模式密度方法在结构异常检测中的应用研究

Application of structural damage diagnosis using density-based outlier detection

结构损伤发生后,结构的动力特性将随之改变,因此可通过检测动力特性的异常来判断结构的健康状况。研究了基于模式密度的异常检测在损伤诊断中的应用,该方法具有无需建立训练模型,能够支持模式动态更新等优点。以ASCE学会提出的基准结构为对象,采用基于相空间重构的奇异值分解技术提取适当个数的奇异值构造特征模式集,应用模式密度方法检测结构异常状况。仿真和试验结果表明,该方法能有效地检测损伤的发生。

The performance of structural vibration varies with the change of the structural condition. Therefore the variation of vibration signals is helpful to detect early damage in the structure. Density-based outlier detection is introduced to diagnosis the damage in a structure. The advantages of the method are that it does not require establishing a training model and is capable of updating dynamically. For demonstration a numerical and experiment study on health monitoring of the ASCE benchmark model is performed. First, appropriate numbers of singular values are extracted using singular value decomposition and phase space reconstruction to act as feature parameter. Then the density-based outlier detection is employed to diagnosis the damage. The analysis results show that the proposed method is effective in damage detection.