| With the development of science and technology, now the working environment of engineering structure and mechanical equipment have changed a lot,which is not only worked in alternating load, but also need to work under high temperature conditions. In many cases, when the engineering structure and mechanical equipment experiene high temperature creep and fatigue, Crack propagation is influeneed by material damage, causing leaking out or explosion.All thesestatus will lead to eeonomie losses and people casualty. So we study the material high-temperature creep reliability, it has a important significance when it is used to evaluate the safety of the structure.This paper described research that high temperature alloy steel GH36 has been tested on high temperature enduring strength and fracture experiments, and draw three factories GH36 steel high temperature enduring strength and fracture time data. And then these randomly distributed material properties of the measured datas are regard as random variables, mathematical statistical methods are used to deal with it. This paper are mainly researched on items listed below:First of all, there is a brief narration about the development of high-temperature creep experiments, and then creep test enginery characteristics and working principle are discussed, and according to the requirements of appropriate testing machine and test pieces.Secondly, the distribution characteristics, parameters estimation and hypothesis testing of the commonly used statistical distribution functions, which includes the normal distribution, lognormal distribution, the two-parameter and three-parameter Weibull are program by MATLAB, and the corresponding linear regression are charted.Thirdly, To handle the time data of GH36 steel high temperature enduring strength and fracture with the distribution characteristics, parameters estimation and hypothesis testing of the commonly used statistical distribution functions, which includes the normal distribution, lognormal distribution, the two-parameter and three-parameter Weibull are program by MATLAB, and draw the best distribution types of each data.Finally, in order to get to meet the credibility of the results of statistical analysis based on the best distribution types of each data, and draw the three factories confidence interval of high temperature enduring strength and fracture, and through the F inspection and t-test contrast analysis we can known that, the dispersion of GH36 steel high-temperature creep data of QiGang production is big than others, and there is no significant difference between erchong and 43chang about theirvariance, but erchong's mean larger. |