Multi-objective exergetic optimization of continuous photo-biohydrogen production process using a novel hybrid fuzzy clustering-ranking approach coupled with Radial Basis Function (RBF) neural network

Abstract

This work was focused on the development of a novel hybrid fuzzy clustering-ranking approach coupled with radial basis function (RBF) neural network for the optimization of the key operational parameters for hydrogen production from photo-fermentation. The bioconversion of syngas to hydrogen via water-gas shift (WGS) reaction was carried out using the light-dependent microorganism Rhodospirillum rubrum and acetate as the sole carbon source. The RBF neural network was used to correlate exergetic outputs (normalized exergy destruction as well as rational and process exergetic efficiencies) to two exogenous input variables (culture agitation speed and syngas flow rate). The developed RBF model was interfaced with the proposed hybrid fuzzy clustering-ranking algorithm to simultaneously maximize the rational and process exergetic efficiencies and minimize the normalized exergy destruction. Moreover, the conventional fuzzy optimization algorithm was applied in order to assess the capability of the proposed approach over the existing methods for multi-objective optimization of complex biofuel production systems such as a continuous photobioreactor with respect to the sustainability and productivity issues. Once the development of the objective functions were carried out using RBF Neural Networks, the proposed algorithm was able to identify the optimum exergetically-sustainable operational parameters of carbon monoxide fermentation to biohydrogen with 20 clusters. This corresponded to the syngas flow rate of 13.68 mL/min and culture agitation speed of 348.62 rpm yielding the process exergetic efficiency of 16.46%, rational exergetic efficiency of 91.59%, and normalized exergy destruction of 2.14. The predicted optimum values by the proposed algorithm in this study were more suitable compared with the conventional fuzzy method. Therefore, the novel algorithm developed in this study may be a promising approach for navigation of the most cost-effective and environmental friendly operational parameters for fermentative hydrogen production from pilot and large scale photobioreactors.