The present invention relates to a method for braking or retarding primarily an induction motor of the kind provided with a short-circuited rotor winding and supplied with current from a controllable inverter. Although the invention has been developed primarily for application with short-circuited induction motors, the invention can be applied equally as well for braking or retarding other kinds of a.c. motors which include a rotor provided with a damper winding.

Short-circuited induction motors and other variable- speed a.c. motors of the aforesaid kind are normally supplied with power from a standard a.c. net or mains, via a frequency converter. This converter includes a rectifier which is connected to an a.c. net and which for the sake of simplicity and for reasons of economy is normally constructed of non-controllable diodes, and further includes a controllable inverter which is con¬ nected to the d.c. side of the rectifier and provided with controllable switches, the delivered a.c. frequency of which can thus be varied by steering the inverter switches for variation of the speed of the motor con- nected to the a.c. side of the inverter.

Since no power feedback can take place from the motor to the a.c. net with a frequency converter of this con¬ struction, a capacitor is connected to the d.c. link between the rectifier and inverter in parallel across the d.c. input of the inverter. This capacitor takes up variations in current and voltage occurring in the d.c. link during normal operation of the motor at substan¬ tially constant motor revolutions, and also during acceleration of the motor. However, when braking or

retarding the motor, which is effected by lowering the inverter frequency down to a frequency level commen¬ surate with the desired lower motor speed, problems arise as a result of the reduction in the kinetic energy of the motor that retardation of the motor demands. As has previously been mentioned, this excess of kinetic energy cannot be fed back in the form of electrical energy to the a.c. net, but must be consumed in the form of losses. Unless special measures are taken, the kine- tic energy excess manifested when retarding the motor will occur primarily in the form of a voltage increase across the capacitor, this voltage increase decreasing solely at the rate of consumption of the excessive kinetic energy in the form of losses in the motor. A large voltage increase across the capacitor cannot be permitted, however, since such an increase is liable to destroy the capacitor and may cause the capacitor to explode. It is a normal requirement that the voltage across the capacitor will not exceed the rated voltage value by more than 20% for instance.

This can be ensured, by monitoring the voltage across the capacitor when retarding the motor, and by reducing the inverter frequency solely at a rate at which the voltage across the capacitor will not exceed the per¬ mitted value. This results, however, in slow retardation or braking of the motor, which is found unacceptable in many cases.

For the purpose of solving this problem and for achie¬ ving rapid retardation or braking of the motor, it is known to connect a resistor in parallel across the capacitor, with the aid of a controlled switch, to the extent required to prevent the capacitor voltage from exceeding the permitted value. This resistor increases

the losses and therewith accelerates consumption of the kinetic energy of the motor, so that braking or retarda¬ tion of the motor can be effected more quickly. This arrangement, however, is relatively bulky and expensive, inter alia because satisfactory cooling of the resistor must be ensured.

Another known method for retarding or braking such motors is the so-called d.c. braking method.. When braking the motor in accordance with this method, the inverter is first disconnected completely, so that no current will flow to or from the motor, whereafter, subsequent to a given waiting time, some of the inverter switches are closed such that solely d.c. current will flow through the motor winding. This braking method is unsuitable in practice, however, other than when the motor is to be slowed to a complete stop. Retardation of the motor down to a desired, lower speed would require the provision of a complicated control system.

A similar method for accelerating the retardation of the motor involves controlling the inverter during retarda¬ tion of the motor in a manner to reduce gradually the frequency of the inverter while, at the same time, enabling a direct current component to flow through the motor winding. This direct current component gives rise to increased losses and therewith to more rapid retarda¬ tion. The method, however, gives a relatively poor result, since it is necessary to restrict the value of the direct current component to the current values permitted for the inverter switches.

Accordingly, the object of the invention is to provide an improved method for retarding a motor of the kind set forth in the introduction which will enable such motors

to be retarded more quickly than known methods, down to any desired low motor speed whatsoever.

This object is achieved in accordance with the inven- tion, with a method according to the following claims.

The invention will now be described in more detail with reference to the accompanying drawing, in which

Figure 1 is a circuit diagram for a motor arrangement in which the invention can be applied; and Figures 2 and 3 are vector diagrams for the interlinked stator flux of the motor, these diagrams being used to explain the invention.

