基于遗传规划的盾构隧道开挖地表最大沉降预测

Genetic Programming Approach to Predicting the Maximum Surface Settlement by Shield Tunneling

由于盾构隧道开挖引起的地面沉降是一个重要而艰巨的任务,许多影响因素都必须考虑,如隧道埋深、盾构直径、盾构掘进时推力、盾构推进速率、注浆填充率、注浆压力、地层的黏聚力、摩擦角、压缩模量等.然而目前没有模型能完全反映各种因素对地面沉降的影响规律,综合考虑各种影响因素,运用遗传规划理论对地表最大沉降进行预测.利用地表沉降实测数据对模型进行测试,建立了确定盾构隧道开挖引起地表最大沉降的遗传规划模型.研究结果表明:预测值与实测值是一致的,误差小于10%.

Prediction of surface settlements due to shield tunneling is a crucial and difficult task, and lots of factors, such as tunnel overburdens, shield diameters, thrusts of shield tunneling, advancement rates of shield, fill factors of grouting, cohesive forces, friction angles and compression modules of the soils, must be considered. So far, however, no models can in- volve all the factors above. A genetic programming （GM） approach was suggested to predict the settlements in this paper, in which all the factors above were considered. The GP was trained by using face settlement data measured in engineering to set up the maximum surface settlement model of GP, and then the GP model was checked by using test samples. The predicting results were in agreement with the measurements with errors of less than 10%.