Data Analytics for Additive Manufacturing
Data collection in the manufacturing system becomes much easier with the advancement of scanning and sensor techniques. However, this also brings obstacles for the interpretation of system dynamics and decision making given such a large dataset. Hence, an intelligent analytical method is necessary to enable a smart manufacturing system. The past decade has witnessed the success of deep learning especially the convolution neural network on a wide variety of tasks. However, rare works have touched the area of manufacturing even though various formats of data especially 3D data are produced in AM. To integrate the state-of-art deep learning methods into the manufacturing area, production problems in design and fabrication phases would be defined, thereafter the patterns and features of the data will be extracted through leveraging statistical properties such as stationarity and compositionality.
