基于SVR的万向节机构可靠性及灵敏度分析

Analyzing the reliability and sensitivity of universal joint mechanism using support vector regression

针对结构隐式极限状态函数的可靠性分析问题,将支持向量机回归理论应用到万向节机构可靠性分析中,提出了一种基于支持向量机回归的万向节机构可靠性分析方法.通过支持向量机回归与一次二阶矩法相结合,利用支持向量机的小样本学习能力,将隐式极限状态函数近似为显式极限状态函数.运用蒙特卡罗法计算出万向节机构的可靠性指标,然后由高精度的显式极限状态方程进行各随机变量的灵敏度分析.最后以某剪切机中万向节机构为对象,进行了机构的可靠性及灵敏度分析.结果表明:该方法具有高精度和高效率的优点,并对其他机械结构可靠性分析具有一定的参考意义.

For the reliability analysis of structure with implicit limit state function,the support vector regression theory is applied to reliability analysis of universal joint mechanism,and a method for reliability analysis of universal joint mechanism is presented through support vector regression. Using the small sample learning capacity of support vector machine,implicit limit state function is approximated to explicit limit state function by combining the support vector regression with first order reliability method. The reliability index of universal joint mechanism is calculated using Monte Carlo methods,and then explicit limit state function could be employed to analyze the sensitivity of each parameter. Finally,the reliability and sensitivity analysis of universal joint mechanism in a shearing machine are carried out using the presented method. The results show that the presented method has two virtues of high accuracy and efficiency,and it has a certain useful value for other mechanical structure reliability analysis.