A power density output prediction method for solid oxide fuel cells based on a radial basis function neural network model with a gaussian function variable radius drosophila optimization algorithm

Due to the high operating temperature of the solid oxide fuel cell (SOFC) and the sealing structure, it is long and difficult to directly measure its operating parameters. Therefore, this study adopts a modeling method based on the radial basis function (RBF) neural network to accurately predict the SOFC system’s output power density. Specifically, firstly, the mechanism modeling method is adopted to model the SOFC system with Aspen Plus software, to collect a large amount of experimental data as the training set and test set of the model. Then, the drosophila optimization algorithm (gaussian Network-based drosophila Optimization Algorithm, GNFOA) based on the gaussian function variable radius is combined with the RBF neural network to optimize the network’s weight and bias to enhance the model’s predictive accuracy. The discoveries indicate that the GNFOA based RBF neural network model shows high precision and dependability when forecasting the SOFC system’s output power density. The R
²
values of the training set and the test set are greater than 0.98, the value of RMSE is less than 0.22, and the value of MSE is less than 0.05.