Docstoc

THERMAL CHARACTERISTICS OF TRACTION MOTORS WITH REGENERATIVE BRAKING

Document Sample
THERMAL CHARACTERISTICS OF TRACTION MOTORS WITH REGENERATIVE BRAKING Powered By Docstoc
					THERMAL CHARACTERISTICS OF TRACTION MOTORS WITH REGENERATIVE BRAKING S.L. Ho H.C. Wong* and S.K. Poon*

Department of Electrical Engineering, Hong Kong Polytechnic University * Industrial Centre, Hong Kong Polytechnic University Abstract This paper describes a computer program developed to simulate the temperature distribution of a traction motor for a journey between two railway stations in Hong Kong. In particular, the authors have used the software to study the temperature distribution of the traction motor running with regenerative scheme along an entire railway network. Special attention are given to evaluate the temperature rise on the winding surface and at specific points where the insulation were reported to be damaged easily. Overall, the traction motor thermal model developed is useful for traction design engineers and timetablers who are required to check whether there are overheating inside the traction motor for a given duty. 1. INTRODUCTION prolonging the working life of the discs and reducing the amount of asbestos dusts escaping into open air. One of the major problems of introducing the electrical braking system in the KCRC system is that regeneration is very difficult and costly. In addition, the regenerated energy after being inverted to ac may have unavoidable harmonic contents which might contaminate the electricity supply of the power company, unless a bulky and expensive harmonic filter is installed. These are thus the main reasons why an electrical braking system has not been installed in the EMUs at the design stage. In this paper, the authors are proposing to feed the braking energy into the 110 V 165 Ah nickel cadmium emergency batteries which are always available in the trains in Hong Kong. Note that with electrical braking, the train retardation is increased and one must control the rate of regeneration to alleviate any discomfort which may appear to the passengers. During the electrical barking period, the coasting speed as well as the cooling effect is reduced. The additional heat generated by the traction motor during regeneration might cause overheating in traction motors. Indeed there are evidence in Hong Kong that regenerative operation might overheat the traction motors to the extent that the service life of the motors were drastically reduced. Hence one must take all these factors into consideration when regeneration is introduced. 2. LITERATURE SURVEY

Over the past 20 years, the passenger and freight traffic of the Kowloon-Canton Railway Corporation (KCRC) is steadily increasing. Saturation is being approached on the cross-border traffic and, together the massive new town developments at Shatin, Tai Po, Fang Lang and Sheung Shui, a new demand for mass transportation is appearing rapidly. A modernization programme in the railway network was thus initiated in the early 1980's. In this modernization programme, the twelve-coach passenger trains which were trailed by diesel-electric locomotives were replaced by a fleet of Electrical Multiple Units (EMUs) running on an electrified rail. The whole nature of this railway has been transformed, from the principally rural, slow frequency service to become a modern, fast inner and outer suburban system. Electrification was at 25 kV AC and the track was doubled throughout the entire route length. The KCRC becomes the provider of the most intensive railway service in the world. In 1988, due to the advent of power electronics, a new fleet of EMU using thyristor-controlled traction system were being brought into service to run alongside with the original tap-changer type of EMUs. Unlike other typical underground mass transit systems, the grading of the track to save energy cannot be used in the KCRC system because of the geographical constraints in Hong Kong. However, the use of electrical braking is possible. Apart from saving energy, the use of electrical braking could also reduce the wear upon the mechanical brake discs, hence

Although DC series motors have been used traditionally in traction applications, regenerative braking has not been used very often in conventional traction equipment. The main reason for this is because the operation of a series generator becomes rather unstable

when connected to a fixed voltage supply. A separate excitation is commonly required for realistic and stable operation. Such an arrangement of series motor, moreover, is very sensitive to supply voltage fluctuations and a fast dynamic response is required in order to provide an adequate brake control. The use of a DC chopper facilitates the regenerative braking of DC series motors due to its fast dynamic responses.
D I a
3