Figure 1 illustrates, by way of example, a short-cir¬ cuited induction motor 1 which is driven from an a.c. net 2 over a frequency converter. The frequency con¬ verter includes a non-controllable diode rectifier bridge 3, which is connected to the net 2, and a con¬ trollable inverter 4 which is supplied from the diode rectifier and the a.c. output of which is connected to the stator winding of the motor 1. The controllable in¬ verter 4 has a control unit 5 which produces control pulses for the controllable switches of the inverter 4, for determining the frequency of said inverter and therewith also the rotational speed of the motor 1 and the firing sequence of the different switches of the inverter, i.e. the sequency in which the inverter switches are closed and opened. Since, in an arrangement of this kind, no current or power feedback to the a.c. net 2 can take place through the diode rectifier 3, a capacitor battery 6 is connected to the d.c. link across the d.c. input of the inverter 4. This capacitor battery 6 takes up the current and voltage variations which

unavoidably occur with normal operation of the motor 1 at substantially constant speed and also when accelera¬ ting the motor.

When retarding or braking the motor 1, which is effected by instructing the control unit 5 to lower the frequency of the inverter 4, the voltage across the capacitor battery 6, however, will increase strongly, as a result of the excess kinetic energy of the motor 1. As before mentioned, this kinetic energy cannot be fed back to the a.c. net 2 through the diode rectifier 3, but must be consumed in the form of losses in the motor 1. Before the kinetic energy has time to be consumed in this way, it will manifest in the form of a large voltage increase across the capacitor battery 6. Such a large increase in voltage cannot be permitted, however, since it will result in destruction of the capacitor battery 6. Nor¬ mally, a voltage increase of, e.g., 20% above the nomi¬ nal voltage value is permitted across the capacitor battery 6. Consequently, unless special measures are taken, the frequency of the inverter , and therewith the speed of the motor 1, can only be lowered slowly, at a rate which will ensure that the maximum permitted voltage across the battery 6 is not exceeded.

For the purpose of retarding the motor more quickly without exceeding the maximum permitted voltage across the capacitor battery 6, the following procedure is applied.

During normal running of the motor at substantially constant revolutions, and also when accelerating the motor, the inverter 4 is controlled by means of the control unit 5 at a frequency and with a firing sequence pattern for the inverter switches such that the fun-

damental frequency of the a.c. current supplied by the inverter will correspond to the desired speed of the motor 1 and such that harmonic currents of higher fre¬ quency will have a small amplitude and therewith result in small losses in the motor.

When retarding the motor, on the other hand, the in¬ verter 4 is controlled by means of the control unit 5 in a manner such that the fundamental frequency of the a.c. inverter will decrease gradually, in a known manner, down to a frequency commensurate with the desired lower motor speed, this decrease taking place at a rate which will ensure that the voltage of the capacitor battery 6 will not exceed the maximum permitted value. At the same time, however, the firing sequence pattern of the in¬ verter switches, determined by the control unit 5, is also changed in relation to the firing pattern used in normal motor operations at constant motor speeds, or when accelerating the motor, such that harmonic currents of higher frequencies than said fundamental frequency and of significant aplitudes will occur in the winding of the motor 1. These over-frequency currents of high amplitude result in considerably increased losses in the motor 1, because of their high amplitudes and because of the current displacement caused by their higher frequen¬ cies. These substantially greater losses in the motor 1 result in corresponding rapid consumption of the kinetic energy of the motor, so that the fundamental frequency of the a.c. inverter 4, and therewith the speed of the motor, can be lowered rapidly, i.e. the motor is retar¬ ded more quickly without exceeding the maximum permitted voltage level across the capacitor battery 6.

The inventive method can be explained most simply, with the aid of the polygon which can be drawn with the aid

of the vector for the interlinked stator flux of the motor 1. This method of illustrating and studying the behaviour of an induction motor is described in a paper entitled "New Methods in the Education of the Dyn. Behaviour of Indue. Motors and other Rotating-Field Machines", written by V. Torok and published by Poli- tecnico di Torino, in connection with the "International Conference on Evolution and Modern Aspects of Induction Machines", July 8-11, 1986.

