Use of On-line Water Quality Monitoring Data to Predict Bacteriological Failures

Variations in continuously monitored on-line water quality data were investigated to establish whether they could be linked to coliform detections at regulatory monitoring points. We focussed on chlorine residual, turbidity and flow rate at water treatment works (WTW)-A. Archived on-line monitoring data from WTW-A were analysed using cross-correlation and self-organising maps in MATLAB® to identify trends in the data running up to coliform detections. The results show that these tools could be developed to help manage WTWs to reduce the number of bacteriological failures. A fingerprint of WTW conditions relating to coliform failures was identified for this case study.