AN EXPERIMENTAL-STUDY OF STRUCTURAL HEALTH MONITORING USING INCOMPLETE MEASUREMENTS

In this paper, algorithmic approaches to enhance structural health mon itoring capability when faced with incomplete measurements are address ed. The incomplete measurement problem has two aspects: (i) experiment al measurement of a lesser number of modes of vibration than that of t he analytical model and (ii) experimental measurement of a lesser numb er of degrees of freedom than that of the analytical model. Studies co mparing model reduction, eigenvector expansion, and a hybrid model red uction/eigenvector expansion to address the second contribution are pe rformed using experimental data. These approaches to the incomplete me asurement problem are evaluated within the frameworks of multiple-cons traint matrix adjustment (both sparsity and nonsparsity preserving alg orithms) and minimum rank perturbation theory, which are both applicab le for model refinement as well as damage location. Experimental evalu ation of the proposed approaches utilize data from the NASA Langley Re search Center 8-bay truss and McDonnell Douglas Aerospace 10-bay truss facilities.