconductor devices, such as transistors, SCR, and GTO thyristors are used in inverters. The input dc voltage may come from the rectified output of a power supply. Alternately, the input dc may enter from an independent source, such as a dc voltage source, a fuel cell, or a battery. In these systems, the supply voltage Ud directly feeds a converter. Typical voltages of dc supplies in vehicles are 12, 24, 48, or 80 V, which are appropriate to make MOSFETs and IGBTs the preferred switching elements. The dc source is able to generate energy for motoring operation of the drive and to sink regeneration energy in the braking operation.
Inverters are usually designed to provide either three-phase or single-phase output. Larger industrial applications require three-phase ac. Low-signal half-wave inverters pass electrical energy during one alternation. These inverters supply loads of 100...200 W powers.
Another classification refers to offline and online inverting. If an inverter is the only source of the load ac line, it is called an offline inverter or autonomous inverter. On the other hand, if an inverter is a part of the common power supply line, it is known as an online inverter or a line -fed inverter.
In accordance with the circuit arrangement classification, the voltage source inverters and the current source inverters are distinguished. A voltage source inverter (VSI or voltage stiff inverter) forms the voltage with required properties: magnitude, frequency, and phase. It is the most commonly used type of inverter. This inverter has the low internal impedance. Generally, it has a capacitor of high capacity connected across the supply source that keeps constant input voltage. The switches of VSI are constructed on the base of the full controlled devices (transistors, GTO thyristors, or MCT). If bidirectional current is required, the freewheeling diodes called also feedback diodes are connected across the switches.
Alternately, a current source inverter (CSI) is the source of the current with the required properties: magnitude, frequency, and phase. As a rule, it has an inductor connected in series with the supply source that keeps the current constant. The switches of the inverter periodically change the output current direction, and the load has very low impedance. Thus, the output voltage of the CSI is shaped according to the voltage drop on the load caused by the output current.
VSI. In Fig. 1, a, a half-bridge midpoint configuration of the single-phase VSI is represented. BJTs, MOSFETs, IGBTs, GTO thyristors, or force-commutational SCRs usually play the role of switches VT1 and VT2. Switches VT1 and VT2 arrange the dc source with the common terminal to supply motor M. Waveforms are shown in Fig. 1, b. During the positive half-cycle of the output voltage, switch VT1 is turned on, which gives Us = +0,5Ud. During the negative half-cycle switch VT2 is turned on, giving Us = –0,5Ud. Note that prior to turning on a switch, the other one turns off; otherwise, both switches will conduct and short the dc supply.
Usually, the load is reactive; therefore, the output current flows as shown in Fig. 1, b. Freewheeling diodes VD1 and VD2 feed the reactive energy of motor M back to the supply line. While Us is positive during 0 < t < 0,5T, either VT1 or VD1 is conducting. However, since Is becomes negative at the beginning of the phase, VD1 must be conducting during this part of the interval. Thus, the feedback diodes conduct when the voltage and current are of opposite polarities.
Figure 10, c, illustrates the single-phase full-bridge VSI. Each of its legs includes a pair of transistors with anti-parallel discharge circuits of reverse current built on the freewheeling diodes. These back diodes provide an alternate path for the inductive current, which continues to flow when a switch is turned off. Again, the diodes return the regenerated power to the dc supply whilst the switches carry the reactive voltage.
Figure 10, d, shows an example of the converter which provides the frequency control of the two-winding induction motor M. The circuit includes a pair of single-phase full-bridge VSIs. The first of them drives the control winding of the motor and the second bridge supplies the exciting winding.
The most frequently used three-phase bridge VSI is shown in Fig. 2, a. It consists of three legs, one per each phase. All inverter legs are equal; therefore, the output of each leg depends only on the dc supply voltage and on the switch status. The output voltage is independent of the magnitude of output load current.
Different modes of the transistors switching on and off may be proposed for this circuit. For example, the possible switching sequence is the next: VT1–VT6–VT2–VT4–VT3–VT5–VT1... In this case, two transistors are together in on state each time span and the output voltages have a rectangle shape, as shown in Fig. 2, b. The firing of the three legs is phase-shifted by 120°. When VT1 is fired, point L1 is connected to the positive terminal of the dc supply, giving UL1 = 0,5Ud. When VT4 is fired, point L 1 is connected to the negative terminal of the dc supply, giving UL1 = –0,5 Ud. Waveforms of L2 and L3 are the same as those of L1, except that they are shifted by 120°. Since each transistor is switched off, its counterpart freewheeling diode passes the tail current in the previous direction. For example, when VT1 turns off, VD4 passes the current until it falls down to zero as the timing diagram shows.
