Computationally Lightweight Method for Campbell Diagram Plotting in High-Speed Electric Machines

Abstract

Campbell diagram is plotted from calculation results to identify a rotating object's critical speeds. The Campbell diagram is formed by calculating the supported natural frequencies in a defined operation speed range. It is an important step when designing rotating machines, e.g., an integrated high-speed electric motor that can be sensitive to the rotor's dynamical behavior. Currently, minimizing unnecessary calculation points is important for rapid design iterations and utilization of physics-based models with artificial intelligence. In cases where large variants of rotor geometry or using high-fidelity models, the calculation burden becomes high. In the research, a methodology based on a minimum number of calculation points and a second-order fitting equation is proposed, i.e., instead of using a high number of fixed calculation intervals, a three-point calculation methodology is proposed. The proposed methodology can be applied with neural network-based methods or implemented with high-fidelity models such as solid element models where the physics-based models can be used to create sensitivity to model parameters and study their influence with the traditional rotordynamics Campbell diagram tool. In the results, a comparison of two case studies is shown, and the computational cost is compared.