Abstract
A solid oxide fuel cell (SOFC) system model can be used to predict the influence of various parameters such as fuel inlet flow and voltage on the system output performance. This article focuses on a typical UAV (Unmanned Aerial Vehicle) SOFC power system and develops a system model that couples the SOFC surrogate model with other BOP (Balance of Plant) models. The SOFC surrogate model is trained using BP (Back Propagation) neural network on a dataset calculated from a multiphysics SOFC model. Based on this model, the effects of system operating conditions on key system performance are studied. The results demonstrate that fast prediction of the SOFC system performance can be realized.
| Original language | English |
|---|---|
| Pages (from-to) | 3820-3830 |
| Number of pages | 11 |
| Journal | International Journal of Green Energy |
| Volume | 22 |
| Issue number | 16 |
| DOIs | |
| Publication status | Published - 2025 |
| Externally published | Yes |
Keywords
- multiphysics model
- neural networks
- Solid oxide fuel cells
- system simulation