Two shape-sensing algorithms, the calibration matrix (CM) method and the inverse Finite Element Method (iFEM), were compared on their ability to accurately reconstruct displacements, strains, and loads and on their computational efficiency. CM reconstructs deformation through a linear combination of known load cases using the sensor data measured for each of these known load cases and the sensor data measured for the actual load case. iFEM reconstructs deformation by minimizing a least-squares error functional based on the difference between the measured and numerical values for displacement and/or strain. In this study, CM is covered in detail to determine the applicability and practicality of the method. The CM results for several benchmark problems from the literature were compared to the iFEM results. In addition, a representative aerospace structure consisting of a twisted and tapered blade with a NACA 6412 cross-sectional profile was evaluated using quadratic hexahedral solid elements with reduced integration. Both methods assumed linear elastic material conditions and used discrete displacement sensors, strain sensors, or a combination of both to reconstruct the full displacement and strain fields. In our study, surface-mounted and distributed sensors throughout the volume of the structure were considered. This comparative study was performed to support the growing demand for load monitoring, specifically for applications where the sensor data is obtained from discrete and irregularly distributed points on the structure. In this study, the CM method was shown to achieve greater accuracy than iFEM. Averaged over all the load cases examined, the CM algorithm achieved average displacement and strain errors of less than 0.01%, whereas the iFEM algorithm had an average displacement error of 21% and an average strain error of 99%. In addition, CM also achieved equal or better computational efficiency than iFEM after initial set-up, with similar first solution times and faster repeat solution times by a factor of approximately 100, for hundreds to thousands of sensors.

Structural health monitoring (SHM) is a growing field of research, as it has the potential to simultaneously improve the reliability of structures and reduce their maintenance cost. SHM requires accurate stress and strain information, preferably for the entire structure. Unfortunately, it is often infeasible to instrument every part of the structure, making it necessary to estimate the stress and strain fields based on data from a limited number of sensors. One promising technique for making this estimate is the inverse finite element method (iFEM), which can be applied to any combination of geometry and loading conditions. In addition, it can also process several different types of sensor data. In this study, benchmark problems based on the MacNeal and Harder linear elastic problem set for FEM algorithms were extended to test the accuracy of iFEM algorithms. As the benchmarks use linear elastic materials, small displacements and strains, the iFEM implementation was also limited to these conditions. Accurate iFEM estimates can be obtained for the benchmark problems for which accurate FEM solutions can be obtained with solid elements, specifically 3-dimensional 20 node hexahedral elements with reduced integration (C3D20R), based on either displacement sensors, strain sensors, or both combined, and provided that a sufficient number of sensors is used. The iFEM algorithms generally produce more accurate estimates of displacements than of strains. The addition of Tikhonov regularization does not result in a significant increase in accuracy for either the displacement or strain distribution estimates and can even deteriorate the results in certain cases.