Sparse strain sensing-based bending-torsion coupled deformation measurement in discontinuous stiffness open-hole CFRP panels using iFEM

Abstract Carbon fiber-reinforced polymer stiffened panels with cutouts are widely used in lightweight aerospace and marine structures, yet the presence of openings leads to discontinuous stiffness and complex deformation mechanisms, particularly under bending-torsion coupled loads. Although the inverse finite element method (iFEM) offers a promising approach for reconstructing full-field deformations without requiring material properties or external loads, its practical application is hindered by the dependency on dense strain sensor networks. To overcome this limitation, this paper introduces a sparse strain sensing framework that synergistically integrates iFEM with a generalized regression neural network (GRNN). The GRNN model is designed to reconstruct the full surface strain field effectively reducing the number of required sensors by adopting a single-sided unidirectional (Y-direction) sensing layout—decreasing from 300 measurement points in conventional dense deployment to 40—while maintaining high reconstruction accuracy exceeding 90% under complex loading conditions. Experimental validation demonstrated that the GRNN-augmented iFEM algorithm achieved reconstruction accuracies of 91.65%, 90.14%, 93.03% and 90.26% under four complex loading scenarios, including combined bending and torsion. This study not only resolves the critical issue of sensor redundancy in conventional iFEM but also provides a robust and efficient strategy for deformation monitoring of structures with discontinuous stiffness and inherent anisotropy.