STUDY OF THE INFLUENCE OF CERTAIN TECHNOLOGICAL FACTORS ON THE RELIABILITY OF TURN INSULATION IN ARMATURE WINDINGS OF DIRECT CURRENT MACHINES

A. Ya. Tsirulin, I. A. Volkomirsky, V. N. Popov,
N. M. Novoselova

(Submitted by the scientific seminar of the Department of Electrical Machines)

Endurance tests of machines of the П series, frame sizes 1–6, conducted over a number of years at the Elektromashina plant (Prokopyevsk), showed that the greatest intensity of failures of armature windings is observed in the first 2–6 thousand hours of operation, that is, a running-in period takes place. All failures were related to turn insulation. It is known that failures during the running-in period are caused by the presence of hidden defects in the winding insulation, due to certain design and technological factors. To increase the reliability of machine windings, it is necessary to identify such factors, assess the degree of influence of each of them, and find ways to weaken their influence.

An analysis of a priori information makes it possible to identify the following main factors that reduce the reliability of turn insulation.

1. Damage (or microdamage) to turn insulation inflicted during winding of sections on a template, placement of sections in the armature slots, alignment of the end parts of the winding, etc. Studies of the influence of technology on insulation damage, which we carried out at the plant, showed that, with the existing production culture of winding manufacture, the damage rate of the turn insulation of enamel wires is low, while for wires with fibrous insulation it is absent.

2. A high slot fill factor, contributing to increased damage to the turn insulation when the winding is laid into the slots.

3. The presence of “crossovers” of conductors in the slot and end portions of sections, leading to pressing-through or abrasion of the insulation during operation of the machine under the action of vibrations and centrifugal forces.

4. Displacement of interlayer spacers in the slot and in the end portion, as a result of which the conductors of both layers come into contact and the turn insulation is subjected to a large electrical voltage.

5. Poor cementation of the winding conductors by the impregnating varnish, making possible mutual displacement of the conductors during operation of the machine under the action of vibrations and centrifugal forces, thermal expansions, etc., and leading to abrasion of the insulation at the “crossovers.” Poor cementation may be caused by insufficiently strict adherence to impregnation and drying regimes and by the non-optimality of the regimes themselves prescribed by the technology.

To study the degree of influence on the reliability of turn insulation, the following main factors were selected:

\(x_1 = K_z\) — coefficient of filling of the free slot area;
\(x_2\) — presence of “crossovers” of conductors;
\(x_3\) — level of cementation of the winding.

The problem of determining the degree of influence of several factors is solved by methods of planning multifactor experiments [1]; therefore, a complete factorial experiment consisting of eight tests for three independent variables was drawn up and carried out (Table 1).

Table 1

The studies were conducted on P32 machines with armature windings made of PETV and PSDKT wire; therefore, the plan in Table 1 was implemented separately for each type of armature-winding insulation. The “large” \((+)\) slot-filling coefficient was equal to 0.757 for the winding made of PETV wire and 0.752 for the winding made of PSDKT wire. The “small” \((-)\) filling coefficient was, respectively, 0.547 and 0.57. The diameter of the winding wire was selected with allowance for obtaining the “small” slot-filling coefficient, while the “large” filling coefficient was obtained by placing additional passive sections with the required number of conductors into the slots.

The number of “crossovers” at the \((+)\) level was provided by artificially tangling the conductors, and at the \((-)\) level by carefully laying the winding, as far as possible without “crossovers.” This work was performed by a very experienced winder.

Cementation at the \((+)\) level was ensured by triple impregnation and thorough drying, and at the \((-)\) level by a single impregnation providing a thin varnish film.

Each test in Table 1 corresponds to one armature manufactured in accordance with the conditions specified by the table. The reliability of the insulation itself cannot be taken directly as the output characteristic, since otherwise it would be necessary to test each machine until armature-winding failure, which requires lengthy tests. Moreover, it would be necessary to test a large number of machines in order to assess the level of reliability objectively. Therefore, parameters of the distribution of the breakdown voltage of the turn insulation of each armature were adopted as the output characteristic, making it possible to evaluate the condition of the insulation and, consequently, factors associated with insulation reliability. In order to test repeatedly the turn insulation of each armature, test turns (four per section) were placed in the winding sections; their ends were brought out to the end-winding part of the winding, bent onto the insulated end-winding part, and banded together with the winding. All machines were subjected to accelerated tests for 500 hours, at an elevated working

temperature (160°C for armatures with windings made of PETV wire and 200°C for armatures with windings made of PSDKT wire), increased vibrations, and the presence of reversals. In these tests, the negative effects on the insulation of the input factors introduced during manufacture of the armatures were to become apparent. The degree of these effects and, consequently, the degree of change in the properties of the insulation in the different experiments were to be different, since the levels of the factors were different.

After completion of the accelerated tests of the machines, the turn insulation of the armature windings was tested for breakdown by an alternating voltage, by applying voltage between the winding (through the commutator) and one of the test turns. The test turns were made of the same wire as the winding; therefore it may be assumed that the turn insulation of a pair of winding turns over a length of approximately six lengths of a section turn was subjected to testing in this manner.

