Prediction of gas fraction in wastewater treatment with ANFIS method for different height of single sparger location

Abstract

Providing more numerical analysis from process engineering can build up big data set for one master machine learning method that can cover many process engineering with different operational conditions. This master computing structure can store huge data set in the form of mathematical equations, and the researchers can avoid occupying computer memory with massive data set about process engineering. In this study, a single sparger at a different level of the 3D (3-dimensional) bubble column reactor (BCR) was simulated with the Eulerian method, representing the bubbly flow process. Then, the computed flow characteristics were trained in the training part of the ANFIS method, and then this method estimates artificial flow characteristics with the prediction ability of itself. The results showed that there was a good agreement between Computational fluid dynamics (CFD) and Adaptive Neuro-Fuzzy Inference System (ANFIS) methods. The ANFIS method with a high number of membership functions can accurately predict the gas fraction when the single sparger location changes. This modeling framework can also optimize the amount of gas fraction by changing the sparger locations. Furthermore, Artificial Intelligence (AI) methods could be beneficial in saving time and money, which can be used in a wide range of research studies to perform different tasks. • Hybrid machine learning methodology for simulation of fluid mechanics. • Numerical simulation of a physical system and implementing the model. • Validation of the model and understanding the effect of process parameters.