An online prediction tool for fatigue strength of steel alloys
- 技術優勢
- Novel application of data-driven approaches in materials science Fast, accurate, and easy-to-use data-driven models
- 詳細技術說明
- An online tool that uses advanced data-driven ensemble predictive models to quickly predict fatigue strength for a given steel based on its composition and processing. #software #analysis #application #manufacturing #materials #metals
- *Abstract
-
BACKGROUND
Fatigue strength is one of the most important mechanical properties of steel. However, fatigue testing is associated with high cost and time for fatigue testing, and the lack of it can result in potentially disastrous consequences. Understanding the properties of specific steel alloys offers important information required for design and failure analysis of mechanical components. Fatigue is estimated to account for over 90% of all mechanical failures of structural components, and hence, fatigue strength prediction is of critical importance.
ABSTRACT
Northwestern researchers have developed advanced data-driven ensemble predictive models for the purpose of predicting fatigue strength with extremely high cross-validated accuracy of >98%. These predictive models for fatigue strength of a given steel alloy are represented by its composition and processing information, and are essentially fast forward models for processing-structure-property performance (PSPP) relationships. The predictive models deployed in the tool are a result of the application of supervised learning techniques on an experimental fatigue dataset from Japan National Institute of Materials Science MatNavi database. A comparison of 40 supervised modeling configurations on the NIMS steel fatigue dataset, including ensemble modeling techniques, makes this the most accurate predictive model over prior works on the same data. They have deployed these models in a user-friendly online web-tool, which can make very fast predictions of fatigue strength for a given steel. Such a tool with fast and accurate models is expected to be a very useful resource for materials science researchers and practitioners who are looking to accelerate new materials discovery and design for steel. The web tool is available at: http://info.eecs.northwestern.edu/SteelFatigueStrengthPredictor
- *Inventors
- Alok Choudhary*AnkitAgrawal
- *Publications
- Agrawal A and Choudhary A (2016). A fatigue strength predictor for steels using ensemble data mining. CIKM 2016 Proceedings of the 2016 ACM
- 國家/地區
- 美國