To carry out the tests, the armature was unbanded, the ends of the test turns were lifted, and conductor electrodes from the test installation were connected to them. The armature, in a vertical position, was immersed in transformer oil so that all the test conductors were immersed in the oil. This proved necessary in order to avoid flashovers outside the winding. Voltage was applied twice to each test conductor, so that it would then be possible to identify cases corresponding to flashovers along microcracks in the insulation, according to the procedure set forth in article [2]. On each armature, 72 paired values of breakdown voltage were obtained. Based on the results of the insulation test, scatter diagrams of the breakdown-voltage distribution were constructed and parameters characterizing the insulation condition of each armature were calculated. The mean voltage values change hardly at all from experiment to experiment and do not characterize the insulation condition. A parameter with good resolving power is the probability of occurrence of low breakdown voltages in comparison with the mean voltage, i.e., the probability of occurrence of insulation defects. For armatures with windings made of PETV wire, the probability \(P_3\) of occurrence of a breakdown voltage less than 3 kV was calculated; for armatures made of PSDKT wire, the probability \(P_{1,1}\) of occurrence of a breakdown voltage less than 1.1 kV was calculated. For PETV wire, the insulation defectiveness \(\lambda\), equal to the number of microcracks per unit length of wire, was also calculated. It was established that the parameters \(\lambda\) and \(P_3\) possess equal resolving power. Table 2 presents the values of the output parameters \(P\) in each experiment for both grades of winding wire.

Table 2

Using the data of Table 2, scatter diagrams of the parameters \(P_3\) and \(P_{1,1}\) were constructed (Fig. 1, a, b), and by the method of regression analysis were calculated

Fig. 1. Scatter diagram of the output parameter when varying the input factors: a) for armatures with a winding made with PETV wire; b) for armatures with a winding made with PSDKT wire

The dependences of these parameters on the input factors also have the following form:

\[ P_3 = 0.255 + 0.05 \cdot x_1 + 0.108 \cdot x_2 - 0.078 \cdot x_3 + 0.019 \cdot x_1x_2 + \]
\[ + 0.033 \cdot x_1x_3 + 0.0094 \cdot x_2x_3; \tag{1} \]

\[ P_{1,1} = 0.422 + 0.212 \cdot x_1 + 0.503 \cdot x_2 - 0.831 \cdot x_3 + 0.151 \cdot x_1x_2 - \]
\[ - 0.057 \cdot x_1x_3 - 0.531 \cdot x_2x_3. \tag{2} \]

On the basis of (1) and (2), the changes in the parameters \(P\) when the input parameters are changed can be determined from the expressions:

\[ \Delta P_3 = 10^{-4} [25 \cdot \Delta x_1 + 116 \cdot \Delta x_2 - 61 \cdot \Delta x_3 + 3.6 \Delta (x_1x_2) + \]
\[ + 10.9 \cdot \Delta (x_1x_3) + 0.88 \cdot \Delta (x_2x_3)]; \tag{3} \]

\[ \Delta P_{1,1} = 10^{-3} [45 \cdot \Delta x_1 + 253 \cdot \Delta x_2 - 690 \cdot \Delta x_3 + 22.7 \cdot \Delta (x_1x_2) + \]
\[ + 3.3 \cdot \Delta (x_1x_3) + 280 \cdot \Delta (x_2x_3)] \simeq 10^{-3} [253 \cdot \Delta x_2 - 690 \cdot \Delta x_3 + \]
\[ + 280 \cdot \Delta (x_2x_3)]. \tag{4} \]

Analyzing expressions (3) and (4) and Figs. 1a, b, one can draw the following conclusions. The value of the slot fill factor \((x_1)\), the presence of “crossovers” in the winding sections \((x_2)\), and the level of cementation of the winding with varnish \((x_3)\) have a substantial influence on the defectiveness of enamel insulation. It should be noted that the smaller the parameters \(P\), the lower the level of insulation defectiveness and the higher its reliability. With a decrease in the slot fill factor and in the number of “crossovers,” and with an increase in the degree of cementation of the winding, the quality improves.

The slot fill factor has practically no effect on the defectiveness of fibrous insulation, but the “crossovers” do have an effect, and the degree of cementation of the winding has an especially strong effect. With poor cementation, this insulation is evidently prone to abrasion at the crossovers, which leads to the appearance of insulation defects during operation of the machine. It is interesting that high-quality cementation of the winding weakens the negative influence of the “crossovers,” as is indicated by the large coefficient at \(x_2x_3\) with a negative sign in expression (2).

A reduction in the slot fill factor for increasing the reliability of enamel turn insulation of windings cannot be recommended, since this reduces the level of utilization of the machines. Increasing the reliability of enamel and fibrous insulation can be achieved by reducing the number of conductor “crossovers” in the winding through improving the standard of winding operations and applying a technology that excludes the formation of “crossovers,” and also by increasing the degree

the cementation of the winding by using varnishes with good cementing ability and by optimizing the impregnation and drying regimes of the windings according to the criterion of cementation quality.

References

1. V. V. Nalimov, N. A. Chernova. Statistical Methods for Planning Extreme Experiments. Moscow, “Nauka,” 1965.

2. A. P. Matyalis, E. K. Strelbitskii. Method for evaluating the defectiveness of insulation in low-voltage electrical machines. Izv. TPI, vol. 229, 1971.