BASIC ELEMENTS OF A DRIVE

A rotational mechanical load in which any one of a wide range of operating speeds may be required is often called an ‘infinitely variable speed drive’ or, more modestly, a ‘variable speed drive’ or ‘adjustable speed drive’. Suitable operating characteristics to provide a given range of load torques TL and speeds N might be provided by a pneumatic or hydraulic drive as well as by several forms of electrical variable speed drive. The basic components of such a drive, Fig. 4.1, are the motor or actuator that delivers torque to the load and controls its speed and the power controller that delivers power to the motor in a suitable form. In this book only those drives in which the drive motor is an electric motor are considered.

The output power developed by an electric motor is proportional to the product of the shaft torque and the shaft rotational speed. The value of the developed torque usually varies automatically to satisfy the demand of the load torque plus any torque associated with friction and windage. Increase of the shaft power due to increase of load torque is usually supplied by automatic increase of the supply current demanded by the motor. Any significant change in motor speed, however, must be obtained in a controlled manner by making some adjustment to the motor or to its electrical supply.

The advances in power electronics since this book was first published in 1987 have chiefly been in the development of more effective semiconductor switching devices. In particular, the future of high power switching applications will involve reduced use of the silicon controlled rectifier (SCR) and gate turn-off thyristor (GTO) and increased use of the metal–oxide–semiconductor (MOS) controlled thyristor (MCT). The most influential development, however, is likely to be due to increased ratings of metal–oxide–semiconductor fieldeffect transistor (MOSFET) devices and, in particular, widespread use of the insulated gate bipolar transistor (IGBT). Design data of these switching devices is widely available from manufacturers. Increases in the range of semiconductor switches and their nonlinear nature has influenced practitioners to move towards computer based design rather than analytical studies. Simulation techniques are widespread and expert systems are under development.

Chapters 1–3 have been extensively revised, compared with the original text, to incorporate much new material, especially concerned with modern semiconductor power switches. With regard to the switching properties of semiconductor devices the authors have adopted an analytically fundamental approach rather than the current industrial standard. This is educationally easier to understand and more conservative in solution than accepted industrial practice.

Some re-organisation of the original text has permitted expansion of the section on Adjustable Speed Drives, now in Chapter 4, to include a brief treatment of various types of synchronous motor.

The previous work on step-wave inverters has been concentrated into the new Chapter 11 and an additional chapter has been included on pulse-width modulation (PWM) control.

INTRODUCTION

The final decision regarding the choice of switching devices and ancillary components must depend on the overall design for total system realisation. Any choice between device options is likely to have an interactive effect on other components. For example, the selection of a certain switching device for use in a particular circuit may or may not require switching aids and/or sophisticated protection. A relatively inexpensive electronic switching device may require ancillary circuitry that is prohibitive in cost, volume requirement, heat generation, excessive control circuit requirements, etc.

A consideration of some of the circuit features listed in the previous paragraph may be grouped under the general heading of ‘circuit protection.’ This, in turn, may be divided between the major classifications of ‘preventive protection’ and ‘abuse protection.’

Preventive protection is required so that when a device is operational it is embedded in an environment which prevents certain secondary device ratings from being exceeded. A typical example of this is the use of snubber circuits, which is adopted to prevent spurious triggering due to transients.

Abuse protection is required to cope with fault conditions which cause the normal primary ratings of the device to be exceeded. Typical examples are overvoltage or current surges arising from short circuiting of the load. In terms of Fig. 2.2 above, the inside or circuit user shell then dilates (i.e. enlarges) to transgress the outer or absolute maximum limit shell of the particular device design.

RATING, SAFE OPERATION AREA AND POWER HANDLING CAPABILITY OF DEVICES

If the modern power control engineer is to make the correct choice of switching device for a given application it is necessary to be aware of the characteristics and limitations of the devices that are currently available. The obvious limitations to any device are its maximum voltage and current ratings, which must not be exceeded. It is hence essential to have some knowledge of device construction and fabrication.

Power handling capability (PH)

The maximum power handling capability PHmax of a semiconductor switch is related to the product of Vbus and Imax. For any semiconductor device the maximum voltage to which the device can be subjected is related to the avalanche breakdown value of the silicon p–n junction, while the maximum device current is limited to the chip current density.The maximum current density is affected by various factors, including temperature and mechanical stress.

Both the maximum voltage and maximum allowable current are affected by the impurity levels present in the chip. Moreover, they are affected inversely in such a way that any increase of one parameter will result in reduction of the other. When endeavours are made to increase a device power handling capability, limiting factors arise as with most engineering problems. A compromise has to be made as invariably the load current flows through the same junctions that have to withstand the supply voltages without unintentional avalanching.

