“…Online sensors at a WRRF are usually prone to anomalies (e.g., noise, failure, drift, and bias), which can dramatically affect the quality and/or the performance of model simulations. Although some studies have investigated anomaly detection and gap‐filling of wastewater data (Alferes et al, 2015; De Mulder et al, 2018; Martin & Vanrolleghem, 2014), they do not focus on the real‐time functionalities that are required in DT applications. In addition, there should be an automated connection between the data processing pipeline, which is mostly implemented in data‐driven programming languages like R and Python, and the process model simulators, which are usually developed using commercial software like WEST (DHI A/S, Denmark), SUMO (Dynamita, France), GPS‐X (Hatch, Canada), SIMBA# (inCTRL, Canada), and BioWin (EnviroSim, Canada).…”