200 km on 1 kWh of electricity. It is worth pointing out that the greatest energy losses in the whole journey are incurred whilst accelerating the trains up to their maximum allowable speed. The use of regenerative braking will normally recover a substantially amount of the kinetic energy. Experiences indicate that the energy consumption on level routes with regenerative equipment would be about 10 % lower than that in trains without regeneration. 3. MACHINE MODELLING

1

5

L

e

0 .1H

R

e

0 .13 6 7

6 4 .7u 82

L

m

1 .64 18 4H

4 C s2

R s2

2

R

m

0 .04 4 4 .7u 2 10 0u 1 10 V

C s1 1 9

C

e

V

s

T

1
8

10 1 0ME G

E

+

0V Rg

82

g
Vg1 R s1

(V ) g

V

T

0

Fig. 1 GTO Regenerative Braking Chopper Circuit for PSPICE Simulation

In a DC chopper controlled series motor with regenerative braking, the braking power P in a constant power braking region can be expressed as a function of  multiplied with a polynomial of Ia. The order of the polynomial is dependent on the accuracy of the traction motor model and it is normally over 6. In order to determine the relationship of Ia = f() for a particular P, an iterative solution using Newton-Raphson method was used in the simulation programme. The speed-time curve for a run between two station stops consists of periods of acceleration, free running, coasting and retardation. The speed-time profiles for all inter station journeys were determined jointly by the timetable schedule and the maximum service retardation. An efficient use of energy is an important factor in the achievement of an operating profit, as energy typically represents 20 % of the operating cost in the local railway. Recent developments with semi-conductor devices have made it possible for regenerative braking to be installed in rapid transit trains having frequent stops, thus maximizing the braking energy with a small weight penalty [1]. For modern, efficient mass transit trains, a typical energy consumption of 24 Wh/t km translates approximately to carrying a passenger for
TL (s) + T f(s) + V (s) 1 Th yris tor Ia (s) + 1 W (S) 2 Km Ia ga in R m(S e + 1 ) Td (s) B (S m + 1) E g(s) + K mW + K m Ia

The shapes of the various parts of large electrical machine are too complicated for exact analysis of the heat flow in different parts of the machine. This situation has led to the use of the water models by Limbora [2], because there are serious difficulties in air models when measuring windage losses and the relative gas velocity over rotor cooling surfaces. Water model are preferred to air model in measuring (i) static pressures and (ii) pressure differences, (iii) hydraulic resistance, (iv) flow rate in duct or in the modelling of hollow conductor, (v) flow velocity vector fields over heat transfer surface, and (vi) windage losses. Water model represents, to scale, all parts of the machine which could have an essential influence upon the flow of the cooling gas. For the mixed flow case, the strength of rotation is identified by the ratio of tangential velocity on the surface of the rotating shaft to the mean axial velocity. The ratio, ( = V/Vm) is called "the rotation ratio". It was found that at a fixed Reynolds number, the Nusselt number increases with an increase in the rotation ratio, . Moreover, the effect of  on Nu is particularly strong at low Reynolds numbers and it becomes negligible above Re = 50,000 for the highest rotation. A physical parameter that correlates the mixed-mode friction coefficient and Nusselt number data by using the normal pure axial flow relation can be defined as the rotation parameter. Moreover, it was common to assume that the rotation of inner cylinder does not significantly affect the Nusselt number until the rotation ratio reaches a value about 0.8. The speed of rotation covers the range of the Taylor number up to about 106, and the range of the Reynolds number based on the axial velocity components and the gap distance can be assumed to be valid up to 7000. It was found that rotation does not affect the Nusselt number at low Taylor Numbers and the heat transfer is determined by the axial Reynolds number [3]. For a smooth rotor the Taylor number for the onset of vortex flow increase firstly with increasing axial flow, and, after reaching a maximum, the Taylor number appears

V (k ph )

a rea A

V1

max . b ra kin g ra te (s lo pe ) = 5 kp h

'

V2'

