ANALYSIS OF TEMPERATURE FIELDS IN P-SERIES DC MACHINES

M. N. Ulyanitskii

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

The broad development of automation requires a considerable increase in the scale of production of low-power DC motors and generators. In connection with this, there is a need to create new, still more economical and reliable low-power DC machines. One of the main directions in solving this problem is the study of temperature distribution in existing machines, with the aim of obtaining the fullest possible understanding of the thermal loading of winding insulation and of identifying possibilities for improving cooling. This work presents the results of an investigation of temperature fields in DC motors of the unified P series, types PB, PO, and P.

1. Experimental procedure

The temperature fields of the motors were investigated using copper-constantan thermocouples made from insulated wires of diameter 0.2 mm. In order to obtain a sufficiently complete and reliable picture of temperature distribution, from 100 to 150 thermocouples were embedded in each of the motors. The thermocouples were placed in the armature, pole coils, pole cores, internal air, frame, and bearing shields.

Thermocouples were embedded in the armature and pole coils during manufacture of the machine. In doing so, no deviations were made from the established winding-manufacturing technology. To place thermocouples in the pole cores, frame, bearing shields, and commutator, holes of diameter 1.5 mm were drilled.

After the thermocouples were inserted, the holes were sealed with aluminum foil. In order to reduce the error caused by heat outflow through the thermocouple, the thermoelectrode leads from the measurement point over the first 1–2 cm were, wherever possible, run along an isothermal surface. In all cases, the arrangement of the thermocouples practically eliminated the possibility of disturbing the normal ventilation cycle. To improve accuracy when measuring the temperature of stationary parts of the machine, differential connection of a reference thermocouple was used (Fig. 1). The thermocouples located in the armature were brought out to an auxiliary current collector with copper contacts.

rings and copper-graphite brushes. The free ends of the thermocouples were soldered to the copper terminals of a round textolite disk, rotating in an annular space insulated from the external environment by a special casing. The uniform arrangement of the terminals around the circumference of the disk ensured practically identical thermal conditions for all free junctions of the rotating thermocouples.

Fig. 1. Diagram for measuring temperature on the stationary parts of the machine.

Fig. 2. Diagram for measuring temperature on the rotating parts of the machine.

The air temperature in the annular gap was determined with a thermometer. The connection diagram for the rotating thermocouples is shown in Fig. 2. Preliminary tests of the current collector showed that the magnitude of the contact thermoe.m.f. does not exceed \(3.5 \cdot 10^{-2}\) mV, or, when converted to temperature for a copper–constantan thermocouple, approximately \(0.8^\circ\mathrm{C}\). To eliminate the error introduced by the transition resistances of the current collector, the thermo-e.m.f. was measured—

…were made according to a compensation circuit with a PPTV1-type potentiometer and a galvanometer with a scale division value of \(2.2\cdot 10^{-3}\ \mu\text{A}\). The average overheating of the windings was determined by the resistance method, using ammeters and voltmeters of accuracy class 0.5. In order to shorten the time from the moment the motor was switched off until the first reading was taken when measuring the hot armature resistance, leads were soldered to the commutator bars and special devices were used for quickly lifting the brushes.

In the experiments, the ambient temperature was measured by three thermometers placed around the operating machine at a distance of \(1.5 \div 2\ \text{m}\).

2. Investigation of temperature fields in enclosed motors of type PB¹)

Investigation of the temperature distribution in the pole windings of PB52 motors (Fig. 3,a) and PB42 motors, which have pole coils with an aspect ratio

\[ \frac{h}{b} > 1, \]

showed that the difference between the overheating values over the width and height of the cross sections of the coils of the main and commutating poles does not exceed, respectively, \(4 \div 6\%\) and \(4 \div 10\%\) of the overheating value at the center of the cross section. In PB12 and PB32 motors (Fig. 4,a), which have pole coils with an aspect ratio

\[ \frac{h}{b} < 1, \]

the difference between analogous overheating values in the main and commutating poles reaches, respectively, \(5 \div 10\%\) and \(5 \div 15\%\).

