Thermoelectric Tracking Control of Solid Oxide Fuel Cell System Using Adaptive Dynamic Programming
Xiao-Long Wu,Hongli Liu,Hong Zhang,Jinting Fu,Bin Xie,Han Hu,Xi Li
DOI: https://doi.org/10.1109/ccdc55256.2022.10034400
2022-01-01
Abstract:An adaptive optimal control strategy for solid oxide fuel cell power generation systems (SOFCs) is proposed in this paper. A new estimator is proposed in a simple form to solve the modeling uncertainty, and express the control of SOFCs as an optimal control problem. Then, a new system thermoelectric parameter tracking control strategy is proposed. The strategy uses data-driven to build the recursive neural network model of the system thermoelectric parameters. In order to ensure convergence, an adaptive law based on parameter estimation error is proposed to update the weights of neural networks online. The simulation results show that the data-driven model can effectively identify the power generation process of solid oxide fuel cell system. At the same time, the control method can realize the optimal tracking control of power, voltage, stack temperature and afterburner temperature with limited input.