Experimental and modeling evaluation of droplet size in immiscible liquid-liquid stirred vessel using various impeller designs

• Published in: Journal of the Taiwan Institute of Chemical Engineers Vol. 100; pp. 26 - 36
• Main Authors: Afshar Ghotli, Reza, Abbasi, Mohammad Reza, Bagheri, AmirHossein, Raman, Abdul Aziz Abdul, Ibrahim, Shaliza, Bostanci, Huseyin
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.07.2019
• Subjects: Anfis–Fuzzy C–Means; Drop size measurement; Empirical correlation; Image analysis; Impeller design; Empirical correlation; Image analysis; Anfis–Fuzzy C–Means; Drop size measurement; Impeller design
• ISSN: 1876-1070; 1876-1089
• DOI: 10.1016/j.jtice.2019.04.005

•Experimentally determination of drop sizes through different designs of impellers.•Drop size analysis through image processing tools.•Develop a model to predict d32.by using ANFIS-Fuzzy C-means (ANFIS-FCM).•Compare the prediction capability of the ANFIS-FCM and the empirical correlations. The present study investigates the effects of impeller design and dispersed phase volume ratio on mean drop sizes (d32) in immiscible liquid-liquid stirred vessel through experimental and modeling approaches. Various impeller designs including conventional and new impeller designs were employed to cover both radial and axial flow impellers. The microscopic method associated with image processing tools was used for the drop size analysis. The results showed the hydrofoil impeller produced the largest drop sizes while the double-curved blade turbine produced the smallest drop sizes, corresponding to about 37% difference. Increasing the dispersed phase volume ratio from 1% to 10%) increased the d32 by approximately 20–40%. Adaptive neuro-fuzzy inference system based on fuzzy C–means (ANFIS-FCM) clustering algorithm was used to develop a model to predict drop sizes, and its validation and accuracy were examined by comparing the results to the experimental data. The results also proved the superior prediction capability of the ANFIS-FCM method over the empirical correlations for the most cases.