This work proposes a modular, configurable, online signal analyzer and event detector which can be applied to the control of SBR plants. The control system was developed using pH, ORP and dissolved oxygen process parameters as input, combining neural networks and fuzzy logic rules in an attempt to identify different operational conditions related to the SBR phases (i.e. oxic, anoxic). A set of training signals was analyzed, identifying general features that are then matched in order of detecting global process events, defined using high-level rules. The detection software achieved a percentage of true positives higher than 90% whereas false negatives were mostly due to noisy or ambiguous sig
L. Luccarini, P. Mello, D. Sottara, A. Spagni (2008). Artificial Intelligence based rules for event recognition and control applied to SBR systems. ROMA : s.n.
Artificial Intelligence based rules for event recognition and control applied to SBR systems
MELLO, PAOLA;SOTTARA, DAVIDE;
2008
Abstract
This work proposes a modular, configurable, online signal analyzer and event detector which can be applied to the control of SBR plants. The control system was developed using pH, ORP and dissolved oxygen process parameters as input, combining neural networks and fuzzy logic rules in an attempt to identify different operational conditions related to the SBR phases (i.e. oxic, anoxic). A set of training signals was analyzed, identifying general features that are then matched in order of detecting global process events, defined using high-level rules. The detection software achieved a percentage of true positives higher than 90% whereas false negatives were mostly due to noisy or ambiguous sigI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