From the overheating values obtained in the experiments on opposite sides of the coils, it follows that, in the pole windings of PB42 and PB52 motors, the point with the maximum temperature is displaced from the center of the transverse section of the coils toward the outer side surface. This is explained by the fact that, in the comparatively narrow intercoil space of four-pole motors, the average overheating of the air exceeds the overheating of the pole cores and differs only slightly (on average by \(5 \div 10\%\)) from the overheating on the outer surface of the pole windings. Two-pole PB12 and PB32 motors have comparatively large dimensions of the intercoil space and of the outer surface of the pole windings. The average overheating of the internal air and of the outer surface of the pole coils in these motors differ by \(17\)—\(20\%\). As a result, convective heat exchange of the pole windings in PB12 and PB32 motors takes place more intensively than in four-pole motors. Experiments show that in the pole windings of two-pole motors the point with the maximum temperature is displaced toward the inner side surface of the coils.

It should be noted that the temperature distribution in the pole coils will depend on the ratio between the average overheating values of the machine windings or, in other words, on the magnitude of the specific losses. For example, with an increase in the losses in the armature windings and in the commutating poles, heat transfer from the outer cooling surface of the shunt-field winding will decrease, which, in turn, will cause a redistribution of temperature in the transverse section of the main-pole coils.

¹) Motors of type PB are produced by the Elektromashina plant without internal fans.

According to the data obtained in studying the temperature distribution in the pole windings in the axial direction, the change in temperature along the length of the pole coils does not exceed \(1^\circ\text{C}\). Such a uniform temperature distribution can be explained by the high thermal conductivity of copper and by approximately identical cooling conditions along the length of the pole coils in PB-type machines. It should be noted that the temperature distribution in the pole

Fig. 3. Temperature distribution in pole coils with side ratio \(\dfrac{h}{b} > 1\): \(a\) — PB52 motor; \(b\) — P42 motor.

windings around the circumference of the frame is nonuniform. The temperature rises of the pole coils located above the horizontal axis of the shaft are \(4 \div 5\%\) greater than in the diametrically opposite coils. An analogous phenomenon is also observed with respect to the temperature rises of the internal air and of the pole cores.

Fig. 4. Temperature distribution in pole coils with side ratio \(h/b < 1\):
\(a\) — PB32 motor; \(b\) — PO32 motor; \(c\) — P12 motor.

Investigation of the temperature distribution in the frame and in the internal air of the machine has established that the average overheating of the air in the interpolar space, i.e., in the zone of the most active heat generation, is 20–30% greater than the average overheating of the air in the region of the bearing shields, and 40–45% greater than the overheating of the frame surface. The ratios obtained convincingly show the advisability of using internal fans in PB-type motors, since mixing the air in the machine and causing it to move along the inner surface of the frame should equalize and, overall, reduce the average air temperature inside the machine and thereby improve the cooling conditions of the windings. The effectiveness of using an internal fan in PB-type machines was checked on the PB42 motor. According to the experimental data, installing an internal fan reduces the overheating of the armature and pole windings by 20–25%.

Analysis of the machine’s temperature fields shows that, in addition to using an internal fan, there are also other possibilities for reducing winding overheating. As follows from the relation of overheating in the cross sections of the pole coils, a considerable part of the heat released in the pole windings is transferred to the yoke through the pole cores. In this case, the most substantial thermal resistances along the path of the heat flow are the resistances between the inner surface of the coils and the pole cores and between the pole cores and the yoke. In the main poles, the temperature drops across these resistances are approximately the same and, in total, amount in PB42 and PB52 motors to about 30%, and in PB12 and PB32 motors to 40–50% of the overheating of the windings at the center of the coil cross section. Relative to the magnitude of this same overheating, the temperature drops between the inner surface of the coils and the pole cores in the commutating poles of PB42, PB52 and PB12, PB32 motors reach, respectively, 20–25% and 25–40%, while the temperature drops between the pole cores and the yoke are 15–20%. In accordance with the data obtained, it may be recommended, in order to reduce overheating of the pole windings, to fill the air gaps between the pole core and the coil with a compound having good thermal conductivity. It should be taken into account that the effectiveness of using the compound will depend on the ratio between the thermal resistances of the insulation of the pole cores and the air gaps. The smaller the thermal resistance of the core insulation, the greater the role played in the heat-transfer process by the thermal resistance of the air gaps, and the greater the cooling effect of filling with compound will be obtained. In this connection, along with using compound filling, improving the thermal conductivity of the pole-core insulation is also of substantial importance.

