Abstract
Real-time scheduling of heterogeneous solid oxide fuel cell (SOFC)-based microgrids is challenging due to startup costs, minimum operating constraints, and uncertainties in renewable generation and load demand. Prediction-based methods may suffer from forecast errors and high computational complexity, especially when multi-unit commitment decisions are considered. This paper proposes a prediction-free retroactive scheduling framework for such systems. The method combines layered decomposition with cumulative economic signals to determine SOFC commitment in real time without relying on future forecasts, thereby transforming the original multi-unit problem into a set of tractable subproblems. An explicit upper bound on the competitive ratio is derived to quantify the performance gap with respect to the offline optimal solution. Case studies based on realistic photovoltaic (PV), wind, and load profiles demonstrate that the proposed method achieves near-optimal economic performance with significantly lower computational burden than prediction-based benchmarks. Real-time validation on an RT-LAB platform further confirms its feasibility under continuous-time operation and its ability to produce system-level and unit-level responses comparable to predictive control. The results verify the effectiveness and practical applicability of the proposed scheduling strategy for real-time SOFC-based microgrid operation.
| Original language | English |
|---|---|
| Journal | IEEE Transactions on Industry Applications |
| DOIs | |
| State | Accepted/In press - 2026 |
UN SDGs
This output contributes to the following UN Sustainable Development Goals (SDGs)
-
SDG 7 Affordable and Clean Energy
Keywords
- Energy Management
- Fuel cell hybrid power systems
- Retroactive Control
- Solid Oxide Fuel Cell
Fingerprint
Dive into the research topics of 'Retroactive Scheduling for Heterogeneous SOFC-Based Microgrids'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver