Title: In-process sensing and machine learning for additive manufacturing
Start: October 2021
Student: Ahmet Koca
Supervisors: Richard Leach, Mingyu Liu
In this project, a sensing technique will be developed to enhance the next generation of metal additive manufacturing (AM). An essential part of future AM is quality control, one aspect of which is the measurement of layer topography. One important measurand is considered to be the micro/nano-scale topography (e.g. see Fig. 1, a circular dendritic microstructure shown in a nickel superalloy 625 AM part), which potentially affects the functionality and performance of the final parts.
It is well known that surface defects can cause the failure of AM parts and lead to a negative impact on the manufacturing cost and the environment. To prevent part failure, it is necessary to detect abnormalities, particularly through real-time surface measurement. AM surface measurement is mainly performed after manufacturing using off-line measurement methods, and sometimes surface measurement is performed in-situ, layer by layer (e.g. using fringe projection methods) but slows down the manufacturing process. This project aims to monitor the fine-scale surface topography (micro/nano-scale structures) of metal AM parts in-process without compromising the manufacturing speed. If any critical defects are detected, the manufacturing process can be stopped to avoid further waste of material and energy. A novel sensor system with high-spatial resolution measurement capability (a scattering technique) will be developed and deep learning algorithms will be developed to mine surface information from the sensor data. The proposed defect detection method, targeted to be integrated into commercial AM machines, can potentially enhance the next-generation intelligent metal AM process to achieve near zero-defect AM parts.
Fig. 1 A circular dendritic microstructure 
 Fox J, Allen A, Mullany B, Morse E, Isaacs R, Lata M, Sood A, Evans C 2021 Surface topography process signatures in nickel superalloy 625 additive manufacturing Proc. euspen Advancing Precision in Additive Manufacturing, Sep.