Introduction

I.1 Generic structure of an industrial variable-speed drive

A variable-speed drive such as the one shown in Figure I.1 is a useful subject for study as it allows us to include most of the functions involved in power electronics; furthermore, devices of this type are widely used in an industrial context. Factories are generally powered by a fixed-frequency alternating network (230–400 V/50 Hz in France), and most of the electrical machines (robots, conveyors, machine tools, etc.), which are used in production lines, with a power requirement of greater than or equal to 1 kW (the most widespread), use a three-phase alternating current (AC) (these are usually induction machines with squirrel cage rotors, or, more rarely, permanent magnet synchronous machines).

In the cases where a variable speed function is required, these machines require a power supply with a variable frequency (and amplitude). However, the electrical network provides a fixed frequency. As we will see, it is easier to produce voltages of variable frequency (and amplitude) using a continuous source (using an inverter) than to transform an AC of frequency f1 into a new AC with frequency f2 ≠ f1 (the device used for this purpose is known as a cycloconverter and will be presented in Chapter 1, but will not be studied in detail as they are no longer widely used). The following chapters will highlight the modular nature of electronic power converters, as in the case of the variable speed-drive shown in Figure I.1. This device uses a rectifier, which allows us to pass from the AC with fixed amplitude and frequency provided by the power network to the continuous supply required by the inverter to produce a network with an AC with variable amplitude and frequency for the electrical device in question. The insulated gate bipolar transistors (IGBT) inverter acts as a rectifier (during machine braking phases), but the diode rectifier used here is unable to return this electrical energy to the network. This means that a brake chopper is needed to consume this excess energy. These different converters will be covered, in turn, over the course of the following chapters.

Before beginning our study of different converter layouts, it is important to consider the specificities of power electronics. This domain involves the control of electrical power, from the power supply to the powered load. The uses of this function are evident, but the methods used need to be covered in more detail.

I.2 Specificities of power electronics

I.2.1 Distinction between signals and energy

Unlike the branch of electronics that concerns information processing equipment (whether analog or digital, using operational amplifiers, A/D or D/A converters, microprocessors, microcontrollers or other programmable or non-programmable digital circuits), power electronics focuses on constructions that process and transform electrical energy to respond to a need. Nevertheless, there are certain similarities between the processing of digital information (digital electronics) and power electronics, due to the way in which electronic components function in switching operations. In power electronics, transistors (among other components) are used in turned off/saturated mode, which renders them equivalent to open or closed switches. While these modes of operation are similar, they do not however concern the same objectives. In digital electronics, this type of function serves multiple purposes. Examples include:

– noise immunity in digital circuits;

– programming new functions in an unmodified physical circuit1;

– improved integration (smaller circuits).

In power electronics, however, the sole aim of using switch functions in components is to maximize efficiency (and therefore minimize loss) in the converter.

I.2.2 Commutations and losses

The interest of switching-based functions may be observed by studying the (simple) construction shown in Figure I.2. This construction is made up of an input voltage source E (e.g. a battery) connected to an association in series with a resistance R (modeling, e.g. a lamp, a heating element, etc.) and a transistor T (in this case, an NPN bipolar transistor).

The equation of this simple single-loop circuit is easy to establish:

=R.IT+VT

The operating point of this circuit is (graphically) the intersection between the characteristic VT(IT) of the transistor T and the load line VT = E − R.IT produced by equation [I.1], as shown in Figure I.3. Note that the characteristic of the transistor is linked to the command signal applied at the base. For example, the operating point M corresponds to a base/emitter voltage VBE = VBE5: we thus obtain a voltage of VT0 between the collector and the emitter of the transistor, and the current IT traveling through the transistor is equal to IT0. The power dissipation of any dipole is the product of voltage × current, in this case VT.IT. The temperature of a component is directly linked to the power dissipation, and constructor documentation for electronic components specifies a maximum junction temperature for operation without damaging the component. Consequently, an electronic component (associated with correctly dimensioned cooling equipment) may be characterized by a maximum power Pmax. The isopower curve PT = VT.IT = Pmax is shown in dotted lines in Figure I.3. We see that point M lies above this curve, and cannot therefore be accessed in steady state. In fact, use of the transistor in the linear zone leads to significant overdimensioning in relation to the power required by the load. This is generally the case for low-power requirements (e.g. class A audio amplifiers) but is not possible for high powers when a high yield is required, notably for equipment with integrated energy, where priority is given to autonomy and/or compactness (using smaller, lighter thermal dissipation elements).

However, the use of the “turned off” and “saturated” transistor modes alone seems to be limiting in terms of the dosage of the power supplied to the load, in which we must either provide no power, in cases where the transistor is operating as an open switch (point B), or maximum power, practically equal to E2/R, if the transistor is saturated (closed switch – point S).

I.2.3 Load inertia and average model

The disadvantage presented by the ON/OFF operating mode for transistors, as discussed above, does not pose significant problems in practice for most applications, due to the inertia of the powered load. The transistor is assimilated to a switch, but, unlike in electromechanical models, this switch may operate using very short open/closed cycles (duration denoted as Td, the switching period) without incurring damage. It is thus possible to make cycles sufficiently short for the powered load to be perfectly insensitive to commutation; it therefore behaves in a manner that is equivalent to a linear power supply with a voltage equal to the average voltage supplied by the switching circuit. The period Td must be short in relation to the time constant of the load:

– thermal time constant for a heating element (generally high – from a number of seconds to a number of minutes in the case of ovens);

– thermal time constant, again, for the filament of an incandescent light bulb (tens or hundredths of a millisecond);

– biological time constant of persistence of vision for a rapid light source2 such as white LEDs3(Td < 40 ms).

An LED powered with a nominal voltage/current 50% of the time will produce a light flow equal to 50% of the nominal flow.

This very general principle also applies to electrical machines, and progress in the design of electronic components means that we can now attain frequencies of the order of around 10 kHz for industrial variable-speed drives with power levels greater than or equal to 1 kW. Evidently, switches commutating low-power levels can reach higher switching frequencies than high-power switches: these components themselves are subject to the same inertia problems as the loads they power. Note, for example, that switching frequencies are often lower than 1 kHz for motors with power levels measured in megawatts, such as those used in railway engines.

I.3 Families of converters

1.3.1 Classification of structures

The most obvious...