基于非参数回归模型轨道风机叶轮的可靠性分析方法  

Reliability Analysis Method of Fan Impeller Based on Nonparametric Regression Model

作　　者：阳吉初 屈小章[1,2] 翟方志 

机构地区：[1]株洲联诚集团有限责任公司,湖南株洲412001 [2]湖南大学汽车车身先进设计制造国家重点实验室,湖南长沙410082

出　　处：《机械设计与制造》2016年第7期100-103,共4页Machinery Design & Manufacture

基　　金：装备预先研究项目(62501036012)

摘　　要：在叶轮设计过程中,其设计变量和实际工艺往往具有一定不确定的差异,在传统设计由于忽略了不确定的影响,导致实际生产产品质量的不稳定性。为了提高轨道风机叶轮设计的可靠性,通过非参数回归模型和6σ,构造了一种适合于轨道风机叶轮设计的可靠性分析方法。详细分析了叶轮结构设计尺寸、材料和载荷的不确定性问题分析,并搭建了叶轮结构可靠性分析方法流程。通过对非参数回归模型的误差分析,得出此拟合模型具有良好的精度,在叶轮强度和振动分析中得出了较好可靠性结果。且在和多项式拟合和神经网络模型对比中,非参数回归模型具有更好的稳健性,因此在叶轮设计工程中具有较好的实用价值。Design variables and actual process have uncertainties during the design process of impeller. Conventional design, however, neglects uncertainties, which always lead to the instability of actual production quality. In order to further improve the reliability of orbit fan impeller design, through the nonparametric regression model and six sigma Latin hypercube algorithm, a robustness analysis method has been constructed on the nanparametric regression model and six sigma, which has a smaller model deviation and good robustness, and the impeller structure reliability catalysis model and the process has been constructed by uncertainty problem analysis of the impeller structure design, dimension, material and load. It is concluded that the fitting model has good accuracy and strength and vibration analysis in the impeller of the better robustness result has been obtained by analyzing the error of the nonparametric regression model. And in contrast to the polynomial fitting and neural network model, nonparametric regression model has better robustness; thereforre the impeller design has good practical value in engineering.

关 键 词：叶轮 6 非参数回归模型 稳健性 可靠性分析 

分 类 号：TH16[机械工程—机械制造及自动化] TH313