6305ELE 2021 - Power El & Drives - S1
6305ELE 2021 - Power El & Drives - S1
6305ELE 2021 - Power El & Drives - S1
School of Engineering
6305ELE
POWER ELECTRONICS, DRIVES AND SYSTEMS
Part 1:
Power Electronics and Drives
E Levi
O Dordevic
2021-2022
TABLE OF CONTENT
© E. Levi, O. Dordevic i
6305ELE Power Electronics, Drives and Systems Power Electronics and Drives
1 INTRODUCTION
Electric energy converters are devices aimed at conversion of either electric energy of
one set of properties into electric energy of another set of properties, or for conversion where
electric energy is at only one side of the conversion chain. Set of properties of electric energy
includes voltage and current level, number of phases and frequency. In the latter case the other
form of energy is most frequently mechanical energy, although it can be chemical (batteries,
for example: chemical energy is converted into electrical energy). While converters that
transform electrical energy into mechanical and vice versa are characterised with moving parts
(for example, rotating electrical machines), converters that convert electric energy with one set
of properties into electric energy with another set of properties are without any moving parts
and are usually called static electric energy converters.
Static electric energy converters fall into two categories: a) static electromagnetic
converters (for example, a transformer) and b) power electronic converters. Energy conversion
that is performed by a static electric energy converter can in general be described with:
𝑓𝑓𝑖𝑖 , 𝑉𝑉𝑖𝑖 , 𝑚𝑚𝑖𝑖 → 𝑓𝑓𝑜𝑜 , 𝑉𝑉𝑜𝑜 , 𝑚𝑚𝑜𝑜 (1.1)
where 𝑓𝑓, 𝑉𝑉 , 𝑚𝑚 stand for frequency, voltage (average in DC case, rms in AC case) and number
of phases, while indices 𝑖𝑖 and 𝑜𝑜 denote input and output side of the converter.
In the case of a transformer frequency at the input and the output is the same, voltage
levels are usually different (although some transformers are used purely for galvanic separation
and input and output voltage levels are then the same), and number of phases at the input and
output is the same. Electric energy conversion is achieved by means of electromagnetic
induction. Transformer circuits at the input and at the output are not connected electrically and
existence of the voltage at the output is a consequence of the magnetic coupling between the
input and the output circuit.
Situation is quite different in power electronic converters (PECs). Frequency, voltage
level and number of phases at the input and at the output can all be very different, and electric
energy conversion from input to output is achieved by means of power electronic
semiconductors. There is a single electric circuit, electromagnetic induction is not involved,
and frequency at both input and output in (1.1) encompasses zero value (i.e., DC case). Electric
energy conversion is achieved by on and off switching of power semiconductors, rather than
by electromagnetic induction.
In broad terms, the task of power electronics is to process and control the flow of electric
energy by supplying voltages and currents in a form that is optimally suited for user loads.
Fig. 1.1 shows a power electronic system in block diagram form. The power input to this
processor is usually (but not always) from the electric utility at a line frequency of 50 or 60 Hz,
single-phase or three-phase. The processed output (voltage, current, frequency and number of
phases) is as required by the load. Normally, a feedback controller compares output of the
power processor unit with a desired (reference) value, and the error between the two is
minimised by the controller. The power flow through such a system may be reversible, thus
interchanging the roles of the input and the output.
The field of power electronics has experienced a large growth in recent years due to
confluence of several factors. The controller in block diagram of Fig. 1.1 nowadays consists of
microprocessors and/or digital signal processors and performs most frequently very complex
control algorithms. Developments in microprocessor and DSP technology have greatly
enhanced capabilities of power electronic systems. Next, developments in semiconductor
technology have enabled significant improvement in voltage and current handling capabilities
and a substantial increase in switching frequency of the power processor of Fig. 1.1. With the
advent of renewable energy generation (wind, photo-voltaic, etc.) use of power electronic
systems has widely spread in the power system area as well.
Power electronic systems exhibit one major difference when compared with linear
electronics. Power semiconductor devices utilised in linear electronics are operated in their
linear region of operation. In contrast to this, power semiconductors utilised in power electronic
circuit are always operated in so-called switch mode. This means that the device is either fully
on, operating with required current carrying capability and almost zero voltage drop, or it is
fully off, carrying zero current and sustaining the required positive/negative voltage. Operation
of a semiconductor device in linear region results in large losses and therefore in low efficiency.
In contrast to this, switch mode operation of a semiconductor device yields operation with
minimum possible loss and thus with maximum efficiency.
