STUDY OF THE INFLUENCE OF DYNAMIC UNBALANCE ON THE VIBRATION OF INDUCTION MOTORS

V. T. Tikhonov

At the present time, researchers and manufacturing plants of electrical machines are faced acutely with the task of eliminating vibrations that arise during the operation of electric motors.

Standards for the permissible vibration and unbalance of induction motors have not yet been included in the all-Union state standard. However, there are a number of draft standards, the introduction of which should be preceded by investigations to identify the causes that generate vibrations, with the aim of their subsequent elimination. In our opinion, the most rational is the second version of the draft standards for permissible vibrations of induction motors of the unified series, with power from 0.6 to 100 kW and higher, recommended in 1957 by the Scientific Research Institute of the Electrical Industry. In the draft standards (Table 1), electric motors are divided into two

Table 1

Table 1 (continued)

class according to vibration requirements and into three groups according to frame size. The first class includes electric motors with high vibration requirements, and the second class those with normal requirements.

The specified standards apply to asynchronous electric motors whose bearings are built into the end shields. Before shipment from the manufacturing plants, the electric motors are tested for permissible vibration at no load, in an unsecured state, on a foundation plate or on a special stand for flange-mounted motors. The tests are carried out without coupling sleeves and pulleys. In the course of the tests, the peak-to-peak value (double amplitude) of the oscillation of the bearing supports is measured in the horizontal, vertical, and axial directions.

In the present work, some results are given that were obtained in the study of vibration of asynchronous electric motors. The task before us was to determine the causes as a result of which vibrations arise, and to outline ways of eliminating them. For this purpose, studies were carried out of the separate influence of imbalance of the rotor, fan, and drive pulley on the vibration of asynchronous electric motors. In addition, vibrations caused by defects in the bearing pair were identified. The results of these studies are of definite interest both for plants of the electrical-engineering industry and for plants of mining electrical-machine building.

The investigation of vibration at various degrees of balancing of the rotor, fan, and drive pulley was carried out on a massive cast-iron plate weighing more than 200 kg. Elimination of the influence of vibration from the external environment was achieved by mounting the installation on a wooden base lying on a layer of wood shavings more than 0.5 m thick. The edges of the base had an air gap and did not touch the walls of the box for the shavings. Vibration was measured with the aid of a vibroprobe with a piezoelectric element, manufactured by us. The vibroprobe sensor during measure-

vibration was in a suspended state, was balanced by a counterweight, and its feeler touched the tested point on the electric motor.

In carrying out the experiments, vibration was measured at no load; motors of the usual type were tested in a free, unsecured state. Flange-mounted electric motors were tested on a special stand. Vibration was measured at the same points of the bearing and the frame of the motor under test. The experiments were carried out with induction motors of types AO31-2, AOF4-2, and ASH3-31-6, manufactured by the Tomsk plant “Sibelektromotor.” The technical data of the electric motors are summarized in Table 2.

Table 2

In preparing an induction motor for testing, the motor rotor was dynamically balanced to a residual unbalance of tenths of a gram-centimeter. At the end face the rotor had threaded sockets for fastening weights of known mass. The sockets were arranged on the right and left sides of the rotor so that it was possible to create artificially both static and dynamic unbalance. In studying the vibration produced by rotor unbalance, the measurement and recording of readings were carried out without the fan and drive pulley. The magnitude and nature of the vibration were determined by analyzing the vibrograms and measuring the peak-to-peak amplitude of oscillations. The vibration was recorded with a loop oscillograph with a time mark of 500 cps and a film speed of 500 mm/sec. The vibration-recording apparatus had a linear characteristic with a sensitivity threshold of not more than 1 μ.

Fig. 1. Vibrogram taken from the housing of ball bearing No. 304 of electric motor AO31-2 with a dynamically balanced rotor. The pulley and fan are removed. Peak-to-peak oscillation amplitude 5 μ.

In Figs. 1, 2, and 3 are shown vibrograms taken from the shields of the bearing supports of electric motors AO31-2, AOF-42-2, and

ASHZ-31-6, for the case when their rotors had been dynamically balanced with a residual unbalance of less than 0.5 g cm.

