Real-Time Site Specific Assessment of Cement Mortar Using a Solitary Wave Based Deep Learning

This paper proposes a real-time non-destructive evaluation technique for site-specific assessment of mortar using highly nonlinear solitary waves (HNSWs). This is achieved by studying a deep learning algorithm based on the convolution neural network (CNN) using HNSWs as input data. Of particular interest is to examine the sensitivity of the pre-trained CNN architectures on hydration process of cement mortar. To collect HNSW datasets for training, validation, and testing of the deep learning algorithm, mortar cube samples with various curing ages are prepared and HNSW datasets are generated from a granular crystal sensor. The pre-trained CNN architectures showed excellent performance for identifying the strength development of mortar based on curing age.