中文摘要：

结构可靠性设计是现代结构设计中能够考虑不确定性因素影响的重要方法之一。针对结构可靠性设计中较为复杂的隐式多概率约束条件问题,本文提出了一种有效的代理模型方法。该方法基于多输入多输出支持向量机方法,并引入分类错误概念,可以同时对多个概率约束条件进行建模,同时对概率约束条件进行了相对保守的结构可靠性分析,得出非零失效概率的两者差值在1%-10%。将该可靠性评估方法嵌入可靠性设计优化框架内,形成了偏于安全的结构可靠性设计方法。文中还提出了一种新偏导数计算方法,用于可靠性优化设计。最后将所提方法应用于减速器可靠性分析和优化设计中,验证了其计算精度和效率。

英文摘要：

Reliability based structural design is one of the important design methods, which can take the effects of uncertain factors into account in modern engineering design. In order to deal with the complex issues of multiple probabilistic constraints in reliability based structural optimization design, an effective surrogate model is proposed in this paper. It is based on multi-input multi-output support vector machine and the consideration of classification error. The proposed surrogate model can deal with multiple probabilistic constraints simultaneously. In addition, reliability evaluations are inclined to be conservative compared with standard support vector machine. By integrating the reliability analysis method into the framework of reliability based design, it forms a conservative reliability based structural optimization design method. A new method is presented for the calculation of partial derivative, which is required in reliability, based structural design. The accuracy and efficiency of the proposed method is illustrated by the reliability analysis and reliability based design of a speed reducer.