Fracture Statistics Dominated by Hot-Cutting Defect and Deviation from Weibull Distribution

Chao Zeng, Chunqing Wang, Yanhong Tian, Chunjin Hang, Wei Liu,Rong An

(State Key Lab of Advanced Welding and Joining, Harbin Institute of Technology, Harbin 150001, China)



Abstract:

This paper characterizes the rarely noticeable hot-cutting defect and statistically models the fracture governed by this new type defect. The morphology of the defect on fired ceramic is examined and quantitatively featured through comparing the fracture strength governed by intrinsic defect and hot-cutting defect. Weibull distribution is utilized to fit the observed strength data and chi-square goodness-of-fit test is conducted to analyze the deviation. Kernel density estimation is introduced to explore the underlying strength distribution dominated by hot-cutting defect. Based on the shape information from kernel density estimating, gamma and lognormal distribution are compared and the hot-cutting defect governing fracture statistics is finally confirmed by chi-square test. Results show that the newly-defined hot-cutting defect is more dangerous than the intrinsic defect and the priori Weibull distribution fails to describe the fracture statistic governed by the edge defect while the lognormal with a slightly right skewed shape fits it very well.

Key words:  hot-cutting defect  deviation from Weibull  Kernel density estimation  lognormal distribution

DOI：10.11916/j.issn.1005-9113.2015.02.006

Clc Number:TB321

Fund: