A combined experimental and artificial neural networks study of distortion in 3D printed beam specimens

Abstract

In this work the distortion (warpage) of 3D printed thin beam specimens was measured as a result of their build orientation and printing position on the platform. The exhibited deflections along the specimens’ length were measured using both a Coordinate Measuring Machine (CMM) and a 3D Scanner. Subsequently, the experimental data were used as input to an analysis aiming at predicting the measured deflections. The analysis involved two Artificial Neural Network (ANN) models which had been trained to predict the specimens’ distortion. The predicted deflections by the two ANN models were compared to the corresponding experimental ones and conclusions regarding their forecasting ability were drawn.