Identification of Membrane Fouling in a Full-Scale Water Treatment Plant Using Data Analytics

Microfiltration membranes are an effective and mature technology for water and wastewater treatment. However, membrane fouling is a critical operational challenge, involving the accumulation of colloids, organics, and the growth of bacterial biofilms on the membrane surface and within its pores, leading to permeability decline. Although membrane fouling has recently been simulated using mechanistic models, these models contain a large number of process variables with high nonlinearity and uncertainty. Data analysis techniques offer alternative effective approaches to describe membrane fouling under different operational conditions. They can also diagnose the causal relationships between the different process variables. In the current study, various data analysis techniques were applied on an industrial dataset, describing 141 full-scale water and wastewater treatment plants. These plants were categorized based on their influent water source, operational conditions, and membrane cleaning management. For example, it was found that most of the plants receiving their influent waters from surface water and wastewater sources are operated under “optimal” conditions. On the other hand, the plants treating groundwater are operated under “conservative” conditions. Moreover, the performance of the membranes in one of the surface water treatment plants was investigated under various operational conditions including several cleaning protocols. In addition, the specific flux decline and slope of specific flux loss were determined and evaluated under different cleaning management schemes to evaluate the membrane’s performance. The interdependencies between the process variables (e.g., cleaning chemical management) and membrane fouling rate are investigated to determine the governing factors in membrane fouling. Based on the results of the data analysis, it was found that an increase in the clean-in-place (CIP) intervals can reduce the slope of specific flux decline. In addition, when the interval of chemically enhanced backwash (CEB) increases, the specific flux recovery was enhanced during the cleaning cycles. The results can be used by decision-makers and water/wastewater treatment plant operators to reduce plant down-time by minimizing the frequency of clean-in-place (CIP) membrane cleaning procedures, while recommending the optimal frequency of membrane flux maintenance to minimize flux loss and rate of flux decline.