CONVERTER-FED MOTOR LOSSES

Transcription

1 CONVERTER-FED MOTOR LOSSES - AN IMPROVED MEASUREMENT METHOD Andrew H. Baghurst CalTest - Port Elliot, South Australia Motor Summit, Zurich, Switzerland 11/12 October 2016

2 INTRODUCTION The undoubted advantages of controlling motor speed with a converter come at the cost of slightly increased motor losses. IEC (TS) : 2013: Rotating electrical machines Part 2-3: Specific test methods for determining losses and efficiency of converter-fed AC machines Clause 6.2: Experience has shown that the additional harmonic motor losses generally increase with load. The work described in this paper set out to determine whether or not that statement is true. Caltest MS'16 Converter fed motor losses 2

3 CLC/WG1 2014/06/04 meeting Feedback of TF3 tests on extra harmonics calculation for induction motor fed by converter. Excerpt from CLC 2014 study CLC/TC2/WG1- Eric Vassent ( Cenelec-TC2 Chairman ) page 3

4 Extra harmonics losses map Conclusion? Is it necessary to measure extra H losses at load? CLC/TC2/WG1 Eric Vassent ( Cenelec-TC2 Chairman ) page 4

5 CENELEC TC2/WG1 STUDY 2014 No conclusions relating to converter-fed motor losses can be drawn from the CENELEC results. It is inconceivable that motors of different ratings could behave so differently when converter fed, as suggested by the Extra harmonic losses map. One of the reasons for this is the inevitable changes in test conditions which occur when the motor supply is changed from nominally sinusoidal to converter. Even small changes in either the values or measurements of the following have the potential to mask changes in motor losses: Voltage Frequency Load torque Motor and/or ambient temperature Useful loss measurements are possible only if made by direct A-B comparison, under almost identical electrical and mechanical conditions, and within seconds of each other. Caltest MS'16 Converter fed motor losses 5

6 ESSENTIAL CONDITIONS FOR A DIRECT A-B COMPARISON The two supplies (sine and converter) have exactly the same frequency. The two supplies have exactly the same (equivalent) voltages. The mechanical load on the motor is exactly the same. The measurements are made under exactly the same motor and ambient air temperature conditions, within seconds of each other, and preferably without stopping the motor and its associated mechanical load, if any. Caltest MS'16 Converter fed motor losses 6

7 A-B COMPARISON USING INDEPENDENT SUPPLIES Motor Converter supply Sinusoidal supply The supplies must have very carefully matched voltages and frequencies. An ability to synchronise the converter with the sinusoidal supply is useful. Switching produces unacceptable electrical and mechanical transients, however. Caltest MS'16 Converter fed motor losses 7

8 A CONVERTER-DERIVED SINE-FILTERED SUPPLY DOES NOT PRODUCE FURTHER MOTOR LOSSES (Diagram courtesy of ABB) Sine-filters commonly consist of 3 x series L-C stages (one per phase). The series-resonant frequency for each phase is usually about 1 khz for a converter switching frequency of 4 khz. The capacitors may be either star or delta connected. Caltest MS'16 Converter fed motor losses 8

9 A-B COMPARISON USING A SINGLE CONVERTER PLUS SINE-FILTER Motor Converter Sine-filter The two supply frequencies are now identical. The changeover switch, and associated transients, could be avoided if it were possible to slide along inductor L from converter (left) to sinusoidal supply (right). Caltest MS'16 Converter fed motor losses 9

10 FIRST SWITCHED SINE-FILTER SKETCH Additional capacitors (red) can improve performance. Caltest MS'16 Converter fed motor losses 10

11 Experimental switched sine-filter: L (each phase) = 4 x 0.76 = 3.04 mh, C = 10 µf Caltest Resonant frequency MS' Converter Hz fed (Capacitors motor losses arrowed, top right) 11

