The Measurement of Additional Losses in Induction Motors: Discussion about the Actually Achievable Uncertainty