Damage Prediction and Estimation in Structural Mechanics Based on Data Mining
Damage in a material includes localized softening or cracks in a structural component due to high operational loads, or the presence of flaws in a structure due to various manufacturing processes. Methods that identify the presence, the location and the severity of damage in the structure are useful for non-destructive evaluation procedures that are typically employed in agile manufacturing and rapid prototyping systems. The current state-of-the art techniques for these inverse problems are computationally intensive or ill conditioned when insufficient data exists. Early work by a number of researchers has shown that data mining techniques can provide a potential solution to this problem. In this paper, they investigate the use of data mining techniques for predicting failure in a variety of 2D and 3D structures using artificial neural networks (ANNs) and decision trees. This work shows that if the correct features are chosen to build the model, and the model is trained on an adequate amount of data, the model can then correctly classify the failure event as well as predict location and severity of the damage in these structures.
- Research Organization:
- Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
- Sponsoring Organization:
- US Department of Energy (US)
- DOE Contract Number:
- W-7405-ENG-48
- OSTI ID:
- 15006200
- Report Number(s):
- UCRL-JC-144764; TRN: US200405%%336
- Resource Relation:
- Conference: 7th International Conference on Knowledge Discovery and Data Mining, San Francisco, CA (US), 08/26/2001--08/29/2001; Other Information: PBD: 23 Jul 2001
- Country of Publication:
- United States
- Language:
- English
Similar Records
Autonomous MHK Monitoring System for Intelligent Converter Health Management
Advanced Composite Wind Turbine Blade Design Based on Durability and Damage Tolerance