Knowledgebased supervision and control of waste-water treatment plant: A real-time application

Abstract The hardware architecture and the software development of a real-time knowledge-based distributed control system for the supervision of a wastewater treatment pilot plant are presented. The operation scheme is based on an A2/O system (anaerobic, anoxic, oxic) and includes organic matter removal, nitrification/denitrification, and Enhanced Biological Phosphorous Removal (EBPR). The hardware architecture involves different supervision levels, including two autonomous process computers (plant control and analysers control) and a PLC. The software architecture includes a Knowledge-Based Expert System (KBES) as the top control system. The KBES has been developed in G2, an expert system development environment based on object-oriented structural design. This system is fed with data obtained from the plant monitoring (in-line, on-line and off-line data) and with actuation data of the lower control levels. Using this information, the KBES supervises the pilot plant in order to help with fault detection and plant maintenance, but also modifies setpoints and other variables of the local control level to adapt the system to the different influent conditions. The overall system has been implemented and validated at pilot scale. Keywords Activated sludge; expert system; knowledge-based expert system (KBES); real-time; wastewater treatmen