Researchers at the Department of Energy’s Oak Ridge National Laboratory (ORNL) have developed a method to improve flaw detection in 3D-printed metal parts. This type of additive manufacturing is used in industries such as energy, aerospace, nuclear, and defense to create highly specialized parts with complex shapes. However, it is challenging to thoroughly inspect and detect flaws in these printed parts.
The ORNL researchers combined inspection of the printed part after it is built with information collected from sensors during the printing process. They used machine learning algorithms to identify flaws in the product, which allows for accurate flaw detection as reliably as traditional evaluation methods.
The most common metal 3D-printing process, laser powder bed fusion, involves selectively melting metal powder with a high-energy laser. Flaws are expected in the material, but it is difficult to qualify and certify parts without a number value for flaw detection.
In collaboration with aerospace and defense company RTX, the researchers designed a part with realistic-looking flaws and monitored the 3D-printing process using cameras. They conducted quality inspections using CT scans. The data collected trained the machine learning algorithm to accurately recognize flaws.
This inspection framework reduces the need for CT imaging and analysis, which is time-consuming and expensive. It also allows for consistent quality analysis and opens up possibilities for mass-producing products like car parts. It provides certainty about flaw size, enabling more design freedom. The researchers plan to further train the algorithm to differentiate between types of irregularities and categorize flaws with multiple causes.
>>Join our Facebook Group be part of community. <<