12 Caltest MS'16 Converter fed motor losses 12

13 Caltest MS'16 Converter fed motor losses 13

14 Caltest MS'16 Converter fed motor losses 14

15 Caltest MS'16 Converter fed motor losses 15

16 Voltage drop across sine filter compensated-for by adjustment of converter output voltage while it is running Caltest MS'16 Converter fed motor losses 16

17 FLUX VOLTMETER Uses a single-pole passive integrator to provide comparable motor terminal voltage measurements for both sinusoidal and converter supplies Caltest MS'16 Converter fed motor losses 17

18 (3 ) (6) (5) (4) (3) (2) (7) (1) EXPERIMENTAL SETUP Converter (1), switched sine-filter (2), torque (3, 3 ), speed (4), waveform (5), flux voltage indicator (6) and power analyser (7) Caltest MS'16 Converter fed motor losses 18

19 Caltest Converter motor lossesand dynamometer (right) 1.1 kw 4 pole 400 V MS'16 motor underfedtest (left) 19

20 THE MEASUREMENT PROCEDURE 1. The motor under test was coupled to a dynamometer, and run at rated load for about an hour with a sinusoidal supply at rated voltage and frequency. 2. The electrical input power was then measured and recorded. 3. The motor was then transferred to converter supply, and the supply (flux) voltage adjusted back to the motor s rated value. 4. The electrical input power was then re-measured and recorded. 5. The mechanical load on the motor was then reduced to 75%, 50%, 25% and 0% (no-load, with the motor uncoupled from the dynamometer), whilst maintaining the motor supply voltage at the rated value, with electrical power input measurements made for both converter and sinusoidal supply, as above, under each load condition. 6. The further motor losses at each load point were calculated by subtracting the power reading in (2) from the value obtained in (4) above. Caltest MS'16 Converter fed motor losses 20

21 RESULTS Motor no. 1: CMG 1.1 kw 400 V 4 pole Load (%) Supply type Speed (rpm) Torque (Nm) Power input (W) ΔP (W) ΔP/ P rated (%) Sine Converter Sine Converter Sine Converter Sine Converter Sine Converter Caltest MS'16 Converter fed motor losses 21

22 Motor no. 2: SEW 1.1 kw 400 V 4 pole Load (%) Supply type Speed (rpm) Torque (Nm) Power input (W) ΔP (W) ΔP/ P rated (%) Sine Converter Sine Converter Sine Converter Sine Converter Sine Caltest MS'16 Converter fed motor losses 22

23 Motor no. 3: Invertek 1.1 kw, 400 V, 4 pole Load (%) Supply type Speed (rpm) Torque (Nm) Power input (W) ΔP (W) ΔP/ P rated (%) Sine Converter Sine Converter Sine Converter Sine Converter Sine Converter Caltest MS'16 Converter fed motor losses 23

24 Further motor losses on converter supply as percent of rating SEW LOSS % OF RATING CMG INVERTEK LOAD (%) Caltest MS'16 Converter fed motor losses 24

25 CONCLUSIONS 1. Feeding a motor from a converter produces further losses which do not change significantly (if at all) with load: the statement in IEC (TS) is therefore incorrect. 2. Converter-feeding a motor of the type used in this study produces further losses of about 1% of the motor s rated power output, when compared with a sinusoidal supply. 3. If the further losses produced by a converter increase with load, that increase is less than 0.1% of the motor s rating. 4. IEC (TS) : 2013 is not capable of providing any useful information about a converter-driven motor s losses. 5. The proposed new method offers a means by which such losses may accurately be determined. 6. The new method may therefore be used as a reference with which any other converter-fed motor loss measurement method may be compared. Caltest MS'16 Converter fed motor losses 25

26 ACKNOWLEDGEMENTS The author acknowledges, with gratitude - Financial assistance from EPFL, Lausanne, Switzerland Assistance with converter programming from Benno Weiss, Siemens, Germany John Yelland, CalTest, who built the experimental switched sine-filter and Andrew Parrott, CalTest, who made the measurements and produced the graphs Caltest MS'16 Converter fed motor losses 26