An Application of a Neural Network to Detection of Deteriorated Steel Structural Members

摘要

The purpose of this study is to confirm the feasibility of a method for detecting deteriorated structural members using three-layer back-propagation neural networks while considering changes in dynamic characteristics, such as higher order natural frequencies and vibration modes. Investigation results for framed structures confirmed that effective detection of deteriorated members is possible, in spite of minor changes in dynamic characteristics due to structural deterioration. Prospects for application of the present method are promising pending further studies on specific problem areas.