Efficient Localisation of Impact Load for Composite Structure Based on Response Similarity Search and Optimisation

Abstract

Background

Composite materials have been extensively used in various industry fields due to their distinguishing characteristics. However, low-velocity impact loads would undermine the mechanical properties of composite structures significantly.

Objective

To improve the integrity and safety of composite structures, it is imperative to unearth the accurate locations of low-velocity impact loads efficiently.

Methods

In this research, a novel approach hybridising response similarity search and optimisation strategy is developed. The innovation of the approach comes from the adoption of a “divide-and-conquer” strategy to alleviate extensive computations for time history reconstruction during the impact load localisation process so as to optimise computational efficiency and accuracy. In more detail, the approach is comprised of two localisation processes: (i) a coarse process to quickly identify several potential positions for an impact load via response similarity measurements based on time-domain and frequency-domain signals; (ii) a precise process to fine-tune the exact location of the impact load by minimising the nominal residual between the reconstructed and actual responses from the above potential positions.

Results

Experiments are conducted on a carbon fibre composite sandwich panel to validate and demonstrate the effectiveness and superiority of the approach in terms of localisation efficiency and accuracy. It indicates that the approach achieves 100% accuracy in impact load localisation. It also shows that the approach only takes approximately 4.0 s to localise 20 impact load cases, which is only about one-eighth of the time taken by the traditional optimisation strategy approach to fulfil the same function.

Conclusions

The hybrid approach designed based on response similarity search and optimisation strategy can greatly improve localisation efficiency and localisation accuracy.