基于聚类算法与碎石桩钻进参数的地层识别

Stratigraphic identification based on different clustering algorithms and drilling parameters of vibro-replacement stone column

在深厚覆盖层地基的振冲碎石桩处理中,有效进行地层辨识对碎石桩成桩中的填料量、控制电流、留振时间、反插深度等施工过程控制参数的确定与实时控制具有重要的指导意义。通过采集的振冲碎石桩施工过程实时监测信息,对不同地层的钻进耗能进行分析,并引进钻进比功概念,利用不同的聚类算法,在进尺深度、电流与钻进比功3个指标上进行了不同深度地层的聚类分析。研究表明:采用轮廓系数、DB指数和CH指数3种聚类精度指标可对不同聚类算法的有效性进行评价。经筛选,凝聚层次(AGNES)聚类算法是最适合振冲碎石桩钻进过程中地层辨识的方法。研究成果为振冲碎石桩施工过程中海量数据挖掘分析提供重要参考,为振冲碎石桩施工质量控制提供技术支撑。

In the treatment for deep overburden foundations by vibro-replacement stone column,effective stratigraphic identification is crucial for the determination and real-time control of construction process control parameters such as filling amount,control current,vibration retention time and reverse insertion depth.This study analyzed the drilling energy consumption in different strata based on real-time monitoring data collected from the construction process of vibro-replacement stone column.The concept of drilling ratio work was introduced and various clustering algorithms were employed to conduct clustering analysis of different depth strata in three dimensions:depth of penetration,current and drilling ratio work.In the analysis of stratigraphic identification of vibro-replacement stone column performed by different clustering algorithms,the performance of the clustering methods can be evaluated using three clustering accuracy indicators:contour coefficient,DB index and CH index.Finally,the ANGENS clustering algorithm was determined as the optimal stratigraphic identification technique for the drilling procedure of vibro-replacement stone column.The results provide an important reference for the massive data mining and analysis during the construction process of vibro-replacement stone column,and supply an important technical support for the quality control of vibro-replacement stone column.