The fundamental harmonic shown by dotted lines predominates here. Other switching combinations are accessible also and will be discussed later.
The problematic mode of an inverter performance concerns obtaining the output voltage, which is higher than the input one. Limitation of the maximum voltage of the switches places the converters either in a restricted area, strongly dependent on the maximum converter voltage, or in the need of associations or more expensive switches, which allow the desired conversion. One of the typical solutions that foresee the use of transformers is given in Fig. 3, a. It is an example of a three-phase VSI built on the three single-phase bridges and three single-phase transformers. Thanks to the star connection of their secondary windings, there is no zero-sequence current in the load. Thanks to transformers, the switching voltage is low enough, thus helping to avoid the switching overvoltages.
Worth attention seems to be the Z-inverter patented by F.Z. Peng in 2003. Its VSI solution is given in Fig. 3, b. In contrast to conventional VSI and CSI inverters, the front-end of the Z-inverter includes a diode and a Z-source of “X” shape, composed of two capacitors and two chokes. In voltage Z-inverter, the diode prevents forbidden reversed current flow, whereas in current Z-inverter – reversed voltage. For this reason, application of the basic Z-inverters is possible only if energy return to the input source is unnecessary, particularly in the fuel cell or photovoltaic cell.
Unlike the conventional bridge inverter, the Z-inverter can assume an additional state defined as a shoot-through state, which occurs when the load is shorted simultaneously by the top and bottom groups of transistors. The main and unique characteristic of the Z-inverter is that the shoot-through state permits to raise the output voltage above the input value. Here, a diode is polarized reversely and does not conduct the input voltage to the inverter therefore energy stored in capacitors is transferred to the chokes. In the non-shoot-through state, when working combinations of the inverter switches are possible, the diode conducts and the voltage changes from 0 to the maximum:
Ud = Ud sup(1/(1— 2q)
where q = ton /Tc - the shoot-through duty factor, satisfying a requirement q<0,5 and ton – the shoot- through time span in the switching period Tc.
The main advantages of the Z-inverters are:
- Increase and decrease of voltage in the one-step energy processing, that is lower costs and decreased loses
- Protection from the short circuits in the inverter branches and from opening of the circuits that improve resistance to failure switching and EMI distortions
- Relatively simple start-up resulting in lowered current and voltage surges
A number of topological variants of the multi-motor applications supply have been proposed. Among them, “single converter – all loads” and “single converter – single load” are the most popular designs. The first topology has often the lowest cost, an average control complexity, good dynamic and static characteristics, and sufficient reliability. At the same time, it requires an additional project development stage and non-standard maintenance arrangement. The second topology is based on standard components and design decisions, though it is more costly, and its control problems arise when the load interconnection is significant.
Today multi-level inverters are the preferred choice for high voltage and high power applications. As the voltage level increases, power circuit complexity will rise as well. A common-mode power circuit of the traction two-motor drive is shown in Fig. 4. To obtain higher voltage possibilities, a couple of series-connected switches built on IGBTs have been added into each leg here. Of course, their introduction results in additional power losses in the circuit, such as conducting losses, blocking losses, and switching losses.
A typical feature of the traction drive is the current unbalance of the phases due to the different loading of the phases, slip, and speed. It results in the leakage currents on the legs and causes control problems. Moreover, due to the switch inequality the current unbalance leads to the voltage unbalance, thus the overheating of the transistors may cause additional power consumption. To avoid unwanted currents and power losses, specific circuit configurations are proposed to drive interconnected motors in the papermaking industry, metallurgical aggregates, traction units, and in other areas.
A five-leg inverter for the coupled-motor system supply is shown in Fig. 5, a. It is constructed by adding two legs to the conventional bridge topology. One of the five legs is a common leg connected with one of the three-phase terminals of each motor, and four other legs are connected with the other two terminals of each motor.
A nine-switch inverter proposed to control the same coupled motor system is shown in Fig. 5, b. This circuit can be considered as an interleaved topology of an upper-side inverter built on switches VT1 ... VT6 and a lower-side inverter, which includes switches VT4 ... VT9. Both inverters may be controlled both jointly and alternately.
