State identification by soft computing techniques during service conditions

摘要

In this paper a double-step strategy for damage detection and localization is applied for the case of a long span suspension bridge.The strategy uses the service response of the bridge under environmental and traffic loads,allowing the continuity of service.In the first step,from the wind response of the bridge,a distributed permanent network of sensors,integrated to a neural network,allows the damage detection and a coarse localization,and produces also an alert signal for the bridge management.In the second step an appositely installed system of sensors,integrated to a neural network,refines the localization to element-level,and produces a damage vector by the mean of a pattern recognition procedure.