Model identification and strategy application for Solid Oxide Fuel Cell using Rotor Hopfield Neural Network based on a novel optimization method

Abstract

The novelty of this paper is to suggest an effective method according to the application of the Rotor Hopfield Neural Network optimized by the Grey Wolf Optimization (GWO) method for the identification of the Solid Oxide Fuel Cell (SOFC) model. In this literature, the basic required metrics to present the transient models of Solid Oxide Fuel Cell are defined. The proposed model is a hybrid model that is composed of the Rotor Hopfield Neural Network (RHNN) optimized by the GWO algorithm. The hybrid RHNN-GWO model, including a steady-state RHNN Neural Network, ensured by an optimization method. The RHNN algorithm is presented to assess the metrics of the RHNN-GWO model. In contrast to the wavering, the Mean Squared Error (MSE) for the RHNN-GWO model is calculated by 0.0017. The presented model results are examined with some well-known model results. The lowest values for Mean Squared Error belongs to the RHNN-GWO model. Also, the proposed model conserves a tremendous value of calculation time compared to the other models. Also, the proposed model shows a good agreement with SOFC results with lower computational difficulty. For 5000 samples, the variation of the voltage is in the [320,360]V interval, which completely follows the reference voltage of the SOFC.