The metal laser powder bed fusion market has an amazing projected compound annual growth rate of 22.5% to 28% over the next 5 years, which is driven by the aerospace and biomedical industries. However, part certification remains an arduous process due to the numerous rouge defects and anomalies which can occur. In situ sensors have promised and been shown to be capable of spotting these defects; however, this comes at the cost of significant portions of data which is unreadable to the average technician for quality control purposes. Furthermore, once a potential defect is spotted, the ability to heal is non-existent in the industry. A proposed solution incorporating intelligent prediction, machine learning, and adaptive machine control is proposed as a reasonable solution to the problem.
Benefit: It is expected that this technology will greatly increase the number of successful prints or at a minimum alert technicians to early failures to allow for critical restarting steps and prevention of waste. Reliable detection in near real time with the ability for automated machine adaption would be completely unique in the industry and allow for critical supply chain improvements. The initial targets will be hard to get parts and the aerospace industry as this will enable the ability to build low volume parts with confidence and avoid costly failures. The initial development will be on LFS' in house manufacturing printers which will quickly serve as a vetting process and provide bottom line improvements. It is expected that this technology could be portable to other OEMs for licensing once fully developed.
Keywords: Defect Detection, Defect Detection, laser powder bed fusion, predictive simulation, beam shaping, additive manufacturing, defect repair, MACHINE CONTROL
