IDENTIFICATION OF ELASTIC PROPERTIES OF ORTHOTROPIC COMPOSITES BASED ON A GENETIC ALGORITHM

Abstract

An inverse approach combining numerical and experimental results with full-field displacement measurements, will allow the identification of all the in-plane elastic properties from experimental tests. The off-axis tensile test was chosen, and in order to calibrate the numerical method an 8-harness satin weave glass fiber reinforced phenolic composite was selected. Instead the use of a non-destructive technique an alternative hybrid approach is proposed to obtain the field displacement. In this way a reference displacement field is generated by finite element method considering the loading and boundary conditions used in tensile tests and the mechanical properties obtained from experiments. In this work effort is made to develop a technique for nondestructive characterization of laminated composites. Elastic properties of the composite are determined through this approach using an inverse technique based on finite element analysis and evolutionary algorithm supported by the experimental results. Mechanical properties of composite develop a link between the load applied in off axis tensile test and displacement field. The implementation of surrogate model ANN (Artificial Neural Network) eliminates the exhaustive calculation of elastic properties and displacement fields. Using UDM (Uniform Design Method) set of design points is generated. This will enables a uniform exploration of domain values that will be used in the development of ANN (Artificial Neural Network) approximation model. Then, Reference displacement field is generated by FEM (Finite Element Method) taking into account the geometry of specimen, considering experimental loading and boundary condition. Experimental output data acquired for displacement field is providing as a reference in the optimization problem. The design variables of the optimization problem are the independent elastic engineering constants E1 (Longitudinal), E2 (Transversal), G (Shear Modulus) and v Poisson’s ratio).