Experimental characterization and probabilistic group velocity mapping of nondestructive impacts in plate structures using simulated annealing algorithm

Identifying low-velocity impacts poses a significant challenge for structural health monitoring (SHM) systems, particularly when measurements are associated with considerable uncertainties. This study presents an optimization framework that combines simulated annealing algorithm with a probabilistic group velocity mapping method to characterize low velocity impacts while minimizing uncertainties. The optimization framework was developed to improve impact localization performance without relying on baseline data or material properties of the structure. Experimental impact tests were conducted on an aluminum plate equipped with a limited number of piezoelectric sensors to evaluate the robustness and efficiency of the proposed approach. Sensor measurements were processed to analyze frequency components, develop full-field group velocity, and identify impact characteristics. The study also evaluated the effect of sensor placement on localization accuracy, demonstrating that optimal sensor configurations can reduce errors and uncertainties by up to 50 %. Impacts occurring near sensors or structural boundaries were found to increase uncertainty, yet the framework successfully localized such impacts with more than 90 % probability. The framework demonstrates the potential for improving the efficacy of SHM systems, enabling accurate impact localization and damage assessment.