ISSN print edition: 0366-6352 ISSN electronic edition: 1336-9075 Registr. No.: MK SR 9/7 Published monthly

Simulation and optimization of a hydrotreating reactor using a new hybrid imperialistic competition algorithm-based adaptive neuro-fuzzy inference system (ICA-ANFIS)

Hosein Eshghanmalek, Habib Ale Ebrahim, and Mohammad Javad Azarhoosh

Department of Chemical Engineering, Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran

E-mail: mj.azarhoosh@urmia.ac.ir

Received: 18 April 2022  Accepted: 27 May 2022

Abstract:

A diesel hydrotreating (HDT) trickle-bed reactor (TBR) with co- and counter-current streams was simulated using heterogeneous models. The simulation results for the output sulfur concentration agree with the pilot data in both co- and counter-current flows. Also, the effects of major operational parameters were examined on the performance of the reactor. The results show the positive effect of counter-current streams direction, temperature, hydrogen pressure, and negative effect of hydrogen sulfide (H2S) pressure and liquid and gas velocities on the hydrodesulfurization (HDS) reaction. The results of the HDT reactor simulation were then modeled using the adaptive neuro-fuzzy inference system (ANFIS) method. According to the results, ANFIS is very powerful in predicting the simulation results. Finally, the reactor operating conditions were optimized to maximize sulfur removal from diesel using a new combining the imperialistic competition algorithm (ICA) and ANFIS, called ICA-ANFIS. The ANFIS was adopted to calculate the cost function in the ICA and reduced the run-time of the optimization program by more than 1000 times. In the optimum result, sulfur removal increased by 33% compared with the baseline. The main novelty of this study is modeling and optimizing the heterogeneous simulation results using the hybrid of ANFIS and ICA methods.

Graphical abstract

Keywords: Hydrotreating; Simulation; Counter-current reactor; Optimization; Imperialistic competition algorithm; Adaptive neuro-fuzzy inference system

Full paper is available at www.springerlink.com.

DOI: 10.1007/s11696-022-02310-0

Chemical Papers 76 (10) 6247–6261 (2022)