A Statistical Prediction Method of AC Copper Loss in Random Wound Windings of Electrical Machines

Due to the random behavior of conductors in machines with random windings, it is difficult to predict and calculate the Alternating Current (AC) copper loss. This paper proposes a practical statistical simulation method to predict the AC copper loss in the random windings of electrical machines. The method is based on a semi-analytical method for evaluating AC copper loss. To feature the random characteristics of windings whilst minimizing complexity, a statistical winding model is established. This model uses only three parameters to control the generation of conductor arrangements with different degrees of mixing. These three parameters are calibrated by the measured data from a repeatedly inserted coil group, rather than a series of wound stator prototypes, thus significantly reducing the cost in a research and development (R&D) process. Thereafter, the statistical model with calibrated parameters is used to predict the AC copper loss of the full machine. A 48-slot 4-pole stator is used to illustrate and validate the method. The proposed method can be used to fast and practically predict the characteristics of the AC copper loss during the early stages of machine design.