2.2 Power Electronics for Transmission Systems
2.2.3 IGBTs and SCC Transmission
The IGBT, first experimentally demonstrated in 1979 [13], combines the voltage-driven high-impedance and low-power gate of MOSFET technology with the power handling capability of bipolar junction transistors (BJTs) and thyristors. However, it wasn’t until the 1990s that its feasibility for high voltage applications was realised. The internal structure of the IGBT, and its equivalent circuit, is illustrated in Figure 2.8. The structure is identical to that of a MOSFET, except for the highly doped p-type layer located at the
collector of the device, which acts as a p-type emitter of a pnp BJT to the drain of the
MOSFET. The BJT is driven by electrons from the MOS channel which provide the base current, and holes are injected from the p-type emitter. Like with a MOSFET, a voltage bias on the gate results in the inversion of the p-type region directly underneath the gate oxide layer; this results in the formation of a shallow channel region which allows current flow from the collector to the emitter.
With reference to the equivalent IGBT circuit shown in Figure 2.8, the most signifi- cant breakthrough in the development of IGBT technology was obtaining a low shorting
resistance (RS) between the emitter and base of the npntransistor (in the p-type regions
underneath the emitters) to minimise the risk of the parasitic pnpnthyristor (formed by
the P+/N−/P−/N+ junctions from the collector to the emitters) latching up. This
latch up condition could occur at high voltages, when the level of electrons injected from the n-type emitters can turn on the parasitic thyristor. The latching up of an IGBT is an undesirable condition, as the collector current is no longer controlled by the gate, and the device can only be turned off in the same manner as the conventional thyristor. However, unlike with the thyristor, this turn-off has to be done quickly in order to prevent excessive power dissipation permanently destroying the IGBT. The required low shorting resistance
Collector (C) Gate (G) E Emitter (E) N+ N+ P+ P+ N- drift region P- P- P+ Oxide C G E RD Cmiller NPN PNP RS
Figure 2.8: IGBT structure and equivalent circuit.
for minimising the risk of latch up is achieved by controlling the doping profiles within
the device to ensure that the npn section of the parasitic thyristor cannot turn on.
As with MOSFET devices, IGBTs are fabricated in a cellular topology, meaning that when the device is turned on, the problem of slow current spreading that is observed in the conventional thyristor is avoided. Furthermore, when the gate voltage is removed to turn the device off, the base drive current is cut off, allowing a depletion region to quickly
form at the emitter pn junction; as such, the turn-off delay of the IGBT is much less
than than the storage time in BJT devices. Another advantage of the IGBT is that it is
possible to exercise linear control through the gate, meaning thatdi/dt and dv/dtcan be
controlled during the commutations. This in turn means that there is no need for adi/dt
With the use of IGBTs, the flexibility of HVDC transmission systems can be greatly enhanced by using SCC VSC-based power conversion, and the problems associated with CSC-LCC transmission, outlined in Section 2.2.2, are eliminated. VSC-based transmission allows the flow of active power and the provision of reactive power in both directions at each end of the HVDC link, which is particularly advantageous for cable transmission, as the absence of polarity reversal means that the cable design is greatly simplified [14]. SCC-VSC also has the benefit of not requiring an AC system voltage source for the commutations, and can be controlled to generate or absorb reactive power independently from the active power flow. Furthermore, the generation of harmonics is greatly reduced in VSC-based systems, due to the relatively poor quality of LCC waveforms, meaning that the requirement for filters is either eliminated or their size allowed to be reduced to absorb only the higher harmonics. Finally, because the direction of current flow can be changed, fast power reversal at each terminal of a HVDC scheme is achievable without the need for switching operations.
Though VSC-based systems have clear benefits in terms of the flexibility of power transmission when compared to conventional thyristor-based LCC systems, they are not without their disadvantages. In the LCC system, due to the combination of peak current- limiting smoothing reactors and fast converter control, which quickly reduces the DC current to zero, the response to DC system faults is rapid, and a restart is possible in less than 300 ms. However, in a VSC system, the presence of the free-wheeling diodes means that the fault current is allowed to continue indefinitely even when the IGBTs are turned off. As such, the fault can only be cleared by the use of circuit breakers, thus meaning that large delays in restoring normal system operation will be realised. Another disadvantage of VSC when compared to LCC is that, due to the high frequency pulse width modulation (PWM) switching of the IGBTs, the power losses in VSC systems are substantially higher
than in LCC systems. However, by using multi-level conversion techniques [15] instead of PWM-based VSC, these power losses can be greatly reduced, as the main IGBTs within the multi-level converter (MLC) are switched at the fundamental frequency, meaning the higher frequency components typical of PWM are eliminated. In [16], where a MLC based on SiC junction field-effect transistors (JFETs) has been investigated, efficiencies of 99.8% have been reported.