Different types of power semiconductors are nowadays available; these constitute the
heart of the power processor of Fig. 1.1 and are dealt with in Chapter 2. However, regardless
of the actual type of the semiconductor, it is always operated as a switch in a power electronic
circuit. The power processor of Fig. 1.1 varies in complexity depending on the application and
on the characteristics of the input and output, but it in general contains always more than one
power semiconductor. The controller is nowadays very often complex, and it may perform
closed-loop control, as shown in Fig. 1.1. The role of the controller is to provide driving signals
for the power semiconductors on the basis of the required control law and the measured
feedback variables. It is worth noting that in many cases power processor, or – as it will be
called further on – power electronic converter (PEC), is operated in an open loop manner,
without feedback. In that case control of power semiconductor devices is performed on the basis
of a pre-programmed control law, without any information about the actual output of the
converter.
Applications of power electronic converters are nowadays numerous and are discussed in
the following section.
From this introductory discussion it is obvious that the study of power electronics
encompasses many fields of electrical engineering, as illustrated in Fig. 1.2. These include
power systems, electric machines, solid-state electronics, analogue and digital control, digital
signal processing, electromagnetics, measurement, and so on. Combining the knowledge of
these diverse fields makes the study of power electronics challenging as well as interesting.
The expanded market demand for power electronics has been due to several factors. First
of all, advances in microelectronics fabrication technology have led to the development of
computers, communication equipment, and consumer electronics, all of which require regulated
DC power supplies and often uninterruptible power supplies. This is a major reason for
expansion in switch-mode DC power supplies and uninterruptible power supplies. Secondly,
increasing energy cost and the concern for the environment have combined to make energy
conservation a priority. One such application of power electronics is in development of new
fluorescent lamps for higher efficiency. Another opportunity for energy saving is in motor-
driven pump and compressor systems. Thirdly, developments in process control and factory
automation increase demand for the enhanced performance offered by adjustable-speed driven
pumps and compressors and for servo drives in robotics. Next, transportation based on electric
vehicles, especially in large metropolitan areas to reduce smog and pollution, is an important
driving force behind developments in power electronic fed electric motor drives. Finally, power
system related applications, including electric energy generation from alternative energy
sources (especially photovoltaic and wind energy systems) and high voltage DC transmission,
are all relatively new areas of power electronics applications with an enormous potential for the
future.
It is nowadays difficult to find any area of human activity in which power electronics is
not present. Table 1.1 lists various applications that cover a wide power range from a few tens
of watts to several hundreds of megawatts. As power semiconductor devices improve in
performance and decline in cost, more systems will undoubtedly use power electronics.
It is more than obvious from these introductory considerations that the area of power
electronics is vast and complex. The emphasis in what follows is placed on a major part of any
power electronic system, power processor of Fig. 1.1. In particular, semiconductor devices, that
are building blocks for any power electronic converter, are at first discussed in Chapter 2.
Chapter 3 deals with a number of topologies of power electronic converters that are nowadays
widely applied. The number of available power electronic converters is enormous, and it is
impossible to cover all of them in this material. A selection is therefore made, in such a way
that the most frequently met converters are elaborated. Remaining chapters are devoted to
applications of power electronic converters in power systems and electric motor drives. The
converter topologies discussed in the chapters on applications are those covered in Chapter 3.
2.1 Introduction
The increased power capabilities, ease of control, and reduced costs of modern power
semiconductor devices compared to those just a few years ago have made converters affordable
in a large number of applications and have opened up a host of new converter topologies for
power electronic applications. The aim of this chapter is to provide a brief overview of power
semiconductor devices that are nowadays in use in power electronics. The emphasis is placed
on terminal characteristics of devices and on their voltage, current and switching speed
capabilities, which are of utmost importance when design of a power electronic converter is
considered.
As already pointed out, power semiconductors are operated as switches in power
electronic converters. Thus, they are either fully on or fully off. If it is possible to consider a
power semiconductor device as an ideal switch, then the analysis of a converter becomes much
easier. It will be shown that in a number of cases semiconductors can indeed be treated as ideal
switches and that such a representation suffices for most application related discussions of
converters.
Presently available power semiconductor devices may be classified into three groups
according to their degree of controllability. Controllability is here used to designate a
semiconductor’s properties with respect to turn-on and turn-off. Turn-on and turn-off may both
be uncontrollable, controllable, or turn-on may be controllable while turn-off is not controllable.
1. Diodes. Neither instant of turn-on nor the instant of turn-off can be controlled
externally. Both are determined by conditions in the power circuit in which diode is placed and
occur naturally.
