多峰分布的岩石单轴饱和抗压强度统计分析方法及应用
Statistical Analysis Method of Uniaxial Saturated Compressive Strength of Rock with Multimodal Distribution and Its Application
张晓 1陈启标 2皮德金 3余荣春 4刘卫今5
作者信息
- 1. 广东省九域工程技术咨询有限公司,广东 广州 510440
- 2. 广东顺建成建筑工程有限公司,广东 佛山 528399
- 3. 广东博意建筑设计院有限公司,广东 佛山 528312
- 4. 广西工业职业技术学院,广西 南宁 530001
- 5. 中国能源建设集团广西电力设计研究院有限公司,广西 南宁 530007
- 折叠
摘要
越南某火力发电厂下伏中等风化岩石单轴饱和抗压强度fr呈现多峰分布,极差大、离散性大,正态分布假设下的统计量与实际情况存在偏差.笔者以fr 样本分位数为切入点,探讨fr 水平、垂直分布规律以及统计成果应用.结果表明:fr在水平方向的分布与地层沉积韵律、层厚、层面倾角等直接相关,fr在垂直方向的分布与取样深度无明显相关性;基于各钻探回次的fr 样本分位数,可通过贝叶斯公式及全概率公式估算钻进成功率,可用于钻进工艺选择以及风险评估;通过线性规划模型可得到钻进工艺配置的最优方案.该统计分析方法对样本分布特性无特殊要求,易于实现,具有一定的工程应用价值.
Abstract
The uniaxial saturated compressive strength fr of moderately weathered rock underlying a thermal power plant in Vietnam shows a multimodal distribution,with large range and large discreteness.There is a deviation between the statistics under the assumption of normal distribution and the actual situation.The authors take the quantile of fr samples as the starting point to explore the horizontal and vertical distribution of fr and the application of statistical results.The results show that the distribution of fr in the horizontal direction is directly related to the sedimentary rhythm,thickness and dip angle of the stratum,and the distribution of fr in the vertical direction has no obvious correlation with the sampling depth.Based on the fr sample quantile of each drilling round,the drilling success rate can be estimated by Bayesian formula and full probability formula,which can be used for drilling process selection and risk assessment;The optimal scheme of drilling process configuration can be obtained by linear programming model.The statistical analysis method has no special requirements for sample distribution characteristics,is easy to implement,and has certain engineering application value.
关键词
多峰分布/岩石单轴饱和抗压强度/统计分析/钻进工艺配置Key words
multimodal distribution/uniaxial saturated compressive strength of rock/statistical analysis/drilling process configuration引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据