Abstract: This research presents a three-phase solar photovoltaic grid-interfaced electric vehicle battery management system (3Ph-SPV-GEVBC) regulated by two synergistic control algorithms. The main control architecture uses a Fractional-Order Super-Twisting Sliding Mode Controller (FO-STSMC) and an Adaptive Gain Nonlinear Extended State Disturbance Observer (AGNESDO) to keep the DC-link voltage and EV battery current in check when power flows in both directions. The secondary framework uses an Adaptive Sparse Kernel Maximum Versoria Criterion (ASKMVC) filter to get three-phase fundamental load-current references from a grid that isn't working properly. The FOSTSMC gets rid of the high-frequency chattering that is common in first-order sliding mode implementations, but it still converges in a finite amount of time in both grid-to-vehicle (G2V) and vehicle-to-grid (V2G) modes. AGNESDO is better than other fixed-gain observers like LIDO, SOSMDO, and LESDO at limiting DC-link transient undershoot and speeding up recovery after a disturbance. The ASK-MVC filter combines a Versoria-kernel loss function with bias-compensated sparsity regularization. This makes it faster to converge and have a lower mean-square error than synchronous reference frame (SRF), LMS, NLMS, NMCC, and SABCAF techniques when there is non-Gaussian impulsive noise, which is common in real-world distribution feeders. The Modified Variable Step-Size Incremental Conductance (MVSI-InC) system allows for maximum power point tracking that is stable even when there is partial shade. The hardware layout has two levels: a DC-DC boost converter for connecting to solar systems and a threephase two-level voltage source converter (VSC) that also works as a DSTATCOM. This setup connects a 5 kW SPV array, a 7.2 kWh LiFePO₄ battery pack, and a 415 V three-phase utility grid. Comprehensive MATLAB/Simulink validation across four operating scenarios, along with hardware prototype experiments, confirm a grid current total harmonic distortion (THD) of 2.41% in V2G mode and 2.87% in G2V mode, both of which meet IEEE 519-2022 limits. The system also runs with a unity power factor and a DC-link undershoot below 3.2% under AGNESDO guidance.

Keywords: Adaptive filtering, battery energy storage, disturbance observer, electric vehicle, FO-STSMC, grid integration, maximum power point tracking, power quality, sliding mode control, solar photovoltaic, ASK-MVC filter, V2G/G2V operation.

Downloads: | DOI: 10.17148/IJIREEICE.2026.14543