Signal monitoring toward an intelligent and automatic control of wastewater treatment plant

In small plants typical operating costs for treated water are higher than those of large plants and this is certainly dependent on the relative higher load variations at small plants compared at the larger, but also on the fact the instrumentation installed on the former is minimal. An appropriate low cost equipment could be installed in small - medium plants (data logger stand-alone with suitable sensors) and the data produced used to implement an intelligent control system capable to monitor the processes continuously, analyse the collected time series and classify the various operational states reachable by the plant. Such a system should be able to recognize known situations, extracting features and patterns from the signals, and apply domain knowledge to choose itself the most appropriate control actions, effectively acting as a Decision-Support System (DSS) with Fault Detection and Isolation (FDI) capabilities. In this paper, the first part of a project where a preliminary analysis of this correlation, comparing trends, range of values and characteristic points, is introduced. The signals used are pH, ORP, DO, measurable by cheap and reliable sensors, which correlation with biological processes is well known in Sequencing Batch Reactor (SBR), but not in Conventional Activated Sludge (CAS) plants.