Reliability of linear losses-to-power scaling method of electric drive systems

This paper assesses the reliability of the linear losses-to-power scaling method of electric drive systems, composed of a permanent magnet synchronous machine and IGBT inverter. The scaling method, which is commonly used in research studies on system-level design, postulates that overall power losses scale linearly to the desired power rating. However, the validity of this assumption is questionable, and it may impact simulation outcomes. To address these issues, the study derives the conditions that must be met for the linear scaling method to be achieved based on the well-known geometric scaling laws of electric machines and information obtained from inverter datasheets. The analysis at component-level indicates that the linear scaling method is effective in predicting efficiency for upscaling cases, with a discrepancy of less than 1%. However, it is found that this method is not appropriate for downscaling scenarios. Despite these limitations, a sensitivity analysis is conducted at system-level to examine the effect of linear losses-to-power scaling on the energy consumption of an electric vehicle, considering power scaling factors of 0.67 and 1.33. The examination compares this effect to different geometric variations of scaled electric machines and different power ratings of inverters, using 12 different standardized driving cycles. The outcomes of the analysis indicate that the difference in terms of energy consumption is within a satisfactory range below 2% for the upscaling and downscaling cases. This demonstrates the feasibility of the method at system-level for fast assessment of energy consumption in the early development phases of electric vehicles.