Experimental Characterization and 1D Simulation of an EV Drivetrain System With a Double-Stator Axial Flux SPMSM

This article addresses the comprehensive experimental characterization and numerical modelling of an axial flux permanent magnet synchronous Electric Vehicle Drivetrain System (AFPMSM). A methodology based on studies from various procedures, both from research and the industrial sector (Original Equipment Manufacturer (OEM) & Tier1), is employed. A Design of Experiments (DoE) approach was employed to guide the experimental procedures, enabling the precise determination of critical parameters, magnitudes, and their interrelationships. Most studies on characterization focus on a single aspect, idealizing or simplifying other influences and nonlinearities, which limits the analysis and does not include the complete drivetrain. An analysis of Field Oriented Control (FOC) and its performance was conducted. This included experimentally quantification of factors such as losses, inertias, phase resistances, temperatures, back electromotive force (BEMF), nominal and peak torque and power, efficiencies, inductances and magnetic flux. Correlations were established through equations or Look Up Tables (LUT), to capture the nonlinearities, determining their effects on other variables. A numerical model of the powertrain was generated and validated in AVL CRUISE<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">TM</sup> M. The prediction error for torque and power of the model is less than 3%, and the mean absolute percentage error (MAPE) for the phase (id) and quadrature (iq) currents are less than 4.74% and 2.51%, respectively. The model reduces the need for experimental testing in the development of electric powertrains. The experimental procedure described in this work enables a comprehensive characterization of the drivetrain, specifically addressing the nonlinearities and associated effects such as magnetic circuit saturation, inductance variation, and temperature impact, thereby providing a more accurate representation of the SPMSM motor's behavior.