New EMS optimizes solar-powered electric bus charging with data-driven method
Handles solar, price, and battery uncertainties using limited data samples
Deep Dive
Wang et al. developed a flexible energy management system (EMS) for solar-powered electric bus charging stations. Their data-driven approach uses polynomial chaos expansion surrogate modeling from limited uncertainty samples and a nonparametric inference method to enrich sparse historical data. Demonstrated on a station with 20 electric buses, the system optimizes charging schedules, renewable integration, and storage while accounting for unpredictable solar output, electricity prices, and bus arrival/departure state of charge.
Key Points
- Uses polynomial chaos expansion surrogate to model system behavior from limited uncertainty data
- Nonparametric inference method enriches input data when historical solar/price/bus SoC data is scarce
- Case study on a 20 electric-bus charging station demonstrates effective cost and reliability optimization
Why It Matters
Helps transit agencies cut energy costs and grid strain by intelligently managing solar+storage for bus fleets.