|
Structure Reliability Calculation Method Based on Finite Element-Artificial Neural Network-Monte-Carlo
XU Yong-jiang, XING-Bing, WU Jin-liang
2008, 27(2):
188-190,216.
According to reliability theory,based on Finite Element, Artificial Neural Network and Monte-Carlo (F-A-M),a new structure calculation method is developed. First,finite element analysis model is built up,inputting values of the structural parameters and loading at random,and then response of the structure is recorded.The number of the record is limited,but it is enough to train the Artificial Neural Network (ANN). After ANN being trained with the training samples,a highly nonlinear mapping relationship between the efficacy and response of the structure is gained. Making use of the generalization capability of ANN,sufficient structural response value is developed at random. Then the reliability of the structure is calculated by the Monte-Carlo method. This approach which not only makes full use of the merit of ANN and Monte-Carlo,also makes the calculation more effective, provides a fresh method for the reliability of structure.
Related Articles |
Metrics
|