To reduce overheating of the pole windings, [1] recommends filling the entire interpolar space with compound. However, it should be noted that, when the air gaps between the pole core and the coil are filled with compound, the additional cooling effectiveness from filling the interpolar space with compound will be substantial only if the resistance of the coil’s heat-flow path to the yoke through the compound in the interpolar space is of the same order as the resistance of the coil’s heat-flow path to the yoke through the pole core. If the latter resistance is comparatively small or practically absent (low-voltage machines with solid-stamped in-

...conductor), additional filling of the inter-coil space with compound will not produce a substantial effect, since in this case the thermal resistance of the heat-flow path from the coil to the frame along the pole core will be considerably less than the thermal resistance of the path through the compound in the inter-coil space.

A reduction in the overheating of pole windings can be obtained by reducing the thermal resistance between the pole core and the frame. Reducing this resistance, for example by means of mechanical machining, will make it possible to lower the overheating not only of the pole windings but also of the armature winding.

Fig. 5. Temperature distribution in the armatures of P-series motors: a — PB32 motor; b — PO32 motor; c — P42 motor.

According to the data obtained in studying the temperature fields in the armatures of PB52 motors (Fig. 5, a), PB32, the values of overheating at various points of the armature windings differ by no more than 10–15%. The difference between the overheating values over the height of the slot does not exceed 5%, and over the height of the end-winding parts 5–10%. The overheating of the teeth and of the copper in the slot portion of the winding differs by 2–3%. The temperature drops over the height of the teeth are also insignificant (1–3°C)

and along the insulation of the armature slots (1.5–2°C). The values of temperature rises at diametrically opposite points of the winding and of the armature iron differ by 2–3°C. The temperature rises of the commutator copper in all motors proved to be 10–12% lower than the average temperature rises of the adjacent end-winding parts. In addition, it was established that the temperature rises of the shaft inside the machine have comparatively smaller values relative to the temperature rises of the tooth bases. Thus, the obtained distribution of temperature-rise values over the armature iron in the radial direction gives grounds for assuming that part of the losses released in the armature is carried off into the surrounding medium through the shaft.

3. Investigation of Temperature Fields in Enclosed Blown Motors of Type PO

According to the results of the tests carried out, the temperature distribution in the pole windings of enclosed blown motors of type PO is more nonuniform than in motors of type PB. In the windings of the main and commutating poles of PO32 motors (Fig. 4, b), PO22, having pole coils with a side ratio

\[ \frac{h}{b} < 1, \]

the temperature drops across the width and height of the cross sections of the coils are approximately equal and amount on average, respectively, to 10–25% and 8–20% of the temperature-rise values at the center of the cross section. In accordance with the temperature-rise values on opposite sides of the coils of the main and commutating poles, the maximally heated point in the cross sections of the pole coils of type PO motors is displaced from the center of the section toward the pole core. The character of the temperature distribution in the cross sections of the pole coils indicates that in type PO motors, which have an internal fan, heat transfer from the pole windings to the internal air occurs more intensively than in type PB motors. The average temperature rises of the air in the intercoil space of two-pole type PO motors are 30–40% lower than the temperature rises of the outer surface of the pole windings.

However, as is evident from the ratio of temperature rises on the side faces of the pole coils, heat transfer to the pole cores also plays an essential role in the heat-transfer process of the pole windings. In the main and commutating poles, the temperature drops between the inner surface of the coils and the pole cores reach, respectively, 30–45% and 35%, while between the pole cores and the frame they reach 20–30% of the temperature rises at the center of the coil cross section. Thus, it is quite obvious that reducing the thermal resistances between the pole coils and cores and between the pole cores and the frame will make it possible to obtain a substantial reduction in the temperature rises of the windings. To reduce the indicated thermal resistances, the methods proposed above for lowering the temperature rises of the pole windings in type PB motors may be used. When these methods were checked on a PO22 motor, a reduction in the temperature rises of the pole windings by 12% was obtained.

