Progetto: Research project

Dettagli progetto


An objective of the research program of the Unit is devoted to structural identification problems on the basis of response
measurements under both static and dynamic loading. Since structural identification problems can be defined as identification of
structural parameter variations, damage diagnosis problems will also be treated, the damage being modeled as decay of suitable
structural parameter properties.
Initially, the research program will be devoted, with particular attention, to set static identification procedures since they allow
highlighting the general properties of the problem, leading to explicit solutions which are of great importance in the engineering
practice. On the basis of the results achieved by the development of static identification procedures, in the subsequent research
phase the extension to the case of dynamic loading will be studied.
The identification procedures, known in the literature, based on the execution of static tests, aim at providing the values of those
parameters influencing the stiffness matrix by means of definition of an error function as the difference between the analytical model
and the real structure responses. Such procedures are usually formulated as constrained, non convex, optimization problems of the
response error function. In particular, they differ from each other for the definition of the error function to be minimized and for the
adopted solution techniques. The procedures based on optimization problems, although able to treat complex structural systems, do
not lead to explicit formulations.
In the research program, here proposed, a procedure currently under study will be adopted for identification of structural
parameters on the basis of displacement measurements by static tests based on a variational formulation of the direct analysis
problem of continuous bodies. In particular, aim of the research is obtaining a consistent and robust formulation for the solution of
large classes of identification problems. The proposed procedure, rather than optimizing an error function measuring the difference
between the model and the real displacements, requires the minimization of the well known Hu-Washizu functional suitably modified.
The modification relies on treating the displacement measurements, given by loading tests, as imposed displacements of fictitious
constraints applied to the measurement points. The formulation of the modified Hu-Washizu principle allows to obtain the values of
those parameters selected for identification by imposing that the reactions of the fictitious constraints are null (compatibility loading
In those cases where the dependence of the stiffness matrix on the parameters to be identified does not allow an explicit identification
solution, the "virtual distortion method" will be adopted. According to the virtual distortion method, variations or imperfections of
structural parameters, or possible damages, can be considered as inelastic distortions imposed on the undamaged structure leading
to an explicit formulation of the inverse problem and an easier solution. In such a way, the known stiffness matrix of a reference
structure can be adopted where the non linear dependence of the original stiffness matrix on the parameters to be identified is
The research program, will be developed in the first semester of the first year as follows:
1) Identification of structural parameters of simple systems, such as straight beams, aiming at obtaining explicit solutions as
functions of displacements measured by loading tests.
2) Applications to cases of damaged beams in presence of both concentrated and distributed damages and study of general
properties of the damage identification problems.
3) Formulation of criteria for an "a priori" identification of the number of damages.
4) Extension to cases of continuous multispan beams and to cases of arch structures.
5) Formulation of sufficient conditions to guara
Data di inizio/fine effettiva1/1/06 → …


Esplora i temi di ricerca toccati da questo progetto. Queste etichette sono generate sulla base dei riconoscimenti/sovvenzioni sottostanti. Insieme formano una fingerprint unica.