PSO-ANFIS modeling for prediction of titanium-pillared clay porous characteristics

پذیرفته شده برای پوستر
: 1038-UFGNSM2021-FULL (R1)
نویسندگان
1Chemical Engineering Department, Fouman Faculty of Engineering, College of Engineering, University of Tehran,Fouman,Iran
2Data Mining Research Group, Fouman Faculty of Engineering, College of Engineering, University of Tehran, Fouman, Iran
3Chemical Engineering Department, Fouman Faculty of Engineering, College of Engineering, University of Tehran, Fouman,Iran
4Chemical Engineering Department, Fouman Faculty of Engineering, College of Engineering, University of Tehran, Fouman, Iran
5Chemical Engineering Department, Fouman Faculty of Engineering, College of Engineering, University of Tehran,Fouman, Iran
چکیده
Pillaring is a modification method to develop a permanent type of porosity in clay. However, pillared clays' main application is in the catalysis and photocatalysis domain; their adsorption ability and selectivity to adsorb a particular pollutant from the liquid and gas phase are also of great importance. In this work, the pillarization of a local type of clay with titanium polyoxycations was investigated. Titanium intercalation solution was prepared by the hydrolysis of titanium isopropoxide in HCl. Titanium pillared clay adsorbents were prepared under different operating parameters to investigate the effects of clay/water ratio (0.5-10 % w/v), H+/Ti molar ratio (2-8), and calcination temperature (300-600℃). The results indicated that the raw clay's specific surface area increased due to the pillaring process from 40 to 164 m2/g. The best specific surface area was obtained at clay/water=0.5% (w/v), H+/Ti=6, and a calcination temperature of 300℃. XRD analysis confirmed the presence of rutile and anatase phases of TiO2 in the pillared clay, and FTIR spectra showed the functional groups of -OH, Si-O-Si, Si-O, and Al-O in this sample. The adaptive neuro-fuzzy inference system (ANFIS) was integrated with the particle swarm optimization (PSO) algorithm to propose a prediction model for estimating the specific surface area of Ti-PILC samples under different preparation parameters. The comparison between predicted and experimental results showed that the PSO-ANFIS model agreed with the experimental data with a correlation coefficient of 0.98. Furthermore, the results showed that the clay/water and H+/Ti ratio are much more effective than the calcination temperature to achieve a larger specific surface area.
کلیدواژه ها