An Structural System Reliability Calculation Method Based on Sample’s Weighting -Artificial Neural Network
This paper presents a new structural system reliability calculation method with artificial neural network adopt by the sample’s weighting and the artificial neural network, which can greatly reduce calculation work. Firstly, according to the reliability theory of total probability, with the concept of sample’s weighting coefficient, the minimum sample size can be obtained according to the numerical characteristic values of random variables and the minor sample t-distribution estimation under a certain expected value. And then, the optimize artificial neural network is set up with the limited training samples based on the analysis result of sample’s weighting coefficient, which has a highly nonlinear mapping relationship between the efficacy and response of the structural system reliability Analysis. By making use of the generalization capability of optimize artificial neural network, sufficient system response value is gained at random. Meanwhile, the weight coefficients of the random sample combinations are determined using the Bayes formula, and different sample combinations are taken as the input for system analysis. According to one-to-one mapping of system by artificial neural network between the input sample combination and the output coefficient, the reliability index of system can be calculated. At last the method provides a new attempt for S structural system reliability analysis and prove to be feasible and effective for practical experience in complex system, which not only makes the artificial neural network calculation more effective based on sample’s weighting coefficient, also makes full use of the merit of artificial neural network instead of the performance function.
Chaohe Chen, Yong Huang and Guangfan Li
W. M. Chen and Z. G. Yang, "An Structural System Reliability Calculation Method Based on Sample’s Weighting -Artificial Neural Network", Advanced Materials Research, Vols. 243-249, pp. 666-671, 2011