2. Thyristors (SCRs – Semi-Controllable Rectifiers). Instant of turn-on of a thyristor is
controllable externally through the port called gate. Thus, the thyristor may be brought into on-
state by means of an external signal. The instant of turn-off is however not controllable. The
thyristor will cease conduction when its current falls to zero (actually, below so-called holding
current). It is therefore necessary to provide means for thyristor turn-off external to the thyristor,
within the power circuit of the converter. A thyristor may be turned-off by natural commutation
(when its current naturally falls to zero due to natural sequence of events within the power
circuit) or by forced commutation (when current is forced to fall to zero by action of additional
part of the power circuit which constitutes so-called commutation circuit). Whether natural
commutation is possible or not depends on the power electronic converter in which thyristor is
placed. Forced commutation requires addition of a number of components in the power circuit
that will enable thyristor turn-off in the desired instant (most frequently combination of L-C
components and additional diodes and thyristors).
3. Fully controllable switches. Both the instant of turn-on and turn-off are fully
controllable externally through an additional control port of the semiconductor. There have
been major advances in recent years in this category of devices. The controllable switch
category includes a number of different types of devices such as bipolar junction transistor
(BJT), metal-oxide-semiconductor filed effect transistor (MOSFET), gate turn off (GTO)
thyristor, insulated gate bipolar transistor (IGBT), etc. The ideal fully controllable switch is
characterised with the following characteristics: it can block arbitrarily large forward and
reverse voltage with zero current flow when off; it can conduct large currents with zero voltage
drop when on; it can switch from on to off and vice versa instantaneously when triggered; and
the power required from control source to trigger the switch is extremely small. Obviously,
actual characteristics will deviate from desired ideal ones, as discussed in subsequent sections.
2.2 Diode
Fig. 2.1 illustrates the circuit symbol for the diode and its steady-state current-voltage
characteristic. When the diode is forward biased, it begins to conduct with only a small forward
voltage drop across it, which is of the order of 1 V. Thus, a diode cannot withstand (block) any
amount of positive voltage. When the diode is reverse biased, only a negligibly small leakage
current flows through the device as long as the negative voltage applied across the diode is
smaller than the reverse breakdown voltage. If reverse breakdown voltage is exceeded, the
diode starts conducting in reverse direction and it is destroyed.
In view of the very small leakage currents in the blocking (reverse-bias) state and the
small voltage in the conducting (forward-bias) state as compared to the operating voltage and
currents of the circuit in which diode is used, the current-voltage characteristic may be idealised
as shown in the lower part of Fig. 2.1. The meaning of this idealisation is that diode carries
current in the forward direction with zero voltage drop and that it blocks reverse voltage with
zero reverse current. In other words, diode behaves as an ideal switch. Such a representation of
diode is usually accurate enough for analyses of power electronic converter circuits (but is
inadequate at the design stage, when for example heat sink is to be designed).
At turn-on, the diode can be considered as an ideal switch because it turns on rapidly
compared to the transients in the power circuit. However, at turn-off, the diode current reverses
for a so-called recovery time, as indicated in lower right-hand corner of Fig. 2.1, before falling
to zero. This reverse recovery (negative current) is required to sweep out the excess carriers in
the diode and allow it to block a negative polarity voltage. The reverse-recovery current can
lead to over-voltages in inductive circuits. In most circuits, this reverse current does not affect
the converter input/output characteristic and so the diode can also be considered ideal during
the turn-off transient.
Depending on the application requirements, two types of diodes are usually used in power
electronic circuits:
1. Fast-recovery diodes. These are designed to be used in circuits where semiconductors
are switched at a high frequency. The switching rate is called switching frequency and it
indicates how many times per second a device is turned on and off. Such diodes have reverse
recovery time ratings of less than a few microseconds.
2. Line-frequency diodes. These diodes are aimed at circuits where switching frequency
is low. Typically, at one side of the converter one has the mains, so that switching frequency is
actually the mains frequency (50 or 60 Hz). Reverse recovery time is much larger for this type
of the diode, but they are available with blocking voltage ratings of several kilovolts and current
ratings of several kilo-amperes.
2.3 Thyristor
The circuit symbol for the thyristor and its current-voltage characteristic are shown in
Fig. 2.2. The main current flows from the anode (A) to the cathode (K). In its off-state, the
thyristor can block a forward polarity voltage and not conduct, in contrast to a diode, as is
shown by the off-state portion of the current-voltage characteristic for positive anode to cathode
voltage.
Fig. 2.2: Thyristor: symbol, current-voltage characteristic, and idealised operating current-
voltage characteristic.
The thyristor may be brought into on-state by applying a pulse of positive current to its
third terminal, the gate (G). In order to bring the thyristor into on-state gate pulse must be
applied when the device is in forward-blocking state (i.e., when anode to cathode voltage is
positive). The forward voltage drop is typically 1 to 3 V.Once the device begins to conduct, it
is latched on and the gate current can be removed. The thyristor cannot be turned off by the
gate and, once in conduction mode, the thyristor behaves like a diode. It will cease conduction
and return to blocking state only when the current falls to zero. When the anode current tries to
go negative, the thyristor reverts to blocking state. This allows the gate to regain control in order
to turn the device on at some controllable time after it has again entered the forward-blocking
state.
In reverse bias at voltages below the reverse breakdown voltage, only a negligibly small
leakage current flows in the thyristor. Usually, the thyristor voltage ratings for forward- and
reverse-blocking voltages are the same. The thyristor current ratings are specified in the same
manner as for a diode, in terms of maximum rms and average currents that it is capable of
conducting.
Using the same arguments as for a diode, the thyristor can be represented by the idealised
current-voltage characteristic that is also shown in Fig. 2.2.
Thyristor turn-off will take place when the current tries to go negative. Similarly as with
diode, the current will reverse (go negative) before becoming zero (Fig. 2.3). However, the
important parameter for a thyristor is not the time for which the current will be negative but
rather the so-called turn-off interval which is defined in Fig. 2.3 with respect to the voltage
waveform. It is defined as interval from the zero crossover of the current to the zero crossover
of the voltage and it denotes the minimum time during which voltage across the thyristor must
be negative after current crossover in order that the thyristor regains its forward voltage
blocking capability. During this time reverse voltage must be maintained across the thyristor.
Only after the lapse of this time the device can again block forward voltage. If forward voltage
is applied prior to the end of this time interval, the thyristor will return to on-state leading to
damage of the device and/or circuit. Thyristor data sheets specify the turn-off interval with a
specified reverse voltage applied during this interval as well as a specified rate of rise of voltage
beyond this interval. This interval is sometimes called the circuit-commutated recovery time of
the thyristor.
2. Inverter-grade thyristors. These are designed to have small turn-off time in addition
to low on-state voltage drop. Their turn-off times are in the range of a few microseconds to 100
microseconds depending on their blocking voltage ratings and on-state voltage drop.
3. Light-activated thyristors. These can be triggered by a pulse of light guided by optical
fibres to a special sensitive region of the thyristor. The primary use is in high voltage
applications (high voltage DC transmission, for example) where many thyristors are connected
in series to make one thyristor string.
As already noted, an ideal switch is characterised with zero current flow in blocking state,
zero voltage drop in on-state, instantaneous turn-on and turn-off and minimum power
requirement for triggering. Real devices are unfortunately not ideal. This gives rise to
development of power loss in a semiconductor. If a device dissipates too much power, it may
overheat and fail. Power dissipation in semiconductors is two-fold. It occurs both in on-state,
due to finite voltage drop, and during turn-on and turn-off because these transitions do not take
place instantaneously. Consider a fully controllable switch which in off-state blocks voltage 𝑉𝑉
and in on-state carries current 𝐼𝐼. Fig. 2.5 illustrates switch control signals and waveforms of
instantaneous voltage and current across the switch during on and off intervals and during
transitions from on to off state and vice versa. Instantaneous switch power loss is also shown
in this figure. Shaded area beneath the power loss curve represents energy lost in the
semiconductor. As can be seen from Fig. 2.5, turn-on interval is composed of three time
intervals: short time delay during which current remains at zero although the control signal has
been applied, current rise time during which current builds from zero up to the value 𝐼𝐼, and
voltage fall time during which voltage falls from initial value of 𝑉𝑉 to on-state voltage drop.
Similarly, turn-off time encompasses three intervals as well: delay time during which there is
no change in either current or voltage although turn-off signal has been applied, voltage rise
time during which voltage rises from on-state voltage drop to the value 𝑉𝑉 , and current fall time
during which current falls from value 𝐼𝐼 to zero. From the power diagram it is obvious that the
large instantaneous power dissipation takes place during turn-on and turn-off interval. Power
loss during on-state interval is significantly smaller than the switching loss. If the switch is
operated at frequency 𝑓𝑓𝑠𝑠 , then there are 𝑓𝑓𝑠𝑠 turn-on and turn-off intervals per second. Hence the
average switching power loss 𝑃𝑃𝑠𝑠𝑠𝑠 in the switch due to these transitions can be approximated
with 𝑃𝑃𝑠𝑠𝑠𝑠 = 12 𝑓𝑓𝑠𝑠 𝑉𝑉𝑉𝑉�𝑡𝑡𝑐𝑐(𝑜𝑜𝑜𝑜) + 𝑡𝑡𝑐𝑐(𝑜𝑜𝑜𝑜𝑜𝑜) �. This is an important result as it shows that the switching
power loss is directly proportional to the switching frequency and the switching times.
