An effective calibration method based on fuzzy network for enhancing the accuracy of inverse finite element method

Abstract

The inverse finite element method (iFEM) is considered as a promising and effective method for shape sensing of structural deformation. However, strain measurement errors and structural installation errors arising from the service conditions seriously affect the reconstructed accuracy. To reduce the system errors, a novel fuzzy-network calibration method for small sample problem is proposed. The error correction model of the displacement-node degrees of freedom (DOF) is established based on the framework of iFEM to reduce deformation errors of the whole structure measured points, and the original correction of DOF is obtained by the Bayesian algorithm. In addition, a non-uniform rational B spline (NURBS) curve is built to extend the sample data set as a result of the lack of the trained original correction. Furthermore, a fuzzy self-construction network (FSCN) is constructed using the extend samples, which can obtain the relation between strain measurement data and the DOF errors. Finally, the power of the method is illustrated through a wing integrated antenna structure model subjected to bending, twist and bending-twist loading, respectively, and the results show that the presented method can significantly improve the accuracy of the reconstruction displacement.