Intelligent monitoring for quality control in a biological nutrient removal wastewater treatment pilot plant

The conflicting demands for tighter environmental quality control on wastewater treatment plants while striving to lower operating costs has lead Sydney Water Corporation to embark on a 'crusade' to understand biological nutrient removal (BNR) processes and develop novel process control strategies to leverage this knowledge. The GIRD Novel Process Control Project is a collaborative attempt to demonstrate this on a pilot scale. An important feature of developing such technology is the implementation of 'intelligent monitoring'. This paper represents a two-pronged approach to an intelligent monitoring strategy, including 1. the successful identification of optimal nutrient sampling locations in order to reduce sensor costs whilst retaining process information, and 2. the premature detection of process abnormalities to prevent uncontrollable operating conditions. PCA is employed to extract features from plant data and a pattern recognition approach is used in the development of fault related training clusters. Copyright (C) 1998 IFAC.