Dynamic Parameter Encoding for Genetic Algorithms
Machine Learning
| Hi-index | 0.00 |
The operating temperature of proton exchange membrane fuel cells stack is a very important control variable, which affects electrochemical reactions and humidity of proton exchange membrane, its variation also has a significant influence on the performance and lifespan of the fuel cells. However, the existing stack models were based on the electrical performance experiments, did not distinctly describe the microcosmic phenomena in the stack. The cell models were unable to consider the thermal conduction between cells, and they were too complicate to design control systems. In this paper, a stack temperature model is developed based on conservation laws. A novel genetic algorithm is presented to identify the model coefficients. Finally, the simulation and experiment results of the temperature distribution and variation are presented, the results show that the model predictions compare well with experimental results.