基于Apriori算法对地面沉降影响因素的数据挖掘及分析

Data Mining and Analysis of Influence Factors for Ground Settlement Based on Apriori Algorithm

为有效利用盾构法施工过程中产生和积累的大量掘进历史数据、挖掘知识和信息,以优化掘进参数、控制地表沉降,采集某地铁相邻两区间的地面沉降数据及掘进参数,基于数据挖掘技术,通过对比优缺点,综合等宽离散和k-means聚类2种方法将数据离散化,并采用Aprioir算法进行关联规则挖掘。通过分析关联规则得到以下结论:出土率、注浆填充率和俯仰角对沉降值影响较大,掘进速度和刀盘转矩对沉降值也有一定影响。根据数据离散区间和关联规则,给出穿越土层为圆砾和卵石时的注浆填充率建议值为185%～190%,注浆压力建议区间为0.16～0.22 MPa,同时建议俯仰角纠偏幅度宜平均每3环小于±0.2°。

In order to effectively utilize the large amount of tunneling historical data, mining knowledge and information generated and accumulated in the process of shield tunneling to optimize tunneling parameters and control surface subsidence. The ground settlement data and tunneling parameters of adjacent sections of a metro are collected. Based on the data mining technology, the data are discretized by two methods: equal width discrete and k-means clustering. By comparing advantages and disadvantages, and using Aprioir algorithm to mine association rules. Through the analysis of association rules, the following conclusions are obtained: The factors including excavated earth rate, grouting filling rate and pitching angle have a great impact on the settlement value. The tunneling speed and cutterhead torque also have a certain impact on the settlement value. According to the data discrete interval and association rules, the suggested grouting filling rate when tunneling through the soil with gravel and pebbles is 185%-190%, the suggested grouting pressure range is 0.16-0.22 MPa, and also suggested that pitching angle correction range on average every three ring is less than ±0.2°.