Uses of Methods with Result Verification for Simplified Control-Oriented Solid Oxide Fuel Cell Models (Artikel)
|Uses of Methods with Result Verification for Simplified Control-Oriented Solid Oxide Fuel Cell Models (Artikel)|
|Autor||Ekaterina Auer, Stefan Kiel|
|In:||Proc. REC 2016|
Solid oxide fuel cells (SOFCs) represent a promising technology in the energy sector. One of their advantages is high efficiency. Improvements are aimed at, for example, reducing their operating temperatures or speeding up starting times. Developing efficient control-oriented models for the SOFC temperature has also been in the focus of ongoing research. The goal here is to devise (global) dynamic models valid for a wide range of operating conditions instead of obtaining (linear) approximations of the behavior only for a certain mode of operation. The use of methods with result verification helps to address reliability of SOFC models and to take into account bounded uncertainty.
Dynamic SOFC models are systems of differential equations, the parameters of which have to be fitted to the available sensor data (usually, thousands of measurements). Normally, such differential equations do not have analytical solutions if we keep in mind the goal of devising models suitable for a wide range of operating conditions. In this paper, we point out two simplification possibilities possessing closed-form solutions along with their areas of validity. Moreover, we study their potential with respect to uncertainty handling with the focus on simulation. We perform a sensitivity analysis for different variants of the suggested simplifications. Finally, we consider uncertainty of different magnitudes in heat capacities of gases, which we propagate through the system using interval, affine and Taylor model based methods.