基于DFOS的边坡多场信息关联规则分析

采用分布式光纤传感技术(DFOS),可获得边坡多场分布信息,但如何对海量的多场信息进行相关分析,正确评价多场耦合作用下的边坡稳定性,仍是一项十分复杂的工作。本文首先介绍了数据挖掘中关联规则分析方法的基本原理,设计了边坡应变场、变形场、渗流场、温度场和环境参量等多场信息DFOS的获取方案,并以三峡库区马家沟边坡为例,对DFOS获得的边坡多场监测数据进行了关联规则分析,揭示了监测时间与库水位涨落、库水位涨落与坡体前缘地下水位间的关联性,并与实际监测数据进行了比较,证明了采用关联规则方法对DFOS边坡多场信息分析的有效性,从而为边坡稳定性评价和滑坡预测预警提供了科学依据。

ANALYSIS OF DFOS-BASED ASSOCIATION RULES ON MULTI-FIELDS DATA MONITORING OF SLOPE

The distribution information of multi-fields in slope can be obtained using distributed optical fiber sensing(DFOS)technology. Nevertheless, it is still a quite complicated work to analyze massive amounts of information about multiple fields relatively, and to evaluate the slope stability under multi-fields coupling effect. This paper introduces the basic principle of analysis method of association rules in data mining. It designs distributed optical fiber monitoring projects on strain field, deformation field, seepage field, temperature field, environmental parameter of slope respectively. Taking Majiagou slope of Three Gorges as an example, this paper analyzes monitoring data of different fields with the method of data mining association rules, and it finds rules on matching attribute between monitoring time and fluctuation of reservoir water level, coordination between fluctuation of reservoir water level and groundwater level of slope's leading edge. Comparing with practical monitoring data, it proves that it is effective to analyze multiple fields of slope by adopting association rules, thereby providing scientific basis for slope stability evaluation and landslide forecast warning.