论文部分内容阅读
由于获得岩石力学参数的大样本的困难性，使得统计岩石力学参数概率分布函数成为一个难题。解决工程问题时，一般认为它们服从正态分布，但其精确度十分有限。对于一个具体工程，如何在有限的小样本条件下，推断岩石的力学参数概率分布函数一直是一个研究热点。针对这一问题，文章引入Ｂａｙｅｓ统计推断方法，建立了一套以Ｂａｙｅｓ最小熵原理和优度比较检验为基础的统计推断体系。其具体做法是：（１）在一定范围内选择岩石力学参数概率分布函数的概型；（２）在小样本下利用Ｂａｙｅｓ最小熵原理估计各种概型的分布参数；（３）采用优度检验，确定最优的岩石力学参数概率分布函数。文中结合实例对方法的有效性进行了分析并指出该方法是有效地解决上述难点的新途径。
Due to the difficulty of obtaining large samples of rock mechanics parameters, making the probability distribution function of rock mechanics parameters a challenge. When solving engineering problems, they are generally considered to be subject to normal distribution, but their accuracy is very limited. How to infer the probability distribution function of mechanical parameters of rocks with a small sample size has always been a hot topic for a specific project. To solve this problem, Bayesian statistical inference method is introduced to establish a set of statistical inference system based on Bayesian minimum entropy principle and excellentness comparison test. The specific methods are as follows: (1) Select the probability distribution function of rock mechanics parameters within a certain range; (2) Use Bayesian minimum entropy principle to estimate the distribution parameters of various models under small sample; (3) Test, determine the optimal probability distribution function of rock mechanics parameters. The paper analyzes the effectiveness of the method with examples and points out that the method is a new way to solve the above difficulties effectively.