Enhanced Calibration and Performance Prediction Method for Entire Propulsion System of eVTOL UAV

The electric propulsion system (EPS) comprises various components, such as a battery, motor, controller, and propeller. Accurate analysis of these components requires many detailed parameters which can be challenging to obtain during the initial design phases. This paper presents a method for constructing a calibration and global prediction model for a family of EPS components, to overcome analysis limitations during these design stages. A low-fidelity analysis method is developed to predict EPS performance with a given set of specifications, and wind tunnel tests are conducted to provide high-fidelity data. A linear calibration model is developed by modifying theoretical equations to include relative additive terms and coefficients. Subsequently, a global prediction model is generated using a radial basis function network. The model predicts calibration coefficients for a set of target EPS combinations without corresponding high-fidelity data. By substituting these coefficients into the modified equations, EPS performance can be predicted with enhanced accuracy, without requiring additional high-fidelity analysis. Three examples of global prediction models are presented to demonstrate the feasibility and current limits of the method.