T = t1 + t2 ' + t3 '

a rea B

a rea A = ar ea B

t ( S) 0 t1 t2 ' t3 '

Fig.2 Speed-time Curve of a EMU with Electrical Braking

to decrease slightly with further increases in axial flow. This tendency towards decreasing critical Taylor number is very observable for slotted rotors. For Pr > 0.6 and Re > r x 103, the thermal boundary layer is thin compared with the transverse dimension of the stream. In closed ducts this means that the main thermal resistance is localized in the region of the viscous and buffer layers. Thus, local imperfections and gaps (e.g. at corners) in the walls of a duct cause a limited amount of deformation of these layers and have comparatively little effect on the thermal resistance of the stream. In practice, it is possible to calculate, to an accuracy of about 10 %, the heat transfer in noncircular ducts with Pr > 0.6 and Re > 7000, by using the value of the so-called equivalent hydraulic diameter (Dh = 4 x internal cross sectional area of a conduit/wetted perimeter of a conduit). The earliest attempts at predicting the temperature rises in machines by relating temperature rise to losses, as a thermal network of lumped conductances and capacitances, was due to Tustin [4]. He had assumed that lumped-resistance networks may be found to reproduce the relationships between the inputs heat power and the mean temperature rises over the various parts of the machine, although the thermal resistivities are distributed. The theory of the analogy is summarized by Simonson [5]. The purpose of his paper is (a) to outline some previous work to show the contribution that can be made by using electrical analogies; (b) to summarize the theory of such analysis, and (c) to show how the technique may be applied in the thermal analysis of a traction motor armature. The finite difference approach to a heat transfer problem consists of defining a mesh or grid to cover the temperature field and determining the temperature at the grid point. Each grid point corresponds to, and is situated at the centre of, a certain area of the field, and the thermal resistances between grid points are equal to the resistances between the centres of the appropriate area. The electrical analogue technique is based on using an equivalent electrical circuit to correspond to the thermal network of the finite difference grid. A resistance network analogue representation of the stator core and winding has been derived by Roberts [6] and is based on the finite difference approximation to the heat flow equations. A solution of the analogue is achieved by matrix methods, on a digital computer. The input data necessary before an analogue solution can be obtained are: (a) the internodal resistances within the packet; (b) the loss inputs at the nodal

points; (c) the thermal capacities of the various air streams determined from air flow and thermal properties of the air, (d) the air temperature which forms part of the input data although corrections to the assumed values can be made during the course of the solution. 3.1 Heat generated by traction motor Heat is generated in traction motors due to the electrical losses in different parts of the machine, primarily because of the copper loss in the conductor and the iron loss in iron. The heat generated by mechanical means, such as frictional losses at the motor bearings and at the interface between the carbon brushes are relatively small compared with that of the electrical losses. The electrical heat generated causes the temperature of the traction motor to rise and the heat generated is dissipated to the ambient air by a combination of conduction, radiation and essentially forced convection. The electric current supplied to the motor varies with the travelling speed. The copper loss equals to I2R and is therefore directly proportional to the square of the current magnitude with the assumption that the electrical resistance of the conductors is constant. The iron loss composed of two components, namely, the eddy current loss and the hysteresis loss. The iron loss is only a minor source of the total heat generated when compared with that generated by the copper loss. It is acceptable to assume that the iron loss is equivalent to a fraction of the copper loss. The copper loss is dependent on the speed of the train and the magnitude of the traction input current, and the iron loss is of similar nature. 3.2 Heat transfer inside traction motor The cooling method adopted by the traction motor is called the self-circulation type in which the circulation of the cooling air is created by means of a fan mounted directly on the armature shaft of the motor. When the armature rotates, the fan will drive the cooling air through the air gap. The heat generated is moved essentially by forced convection. Radiation will also take place at the machine surface but will not be regarded as significant at the level of temperature encountered by the motor. From the Newton's Law of cooling, the rate of heat transfer by heat convection (q) between a heated surface (with surface area A) can be determined from q = hA(w -f). The heat transfer coefficient h is a very complicated function which depends on the fluid flow