Let us compare the frequency of the vibrogram recordings (Figs. 1, 2, 3) with the frequency of the time-marker curve. The lower curves on the vibrograms correspond to the vibration frequency of the marker, equal to 500 Hz. During the experiments, the rotational speed of the AO31-2 and AOF-42-2 electric motors was 2860 rpm, and that of the ASHZ-31-2 was 1000 rpm, which corresponds to a rotor rotational frequency of 47.66 Hz for the first two motors and 16.66 Hz for the ASHZ-31-6 motor. Thus, one period of rotor rotation for the first two electric motors corresponds to a vibrogram segment of 10.5, and for ASHZ-31-6 to 30 periods of the time marker. Analysis of the results indicates that the frequency of the recorded vibrograms considerably exceeds the rotational frequency of the electric-motor rotors. The vibration is caused by reasons not associated with rotor unbalance.

Fig. 2. Vibrogram taken from ball-bearing housing No. 306 of the AOF42-2 electric motor with a dynamically balanced rotor. The pulley and fan have been removed. Vibration amplitude 7 microns.

Fig. 3. Vibrogram taken from the sliding-bearing housing of the ASHZ-31-6 electric motor with a dynamically balanced rotor. The pulley and fan have been removed. Vibration amplitude 3 microns.

For the vibrograms shown in Fig. 1, in a recording segment equal to one period of rotor rotation, seven vibration peaks are counted. This is especially clearly visible on the enlarged vibration curve shown in Fig. 4. Evidently, in this case the peaks are associated with jolts and impacts that occur in the bearing pair as the balls move along the raceways of the electric motor’s ball bearings. This is also confirmed by analysis of the vibrograms taken from the bearing shield of the AOF42-2 electric motor (Fig. 5). In the vibrogram segment corresponding to the rotor rotation period, there are eight peaks of maximum vibration amplitude associated with defects in the bearing pair. It should be noted that the number of peaks on the vibrograms cor—

decreases with the number of rolling elements in ball bearings. Ball bearing 306, on which the AOФ42-2 electric motor is mounted, has eight balls, while ball bearing 304 has seven balls.

A vibrogram taken from the housing end shield of the sleeve bearing of the AШ3-31-6 electric motor has a different character. The vibrations here are very small and are explained by defects in the liners and the shaft journal (Fig. 3).

Fig. 4. Enlarged image of Fig. 1. Over an interval of 10.5 periods of the time marker there are seven vibration peaks, which corresponds to the number of balls in ball bearing No. 304.

Experiments on measuring the amplitude of oscillations in electric motors whose rotors had been dynamically balanced (residual unbalance less than 0.5 g·cm), with the fan removed and

Fig. 5. Enlarged image of Fig. 2. Over an interval of 10.5 periods of the time marker there are eight vibration peaks, which corresponds to the number of balls in ball bearing No. 306.

the drive pulley, showed that vibrations associated with defects in the bearing pair lie within the range from 3 to 10 μm.

As the rotational speed of the rotor increases, the vibration magnitude of a motor mounted on rolling bearings increases somewhat faster than the vibration of a motor assembled on sleeve bearings.

On the basis of the foregoing, it follows that, for the manufacture of induction electric motors according to the first class with a raz-

With an oscillation amplitude of 3–5 microns, in addition to careful dynamic balancing of the rotor, it is necessary to eliminate the causes that produce vibrations in the bearing assembly. At first glance it seems that vibrations produced by defects in the bearing assembly are very small and do not have any substantial effect on the operation of electric drives. But this is not so. In some cases they can cause very considerable oscillations of the entire system.

