摘要
岩土体抗剪强度参数常由剪切试验数据采用拟合回归方法获得,并可采用最小二乘法、点群中心法、随机-模糊法、分位数法、优定斜率法、模糊拟合法与非对称三角模糊数法等7种方法进行分析。为探讨各种方法的适用性与合理性,首先,归纳总结7种拟合回归方法的基本原理与分析过程;然后,采用7种方法对4个工程算例进行拟合回归分析,并通过计算结果探讨了各种方法的差异性,进而分析了异常数据与取值界限对不同方法所得岩土体抗剪强度参数的影响规律。结果表明:较少的试验数据点易于点群中心法的实施,但其适用性不强,受主观因素影响较大,优定斜率法与此相似;离散性与异常数据对随机-模糊法、模糊拟合法与非对称三角模糊数法最可能值的拟合回归结果影响较小,实际工程中可优先采用随机-模糊法或模糊拟合法确定岩土体抗剪强度参数最可能值;确定岩土体抗剪强度参数标准值可优先采用0.2分位值的分位数法所确定的拟合回归结果,其次也可采用模糊拟合法确定的置信区间下限值,两者的分析误差约为5%~10%。
The shear strength parameters of rock or soil are usually obtained by the shear testing data with fitting regression methods.The methods include the least square method,center point group method,random-fuzzy method,quantile regression method,optimal gradient method,fuzzy fitting method and asymmetric triangular fuzzy number method.In order to explore the applicability and rationality of various methods,the basic principle and analysis process of these seven fitting regression methods are firstly summarized.Then,four engineering examples are analyzed with different methods.The differences of these analysis results are discussed.Furthermore,the influence rules of abnormal data and value boundary to the shear strength parameters of rock or soil,obtained by different methods,are analyzed.The analyses results show that the center point group method is suitable for the regression analysis with less testing data.Its applicability is limited and it is easy affected by the subjective factors.The optimal gradient method is similar to the center point group method.The fitting regression results obtained by the random-fuzzy method,fuzzy fitting method or asymmetric triangular fuzzy numbers method are less affected by the discrete or/and abnormal data.In practical engineering,the shear strength parameters of rock or soil can be firstly calculated by random-fuzzy method or fuzzy fitting method.The standard values of the shear strength parameters of rock or soil can be preferentially determined by the quantile regression method with the 0.2-quartile.The lower boundary value of the confidence interval determined by the fuzzy fitting method can also be selected as their values.The error between the results of these two methods is about 5%to 10%.
作者
蒋胜波
田军
张永杰
JIANG Shengbo;TIAN Jun;ZHANG Yongjie(Hunan Provincial Communications Planning Survey & Design Institute Co. ,Ltd. ,Changsha,Hunan 410009,China;Hunan Zhangsang Highway Construction and Development Co. ,Ltd. , Zhangjiajie,Hunan 427000,China;School of Civil Engineering,Changsha University of Science and Technology,Changsha,Hunan 410114,China)
出处
《公路工程》
2021年第4期59-65,共7页
Highway Engineering
基金
国家自然科学基金资助项目(51208063,51178063)。
关键词
抗剪强度参数
拟合回归方法
概率统计
模糊数学
标准值
shear strength parameters
fitting regression methods
probability statistics
fuzzy mathematics
standard value