condition, fluid properties and geometrical configuration of the heat surface. For forced convection, the heat transfer coefficient may be determined by the non-dimensional expression, i.e. Nu = f(Re, Pr). Different boundary surfaces will have different value of heat transfer coefficients. However, due to geometrical similarity, two heat transfer coefficients can be reasonably assumed. The first one is required on the surface of the armature and the laminated iron, whereas the second one is required for the surfaces of the field coil. It is assumed in this simulation that convective heat transfer on these surface occurs because of the air flowing pass through a concentric annulus with the inner cylinder rotating. 3.3 Heat transfer at the field coil surfaces Geometry around the field coil surfaces is very complicated and cannot be simulated exactly by a single shape. Thus the following three configurations have been considered in order to arrive at a better approximation. These are, namely, (a) flow over a flat plate; (b) flow through a parallel plate and (c) flow through a conduit with defined cross section. The air flow through a triangular duct is considered as a better approximation of the geometry under investigation. Effect of the duct length and the cross-sectional dimension of the duct on the heat transfer coefficients have been taken into account in the heat transfer analysis by the authors. It is worth pointing out that turbulent flow will occur inside the airgaps because of the relatively high fluid velocity. In fact, there is no suitable non-dimensional expression to predict the heat transfer in a triangular duct in which turbulent flow exists. Noting that the Prandtl number of air is about 0.697 at 350 K and is greater than 0.6, the concept of hydraulic diameter can be applied.

irregular geometries and complicated boundary conditions. Finite element method is not as efficient as the finite difference method in manipulating the iteration process and in formulating the model. Since the main heat generation and heat dissipation are on the coil surface which is a relatively plane surface and not a complicated irregular 3-dimensional object, the finite difference method is very suitable to handle such geometry. In addition, the finite difference method is more stable when sets of simultaneous equations are solved by sequential iteration compared with the finite element method. Therefore finite difference method has been employed by the authors for the development of numerical models. 4.1 Assumptions of the numerical model The following assumptions have been adopted in developing the numerical model: i) There is no heat transfer in the axial direction of the motor. It is because the core is laminated and therefore the heat transferred between laminated plates is assumed to be negligible. The axial conduction is also ignored. Since the gap between the two cylinders is very small when compared to their diameters, the developing region is negligible when compared to the well-developed region. Variation in temperature and velocity profiles of the fluid boundary layer along the axial direction can be assumed negligible and heat convection is assumed to be a 2-dimensional process in the radial direction. The heat transfer model can therefore be represented simply by a 2-D model. ii) The fluctuation in the thermal properties i.e. , k, Cp and , caused by changes in the fluid temperature is negligible. iii) Heat generation is due to the winding loss and core loss only. During the whole journey, the internal energy of the motor will be increased each time after every motoring and braking cycle.

4.

COMPUTER SIMULATION OF TEMPERATURE OF TRACTION MOTOR

Many models could be developed for predicting the temperature distribution inside the traction motor. Both finite element models and finite difference models have been considered by the authors. The principal advantages of the finite element approach for convective heat transfer is that it can handle

iv) The system of equation shall be expressed in the matrix form of [A][x] = [B]. After the first iteration, the error C1 = [A][X] - [B] is calculated. The iterations will be continued until the accumulative error (C12 + C22 + .... ) is less than 0.01 and the temperature solutions will then be assumed to have converged. 5. SIMULATION RESULTS

As the existing KCRC fleet of trains do not have regenerative features, one must validate the simulation results by considering the normal mode of operation with the traction motors operated without regenerative braking as shown in Fig. 4.
8 15 16 22 14 17 23 13 18 24 19 25 31 32 33 34 43 44 45 46

