Design and benchmarking of a robust strain-based 3D shape sensing system

Shape sensing can provide insight into the structural health and operating conditions of slender engineered lifting and control surfaces in aerospace and maritime applications. Shape sensing often relies upon digital image correlation, inertial measurement, or inverse finite element methods, which can be impractical in applications involving real-time reconstructions, static and dynamic deformations, or dynamic mode shape identification, especially outside of controlled laboratory environments. This work describes ongoing efforts to design, build, and validate a low-cost and robust tool for real-time shape sensing. The sensor consists of a simple aluminum spar, instrumented with strategically placed strain gauges. A kinematic model is used to reconstruct bi-axial bending and torsional displacements along the spar. The model is validated against FEM simulations and canonical analytical solutions. A prototype of the spar is benchmarked using a motion capture system. The pre-calibrated errors are correlated with the direction of bending, permitting a directionally compensative calibration scheme. After calibration of the spar and kinematic model, validation errors are 2.18% in bending magnitude and 0.97° in bending direction. This work addresses the emerging need for new low-cost sensors for structural health monitoring in environments where other sensing methods do not perform well. •A kinematic model is used to reconstruct displacements using strain measurements.•The model is validated against FEM and canonical analytical solutions.•A prototype of the spar is benchmarked against motion capture data.•A directionally compensative calibration is applied to the validation dataset.•Validation errors are 2.18% in bending magnitude and 0.97° in bending direction.