Model-Based Optimization of Drive Cycle Efficiency in Electric Two-Wheelers

Improving the energy efficiency of electric vehicles (EVs) is essential for extending range and reducing environmental impact. This presentation introduces a model-based approach to optimize drive cycle efficiency using a system-level model of an electric two-wheeler and a variable flux permanent magnetic synchronous machine (VF-PMSM) model developed via finite element analysis. A Simulink^® model captures electromechanical behavior across varying conditions, enabling targeted improvements of up to 10% in average drive cycle efficiency by dynamically adjusting machine parameters based on drive cycles like FTP75. The second part discusses advanced control techniques for permanent magnet synchronous motors (PMSM). We present a data-driven approach for safe control optimisation in PMSM, overcoming challenges of traditional methodologies. By replacing the single Gaussian kernel in Bayesian optimisation with additive Gaussian kernels, our method enhances the ability to model high-dimensional functions and shortens the iteration process. Experiments validate the proposed algorithm in field-oriented control (FOC) of PMSM using MATLAB^®, Simulink, and real-time hardware from Speedgoat^®.