A rm

atu

re
83 84 85 86 10 0 10 1 10 2 11 5

67 68 69 70 71 87 88 89 72

10 3 11 6 1 3 2 13 1 13 2 12 9 13 0 11 8 11 7 10 4

73

90

74

10 5

91

75

4 7

5

6

11 9

10 6

92

76

13 3 9 10 11 12 13 4 20 26 21 27 28 29 30

12 0

10 7

93

77

12 1

10 8

94

78

13 5

12 2

10 9

95

79

13 6

12 3

11 0

96

80 81

35

36

37

38

39

40

41

42

13 7

12 4

11 1

97 82 98

13 8

12 5

11 2 99 14 5 14 8 11 3 11 4 12 7 14 2 12 8 14 3 14 4 14 7 15 0 15 1 15 9

47

48

49

50

51

52

53

54

13 9 14 1

12 6

From Fig. 4 it could be seen that the simulated results and the temperature measured at the main pole of the traction motor are in good agreement. Other parts of the motors being studied also have more or less the same order of agreements as well. To compare the temperature distribution of the traction motor with and without electrical braking, the temperature of the traction motor along the journey from Kowloon to Lo Wu are continuously monitored and analyzed. A typical result for the temperature profile of the interpole for one journey without regenerative braking is shown in Fig. 5.

14 0 55 56 57 58 59 60 61 62 63 64 65 66 15 2 15 3

14 6 14 9

15 4

15 5

15 6

15 7

15 8 18 0 17 7 17 8 17 9 18 1

16 0

16 1

16 2

16 3

16 4

16 5

16 6

16 7

16 8

16 9

17 0

17 1

17 2

17 3

17 4

17 5

17 6

18 2

18 3

18 4

18 5

18 6

18 7

18 8

18 9

19 0

19 1

19 2

19 3

19 4

19 5

19 6

19 7

19 8

19 9 20 2

20 0 20 1

20 3 20 4 20 5 20 6 20 7 20 8 20 9 21 0 21 1 21 2 21 3 21 4 21 5 21 6 21 7 21 8 21 9 22 0 22 1 23 7 22 2 22 3 22 4 22 5 22 6 22 7 22 8 22 9 23 0 23 1 23 2 23 3 23 4 23 5 23 6

23 8

Tra cti on Moto r Gri d L in e a nd N od al Po in t

Fig. 3 Traction Motor Temperature Distribution

Temperature profile of traction at the main pole electrical braking in a journey can be typically explained in three stages as: State 1 Temperature of the motor rises quickly when the train starts to accelerate from a station. The temperature rise occurring during this stage is due to the very high current input of 650 A into the motor. The heat generation rate is higher than the rate of heat dissipation from the motor by convection because the rotational speed of the shaft fan is relatively low during this period and hence the corresponding forced convective heat transfer is relatively ineffective. During this period, a rapid temperature rise of the motor occurs and a considerable amount of heat is stored in the motor coil. Stage 2 When the train speed runs in excess of about 45 kph, the rate of heat dissipation from the motor by forced convection increases rapidly because of increases in both the rotational speed of the motor and the cooling air velocity through the annulus gap of the motor. When the speed of the train reaches about 45 kph and above, the weak field control starts to function and the input current is reduced from 650 A to 350 A, and hence a lower heat generation rate results immediately. The lower rate of heat generation plus the rate of heat conduction from the coil centre to the coil surface of the motor due to the heat storage during the previous stage could be dissipated to the convective air rapidly by the enhanced forced convection. Temperature of the motor starts to decrease and the cooling of the motor is

Fig. 4

Temperature profile of traction motor at the interpole The temperature simulation program is used to simulate the temperature distribution of the traction motor during the journey. It requires the input of the speedtime profiles and current-speed profiles for the journey. As we have seen from the previous discussion, the speed-time profile and current-speed profile of the traction motor with regeneration were different from the original case without electrical braking. This will give different temperature-time profiles for the traction motor being studied. The authors can also use the software to evaluate whether there are hot-spots inside the machine being studied. 5.1 Discussion on simulation results The macroscopic thermal behaviour of a traction motor is the temperature rise of the machine resulting from a change in train speed and motor current. The formation of the temperature-rime profile for a EMU without

