Machine Learning Prediction Of The Load Evolution In Three-point Bending Tests Of Marble

Machine learning in the form of artificial neural networks was applied to investigate whether specimen load evolution can be predicted as a function of acoustic emission (AE) signals in the case of three-point bending (TPB) marble specimens instrumented with piezoelectric sensors. The ultimate objective of this study is to develop a model that can quantify rock behavior under loading that can lead to rock-failure prediction in underground structures subjected to bending, such as roof failure in development or production openings.