Speaker
Description
This contribution addresses the identification of concentrated static loads from strain measurements in geometrically nonlinear composite plates instrumented with embedded Fiber Bragg Grating (FBG) sensors. Since a rigorous investigation of this problem is missing in the literature, a mathematical framework is formulated from first principles.
A data-driven problem inversion algorithm is developed in which the external load is parametrized by a finite-dimensional vector, while the structural response is modeled as a nonlinear mapping from the load parameter space to the strain measurement space. This mapping is reconstructed from calibration data using interpolation and regression techniques. The unknown load parameters are then estimated by minimizing a distance functional in the output space.
The performance of the algorithm is first assessed through analytical and numerical case studies to identify the critical factors governing accuracy. Subsequently, the method is validated experimentally through the identification of static concentrated loads applied to a fully clamped carbon fiber plate.
Results indicate that data-driven load identification from strain measurements is feasible only under well-defined conditions. In particular, identification accuracy is primarily governed by sensor placement, calibration sampling density, and data quality. Sensors should be positioned outside the region of interest to reduce local response complexity and improve reconstruction robustness. Moreover, the experimental calibration data must be adequately processed to approximate the system response expected during real world deployment of the load identification algorithm.