A 2D neural network approach for beam damage detection

The aim of this contribution is to propose a 2D Hopfield Neural Network for parameter identification and consequent damage detection in 2D processes. This method is applied for online damage detection in Euler-Bernoulli beams subjected to external forces. Damage is associated with significant change in the parameter values of the model. The 2D Hopfield Neural Network presented here is an extension of a 1D Hopfield Neural Network recently proposed in the literature. At each time instant, the network produces an estimate of the parameters at a certain beam point based on the previous estimates in the point itself and in its neighbours. The network tracks the change in the beam parameters and the results obtained are very satisfactory.