Information Retrieval from Mathematical Models for Process Optimisation in Waste-Water Treatment

Abstract

For the operation of a sewage plant it is of great importance always to know the current state of the biological processes inside the plant as precisely as possible. For supporting the operator in this task we present in this paper a new approach of information retrieval from mathematical simulation models combined with knowledge-based techniques. In this approach a well-known and widely used mathematical model of the biological processes inside a sewage plant, the „Activated Sludge Model“ has been implemented together with a model of the plant’s sludge settling processes in an existing real-time simulator. This allows a clear structuring of the model and thus it is more understandable to the operators. For a computer-aided identification of the process-state within the biological part of the sewage plant, the simulation is run on-line in parallel to the real cleaning process. The measured and simulated process-data are fed into an on-line realtime diagnosis system that evaluates the state of the cleaning process and gives suggestions to the operator for an optimal process control. In the simulation system these suggestions again can be evaluated by simulating the future behaviour of the cleaning process with the changed control parameters.