STRUCTURAL DAMAGE DETECTION BASED ON ENERGY TRANSFER BETWEEN INTRINSIC MODES

摘要

In this study, a new damage detection algorithm for specific types of damages such as breathing cracks, which are called "active discontinuities" in this paper, is proposed. The algorithm is based on the nonlinear behavior of this class of damages and hence, is more precise and sensitive to damage compared to other common linear methods. The active discontinuities can be regarded as additional degrees of freedom (DOFs) which need energy to be excited. Because the input energy of both the intact and the damaged structures is finite, the energy content of vibrating modes will be changed due to damage. Thus, the properties of distribution of energy between vibrating modes can be used as indices for detecting damage. An essential detectability condition using this concept is decomposing a signal such that no spurious mode imposed to its expansion. In order to satisfy this condition, Empirical Mode decomposition (EMD) is used to extract the vibrating modes since all nonlinearities in a signal are preserved while no spurious mode or assumption of stationarity is imposed on the problem. Prevention of mode mixing, which is an important drawback of EMD, is another necessary condition for robustness of the algorithm. A solution is proposed in this paper to satisfy this condition in which special constraints are imposed on the normal procedure of EMD. Then, the fourth central moment, kurtosis, is used to compare the distribution of energy between the modified vibrating modes. The algorithm is verified through experimental testing of a simple steel cantilever structure under various damage scenarios. Results demonstrate the efficacy of the method for detecting discontinuities in a real structure.