Figure 2 illustrates schematically and by way of example one such polygon formed by the vector for the interlinked stator flux V^ of the motor 1. This polygon is obtained for a firing sequence pattern of the in- verter switches used in normal running of the motor at a substantially constant motor speed, and also when ac¬ celerating the motor. The vector for the interlinked stator flux ~ψ" _s of the motor can assume any one of six different directions displaced mutually through 60°. When the motor runs normally, at a substantially con¬ stant speed, or when the motor is accelerated, the firing pattern applied to the switches of the inverter 4 is such that the poplygon formed by the vector"UT will only deviate to the smallest possible extent from a circular configuration. As a result, the over-frequency currents in the motor winding, i.e. currents whose frequencies exceed the fundamental frequency correspond¬ ing to the prevailing speed of the motor, will obtain low amplitudes and therewith give rise to small losses. It will be understood that many different polygons for the vector"\JT are able to fulfill this condition to a greater or lesser extent, and that the polygon can be given an essentially circular configuration, by using a corresponding complicated switch firing sequence in respect of the inverter 4. It can be mentioned in this

respect that each directional change of the vector ^{" *} _s through 60 ^* requires a change in the operational state of two switches in the inverter 4.

When retarding the motor, on the other hand, the switch firing sequence applied is one in which the polygon of the vector for the interlinked stator flux of the motor 1 will deviate markedly from a circular configura¬ tion, as illustrated, for example, in Figure 3. As a result, the over-frequency currents in the motor wind¬ ing, i.e. currents having frequencies which exceed the fundamental frequency corresponding to the speed of the motor, will obtain significant amplitudes and therewith give rise to a substantial increase in losses in the motor, so that the kinetic energy of the motor is con¬ sumed more rapidly, therewith enabling the speed of the motor to be reduced more quickly. Several different sequence patterns for the firing of the inverter switches which will generate vector polygon pathways which deviate markedly from a circular configuration and therewith produce the result desired are also conceiv¬ able in this instance. It has been found particularly advantageous, however, to use a switch firing sequence pattern which produces a vector polygon of a substan- tially six-pointed star configuration, as illustrated in Figure 3.

As will be seen, the polygon illustrated in Figure 3 deviates from a true, symmetric six-pointed star, in that the points of the star are truncated or blunted slightly. This is because if an ideal six-pointed star having "sharp" points is used, these points would give rise to over-frequency currents which although having high amplitudes are of short duration, and consequently such currents will not contribute appreciably to in-

creasing motor losses. On the other hand, these short duration currents of high amplitudes would contribute to the current load on the inverter switches. In addition, a change in the direction of the vector of the kind required at the points of a true six-point star requires a change in the operational state of four switches in the inverter, i.e. a change commensurate with two se¬ quential changes in the direction of the vector each of 60 ^* . Consequently, a switch firing pattern which pro- vides a polygon according to Figure 3 is preferable to a firing pattern which gives a true, symmetric, six- pointed star configuration having sharp points. This latter firing pattern can also be used, however.

It has been found that the over-frequency, loss-genera¬ ting currents are orthogonal in relation to the current of fundamental frequency, such that the peak values of the different currents do not occur simultaneously. One important advantage of this is that the over-frequency currents can be given a significant amplitude, so that said currents will result in corresponding heavy losses, without overloading the inverter switches with current. The over-frequency currents can therefore have an ampli¬ tude of the same order of magnitude as the nominal current value of the inverter switches.

The extent to which the polygon of the vector deviates from a circular configuration can be given quantita¬ tively with the aid of its degree of irregularity according to the expression

Degree of irregularity -= where R _maχ is the largest radius of the polygon and R _min is its smallest radius. It can be mentioned by way of

example that a regular hexagon has a degree of irregu¬ larity of 0.072, whereas a true, symmetrical, six- pointed star with fully pointed tips has a degree of irregularity of 0.268. Trials have shown that a switch firing pattern for the inverter, which produces a vector polygon having a degree of irregularity of at least 1/8 or 0.125, produces a result which is satisfactory from the aspect of the present invention when retarding or braking the motor.