Enabling high-fidelity electrothermal modeling of electric flying car batteries: A physics-data hybrid approach

This article proposes a novel control-oriented hybrid electrothermal model (HETM) for large-format electric flying car batteries in urban air mobility (UAM) applications. The model accurately predicts the detailed thermal distribution and the terminal voltage of the battery using only current and ambient temperature as inputs. This model includes a physics-data hybrid thermal model (HTM) and an equivalent-circuit electrical model (ECEM), which interact by heat generation calculation and temperature dependence of electrical parameters. The HTM originates from a physics-based spatially-resolved thermal model and is enhanced by a gated recurrent unit-based thermal calibration network. This thermal model combines the strengths of physics-based and data-driven models, ensuring generalizability while significantly improving prediction accuracy. Temperature-dependent capacity compensation and the best parameterization scheme are considered to improve the ECEM's fidelity. A typical UAM cycling profile is generated by a fixed-wing vehicle and includes a flying car flight mission profile and a multistage constant-current charging segment. Various UAM profiles are conducted on an 8-Ah pouch-type NCM lithium-ion battery in a wide temperature range of 0 °C ∼ 40 °C. The HETM and its submodels are parameterized/trained and validated experimentally and the results demonstrate that the HETM achieves estimation deviations of less than 18 mV for terminal voltage and less than 0.8 °C for all key temperatures under various UAM profiles at 25 °C. The fidelity comparison results of HETM with the state-of-the-art thermal/electrothermal models further confirm its superiority. Finally, the practical significance and potential applicability of HETM are discussed to highlight its value.