Posted on 01 October 2019

Switched Reluctance Drives in Weak Supply Nets

Free Bodo's Power Magazines!




Electronic Stabilization Methods

As a low rate of torque ripple usually is one of the major design targets of switched reluctance drives it is necessary to have more than one torque productive phase at a time.

By Andreas Schramm, Universität der Bundeswehr, München


Particularly when one phase current is being switched off when the respective rotor teeth reach the aligned position relative to that phase, there already has to be a second torque productive (and therefore excited) phase to prevent the output torque from decreasing. That implies that for a certain period of time the two phases have to be supplied with current simultaneously. In [1] and [2] the design of switched reluctance machines with two phases excited at a time and their high torque quality is discussed. The positive effect of phase current profiling on the machine’s performance is described in [3] and [4]. In such cases the DC-link current of the power converter can reach up to twice the value of the desired phase current during commutation. In weak supply nets this can cause a drop of the DC-link voltage. In some applications, e.g. auxiliary drives in aircrafts or automobiles, this voltage drop can not be accepted, because the voltage of the on-board electrical system has to follow certain standards and therefore must not drop below some particular value. Low torque ripple on the one hand and only one torque productive phase at a time on the other hand are contradicting targets. But as a larger voltage drop could affect other indispensable functions of the system under consideration, the prevention of such a deep voltage drop is even a higher aim than a high quality of the output torque of the drive.

In the following, different electronic means for reducing the voltage drop in a system of fixed on-board net and fixed switched reluctance drive will be investigated and compared. A variety of mechanical measures for optimising a switched reluctance drive with regard to a stable DC-link voltage are discussed in [5] and [6].

Simulation model

Figure 1 shows an equivalent circuit diagram of an on-board power supply network with a multiple phase switched reluctance drive connected.

Multiple phase switched reluctance drive and on-board power supply network

The displayed circuit is used for simulating the system’s behaviour. It consists of a synchronous generator with a rectifier, that is in fact simulated using a DC voltage source, which is set to Vsource = 22V. This is a reasonable simplification of the real system, being introduced to combine both, realistic simulation results and short computation time. The on-board power supply network including numerous loads that are not investigated in detail is represented by a number of resistors, inductivities and a capacitor. In order to keep the diagram as simple as possible, only two of the four phases of the simulated drive are shown. For the same reason the ohmic resistances of all connecting cables as well as their inductivities and capacities (if they can not be neglected anyway) are not displayed. The inductivity in each phase finally represents the respective phase winding of the reluctance machine.

Examined voltages

The DC-link voltage, VDC,is one of the characteristic terms of switched reluctance drives. It limits the gradient of the phase currents during commutation and hence acts as a limitation of the maximum possible speed of the machine. The output voltage of the on-board power supply network, Vnet, is the voltage that also has to follow the standards mentioned in Chapter 1. These standards demand for Vnet not to drop below 18V. Figure 2 presents the results of a simulation for both voltages.

DC-link voltage and output voltage of the on-board power supply network.

Herein the drive is operating at low speed with a high torque demand. The firing angles of the semiconductor switches are set to values that enable the highest torque quality to be achieved, i.e. the torque ripple is as small as possible. Therefore the DC-link has to conduct up to twice the phase current during commutation, which causes the voltage drops that can be clearly seen. The fact that the two displayed voltages differ results from the existence of the choking coil (see Figure 1). Although the choking coil diminishes the voltage fluctuations, it does not prevent Vnet from dropping below the claimed value.

Electronic measures reducing the voltage drop

a) firing angles

By switching off the leading phase earlier while not changing the instant of switching on the succeeding phase, the amplitude of the DC-link current can be reduced. Figure 3 shows the simulation results for VDC and Vnet with a slightly (0.4°mech) advanced firing angle for switching off. Obviously now Vnet complies with the standard.

DC-link voltage and output voltage of the on-board power supply network for altered firing angle

As far as the voltage is concerned, this method appears to be successful, but it has a remarkable negative effect on the torque quality (see Figure 4). This large disadvantage of the proposed method leads to the development of the second proposal.

