Improvement of operational performance of submerged membrane photocatalytic ultrafiltration for industrial oily wastewater treatment using AL/ML technique and statistical optimization methodology

Abstract

Water scarcity is the next big problem the world is about to face as 70% of the Earth’s surface is covered in water of which only 2.5% is fresh water. Oil refineries are a major source of water pollution. They produce a significant amount of oil wastewater. The disposal of this contaminated water represents a significant challenge for the entire petroleum industry. The treatment of municipal and industrial wastewater is another critical issue. From 2018 to 2025, the global membrane filtration market is expected to reach USD 19.6 billion, growing at a CAGR of 6.4%. The increasing population, growing awareness of wastewater reuse, rapid industrialization, high-end products and efficiency offered by membrane filtration technologies, shift from chemical water treatment to physical treatment, and stringent regulations regarding treatment water and water discharge are the key drivers of the membrane filtration market. This research focuses on the wastewater treatment method of submerged membrane photocatalytic ultrafiltration. The membrane is used to clean oily wastewater in this method. Membrane performance evaluation projects entail trial-and-error methods under a wide range of process operating conditions. The majority of previous research projects on the submerged membrane ultrafiltration (SMUF) system used an experimental approach with one process variable at a time. This method is time-consuming and expensive. Principal component analysis (PCA), design of experiments (DOE), and Artificial Intelligence/Machine Learning (AI/ML) appear to be capable of circumventing the limitation. Through a systematic experimental strategy based on PCA, DOE, and AI/ML, this study improved the operational performance of submerged membrane photocatalytic ultrafiltration operations technology in industrial oily wastewater treatment. This strategy focuses on tuning and simultaneous optimization of membrane operating variables under various controllable conditions. Meanwhile, the lack of a quick numerical simulation model capable of recognizing and predicting the effects of various process operating conditions on the final performance of the wastewater treatment system has prompted the utilization of statistical and artificial intelligence (AI) methodologies. This study’s ultimate goal is to optimize and improve the operational performances of the Submerged Membrane Photocatalytic Ultrafiltration (SMPUF) system in industrial oily wastewater treatment. To accomplish this goal, statistical methods, principal components analysis (PCA), design of experiments (DOE), and AI/ML techniques are employed to investigate and improve the system’s performance.

Date	31 Jan 2023
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	Claudiane Ouellet-Plamondon (Supervisor) & Benoit Barbeau (Co-supervisor)