一种基于威布尔分布的金属疲劳试验异常数据判别方法

金属疲劳试验的数据符合威布尔分布,是非对称的统计分布,而现有的异常数据判别方法均基于对称性分布(正态分布、t分布)。威布尔分布与对称性分布(正态分布、t分布)主要差异体现在低寿命区和高寿命区,这正是异常数据出现的区域。基于威布尔分布推导了一种异常数据判别方法,推导过程中进行了合理地工程处理。经比较,该方法相对于基于对称性分布的判别方法更可靠。

An Outlier Detection Method of Metal Fatigue Tests Based on Weibull Distribution

Data of metal fatigue test is described with Weibull distribution which is asymmetrical. But general methods for outlier detection in metal fatigue tests are based on symmetrically statistical distribution. The difference between Weibull distribution and symmetrically statistical distribution is often found in the areas of small life and high life where outlier appears. A method of outlier detection is derived based on Weibull distribution in this paper. And some reasonable measures are used in the derivation. By comparison, the method in this paper is more reliable than those based on asymmetrical distribution.