Detection of Structural Damage in Medium Density Fiberboard Panels using Neural Network Method

Abstract

This research assessed the feasibility of using a neural network to detect low levels of damage in small samples of medium density fiberboard (MDF). The neural network was a three-layer back-propagation network. The undamaged stress wave frequency spectrum patterns were trained by the neural network. The trained patterns were then compared to stress waves patterns taken from MDF samples loaded to various percentages of their estimated failure load. In this experiment, if an application load is below the proportional limit, a small change in wave patterns occurs. The neural network has the unique ability to train data to recognize spectral patterns and has been used with success for the detection of structural damage.