Abstract
In automotive engineering, spot welds on assembled structures such as Body in White (BiW) have a significant effect on the vehicles' dynamic characteristics. Understandably, imperfections in the spot welds will cause variations in the dynamic properties such as natural frequencies and mode shapes of the structure. In this paper, a complicated welded structure which is a simplified Natural Gas Vehicle (NGV) platform is investigated. The structure fabricated from thin metal sheets consists of ten components. They are jointed together by a number of scattered spot welds. NASTRAN Solution 200 based on sensitivity analysis is used to identify the most sensitive parameters to natural frequencies. The numerical model of the undamaged structure is initially updated in order to minimise the discrepancies between the measured and numerical data using NASTRAN optimisation code. The initial updated model serves as a benchmark for the subsequent structural damage identification. The numerical data of the benchmark model is then compared with the measured data obtained from the damaged structure. The same updating procedure is applied to the benchmark model in order to bring the numerical data as close as possible to the measured data of the damaged structure. The disparity in certain parameter values from the parameter values used in the benchmark model shows a fault or damage in the location of a particular joint, depending on the severity of this disparity. The challenge in this work is to localise damaged area and quantify the damage of the complicated structure with multiple spot welds in the presence of uncertainty in the location and material properties of the welds.
Export citation and abstract BibTeX RIS