Some Reflections upon the Design by Experiment of Mechanical Structures

The design of a structural system or the control of an existing one need the knowledge of a large number of parameters that justify an experimental approach, even at this time of great development of numerical simulation tools. In the past, the main task of experimental stress analysis was to determine or verify theoretical or numerical calculations of states of stress in geometrically complex structures. Today, however, due to the power of modern Finite Elements Methods, it is not so important to use experiments to control the numerical solutions of forward problems- i.e. problems which start from the knowledge of the forces and of the structural characteristics of the body (i.e. “the system parameters”), to determine the state of displacements or stress/strain in critical points, because the real uncertainties are more often connected with the actual loadings, temperature distributions, constraints and/or the system parameters in terms of stiffness, boundary conditions, geometry and material properties. In order to consider the structure in this more general perspective, it is better to speak of experimental analysis (more than experimental stress analysis), because the concluding task of the experiment in structures engineering is to identify the loadings and/or the system parameters by measurements in order to assess the system’s reliability. But every experimental analysis starting from the knowledge of some measured entities in order to deduce the properties of the applied forces and structure constitutes an inverse problem. In other words, inverse problems are posed either for cause identification or for parameter identification or also for a mix of both.