基于Bootstrap与SVM集成的可靠性评估方法

Reliability assessment method based on Bootstrap and SVM

针对传统的可靠性评估方法应用于数控装备时遇到的非线性与准确度低等问题,提出一种适用于小子样情况的数控装备可靠性评估新方法。综合采用Bootstrap和支持向量机等方法建立了小子样条件下的可靠性评估模型,给出了可靠性评估全过程的统一建模语言序列图,并进行了实例仿真分析。结果表明,该方法能够较好地解决小子样条件下可靠性评估遇到的非线性和评估准确度低等问题,并能够给出可靠性评估结果的可信性度量。

To tackle characteristics of nonlinear and low accuracy for traditional reliability assessment,a novel machine tool reliability assessment method based on Bootstrap and Support Vector Machines（SVM） was proposed.The reliability assessment models for point estimation and confidence interval were respectively set up under the condition of the small sample.A sequence diagram of unified modeling language was adopted to describe the entire process of machine tool reliability assessment.A case was implemented by simulation analysis to validate the method.It suggested that the nonlinear and low accuracy encountered in classic machine tool reliability assessment might be solved by using the proposed method when only small sample was obtained.