Methods and guidelines for the choice of shell theories

This paper provides an overview of the modeling approaches adopted over the years to develop shell theories for composite structures. Furthermore, it presents a method to assess any structural theory concerning the accuracy and computational efficiency and trigger informed decisions on the structural theory to use for a given problem. This method exploits the synergies between the Carrera unified formulation and the axiomatic/asymptotic method. Typical outcomes are the best theory diagrams or the estimation of accuracy of a theory as compared to quasi-3D solutions. The proposed framework can be useful to provide guidelines on the construction of structural theories and can serve as a trainer for the deep learning of neural networks.