The evaluation of the state of structural components and the detection of occasional material flaws are extremely important to ensure the safe operation and the lifetime of the structures. The present B.Sc. thesis aims at optimizing the nondestructive material evaluation technique of the ferromagnetic devices, which is used in the detection of material flaws of the specimens. In this test an electromagnet is usually swept on the surface of the specimen. The method is based on the measurement of the change of the magnetic flux in the magnet caused by the material flaw.
According to the given configuration, the results can be very sensitive to the uncertainties of the model input parameters. The purpose of the optimization is to separate the most and less important parameters according to their influence on the uncertainty of the model output. This kind of test can be performed by means of sensitivity analysis, which quantitatively characterizes the uncertainties caused by each model parameter, respectively. The chosen sensitivity analysis method calculates the so-called Sobol-indices, which aims at partitioning the variance of the output among the input parameters and their groups according to their contribution to the total output variance. The evaluation is performed by means of a polynomial chaos metamodeling technique.
The results revealed the importance of the air gap between the magnet and the specimen, so this parameter should be considered first to improve the reproducibility of the measurement.