Figure 5, b presents the results of an investigation of the temperature distribution in the armature of a PO32 motor. As is evident from the data presented, the temperature field of the armature winding has a comparatively uniform character. The temperature of the armature teeth is practically equal to the temperature of the copper in the slot portion of the winding. The shaft temperature is somewhat lower

temperatures of the armature iron. Thus, part of the heat generated in the armature, as in motors of type PB, is removed through the shaft.

4. Investigation of temperature fields in motors of protected design, type P

By investigating temperature fields in the pole windings of motors P12 (Fig. 4, c), P22, having pole coils with an aspect ratio

\[ \frac{h}{b}<1, \]

it was established that the ratios between the temperature rises on the surface and at the center of the transverse sections of the coils have approximately the same values as in pole coils of similar shape in motors of type PO. According to the data obtained, the temperature differences across the width and height of the transverse sections of the pole coils reach, respectively, 25–30 and 20% of the value of the temperature rise at the center of the cross section. In pole coils with an aspect ratio

\[ \frac{h}{b}>1 \]

(Fig. 4, c), the analogous temperature rises differ by 20–25% and 25–35%. In accordance with the values of the temperature rises on the surface of the pole windings, the point of maximum heating in the transverse section of the pole coils of type P motors, regardless of the aspect ratio of the coil, is located above the horizontal axis of the cross section and is displaced from the center of the section toward the pole core.

The temperature difference between the pole core and the inner surface of the pole coils with aspect ratios

\[ \frac{h}{b}<1 \]

and

\[ \frac{h}{b}>1 \]

reaches, respectively, 40–50% and 20–30%, while between the pole cores and the frame it reaches 10–15% and 20–25% of the value of the temperature rises at the center of the transverse section of the coils.

The temperature distribution in the armature windings of type P motors is very nonuniform (Fig. 5, c). According to experimental data, the values of the temperature rises at individual points of the windings differ by more than 30–40%. The differences along the height of the end portions reach the greatest magnitude (30–40%). The values of the temperature rises along the height of the slots and in the axial direction of the windings differ by 10–15%. The maximum temperature in the radial direction was shown by thermocouples located between the winding layers. In all the motors studied, the temperature rises of the commutator copper are 40–50% lower than the average temperature rises of the adjacent end portions, and the temperature rises of the shaft are approximately 25–30% lower than the temperature rises in the tooth zone of the armature.

Conclusions

1. A comparative analysis of the temperature fields of the motors studied shows that the most uniform character of temperature distribution, and consequently the most uniform thermal loading of the winding insulation, is found in motors of type PB. According to experimental data, the values of the temperature rises at various points of the windings of these motors differ from the average temperature rises of the windings, obtained by the resistance method, by no more than 5–10%. The most nonuniform thermal loading of the insulation is found in the windings of motors of type P. The values of the temperature rises at individual points of the windings of motors of protected design may exceed the values of the average temperature rises of the windings by 15–20%.

1. Analysis of the temperature fields indicates the possibility of reducing the overheating of the pole windings of motors by decreasing the thermal resistance between the pole cores and the frame and by filling with compound the air gaps between the pole cores and the coils. It should be noted that the application of the indicated methods in motors of the PO and P types, as well as in motors of the PB type having pole coils with the ratio $\frac{h}{b}<1$, will make it possible, simultaneously with lowering the average overheating of the windings, to reduce the difference between the values of overheating in the transverse sections of the coils, i.e., to obtain a more uniform temperature distribution in the pole windings.

2. The experimental data obtained may be used for developing a methodology for thermal calculations and for a comparative assessment of the reliability of the types of motors investigated.

REFERENCES

1. O. P. Sidorov. Cooling of enclosed direct-current machines. Elektrichestvo, No. 9, 1962.

2. G. P. Zedginidze. Measurement of the temperature of rotating machine parts. Mashgiz, 1962.

3. G. K. Zherve. Industrial testing of electrical machines. Gosenergoizdat, 1959.