AI and machine learning are making waves to improve industrial and business processes is rightly said. Now, a newly developed machine learning-based approach can predict the elements and manufacturing process required to create an aluminum alloy with specific wanted mechanical properties. The technique is published in the journal Science and Technology of Advanced Materials.
Physically, aluminum alloys are lightweight, energy-saving substances that predominantly have aluminum. It also contains other elements such as silicon, manganese, magnesium, zinc, and copper. This combination of elements and manufacturing processes determines how much stress the alloy can withstand. For example, aluminum alloys categorized as 5000 series contain magnesium and other elements. They find use as welding material in cars, buildings, and pressurized vessels. Aluminum alloys categorized as 7000 series contain zinc, and usually copper and magnesium, and are mostly used in bicycle frames.
Meanwhile, to experiment various combinations of materials and manufacturing processes to obtain aluminum alloys is time-intensive and expensive.
Past Aluminum alloy database knowledge fed into Machine Learning approach
To overcome this, researchers at the National Institute for Material Science, Japan and Toyota Motor Corporation have developed a materials informatics technique. The approach feeds available data from aluminum alloy repository into a machine learning model.
Such information provides useful knowledge to the model to understand the relationship between mechanical properties of alloys, different materials they are composed of, and the type of heat applied during manufacturing. Once the model receives enough data, it can predict what is required to fabricate a new alloy with desired mechanical properties. This is executed without the need of input or supervision from a human.
For example, the model found, aluminum alloys of 5000 series are highly resistant to stress, and deformation of the alloy is possible by increasing the magnesium and manganese content and decreasing the aluminum content.