Damage Estimation Method Using Committee of Neural Networks

摘要

The committee technique for neural networks has been widely used for pattern recognitions in speech and vision studies. In this study, the committee technique is applied to damage estimation of structures for the purpose of structural health monitoring. The input to the neural networks consists of the modal parameters, and the output is composed of the element-level damage indices. Multiple neural networks are constructed and each individual neural networks is trained independently with different initial synaptic weights. Then, the estimated damage indices from different neural networks are averaged. Several committee methods were investigated and used to estimate the element-level damage locations and severities. The validity of the committee technique for damage estimation was examined on a frame structure through numerical simulation. Then experiments were carried out on a bridge model with a composite cross section subjected to vehicle loadings. It has been found that the estimated damage indices improve significantly by employing the committee technique.