Evaluation of qualitative trend analysis as a tool for automation

Abstract Ammonium N H 4 + load based aeration control on biological wastewater treatment plants saves costs and enhances nitrogen removal. However, the need for maintenance intensive N H 4 + sensors hamper the controls application in practice. Alternatives, in the form of soft-sensors are broadly discussed in academia. A soft-sensor recently described in literature exploits the pH effects induced by biological N H 4 + oxidation. This concept is now further developed by means of qualitative trend analysis (QTA). Previously, the qualitative path estimation (QPE) algorithms was proposed as a fast and reliable QTA algorithm for batch process data analysis. It does not allow online application in continuous flow systems however. In this work, a modification of QPE, call qualitative state estimation (QSE), is proposed as a suitable algorithm for continuous-flow systems. Initial tests indicate that the QSE algorithms is a robust technique for extraction of relevant information in a full-scale environment. At the WWTP Hard in Winterthur, this resulted in cost-saving automation of the aeration system. This contribution summarizes these first results.