EFFECT OF LOAD ON VIBRATION AND NOISE OF LOW-POWER DC ELECTRICAL MACHINES

Yu. V. GORST, P. T. MALTSEV, A. T. MOLODYKH

(Presented by the scientific seminar of the Departments of Electrical Machines and General Electrical Engineering)

The rapid growth of automation and the ever-increasing use of automatic regulation and control systems in various branches of industry, where DC electrical machines are used as actuating elements, impose heightened requirements on their vibroacoustic characteristics. The increase in rotational speeds, together with a simultaneous decrease in weight per unit power in modern electrical machines, leads to an increased influence of oscillatory phenomena both on the operation of the machines themselves and on the accuracy of the functions they perform, as well as on the maintenance personnel. In this connection, the study of vibration and noise in electrical machines becomes of current importance for improving the quality and increasing the reliability of the latter.

In the vibration and noise spectra of DC electrical machines, magnetic vibrations and noise often dominate over the other components and, as a rule, determine their overall level. A number of works [1, 2] have been devoted to the study of magnetic noise in DC electrical machines. However, the calculated values of the magnetic-noise level obtained according to the existing recommendations for low-power electrical machines (up to 600 W) differ substantially from the measured noise-level values. Apparently, because of their small linear dimensions, low-power machines should be regarded as a spherical rather than a cylindrical sound source. As a spherical sound source, an electrical machine is considered in the method developed by B. B. Voronetskii and E. R. Kucher [3] for calculating the magnetic noise of asynchronous electric motors. Meanwhile, DC machines, without indication of their overall dimensions, are considered in [1] as a cylindrical sound source. In addition, up to the present time no reliable method has yet been developed for calculating the magnetic vibrations of DC machines. Therefore, the measured levels of vibration and noise give a more correct idea of the vibroacoustic characteristics of an electrical machine. Since the creation of a vibration- and noise-insulated load for motors and generators presents known difficulties, it is currently accepted practice to measure their vibration and noise in the no-load mode. At the same time, searches are being conducted for more objective methods of measuring and evaluating vibration and noise.

Experience shows that the levels of magnetic vibrations and noise of electrical machines operating under rated-load conditions and in

in no-load operation, differ significantly from one another. Meanwhile, the customer is primarily interested in the levels of vibration and noise under rated-load conditions.

The purpose of the present work is to devise a method for objectively estimating the vibration and noise level of DC electrical machines at rated load from the results of no-load measurements.

When a DC machine operates under rated load, a considerably larger current flows through the armature winding than in no-load operation. As a result, the linear loading increases. An increase in linear loading causes more intense vibrations of the machine frame and, consequently, leads to an increase in the level of magnetic noise [1].

Research Method

The difficulty of the investigation lies in the fact that, when the motor is coupled to a loading device, the vibration and noise of the latter may substantially affect the vibration and noise level of the electrical machine being tested, and may even distort their spectral composition. Therefore, to investigate the effect of load on the vibration and noise of an electrical machine, a special noise- and vibration-isolated loading device was constructed (Fig. 1). As the load, a DC electrical machine operating in generator mode is used; in what follows we shall call it the loading machine (LM).

Fig. 1. Loading device.
1 — microporous-rubber gasket, 2 — enclosure of the silenced chamber, 3 — felt gaskets, 4 — flexible coupling, LM — loading machine, TM — tested machine.

Since low-power electrical machines were subjected to investigation, it proved possible to make the silenced chamber of small dimensions. Materials with the highest possible sound-absorption coefficient $\alpha$ were used. The chamber is made of plywood 10 mm thick and is lined on the inside with two layers of felt 5 mm thick. There is an air gap between the felt layers. Vibration isolation is achieved by means of a 35 mm thick microporous-rubber gasket between the loading machine and the chamber. The loading machine is connected to the motor under test by means of a flexible coupling. For convenience, the end faces of the silenced chamber

made removable, and a layer of felt is fastened in the doorway. Owing to this, when the end faces are closed by means of locks, they fit tightly against the chamber walls and good sound insulation is ensured. Figure 2 gives the noise spectrum of the loading machine outside the chamber (curve 1) and the noise spectrum of the entire loading device (curve 2), from which it is evident that the damped chamber has a sufficient degree of sound insulation. The loading device manufactured and described above can also be used as a vibration- and noise-isolated drive in investigations of the vibration and noise of low-power generators.

Fig. 2. Noise spectra of the loading device:
1 — loading machine,
2 — loading device as a whole.

Measurement of vibration levels was carried out with an IVPSh-type vibrometer; the VZCh-type vibration transducer was fastened to the motor housing. The noise level was measured with an Sh-63-type sound-level meter; the microphone was placed at a distance of 0.5 m from the machine being tested. In cases where the outside diameter did not exceed 150 mm \((D_{\text{н}} = 150\ \text{mm})\), the microphone was placed, in accordance with the industry standard, at a distance of 0.25 m from the outer surface of the product. Analysis of the spectral composition of vibration and noise was carried out by means of a PF-1 half-octave filter. The motor load was gradually increased from zero to the rated value.

