[19p-231B-1~9] 4.7 Laser Material Processing and Manipulation

▲ [19p-231B-9] [INVITED] Automated data acquisition and deep learning in a laser processing

〇Yohei Kobayashi¹, Shuntaro Tani¹ (1.University of Tokyo)

Keywords:laser processing, deep learning

Productivity growth by use of internet of things (IoT) is highly demanded in order to realize a smart society. Laser machining is a promising candidate of the key technology for mass customization e.g. high-mix low volume production, because it could realize a cutting, drilling, welding, and even coating, only by changing the parameters of the laser and the scanner. In order to realize this kind of system, we have to know how the parameter change of the lasers or scanners influence on the laser processing. We could realize a simulator of the laser machining if we knew all the results of all the parameter sets. Back in reality, a knowledge of the laser processing is acquired by hand and accumulated in mind, which is called as “experience.” Under this condition, the amount of the data is hard to exceed some critical volume to adopt the data for the big-data analysis including a deep learning.
We will discuss some automated data acquisition systems and how a deep learning technology is combined to the laser processing.