Impact location and load identification through inverse analysis with bounded uncertain measurements

The growing demand for real-time damage assessment necessitates development of an efficient inverse analysis algorithm with consideration of practical issues such as uncertainty in measurement. A mathematical model-based inverse analysis scheme is proposed to identify impact locations and reconstruct impact load time history of a simply supported plate through multiple levels of analysis. The proximity of the impact location is first determined by the triangulation method and the impact location is then refined by minimization of an objective function through the particle swarm optimization method (PSO). Loss of data due to filtration is addressed in a further level by performing an interval analysis based on extreme measurement errors. The outcome of the analyses is a mean impact location, a load time history, and a range of likely deviations. The extreme deviation in impact location is shown by bounding lines, which form a rectangle. The deviation in load time history is also shown by upper and lower bounding sinusoidal curves. The results of the analyses indicate that the proposed method can effectively locate the impact point and reconstruct the load time history even with the existence of noise in the measured response.