Results of the Investigations

DC motors of various types and designs were subjected to investigation; they differed from one another in rotational speed \((n = 1000—3000\ \text{rpm})\), in power (from 100 to 500 W), and in voltage (from 60 to 220 V). In addition, the motors tested differed substantially in their dimensions, and in design they included both flange-mounted and foot-mounted versions, both without a skew of the armature slot and with a skew by one slot pitch. More than 10 different types of DC electrical machines (up to 10 units of each type) were investigated, for which a characteristic feature is the absence of additional poles. As a result of mathematical processing of the experimental data [4], a linear dependence of the level of magnetic vibrations and noise on the load magnitude was established. Analysis of the data obtained shows that the increase in the levels of magnetic vibrations and noise of electrical machines depends not only on the load but, to some extent, also on the motor power. In this connection it became necessary to divide the electrical machines into a number of groups according to power.

Fig. 3. Change in magnetic vibrations and noise under the influence of load.
1 — change in magnetic vibrations for machines with power \(P < 200\ \mathrm{W}\);
2 — change in magnetic noise for machines with power \(P < 200\ \mathrm{W}\);
3 — change in magnetic vibrations for machines with power \(P = 200 \div 500\ \mathrm{W}\);
4 — change in magnetic noise for machines with power \(P = 200 \div 500\ \mathrm{W}\).

The machines investigated were divided into two groups: group I includes machines with power up to \(200\ \mathrm{W}\), and group II those above \(200\ \mathrm{W}\).

The results of the investigations are presented in Fig. 3, where curves 1 and 2 show the influence of load on the level of magnetic vibrations and noise of machines of group I, and curves 3 and 4 for machines of group II. It is evident from the graphs that, for dc motors with power up to \(200\ \mathrm{W}\), the levels of magnetic vibrations and noise at rated load increase by \(5\text{–}6\ \mathrm{dB}\) compared with the levels obtained at no load. For machines with power above \(200\ \mathrm{W}\), this difference is \(8\text{–}10\ \mathrm{dB}\). A slight deviation from these data is observed in motors having a slot skew of one slot pitch, where the difference in levels at no load and at rated load is \(4\text{–}5\ \mathrm{dB}\) for group I and \(7\text{–}8\ \mathrm{dB}\) for group II.

However, it should be noted that the level of vibration and noise at the tooth frequency for machines of group I, measured in the no-load mode, is \(8\text{–}10\ \mathrm{dB}\) higher than for machines of group II. Therefore, an increase in the level of magnetic vibrations for machines of group I by \(5\ \mathrm{dB}\), and for group II by \(9\ \mathrm{dB}\), expressed in absolute units of vibration velocity, is approximately the same value, \(\Delta_v \approx 0.01 \div 0.012\ \mathrm{cm/sec}\). Similarly, for airborne noise of magnetic origin this value is \(\Delta_p \approx 2.2 \cdot 10^{-1} \div 2.4 \cdot 10^{-1}\ \mathrm{bar}\). Apparently these values will also be valid for more powerful dc machines; however, verification of this assumption requires carrying out a broader experiment.

Conclusions

Analysis of the results of the experimental investigations carried out makes it possible to draw the following conclusions.

1. The levels of magnetic vibrations and noise of low-power DC electrical machines operating under rated-load conditions and under no-load conditions differ from one another by a certain constant value.

2. The magnitude of the change in the levels of magnetic vibrations and noise depends on the power of the machine.

3. An estimate of the levels of magnetic vibrations or noise of low-power electrical machines operating under rated-load conditions can be made from the results of measurements during no-load operation of the machine by calculation according to the following formula:

\[ L_{\mathrm{nn}} = L_{\mathrm{xx}} + \Delta, \]

where \(L_{\mathrm{nn}}\) is the level of magnetic vibrations or noise of the electrical machine under rated-load conditions;

\(L_{\mathrm{xx}}\), \(L_{dx}\) is the level of magnetic vibrations or noise of the electrical machine under no-load conditions;

\(\Delta\) is a constant addition in dB, obtained experimentally.

The value of \(\Delta\) depends on the power of DC electrical machines and is: for the level of magnetic vibrations, \(\Delta = 5 \pm 0.5\) dB for machines with power \(P \leq 200\) W, and \(\Delta = 9 \pm 0.5\) dB for machines with power \(P = 200 \div 500\) W; for the level of magnetic noise, \(\Delta = 6 \pm 0.5\) dB and \(\Delta = 9.5 \pm 0.5\) dB, respectively.

Literature

1. I. G. Shubov, “Determination of the magnetic-noise level of DC machines by the method of electromechanical analogy,” Elektrichestvo, No. 4, 1958.

2. N. P. Ermolin, “Magnetic noise of DC machines,” Izvestiya LETI, issue XXVIII, 1955.

3. B. B. Voronetskii and E. R. Kucher, Magnetic Noise of Asynchronous Electric Motors, Gosenergoizdat, 1957.

4. A. M. Dlin, Mathematical Statistics in Engineering, Soviet Science Publishing House, 1958.