Proactive Alerts, Rapid Response: Event Detection for Membranes in DPR Erik Vosburgh, March 5, 2024 Billy Raseman, PhD, PE1, Troy Walker, PE, MIE1, Brian Pecson, PhD, PE2, Aleks Pisarenko, PhD, WTPO Grade 52 1: Hazen and Sawyer 2: Trussell Technologies Water Research Foundation Project 4954 Goal: Create a practical blueprint for enhanced DPR monitoring and control. Outcomes: Developed framework for process engineers to design site-specific event detection systems (EDS) Open-source code for a working EDS prototype. Project Team: Core strengths: process engineering and operations. Additional expertise: data engineering and analytics. Membrane mediated treatment processes generate massive amounts of data Surveillance Many operators There is need to systems often fail only analyze and cut through the fog to alert operators report the data of data and find until a critical limit once a month early signs of has been crossed critical control point failures Gap Gap Gap 3 EDS Applied to Microfiltration Pre-requisite for EDS: Brainstorm potential CCP failures and data required Monitoring Point Event – Are sensors reporting the truth? Sensor left on hold Water Quality Event – Are there upstream WQ changes requiring attention? N/A Process Event – Are the treatment units functioning properly? MF filtrate turbidity as surrogate measure for pathogen LRV MF filtrate turbidity and MF process status indicators were Integer value used to indicate the state of MF: 0 = Off prioritized. 1 = Forward Flow 2 = Reverse Filtration MF Process Status 3 = Air Scrub 4 = Strainer BW Data from periods where the 5 = MIT 6 = CIP system is backwashing, 7 = Flush 8 = Enhanced Flux Maintenance (EFM) performing integrity tests, or 9 = Fill 10 = Paused undergoing chemical cleans are not representative of system performance. EDS Applied to Microfiltration Data Flagging & Event Detection Stagnant data Abrupt change Upper bound Lower bound eulaV Stagnant test Range test (bound violation) Time Process event: Turbidity measurements exceeding a 0. NTU threshold for at least 15 consecutive minutes categorized as a process failure event. KPIs Total Organic Carbon (TOC) CCP Performance and Electrical Conductivity (EC) Specific Flux considered as candidate CCP Feed Permeate Feed EC Normalized TOC TOC monitoring parameters. Differential Pressure TOC % Pe F r e m e e d a te EC TOC Removal Recovery TEOCC Removal Removal TOC selected as surrogate for pathogen removal. EDS Applied to Reverse Osmosis RO Data Selection Plant & RO tags: 8000 Data Tags Greater number of data tags compared to microfiltration. EDS Applied to Reverse Osmosis Data Flagging & Event Detection Process event: Membrane breach Indicated by effluent TOC spike above historical levels.

Details
First NameErik
Last NameVosburgh
KeywordsRaseman, Walker, Pisarenko, Pecson, Total Organic Carbon, Reverse Osmosis, Microfiltration, Conductivity, Turbidity, Osmosis, Sensor, Lead
Year2024
FileTUE07-01_Vosburgh_Erik.pdf