Time-alignment-free impact load localisation via dynamic time warping

Impact load localisation in engineering structures experiences significant challenges when measurement data are incomplete or asynchronous. Conventional similarity-based methods, which rely on strict point-to-point temporal alignment, exhibit reduced reliability under such non-ideal conditions. To overcome this limitation, this paper proposes a time-alignment-free framework termed ‘Dynamic Time Warping-Based Localisation and Time-Domain Identification (DTW-LTI)’. The primary novelty of this approach is the utilisation of the DTW algorithm to quantify signal similarity based on morphological features rather than rigid temporal correspondence. This strategy effectively decouples localisation accuracy from the requirement of precise time synchronisation. The identified impact location is subsequently integrated with a regularisation-based inversion method to reconstruct the load time history. Validation was conducted through numerical simulations and physical experiments on stiffened and curved plates. The results demonstrated that, under a condition of 5 % random data loss, where standard Euclidean and cosine similarity metrics yielded 0% accuracy due to misalignment, the proposed DTW-LTI framework maintained a localisation accuracy of 87.5% in simulations and more than 70 % in experiments. These findings confirm the robustness and suitability of the proposed method for structural health monitoring in environments where data integrity cannot be fully guaranteed.