A thyristor single-phase forced commutated CSI is shown in Fig. 6, a. Here, the single-phase bridge plays the role of the commutator. For the current source mode, an inductor is included in the input circuit of the inverter. A capacitor is placed in the output as an energetic buffer between the pulsing inverter and the load. In addition, the capacitor is the instrument of the thyristors forced commutation. While the thyristors VS2 and VS3 conduct current, the input voltage charges the capacitor. Since thyristors VS1 and VS4 switch on, VS2 and VS3 obtain the reverse voltage of the charged capacitor, which helps them to close immediately. The capacitor begins recharging to the other polarity, finishing it before the next switching instant. The higher is the current, the faster the capacitor recharging and the forced commutation intervals are shorter.
The thyristor forced commutated CSI are the most widely used systems at power levels in the range 50 to 3500 kW at voltages up to 700 V. The high-voltage versions 3,3/6,6 kV have been developed, however, they have not proved to be economically attractive. Figure 15, b shows a thyristor forced commutated three-phase CSI. The dc current taken from the current source is sequentially switched with the required frequency into the load. The circuit commutation transient may be described as follows. With no commutation in progress, two thyristors, for example, VS1 and VS6 carry the dc while capacitor C1 is positively charged because of the preceding commutation. If thyristor VS2 is now turned on, VS1 is extinguished in a rapid transient and VS2 assumes the dc. This is the starting condition of the commutation transient. While the current in the L 1 phase is now reduced towards zero, the current in the L2 phase is rising. During this interval, the L1 phase is fed through the capacitor C1 as well as the series connected capacitors C2 and C3. Eventually, diode VD1 is blocked and the commutation is completed while VS6 and VS2 are conducting. The diodes are required for decoupling in order to prevent the capacitors from losing their charge necessary for the next commutation. Without these diodes, a capacitor would discharge through two phases of a load.
Idealized waveforms of the output currents are shown in Fig. 6, c. Each thyristor conducts in 60 electrical degrees. When a thyristor is fired, it immediately commutates the conducting thyristor of the same group (top group VS1, VS2, VS3 or bottom group VS4, VS5, VS6).
The first harmonic of the load current waveform is predominant, except when the superposition of voltage spikes caused by the rise and fall of the load current at each commutation. The operating frequency range is typically 5 to 50 Hz, the upper limit being set by the relatively slow commutation process. This system is used to feed the single-motor ac drives of fans, pumps, extruders, compressors, etc., in which good dynamic performance is not required and an inferior power factor, which comes down along with the speed falling, is acceptable.
Conventional three-phase CSI built on IGBTs is shown in Fig. 7. Commonly, it has the large inductor connected in series to the supply source, which keeps the current constant as well as the capacitor bank across the output. Capacitors are required to filter the current harmonics, to make the load current essentially sinusoidal, and to reduce voltage spikes.
In all topologies discussed above, the electronic devices operate in a switch mode where they are required to turn on and off the entire load current during each switching. In these operations, switches are subjected to high stresses and high power loss that increases linearly with the switching frequency. During turn-on, simultaneous current growth and voltage extinction occur in the switches, whereas in the case of turn-off the exact opposite occurs – simultaneous current extinction and voltage growth. In both situations, in real power switches, significant switching losses occur. Another significant drawback of these operations is an electromagnetic noise produced due to the large current and voltage transients. For that reason, the typical switching frequency of the hard switching applications is limited to a few tens of kilohertz (depending on the type of power). These shortcomings of switching converters are exacerbated if the switching frequency is increased in order to reduce the converter size and weight and hence to raise the power density, while maintaining high efficiency.
Contemporary trend is to design DC/AC converters operating at as high switching frequency as possible (from 2 kHz in high power to 200 kHz and more at low power ratings) using fast-switching MOSFETs. As long as at high switching frequency the mentioned negative effects rise to an inadmissible level, dissipative passive snubbers and active snubbers with energy recovery were introduced. Particularly, the turn-on L-R-VD snubber built in series with the transistor allows decreases maximum value of the transistor current and current stress, limiting the turn-on switching losses and transferring them to resistor R. The turn-off snubber C-R-VD built across the transistor allows one to transfer turn-off switching losses from the transistor to the resistor R. Therefore, it decreases the maximum voltage in the transistor. In such a manner, the snubbers produce a more secure switching trajectory of transistors.
In the modern high power electronic converters, the number of snubbers is minimized or they are not used at all. This results from the fact that new power switches are developed as well as from the pursuit of cost decreasing. Obviously, semiconductor devices are then more loaded and should be over-dimensioned.
