Influence of the Machine Design on the Indirect Efficiency Determination of Permanent-Magnet Synchronous Machines

For high-power permanent-magnet synchronous machines, the efficiency is usually in the range above 95%, where for a direct efficiency determination the measurement error gets too big. An alternative approach of indirect efficiency determination allows a bigger accuracy. It is done by the summation of individual losses, i.e., current-depending and additional load losses, iron losses, and additional losses caused by inverter feeding. These loss groups are determined at no-load and removed rotor operation. At power ratings below 200 kW the efficiency usually not exceeds 95%, so the proposed indirect method can be compared by the direct method experimentally. This comparison can be done also by numerical methods theoretically without an efficiency limit. In this context, five different permanent-magnet test machines, within the power range from 45 kW to 160 kW, were used to investigate the influence of the stator and rotor design on the losses and on the efficiency determination. For the test machine with a rated power of 160 kW and distributed fractional-slot stator winding, the indirect procedure is shown in detail via no-load, removed rotor, and load measurements for different rotor speeds. For this test machine, a quite good agreement between direct and indirect efficiency values was achieved for motor and generator operation, with deviations of less than 1 percentage point. For the other four test machines, measured loss and efficiency values were compared to simulation results at sinusoidal current feeding. Especially for those permanent-magnet synchronous machines with distributed stator winding without any sub-harmonic stator field waves, the proposed indirect efficiency determination method demonstrates to be a useful alternative to direct efficiency determination. For special machine designs with sub-harmonic stator field waves, caused by fractional slot windings, higher additional rotor losses may occur, which are not considered by the presented method. Therefore, the indirectly determined efficiency values for those machines show higher deviations up to 2 percentage points compared to the direct method.