振動台実験のランダム応答データを用いた多層建築物の損傷検出

概要

論文の詳細を見る
To detect the local damage of multi-storey buildings, an efficient method using random vibration measurement is presented. First of all, the change in modal properties due to damage is identified from ambient vibration time series by a multivariate ARMA model. Then, the damaged storeys containing damaged members are extracted from the whole structural system in terms of the change in curvature of mode shapes. Lastly, the location of damaged members is pinpointed and the extent of damage is quantitatively evaluated within the damaged storeys in terms of the change in member stiffness and damping by an inverse perturbation method with weighted least squares technique, taking account of uncertainties of identified modal properties. To verify the applicability of the scheme, a series of shaking table tests of a multi-storey building model have been performed. Based on the results identified by using random data from the vibrational experiment, the effectiveness and limitation of the damage detection scheme is discussed.
社団法人日本建築学会の論文
1995-07-30