Fig. 5

ensued. It also led to the removal of part of the heat stored up in the motor during stage 1. Stage 3 A second rise in temperature (positive slope) or fall in temperature with smaller slope (negative slope) is observed when the motoring command is retreated and a braking command is called upon before the train approaches a station. This temperature profile during the coasting period is dependent on the following four factors: Firstly, there was a significant overall temperature rise of the traction motor after every trip which indicates that the rate of heat generation is as a whole greater than the rate of heat dissipation, especially when the cooling period of the motor during the trip is not long enough. Note that the copper field coils of the motor are wrapped by insulation materials and reinforced by varnish, resulting in a very low thermal conductivity and thus the heat dissipation from the coil centre towards the surface is carried out at a relatively low rate. The heat generated is therefore accumulated in the coil and a great temperature gradient is created between the internal region and the surface of the motor. Secondly, the motor is very close to the braking equipment such as the brake disc and the brake pad. The braking system employed by the train is a simple frictional braking system which converts all the kinetic energy of the train into heat at the brake discs and pads. The local ambient temperature around this area increases suddenly because of such heat dissipation. Thirdly, as the train speed decelerates quickly to zero, the convective heat transfer coefficient decreases rapidly to result in a rapidly decreasing convective heat transfer. The convection is further suppressed by the high ambient temperature which reduces the temperature differential between the motor surface and the ambient air. Finally, the under frame equipment is partially enclosed by the platform of station which would give further reduction in the heat dissipation from the motor. The sign of slope (positive of negative) of the profile at braking period therefore depends on the journey distances between stations. For long distance journeys, the maximum travelling speed is much higher and the time taken for coasting is longer, therefore the effect stated in stage 3 would be more significant. Hence the amount of heat generation is higher to result in a

positive slope profile. For short distance journeys, the slope of the profile became negative. 6. CONCLUSIONS

From the simulation results it was found that there was no hot spots inside the traction motor. Subsequent investigation also revealed that the motor failures were indeed not caused by overheating. The studies as presented in this paper also suggests that it is feasible to feed the regenerative energy back to the emergency batteries without the requirement of complicated controllers. 7. 1. REFERENCES Wilkinson, D.T., Electric Braking Performance of Multiple Unit Trains, Proc. IMechE, 1985, Vol 199, No.D4. Limbora, K., Water Models for the Investigation of Cooling Gas Flow in Large Rotating Electrical Machines, Proc. IMechE, 1969-70, Vol. 184, Pt. 3E. Lee, K.Y., Effects of Harmonics on Thermal Performance of Traction Motors, Mhil. Thesis, H.K. Polytechnic University, 1992. Bates, J.J and Tustin, A., Temperature Rises in Electrical Machines as Related to the Properties of Thermal Networks, Proc. IEE, 1956, No. 2026U. Simonson, J.R., The Use of Electrical Models in Cooling Studies of Electrical Machines, Proc. IMechE., 1969-70, Vol 184 Pt. 3E. Roberts, T.J., The Solution of the Heat Flow Equations in Large Electrical Machines, Proc. IMechE., 1969-70, Vol. 184 Pt. 3E.

2.

3.

4.

5.

6.

THERMAL CHARACTERISTICS OF TRACTION MOTORS WITH REGENERATIVE BRAKING S.L. Ho H.C. Wong* and S.K. Poon*

Department of Electrical Engineering, Hong Kong Polytechnic University * Industrial Centre, Hong Kong Polytechnic University

In this modernization programme, the twelve-coach passenger trains which were trailed by diesel-electric locomotives were replaced by a fleet of Electrical Multiple Units (EMUs) running on an electrified rail. Electrification was at 25 kV AC and the track was doubled throughout the entire route length. The KCRC becomes the provider of the most intensive railway service in the world. Apart from saving energy, the use of electrical braking could also reduce the wear upon the mechanical brake discs, hence prolonging the working life of the discs and reducing the amount of asbestos dusts escaping into open air.

