Online Winding Hot Spot Temperature Prediction for Electrical Machines Based on Generic Relationship between Average and Maximum Temperatures

In this paper, the generic relationship between the average and maximum stator slot winding temperatures in electrical machines is analytically revealed for the first time. It is shown that the difference between average and maximum temperatures can be presented by the product of copper loss with a function related to basic winding geometric and thermal parameters. For machines with non-overlapping concentrated windings, the thermal effect of end windings is relatively less critical, and the average winding temperature in slots can represent the average temperature of the entire winding. Therefore, the winding hot spot temperature can be easily predicted based on the knowledge of average temperature, as it can be obtained by any stator resistance estimation method. The proposed method has the merits that it does not require complex thermal modeling and overcomes the significant limitation of stator resistance estimation methods in conventional thermal condition monitoring, i.e., only the average temperature can be predicted. The accuracy and effectiveness of the proposed method are verified by both the thermal finite element method and the experiments based on a 12-slot/10-pole permanent magnet synchronous machine.