Measurements of a 37 kw induction motor. Rated values Voltage 400 V Current 72 A Frequency 50 Hz Power 37 kw Connection Star

Transcription

1 Measurements of a 37 kw induction motor Rated values Voltage 4 V Current 72 A Frequency 5 Hz Power 37 kw Connection Star

2 Losses of a loaded machine Voltage, current and power P = P -w T loss in Torque m Power supply Machine under test Load machine Problems: Measuring the power difference requires accurate methods Segregation to loss components is difficult

3 Calorimetric measurements Voltage, current and power Power supply q, r, c T i T o Torque ( ) P = qrc T -T loss o i Machine under test Load machine A more accurate method than measuring the difference between the input and output powers. Does not solve the problem of loss segregation.

4 No-load loss as a function of the voltage 2 IEC Measuring set-up Variable voltage source Voltage, current and power Input power [W] Result Voltage [V]

5 No-load loss as a function of the voltage II 12 P = P - Ri in n 2 n P [W] P = 152 W Loss components [W] Pcore Pres Pfric u^2 [V^2] Voltage [V]

6 No-load loss as a function of the voltage III Problem (at least in this 37 kw machine): When the same test was repeated, significantly lower friction losses were obtained. Probably the bearing losses vary somehow stochastically.

7 Retardation test T P fric fric dw =-J dt dw =-Jw dt Speed [1/s] Speed measurement Time [s] Grid Friction loss [W] Speed [1/s]

8 Running at no-load at a small voltage Voltage, current and power Speed measurement Variable voltage and frequency source The rotation speed was kept at 15 rpm by adjusting both the frequency and voltage. Power [W] Voltage [V]

9 Friction losses at speed 15 rpm Summary of friction losses [W] Minimum Maximum Difference A) No-load test at variable voltage B) Retardation test C) Running at small voltage and 15 rpm Pfric [W] A B C

10 Loss measurement at synchronous speed Grid Voltage, current and power measurement The test machine was coupled to a synchronous machine that forced rotation at the synchronous speed. u f i f

11 Loss measurement at synchronous speed II Power [W] Power [W] Measurement Successive synchronisations and loss measurements at the rated voltage 4 V Voltage [V] Loss as a function of supply voltage. The magnetisation of the synchronous machine was reversed at zero voltage to obtain the two branches.

12 A refined machine model Fundamental machine Hysteresis machine Harmonic machine Friction load 1.. Torque [Nm] Speed [rpm]

13 Refined machine model II Ideal induction machine Hysteresis machine Torque [Nm]. -1. Torque [Nm] Speed [rpm] Speed [rpm] Harmonic machine 1. Friction load Torque [Nm]. -1. Torque [Nm] Speed [rpm] Speed [rpm]

14 Tests at small slips Grid Voltage, current and power measurement u, f r Induction machine Slip-ring machine

15 Tests at small slips Slip =.2 %; Pave = 788 W P [W] Slip = -.2 %; Pave =462 W P [W] Rotor angle [] Rotor frequency [mhz] Input power at two opposite slips ±.2 as a function of the position angle of the rotor. The average powers as a function of the rotor frequency. The powers at the synchronous speed are extrapolated.

16 Tests at small slips P [W] 8 T [Nm] Voltage [V] Voltage [V] The average power at synchronous speed as a function of the supply voltage, f = 5 Hz Hysteresis torque as a function of the supply voltage, f = 5 Hz

17 Tests at small slips P [W] 1 P [W] Voltage [V] Voltage [V] Slip control Standard no-load tests Comparison of losses obtained from the standard no-load test and from the slip-control tests. The difference of the loss curves of the previous figure. This difference is associated with higher harmonics and friction.

18 Segregation of the losses P [W] Voltage [V] Pcore Pfri c Pres Pharm

19 Losses from FE analysis 1 A A + s + s f = Łm ł t u b b = 1 f d = d s + A l J l Ł t ł a a

20 Power and resistive losses Power taken from the supply p in m = i=1 ss i i ui Losses in a filamentary winding (no eddy currents) s m res= s i i s i=1 p R i ( ) 2 Losses in a solid conductor r A pres =- J dv t V

21 Eddy-current loss in thin conductors x B d J = s E db E=-x dt P 4 4s db p = = JEdV = x dv V 2 2 pld pld Ł dt ł V 4s db 2 d 2 d db = x l2 - x dx = s 2 pld Ł dt ł Ł2 ł 16 Ł dt ł d - 2 V d

22 Circulating currents between parallel wires Sometimes just a few effective turns in a slot (three in the figure) Tens of thin parallel wires to reduce the eddy currents The leakage flux over the slot causes different electromotive forces in the parallel wires B Circulating currents are induced between the parallel conductors

23 Losses in electrical steel sheets Conventional core loss model P = P + P core hyst eddy = C fb h x C f B e Hysteresis losses in FEM-model Eddy-current losses in FEM-model hyst N n= 1 hyst n n n V c ( ) ( ) P = C f B x, y dv eddy N n= 1 eddy 2 n n n V c 2 ( ) ( ) P = C f B x, y dv 2 Fourier-coefficient of flux density (, ) Ø (, ) Bn xy = º An xyezø ß

24 Losses in the magnetic core B H Welded seam B B B J Skin effect in the sheets Insulation fault Insulation faults in the sheets lead to currents circulating from sheet to sheet

25 Losses in the magnetic core Punching the sheets deteriorates the magnetic characteristics and causes burrs that enable galvanic contacts between the sheets. Additional load losses in % of input power 3, 2,5 2, 1,5 1,,5,, Rated output / kw As losses are difficult to estimate, standards allow loss estimation based on the input power of the machine.

26 More comprehensive core loss models A hysteresis model can be combined with a FE model. The effect of core losses is included in the field solution B [T] H [A/m] Iron sample in an alternating field Fields in the core of a machine

27 Computed electrical losses

28 Losses related to inverter supply High-frequency flux and distribution of average resistive loss

29 Sinusoidal supply and inverter supply kw machine at no load 37 kw machine at rated load Loss [W] Loss [W] Sinusoidal 1 Inverter 2 Sinusoidal 1 Inverter 2 Resistive stator loss Resistive rotor loss Resistive stator loss Resistive rotor loss Core loss in stator Core loss in rotor Core loss in stator Core loss in rotor

30 Validation of loss models 2 18 Measured losses 16 P [W] Voltage [V] Pcore Pfri c Pres Pharm

31 Validation of loss models II Total electromagnetic loss [W] Resistive stator loss [W] Voltage [V] Measured Computed Voltage [V] Measured Computed

32 Validation of loss models III Core loss [W] Voltage [V] Measured Computed Losses due to braking harmonic torques [W] Voltage [V] Measured Computed

33 Summary of loss analysis Segregating the loss distribution from the measured results is a challenging task Computing the resistive losses is quite well under control some problems remain related to circulating currents Analysis of core losses requires better models hysteresis losses in rotating fields eddy-current losses in non-linear steel sheets Insulation faults causing inter-sheet currents A practise accepted in standards (induction motor): Core losses are measured at no load The losses related to loading are calculated