A fuzzy-based expert system (ES) for the diagnosis and supervision for anaerobic digesters is presented. The system was developed in a Microsoft Windows support using fuzzy logic inference together with a rule base for the implementation of expert knowledge. The ES runs on-line through three main modules, which determine the state and trend of the process, and the best set points for the actuation on the final control elements of the plant. Two further modules run in parallel, when they are required by the operator, using off-line and on-line information for the detection of inhibition due to toxic compounds in the process and for the validation of the on-line diagnosis.
The diagnosis and supervision ES was tuned up in order to adjust the membership functions describing the process, and lately tested, running on-line, to study the response of the rule base.