In this paper, the authors are proposing to feed the braking energy into the 110 V 165 Ah nickel cadmium emergency batteries which are always available in the trains in Hong Kong. During the electrical barking period, the coasting speed as well as the cooling effect is reduced. The additional heat generated by the traction motor during regeneration might cause overheating in traction motors. Indeed there are evidence in Hong Kong that regenerative operation might overheat the traction motors to the extent that the service life of the motors were drastically reduced. Hence one must take all these factors into consideration when regeneration is introduced.

D I a
3

1

5

L

e

0 .1H

R

e

0 .13 6 7

6 4 .7u 82

L

m

1 .64 18 4H

4 C s2

R s2

2

R

m

0 .04 4 4 .7u 2 10 0u 1 10 V

C s1 1 9

C

e

V

s

T

1
8

10 1 0MEG

E g (V ) g

+ Vg1 Rg

0V

82

V

R s1

T

0

GTO Regenerative Braking Chopper Circuit for PSPICE Simulation

V (k ph )

a rea A

V1

max . b ra kin g ra te (s lo pe ) = 5 kp h

'

V2'

T = t1 + t2 ' + t3 '

a rea B

a rea A = ar ea B

t ( S) 0 t1 t2 ' t3 '

Speed-time Curve of a EMU with Electrical Braking

An efficient use of energy is an important factor in the achievement of an operating profit, as energy typically represents 20 % of the operating cost in the local railway. Thus it is important to regenerate the braking energy back to the system. In the KCRC system it will be rather difficult to regenerate the energy direct into the 25 kV system. There could be instability problems indeed when one is trying to feed the regenerative energy using dc series motors. However it is possible to regenerate the energy into the emergency battery systems.

From the Newton's Law of cooling, the rate of heat transfer by heat convection (q) between a heated surface (with surface area A) can be determined from q = hA(w -f). The heat transfer coefficient h is a very complicated function which depends on the fluid flow condition, fluid properties and geometrical configuration of the heat surface. For forced convection, the heat transfer coefficient may be determined by the non-dimensional expression, i.e. Nu = f(Re, Pr). Different boundary surfaces will have different value of heat transfer coefficients.

A rm

a tu

re
83 84 85 86 10 0 10 1 10 2 11 5

67 68 69 70 71 87 88 89 72

10 3 11 6 1 3 2 13 1 13 2 12 9 13 0 11 8 11 7 10 4

73

90

74

10 5

91

75

4 7 8 15 16 22 14 17 23 13 9 10 11

5

6

11 9

10 6

92

76

13 3 12 13 4

12 0

10 7

93

77

12 1

10 8

94

78

18 24

19 25

20 26

21 27 28 29 30 13 5 12 2 10 9 95 79

13 6

12 3

11 0

96

80 81

31

32

33

34

35

36

37

38

39

40

41

42

13 7

12 4

11 1

97 82 98

13 8

12 5

11 2 99 14 5 14 8 11 3 11 4 12 7 14 2 12 8 14 3 14 4 14 7 15 0 15 1 15 9

43

44

45

46

47

48

49

50

51

52

53

54

13 9 14 1

12 6

14 0 55 56 57 58 59 60 61 62 63 64 65 66 15 2 15 3

14 6 14 9

15 4

15 5

15 6

15 7

15 8 18 0 17 7 17 8 17 9 18 1

16 0

16 1

16 2

16 3

16 4

16 5

16 6

16 7

16 8

16 9

17 0

17 1

17 2

17 3

17 4

17 5

17 6

18 2

18 3

18 4

18 5

18 6

18 7

18 8

18 9

19 0

19 1

19 2

19 3

19 4

19 5

19 6

19 7

19 8

19 9 20 2

20 0 20 1

20 3 20 4 20 5 20 6 20 7 20 8 20 9 21 0 21 1 21 2 21 3 21 4 21 5 21 6 21 7 21 8 21 9 22 0 22 1 23 7 22 2 22 3 22 4 22 5 22 6 22 7 22 8 22 9 23 0 23 1 23 2 23 3 23 4 23 5 23 6

