Approach to complex hydrogen reactor optimization modeling based on ANFIS

An innovative approach to hydrogen reactor modeling based on Adaptive Neuro-Fuzzy Inference System (ANFIS) is proposed to improve the approximating and self-adaptive ability of existing models. To exert the information of the hydrogen reactor operation data for constructing the optimization model reasonably, the adaptive neural network algorithm is combined with the fuzzy logic inference mechanism and the subtractive clustering method. There are good agreements between the actual values and the ones obtained using the proposed model. The results show that the ANFIS model with high accuracy and few convergence epoches have the capability to be applied to engineering simulation applications.