基于贝叶斯定理的异重流水卷吸经验式不确定性研究

Study on the Indeterminacy in the Empirical Relation for Turbidity Currents Water Entrainment Based on Bayesian Theory

为获取了异重流水卷吸公式中经验系数(A、B)的大量样本,本文基于贝叶斯定理的Metropolis-Hastings随机采样方法,结合现有异重流水卷吸实验数据,经对样本统计分析表明:概率最大的A、B样本取值与原公式取值相近。95%A、B样本相对于概率最大的样本取值范围分别为[51%,630%]和[63%,1200%],比25%样本的经验系数取值范围明显扩大,因此,经验系数具有较大的不确定性,且B远大于A,当确定经验式时,尤其要注意B的取值。

To get a large number of empirical coefficients(A,B)in the formula of turbidity currents water entrainment,This paper randomly sampled on Metropolis-Hastings in Bayesian theory,combining with current experimental data,the analytic results showed that the maximal probability A,B sample values were similar to the original formula.The threshold of 95%A,B was[51%,630%]and[63%,1200%],respectively,than maximum probability values and their thresholds were very larger than 25%samples,therefore,the empirical coefficients had a large number of indeterminacy and B was very larger than A.When the empirical relation was established,B should be especially attended.