23 8

Tra cti on Moto r Gri d L in e a nd N od al Po in t

Traction Motor Temperature Distribution

Temperature profile of traction at the main pole

Temperature profile of traction motor at the interpole

The macroscopic thermal behaviour of a traction motor is the temperature rise of the machine resulting from a change in train speed and motor current. The temperature-time profile for a EMU without electrical braking in a journey can be typically explained in three stages as: State 1 Temperature of the motor rises quickly when the train starts to accelerate from a station. The temperature rise occurring during this stage is due to the very high current input of 650 A into the motor. The heat generation rate is higher than the rate of heat dissipation from the motor by convection because the rotational speed of the shaft fan is relatively low during this period and hence the corresponding forced convective heat transfer is relatively ineffective. During this period, a rapid temperature rise of the motor occurs and a considerable amount of heat is stored in the motor coil.

Stage 2 When the train speed runs in excess of about 45 kph, the rate of heat dissipation from the motor by forced convection increases rapidly because of increases in both the rotational speed of the motor and the cooling air velocity through the annulus gap of the motor. When the speed of the train reaches about 45 kph and above, the weak field control starts to function and the input current is reduced from 650 A to 350 A, and hence a lower heat generation rate results immediately. The lower rate of heat generation plus the rate of heat conduction from the coil centre to the coil surface of the motor due to the heat storage during the previous stage could be dissipated to the convective air rapidly by the enhanced forced convection. Temperature of the motor starts to decrease and the cooling of the motor is ensued. It also led to the removal of part of the heat stored up in the motor during stage 1.

Stage 3 A second rise in temperature (positive slope) or fall in temperature with smaller slope (negative slope) is observed when the motoring command is retreated and a braking command is called upon before the train approaches a station. This temperature profile during the coasting period is dependent on the following four factors: Firstly, there was a significant overall temperature rise of the traction motor after every trip which indicates that the rate of heat generation is as a whole greater than the rate of heat dissipation, especially when the cooling period of the motor during the trip is not long enough. Note that the copper field coils of the motor are wrapped by insulation materials and reinforced by varnish, resulting in a very low thermal conductivity and thus the heat dissipation

from the coil centre towards the surface is carried out at a relatively low rate. The heat generated is therefore accumulated in the coil and a great temperature gradient is created between the internal region and the surface of the motor. Secondly, the motor is very close to the braking equipment such as the brake disc and the brake pad. The braking system employed by the train is a simple frictional braking system which converts all the kinetic energy of the train into heat at the brake discs and pads. The local ambient temperature around this area increases suddenly because of such heat dissipation. Thirdly, as the train speed decelerates quickly to zero, the convective heat transfer coefficient decreases rapidly to result in a rapidly decreasing convective heat transfer. The convection is further suppressed by the high

ambient temperature which reduces the temperature differential between the motor surface and the ambient air.

Finally, the under frame equipment is partially enclosed by the platform of station which would give further reduction in the heat dissipation from the motor. When electrical braking is introduced, the temperature-time profile will be slightly changed because motor current will flow in the windings during the regenerative braking period. Hence the rise in temperature at stage 3 is higher.

The overall temperature-time profile with and without regenerative braking was however roughly the same.

CONCLUSIONS

From the simulation results it was found that there were no hot spots inside the traction motor. Subsequent investigation also revealed that the motor failures were indeed not caused by overheating. The studies as presented in this paper also suggests that it is feasible to feed the regenerative energy back to the emergency batteries without the requirement of complicated controllers.


				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:40
posted:12/23/2009
language:English
pages:26
Jun Wang Jun Wang Dr
About Some of Those documents come from internet for research purpose,if you have the copyrights of one of them,tell me by mail vixychina@gmail.com.Thank you!