In such cases, more and more novel solutions are applied in which soft switching is employed instead of traditional hard switching. This concept consists in utilization of resonant tanks in the converters in order to create oscillatory voltage and/or current waveforms. In such a case, Zero Voltage Switching (ZVS) or Zero Current Switching (ZCS) conditions can be created for the power switches.
To realize high frequencies, the switching processes should be produced when the voltage across the switch and/or current through it is zero at the switching instant. With miniaturization being one of the main driving forces for the development of innovative power electronics, resonant converters have attracted considerable attention. Resonant inverters are the switching converters, where controllable switches turn on and off at zero voltage and/or zero current. In that way, high switching frequencies can be realized without pushing switching losses to desirable levels. Therefore, less volume has to be provided to decrease dissipation losses and reactive components can be reduced in size, resulting in designs that are more compact.
Typically, resonant inverters are defined as the combination of different inverter topologies and switching strategies. The power flow to and from the load is controlled by the resonant impedance, which in turn is adjusted by the switching frequency. The following three compositions of the resonant circuit and the load are widespread:
- series resonant inverters using the series connection of the load and the series tank circuit - parallel resonant inverters having a parallel connection of the load and the inductor or the capacitor of the tank circuit
- Series-parallel resonant inverters with the load connection across the part of the tank circuit
Another classification is based on the place of the inductor in the inverter circuit:
- supply-resonant inverters having an inductor in the dc side
- in the load-resonant inverters, an inductor is placed in the ac side
Parallel and series-parallel ZCS resonant inverters are shown in Fig. 8, a, b. They are similar to the CSI although their parameters are strongly different. Here, the input inductor and the output capacitor arrange the resonant circuit with the switching bridge placed between them. Parameters of the tank circuit and the bridge switching frequency are selected by such that the input current has a discontinuous behavior, as Fig. 8, c shows. Thanks to this choice, bridge transistors switch off when the current falls down to zero. When the transistors switch on (t0), the capacitor charges through the reactor to the voltage UC, which is higher than the supply voltage Ud. However, in the t1 instant current falls to zero, thus, transistors get inverse voltage. During the current delay (t1 t2) the capacitor discharges. Then (t2) the next pair of the transistors switches on without current again. From t3, the capacitor voltage changes the polarity. Thanks to zero-current switching, switching losses are low, thus the frequency may be increased significantly.
Half-bridge, midpoint, and full-bridge ZCS series resonant inverters are shown in Fig. 9, a, b, c. Unlike the parallel inverters, the capacitor voltage of the tank circuit does not fall down during the zero-current delay, but the load current is discontinuous here.
The series-parallel ZCS resonant inverter and its timing diagrams are given in Fig. 9, d, e. Since VT1 switches off, current flows through VD2. When current decreases to zero point, the capacitor begins to discharge. Further, the current flows via VT2. On that interval, the capacitor charges with the other polarity. Further, the processes repeat similarly.
As usual, the limitation of the maximum MOSFET voltage restricts converter effectiveness;
nevertheless, the transformers help to solve the problem. This situation is more and more frequent, like in fuel cell systems, in which the input voltage is low, normally between 25 V and 60 V, and the required output voltage is usually compatible with standard ac values, normally between 110 V and 230 V. It means that the voltage ratio is between 5 and 9 when the fuel cell is operating at the rated power.
The resonant inverters displayed in Fig. 10 consist of switching circuits VT1 ... VT4 and LC resonant circuits, thus forming an alternating low voltage. The maximum frequency of the tank circuit LC is near the communication frequency of the switches. Further transformers step the voltage up to the value required by the load motor.
In practice, single-phase inverters are used when the power of the load is 100...200 W. On the output side, this most commonly used inverting approach provides the functions of a voltage source. An effective VSI implementation method involves the use of the transistor bridges with freewheeling diodes.
CSIs can be used for such electrical equipment that needs the control of the current value, particularly in controlled-torque drives. As compared to the VSIs, they are not so popular because of the large input inductor and the requirement in a resistive-capacitive load. Switching frequency of a CSI is smaller, so the load current waveform is distorted, leading to larger derating of the load to avoid overheating. Thereat, instead of a CSI, in electric drives the VSI are used as a current source in which an appropriate current feedback arrangement is used.
In resonant inverters, the controllable switches turn on and off at zero voltage and/or zero current that opens the way to the frequency increasing. Therefore, less volume is required for the converters and reactive components can be reduced in size, resulting in design that is more compact and attracting the attention of portable drive designers. Unfortunately, the frequency of the resonant inverters cannot be changed by the reference signal of the control system.