This book is intended as a teaching textbook for advanced undergraduate and postgraduate courses in power electronics. The reader is presumed to have a background in mathematics, electronic signal devices and electric circuits that would be common in the early years of first degree courses in electrical and electronic engineering. It is the writers' experience that engineering students prefer to learn by proceeding from the particular to the general and that the learning route be well illustrated by many worked examples. Both of these teaching practices are followed here and a lot of problems are also included for attempt and solution at the ends of most chapters.

About one half of the text was written while the principal author (W.S.) was on study leave at the Department of Electrical and Computer Engineering, University of Wisconsin, Madison, Wisconsin, USA. His grateful thanks are acknowledged to the stimulating company of Professor Donald Novotny and Professor Tom Lipo during this period of sabbatical scholarship, sponsored by the Fulbright Commission.

It has become evident in recent years that the reign of the silicon controlled rectifier member of the thyristor family, as the universal semiconductor power switch, is drawing to a close. Except in very high power applications the technology of the immediate future lies with three-terminal, control electrode turn-off devices such as the gate turn-off thyristor (GTO), the bipolar power transistor and the field effect power transistor (FET).

THE PRINCIPLES OF RECTIFICATION

The process of electrical rectification is where current from an a.c. supply is converted to a unidirectional form before being supplied to a load. Although unidirectional, the load current may pulsate in amplitude, depending on the load impedance. With resistive loads the load voltage polarity is fixed. The polarity of the voltage across series-connected load inductance elements may vary during the load current cycle.

In a rectifier circuit there are certain electrical properties that are of interest irrespective of circuit topology and impedance nature. These properties can be divided into two groups, (i) on the supply side, and (ii) on the load side of the rectifier, respectively. When the electrical supply system has a low (ideally zero) impedance, the sinusoidal supply voltages remain largely undistorted even when the rectifier action causes nonsinusoidal pulses of current to be drawn from the supply. For the purposes of circuit analysis one can assume that semiconductor rectifier elements, such as diodes and thyristor devices, are ideal in that they are dissipationless and have zero conducting voltage drop.

A study of rectifier circuits is basically a study of waveforms. No energy is stored within a rectifier so that there is a constant connection between the currents and voltages on the a.c. side and the current and voltage on the d.c. side. In rectifier calculations the essential requirement is to obtain an accurate physical picture of the operation and then establish circuit equations that are valid for the particular condition.

THREE-PHASE INDUCTION MOTOR WITH SINUSOIDAL SUPPLY VOLTAGES

A three-phase induction motor contains a three-phase distributed winding that is housed in slots on the stationary part of the motor, usually called the stator. The rotating part of the machine, or rotor, also contains either a distributed three-phase winding or a cage of interconnected copper bars that serve as rotor winding conductors. When the rotor contains a distributed winding the three phases of this winding are connected to three slip rings on the motor shaft and the motor is known as a wound-rotor machine or slipring machine. When a cage of copper bars is used these bars are electrically connected by end rings inside the rotor, no electrical connection can be made to them and the motor is known as a squirrel-cage motor or, more simply, a cage motor.

One set of three-phase windings is connected to a three-phase voltage supply and this set becomes the primary or excitation (field) windings. With a slip-ring motor either the stator or the rotor windings may act as primary windings, although invariably the stator is used. With a cage motor only the stator windings can be used as primary windings. The other set of motor windings, known as secondary windings, is not connected to the electrical supply but is closed on itself. There is no electrical connection between the primary windings and the secondary windings but these are linked magnetically, as in a transformer.

POWER FLOW CONTROL BY SWITCHES

The flow of electrical energy between a fixed voltage supply and a load is often controlled by interposing a controller, as shown in Fig. 1.1. Viewed from the supply, the apparent impedance of the load plus controller must be varied if variation of the energy flow is required. Conversely, seen from the load, the apparent properties of the supply plus controller must be adjusted. From either viewpoint, control of power flow can be realised by using a series-connected controller with the desired properties. If a current source supply is used instead of a voltage source supply, control can be realised by the parallel connection of an appropriate controller. For safety reasons the latter technique is rarely adopted.

The series-connected controller in Fig. 1.1 can take many different forms. In a.c. distribution systems where variability of power flow is a secondary requirement, transformers are often the prevalent interposing elements. The insertion of reactive elements is inconvenient because variable inductors and capacitors of appropriate size are expensive and bulky. It is easy to use a series-connected variable resistance instead, but at the expense of a considerable loss of energy. Viewing from the load side, loads that absorb significant electric power usually possess some form of energy ‘inertia’. This allows amplitude variations created by the interposed controller to be effected in an efficient manner.

Amplitude variations of the controller may be exchanged for a fractional time variation of connection and disconnection from the supply.