An Inverse Vibration Problem Solved by an Artificial Neural Network

Abstract

Inverse problems in vibration is a process of determining parameters based on numerical analysis from a comparison between measured vibration data and its predicted values provided by a mathematical model. In this work the dis- placement data have been chosen in order to identify the stiffness matrix which will cause a changing in the time-history of the system displacement. This is an inverse problem, since the stiffness matrix evaluation is obtained through the de- termination of the modified stiffness coefficients. In this work, the artificial neural network technique is applied to the inverse vibration problem where the goal is to estimate the unknown time-dependent stiffness coefficients simultaneously in a two degree-of-freedom structure, using a Multilayer Perceptron Neural Network model. Numerical experiments have been carried out with synthetic experimental data considering a noise level of 1%. Good recoveries have been achieved with this methodology.