Automated quality inspection of die cast and hot-pressed metal components
|Coordinator||Swerea SWECAST AB|
|Funding from Vinnova||SEK 1 390 909|
|Project duration||November 2018 - November 2020|
|Venture||The strategic innovation programme for Metallic material|
Purpose and goal
The project aims to develop smart, flexible and user friendly machine vision systems that have the capacity to detect surface defects such as porosities on die cast and hot-pressed metal components. The software of the systems developed in this project are based on machine learning algorithms, or, as these algorithms are popularly known, artificial intelligence. The goal is to develop solutions that are close to a state where they can be implemented in die casting foundries and at hot-pressed metal manufacturers.
Expected results and effects
It is expected that at least one defect type on at least one component from each manufacturer will become detectable under various conditions (varying light intensity, component angle etc.) by at least one of the camera systems developed in the project. A short term effect is that casting foundries and hot pressing companies understand the potential of AI systems for quality inspection. In the long term, smart cameras will constitute parts of powerful, automatic production control systems that significantly decrease consumption of energy and resources, and environmental impact.
Planned approach and implementation
The project is carried out in two main steps. In the first step, a test bed is set up at Swerea SWECAST, where the two vision system companies can gather image data and test their systems under production-like conditions. A third system that is based on cheap web cameras and open source code is also developed in this project; this is done by RISE SICS, under some guidance from University of Skövde, who are currently running another project where they are developing a similar system aimed at other businesses. In the second step, tests are carried out on the factory floor.