FRACTIONAL-ORDER SUPERCAPACITOR MODELING FOR HIGH-POWER EVTOL APPLICATIONS USING GCD-BASED CHEBYSHEV-GALERKIN SPECTRAL ANALYSIS
DOI:
https://doi.org/10.26577/ijmph.20210126Abstract
Abstract. The rapid development of electric vertical takeoff and landing (eVTOL) vehicles has increased the demand for high-power, lightweight, and fast-response energy storage systems capable of handling severe transient loads. Supercapacitors have emerged as a promising solution in hybrid energy storage systems due to their high-power density, rapid charge--discharge capability, and long cycle life. However, their electrochemical dynamics exhibit strong fractional-order behavior caused by anomalous ion diffusion and the fractal electrode--electrolyte interface, making accurate numerical modeling challenging for conventional methods. This paper presents an improved GCD-based Chebyshev--Galerkin spectral framework for solving the fractional integro-differential equations (FIDEs) governing electric double-layer capacitor (EDLC) dynamics under the constant phase element (CPE) model. The proposed method constructs a generalized fractional Chebyshev basis that spans a truncated fractional-power trial space adapted to the Caputo derivative, thereby improving representation of fractional-power behavior relative to standard polynomial bases. In addition, a graded Gauss--Legendre quadrature strategy is introduced to accurately resolve the singular behavior near the initial charging stage. Three numerical approaches are comparatively investigated: the finite-difference L1 scheme, the standard Chebyshev--Galerkin method, and the proposed GCD-based spectral method. For the representative supercapacitor model with fractional order \texorpdfstring{$\alpha=0.75$}{alpha = 0.75}, the proposed approach achieves a maximum error of \texorpdfstring{$4.2\times10^{-3}$}{4.2e-3} using only 7 degrees of freedom, whereas the standard Chebyshev--Galerkin method has a larger reported error even at higher basis orders, and the finite-difference method requires more than 500 degrees of freedom for comparable performance. The reported results demonstrate improved error per degree of freedom for the tested case. The measured error reaches a plateau at higher basis orders, and real-time suitability remains to be established by processor-specific runtime benchmarking.
Keywords: Supercapacitor modeling; Fractional-order systems; Chebyshev--Galerkin method; Spectral numerical methods.












