【作者】 张世银；

【Author】 ZHANG Shi-yin (School of Civil Engineering and Architecture,Anhui University of Science and Technology,Huainan Anhui 232001,China)

【机构】 安徽理工大学土木建筑学院；

【摘要】 土层冻结温度的高低直接影响着形成有效冻结壁厚度的时间和冻结壁厚度的取值,它与土的组成成份、含水率、密度、液限含水率、压力、含盐量等有一定的关系。针对新建矿井实测样本较少的特点,利用基于结构风险最小化原理与小样本学习方法-支持向量机算法,建立了人工冻土冻结温度支持向量机计算模型。采用不同的核函数的支持向量机对人工冻土冻结温度计算进行对比分析,确定了适合于人工冻土冻结温度计算的核函数。人工冻土冻结温度支持向量机模型计算结果表明,该方法是一种有效的方法,为人工冻土冻结温度的计算提供了一条新途径。更多还原

【Abstract】 The level of soil freezing temperature has a direct impact on the time of formatting effective frozen wall and the value of freeze wall thickness,and it has a certain relationship with the soil composition,moisture content, density,liquid limit moisture content,pressure and saltness.For the feature of less measured sample of new mines,using the theory that is based on structural risk minimization and the method of small sample study - Support Vector Machine Algorithm,the support vector machine computation model of artificial frozen soil temperature has been established.Using support vector machines of different kernel functions to analyze the frozen temperature calculation of artificial freezing soil,the kernel function that is fit for the frozen temperature calculation of artificial has been determined.The results of support vector machine calculation model shows that this method is an effective method,it has provide a new approach for the frozen temperature calculation of artificial freezing soil.更多还原