The development in sensor technology has made many wastewater treatment systems data rich but not necessarily information rich. To extract the adequate information from several sensors is not trivial, and it is not sufficient to consider only the time series. Different tools for detecting unusual on-line measurement data and deviating process behaviour are discussed. In this paper various dimension reduction as well as advanced filtering methods are considered in order to extract adequate information for fault detection and diagnosis. Both the operator and the process engineer can take advantage of such methods for proper monitoring of the plant, in particular extreme events and their causes.