Development of a case-based system for the supervision of an activated sludge process

The development of a case-based reasoning system for the supervision of an activated sludge process is presented here. The methodology proposed permit s the use of past experiences to solve new problems that arise in the proce ss. These experiences are classified as cases or situations. The adaptation of cases and the generation of new cases are used to tune the response of the system and to learn from the new information generated by the process. The case and the case library definition the initial seed, the search and r etrieval process, the adaptation, the action, the evaluation and the learni ng steps are presented and outlined. The process studied is the wastewater treatment plant of Girona, Spain. Two examples of the response of the syste m to two different operational situations are presented. The paper also out lines the integration of different fields in a multidisciplinary approach a s the most optimal solution to ensure the successful control and supervisio n of complex processes like the activated sludge process. With this aim the integration of an array of specific supervisory intelligent systems (for t he logical analysis and reasoning) and numerical computations for detailed engineering is suggested.