基于LSSVM的重力坝地震稳定易损性分析

提出了一种基于最小二乘支持向量机的易损性分析方法，将有限元数值计算的结果作为学习样本，建立最小二乘支持向量机模型，进而进行Monte Carlo仿真，获得结构的易损性曲线，在保证计算精度的同时，提高了计算效率。采用该方法对某重力坝挡水坝段进行了动力稳定易损性分析，并对主要参数进行了敏感性分析。结果表明，该坝段能够保证设计地震作用下的动力抗滑稳定性，且有较大的安全裕度。

Fragility analysis for seismic stability of a gravity dam based on least square support vector machine

A fragility analysis methodology based on least squares support vector machine was presented. Using the results of finite element analysis as learning samples, the least squares support vector machine could be established, and integrating the learned model with Monte Carlo simulation, the fragility curve could be obtained with high accuracy and efficiency. Using the proposed methodology, the anti-sliding stability fragility of non-overflow section of a concrete dam was analyzed and the sensitivity of the key parameter was studied. The results show that the dam could ensure the dynamic stability under design earthquake and possess high margin in safety.