Preventive maintenance of sensors assures their correct operation for a certain period of time, but do not guarantee that they remain faultless for other periods, besides being occasionally unnecessary. In industrial plants the analysis of signals from sensors that monitor a plant is a difficult task due to the high-dimensionality of data. Therefore an on-line strategy for monitoring and correcting multiple sensors is required. This work proposes the use of Auto-Associative Neural Networks with a Modified Robust Training and the Sequential Probability Ratio Test (SPRT) in an on-line monitoring system for self-correction and anomalies detection in the measurements generated by a large number of sensors. Unlike existing models, the proposed system aims at using only one AANN for each group of sensors to reconstruct signals from faulty sensors. The model is evaluated with a database containing measurements of industrial sensors that control and carry out the monitoring of an internal combustion engine installed in a mining truck. The result shows the ability of the proposed model to map and correct signals, with a low error, from faulty sensors.
auto-associative neural networks; sensors; maintenance; signal monitoring system