Therefore, if devices with shorter switching times are available, it is possible to operate them
at higher switching frequencies. The average power loss during on-interval is proportional to
the on-state voltage drop and can be given as 𝑃𝑃𝑜𝑜𝑜𝑜 = 𝑉𝑉𝑜𝑜𝑜𝑜 ⋅ 𝐼𝐼 ⋅ 𝑡𝑡𝑜𝑜𝑜𝑜 /𝑇𝑇𝑠𝑠 .
Leakage current during off-state is negligibly small and therefore power loss in off-state
may be neglected (𝑃𝑃𝑜𝑜𝑜𝑜𝑜𝑜 ≈ 0).
Fig. 2.5: Conduction and switching losses in fully controllable semiconductor devices.
Fig. 2.6: Bipolar junction transistor: symbol, actual and idealised current-voltage characteristics,
and multiple connection of BJTs into standard and triple Darlington.
Fig. 2.7: N-channel MOSFET: symbol, current-voltage characteristics and idealised current-
voltage characteristics.
Fig. 2.8: GTO thyristor: symbol, current-voltage (𝑖𝑖-𝑣𝑣) characteristic, idealised operating 𝑖𝑖-𝑣𝑣
characteristic and connection of snubber circuit for d𝑣𝑣/d𝑡𝑡 limiting in inductive circuits.
A couple of remarks can be given as guidelines when selecting the switches. These are
summarised as follows:
1. Very high switching frequencies, 50 kHz and above: MOSFETs (with voltage and
current ratings typically up to 1 kV and 0.5 kA).
2. Very high voltage and current ratings: thyristors (typically 5 kV, 3 kA, but with low
switching frequencies, up to 0.5 kHz).
3. Medium voltage and current ratings, medium switching frequencies: IGBT (typically up
to 2 kV, 0.5 kA, 30 kHz).
As is obvious by now, each semiconductor, regardless of the type, allows current flow
through a branch of the circuit in which it is inserted, in one direction only. As power electronic
converters are in vast number of cases operated with AC at input or AC at output, or with AC
at both input and output, it is often necessary to allow for bi-directional current flow through a
branch of the circuit. In order to achieve possibility of bi-directional current flow two
semiconductor devices, of same or different type, have to be connected in anti-parallel (back-
to-back connection).
One frequent anti-parallel connection of two semiconductors is shown in Fig. 2.10. Two
thyristors are used. Current flow in both directions (from A to B and from B to A) is controllable,
so that voltage 𝑣𝑣 may take both positive and negative values, as indicated in the current-voltage
characteristic.
From the point of view of current-voltage characteristic situation remains similar to the
one of Fig. 2.10 if two fully controllable switches are connected in anti-parallel. The only
difference is that in such a case it is possible to revert from conducting state to non-conducting
state (i.e., from zero voltage, non-zero current to zero current, non-zero voltage) by means of
the turn-off switch control.
The back-to-back connection, illustrated in Fig. 2.10, is very frequently used in practice,
as will be shown in the next Chapter.
Fig. 2.10: Back-to-back connection of two thyristors and associated 𝑖𝑖-𝑣𝑣 characteristic.
Q1. Explain the desirable characteristics of semiconductor devices that are used in power
electronic converters. How is the reality different from the ideal situation?
Q2. Explain the role of a power electronic converter as an interface between an electric source
and a load.
Q3. a) Explain the difference between various semiconductors with respect to their turn on
and turn off capability.
b) Draw the symbol and idealised current-voltage characteristic for a diode, a thyristor, a
BJT, a MOSFET, a GTO and IGBT.
c) Select the most appropriate semiconductor component for the following application
requirements:
i) 500 kHz switching frequency, device rating 500 V, 250 A;
ii) 500 Hz switching frequency, device rating 4,000 V, 2,500 A;
iii) 25 kHz switching frequency, device rating 1,000 V, 400 A.
d) Explain, using appropriate connection of two thyristors, how bi-directional current
flow can be obtained in one branch of the circuit.