To determine the influence on vibration of static and dynamic unbalance of the rotor, experiments were carried out on AO31-2 and AOF42-2 electric motors; the pulleys and fans were removed. Vibration was measured, as indicated above, at permanently selected points. The motor under test was installed on a cast-iron plate in an unsecured state, and static and dynamic unbalance was artificially introduced into the rotor of the electric motor. With an artificially introduced static unbalance of the rotor of less than 2 g·cm, the vibration curves had the same character as the curves shown in Figs. 1 and 2, and the magnitude of the oscillation amplitude practically did not change. With a static unbalance of 2 g·cm, the influence of the rotor unbalance became noticeable on the vibration curve. In form, the curve began to approach a sinusoid. With a further increase of the static unbalance to 5.5 g·cm, the curve took the form of a sinusoid with a vibration period equal to the period of rotation of the rotor.

A study of the influence on vibration of dynamic unbalance of the rotor of the AO31-2 electric motor showed that the vibration curves have a nonstationary periodic character. The oscillation amplitude at unbalances of 8.5 and 12 g·cm reached more than 40 microns and significantly exceeded the permissible standards.

On the basis of the investigations carried out, one may conclude that the unbalance value of 1 g·cm adopted in the draft standards of the All-Union Scientific Research Institute of the Ministry of Electric Power Stations is unjustified. In order to reduce the time and costs of dynamic balancing (without detriment to quality), it is advisable to establish the permissible unbalance of rotors within 2 g·cm for Class I machines and 5 g·cm for Class II machines.

In studying the influence of the fan on the vibration of an electric motor, statically balanced fans with various magnitudes of end runout were mounted on a dynamically balanced rotor whose oscillation amplitude did not exceed 5 microns. After this, the vibration was recorded on film by means of an oscillograph. The vibration curve of the AO31-2 motor with a statically balanced fan mounted on the rotor, in the absence of end runout of the latter, indicates a significant increase in vibration due to the action of the fan.

The amplitude of vibration does not exceed 8 µm. In the case where a fan having axial runout relative to the housing of up to 2 mm or more was mounted on the rotor of the AOF42-2 electric motor, the amplitude of vibration of the motor increased to 40 microns or more. The vibration of an electric motor also increases significantly in the case where a statically balanced, but not dynamically balanced, drive pulley was mounted on the rotor.

Fig. 6 shows a vibrogram from the housing of the AO31-2 motor, whose rotor had previously been dynamically balanced and whose vibration was recorded in Fig. 1. The vibrogram was recorded after a statically balanced, but not dynamically balanced, drive pulley had been mounted on the rotor. The amplitude of vibration increased from 5 to 50 µm and exceeded

Fig. 6. Vibrogram taken from the housing of the AO31-2 electric motor when a drive pulley, dynamically balanced with the fan removed, had been mounted on the dynamically balanced rotor. Vibration amplitude 50 µm.

the limits of the permissible vibration standards. The vibration was of a nonstationary-periodic character. After dynamic balancing of the rotor together with the drive pulley on a DB-4 balancing machine within an unbalance of up to 2.5 g·cm, the amplitude of vibration of the motor housing was reduced to 10 µm and, at some points, even less.

The conducted studies of the vibration of induction electric motors make it possible to draw the following conclusions:

1. The vibration of an electric motor is determined by the dynamic unbalance of the rotating masses.

2. The standards recommended by the Scientific Research Institute of the Ministry of Electric Power Industry for the permissible vibration of induction electric motors of the unified series are stringent. It is advisable to increase the permissible vibration amplitude, especially for electric motors of the first class. The values of permissible unbalance of rotors and fans are not substantiated and require additional investigation.

3. The end runout of the fan relative to the housing must be reduced to a minimum, since it is one of the causes of an increase in the vibration amplitude of an electric motor and at times nullifies high-quality dynamic balancing of the rotor.

1. The drive pulley, balanced statically, may prove to be dynamically unbalanced, as a result of which a significant increase in motor vibration will occur.

2. To reduce the vibration of electric motors, it is necessary to proceed to joint dynamic balancing of the rotor, fan, and drive pulley.

REFERENCES

1. Draft standards for permissible vibrations of induction electric motors of the unified series with power from 0.6 to 100 kW and above. Research Institute. Ministry of the Electrical Industry of the USSR, 1957.

2. Yu. I. Iorish. Measurement of Vibration, Mashgiz, 1956.

Noted Misprints

Order 39