Fuzzy-neuro system for bridge health monitoring

Many civil and mechanical systems are in continuous use despite aging and associated potential risk for damage accumulation. Hence, the ability to monitor the structural health of these systems on a real-time basis is becoming very important. This paper describes a practical real-time structural health monitoring system using soft computing tools and its application to the structural health monitoring of a steel bridge located in Missouri. Vibration data collected from this bridge was processed and fed to the fuzzy logic decision system. The fuzzy logic decision system makes use of fuzzy clustering to determine the possible presence of damage in the bridge. A neural network prediction system which makes use of backpropagation algorithm predicts the amount of actual damage in the members which were predicted damaged by the fuzzy logic.