Development of a dual-metric operational decision-support model for full-scale submerged membrane bioreactors (MBRs)

Membrane bioreactor (MBR) operations are often characterized by reactive and suboptimal responses to complex fouling phenomena, leading to high operational costs. To facilitate a shift toward proactive management, this study developed and evaluated a novel dual-metric operational decision-support model. We analyzed year-long operational data from a full-scale MBR facility, integrating the hydrodynamic index (H) with detailed resistance component analysis (back-pulsing resistance, R_BP; cake resistance, R_cake). The study identified two critical hydrodynamic thresholds: H_Target, representing sustainable capacity and the onset of accelerated fouling, and H_Lim, defining the absolute operational limit. Detailed trajectory analysis revealed two primary fouling pathways: Pathway A (R_BP-initiated cyclical degradation) and Pathway B (R_cake-dominant response to acute stress). A key finding is that the system's fouling maturity (baseline R_cake, R_BP) influences its sensitivity and response pathway. Based on these insights, a control matrix was developed, categorizing the operation into nine statistically distinct fouling states defined by H values and resistance thresholds. This was further simplified into a practical action matrix to guide targeted operator interventions. The proposed model aims to improve system performance, most notably through dynamic air scour control using H_Target as a practical setpoint and the diagnosis of resistance components. This physics-informed, data-driven methodology is expected to bridge the gap between field operations and future machine learning-based control policies, facilitating reduced energy and chemical consumption through optimized operational decision-making.