摘要
在可靠度计算中,任意对立随机变量往往为不同概率分布,不利于可靠度计算。通常采用正态转换法将其转换成具有统一形式的正态分布,以便进行可靠度计算。但是,正态转换并不能保证原有随机变量概率分布特性,从而一定程度上造成可靠度计算误差。针对上述问题,本文提出在可靠度计算中利用最大熵法将任意随机变量概率密度函数转换成近似精度更高且具有统一标准形式的最大熵概率密度函数(MEPDF),然后提出对应的一阶可靠度计算方法进行可靠度求解,最后通过实例分析证明该方法的有效性。
As for reliability computation, arbitrary independent random variables always follow with different types of probability distribution, which bring in difficulties for computing reliability. Generally, arbitrary independent random variables are always transformed into the unified form of normal variables by means of normal transformation for the convenience of reliability analysis. But normal transformation can not definitely ensure the probability distribution feature of the original probability density function （PDF） and to some degree affects the accuracy of the computed reliability results. Aiming at this prob- lem, the paper transforms the PDF of arbitrary independent random variables into the the unified form of the maximum entropy PDF （MEPDF） by means of Maximum entropy method,which are more accurate and then exploits the corresponding the first and second order moment approach. The proposed method has been validated by practical analyses.
出处
《计算力学学报》
CAS
CSCD
北大核心
2015年第1期41-47,共7页
Chinese Journal of Computational Mechanics
基金
国家自然科学基金青年科学基金(51305143
51205141
51305142
51405168)
华侨大学高层次人才科研启动(12BS204)资助项目
关键词
最大熵
可靠度
概率密度函数
maximum entropy
reliability
probability density function
