Modelagem estocástica e validação experimental de uma viga Euler-Bernoulli

Abstract

This paper presents a stochastic modeling of the Euler-Bernoulli beam and, here, it was used a technique known as Stochastic Finite Elements, this technique has been widely used in the last years due to large evolution of the capacity of the processors since it has a high computational cost. Among the design variables, the possible uncertain parameters are the modulus of elasticity, specific mass and the proportional damping coefficients, α and β. By the sensibility analysis, it was possible to verify which parameter has the biggest influence on the system response. Optimization techniques were also used to identify the uncertain parameters. The uncertainties are modeled as homogeneous Gaussian stochastic fields and discretized according to the spectral method by using Karhunen-Loève expansions. The Monte Carlo Simulation method combined with the Latin Hypercube Sampling is used as stochastic solver. Finally, an experiment was performed with a vertical beam, to obtain the parameters by means of the inverse problem, using optimization techniques in this identification process, and then, to use them to obtain the envelope and check how much the experimental result is inside it.