Output torque for different firing angles

b) feed-forward control of the DC-link voltage

To implement such a control mechanism it is necessary to measure the voltage of the DC-link capacitor, VDC, during operation of the drive. When the drop of the measured voltage exceeds a certain threshold the firing angle for switching off is adapted in that way, that the subsequent phase is being switched off a bit earlier. The instant when the capacitor voltage drops down to the lowest possible value is when the DC-link current reaches it’s peak value. This happens right before the moment when the phase which approaches the aligned position is being switched off, because at this very moment both phases are definitely excited. The exact point of time when this happens can be gained from the control parameters of the drive.

Of course, as the functional principle of this method is the same as the one of the method described in a), this also has a negative effect on the torque ripple rate as well as on the mean output torque. Both, the reduction of the mean output torque and the rise of the torque ripple can at least partly be compensated by increasing the nominal value of the phase current.

Figure 5 displays the behaviour of the drive using the proposed control method. The four traces show from top to bottom: output torque, phase currents, DC-link voltage, and electrical input power.

Switched reluctance drive with active firing angle adaptation

Again the drop of the output torque can be recognised always when the phases commutate. The fact that the drop of torque always differs from one commutation period to the next indicates the active state of the firing angle adaptation. The positive effect of this control method on the DC-link voltage can be seen in the third trace. The voltage drop can be prevented almost completely.

c) DC-link capacitor

The third alternative to reduce the voltage drop is to increase the DClink capacitance. The default value of the capacitor is CDC = 8,000µF. Simulation results for a capacity of CDC = 130,000µF are illustrated in Figure 6.

DC-link voltage and output voltage of the on-board power supply network for CDC =130,000µF

This high value results from the thought, that the drop of the DC-link voltage should not exceed (). To realise such a large capacitor in an environment with limited space, the use of Supercaps is recommended. As Supercaps can easily be damaged when they are exposed to higher voltages than (), several of them have to be connected in series. The higher serial resistance of this arrangement has been considered in the simulation.


All alternatives to lessen the drop of the power supply net voltage appear to be successful, as the simulation results look very promising.

The first method of coping with the problem seems to be very cost-effective at first sight, as no physical modifications, neither of the machines, nor of the power electronics have to be implemented. When taking a closer look at this proposal one should recognise, that a very exact monitoring of the angular rotor position is required to make this mechanism work properly. An expensive high resolution resolver is needed to provide that exact data, which raises the costs significantly. Moreover the large disadvantage of this method, the growing torque ripple and the reduction of the mean torque, can not be denied.

The second alternative has the advantage that no precise measurement of the rotor angle is necessary and that always the maximum possible overlap of phase excitation (evaluated against the voltage drop) is guaranteed. Thus the negative effects on the torque performance resembling those of the first method do only occur when the decreasing DC-link voltage really requires an adaptation of the firing angles.

The third possibility should be the one to find high acceptance in connection with the most applications, as it has no negative effect on the torque performance of the drive. Of course also in this case the costs are high, which contradicts one of the main goals of developments for example in the automotive industry.



1) M.E. Zaim, K. Dakhouche, and M. Bounekhla, "Design for torque ripple reduction of a three-phase switched reluctance machine," IEEE Transactions on Magnetics, Volume 38, Issue 2, March 2002, Pages: 1189 – 1192.
2) A. M. Michaelides, and C. Pollock, "Modelling and design of switched reluctance motors with two phases simultaneously excited," Proc. Inst. Elect. Eng. Electrical Power Applications, vol. 143, Sept. 1996, pp. 361–370.
3) A. Greif, "Untersuchungen an geschalteten Reluktanzantrieben für Elektrofahrzeuge," PhD-dissertation, 2000, Universität der Bundeswehr München, Germany, (in German).
4) T. J. E. Miller, "Switched reluctance motors and their control," Clarendon Press, Oxford, 1992.
5) D. Gerling, and A. Schramm, "Mechanical Optimisation of Switched Reluctance Machines in Weak Supply Nets," Power Conversion Intelligent Motion Europe, 2005, Nürnberg, Germany, in press.
6) D. Gerling, and A. Schramm, "Überprüfung der Auslegung einer Reluktanzmaschine hinsichtlich Einbrüchen der Zwischenkreisspannung während der Kommutierung," Technical Report No. 10, 2004, Universität der Bundeswehr München, Germany, not published, (in German).



VN:F [1.9.17_1161]
Rating: 0.0/6 (0 votes cast)

This post was written by:

- who has written 791 posts on PowerGuru - Power Electronics Information Portal.

Contact the author

Leave a Response

You must be logged in to post a comment.