Model-free adaptive control for the PEMFC air supply system based on interval type-2 fuzzy logic systems
Gang Luo,Bingxin Ma,Zezheng Wang,Ling Yin,Yongfu Wang
DOI: https://doi.org/10.1063/5.0014788
IF: 2.847
2020-11-01
Journal of Renewable and Sustainable Energy
Abstract:Control aims to avoid oxygen starvation and maximize the net power output by maintaining the optimal oxygen excess ratio (OER), which varies between 1.8 and 2.5. Because of the nonlinearity of the proton exchange membrane fuel cell (PEMFC) air supply system and the different conditions, ensuring an optimal OER is still a challenge. In this study, a model-free adaptive controller is presented for the PEMFC air supply system based on feedback linearization and interval type-2 fuzzy logic systems (IT2 FLSs). Theoretical analysis and experimental results verify the effectiveness of the proposed control scheme. For the theoretical analysis, first, the PEMFC air supply system is transformed into a canonical form with the feedback linearization technique. Then, zero-dynamics stability is discussed in detail to determine the stability of the internal dynamics. Finally, an adaptive interval type-2 fuzzy logic system controller (AIT2FLSC) is designed on the basis of the Lyapunov stability theory, which does not require complete <i>a priori</i> knowledge of the system dynamics. For the experimental results, the root mean square error (RMSE), variance, and standard deviation (SD) of the tracking error are used as tracking performance metrics to evaluate the control accuracy of the proposed AIT2FLSC, which are 0.0968, 0.0093, and 0.0962, respectively. Compared with the traditional proportion integration differentiation controller (RMSE 0.1119, variance 0.0122, and SD 0.1105), this proposed algorithm obtains better adaptability and the RMSE of the tracking error improves 13.48%. Compared with the adaptive type-1 fuzzy logic system controller (AT1FLSC) (RMSE 0.1076, variance 0.0113, and SD 0.1063), this AT2FLSC has a stronger ability to deal with uncertainty and the RMSE of the tracking error improves 10% when the stack temperature is fixed (353.15 K). Furthermore, when the stack temperature is time-varying, the RMSE, variance, and SD of the tracking error under the AIT2FLSC are 0.0966, 0.0092, and 0.0960, respectively, which is less than AT1FLSC (0.1085, 0.0115, and 0.1073) and the RMSE of the tracking error improves 10.99%.
energy & fuels,green & sustainable science & technology