基于ANN-GA的盾构施工参数优化研究  

OPTIMIZATION OF SHIELD CONSTRUCTION PARAMETERS BASED ON ANN-GA

作　　者：修春松 XIU Chunsong(Power China Railway Construction Investment Group Co.,Ltd.,Beijing 100070,China)

机构地区：[1]中电建铁路建设投资集团有限公司,北京100070

出　　处：《低温建筑技术》2023年第9期92-94,共3页Low Temperature Architecture Technology

摘　　要：隧道盾构施工过程中最易出现土方超挖的现象。为了减小土方超挖带来的影响,文中以成都某地铁隧道盾构区间为背景,将神经网络和遗传算法相结合,对土压平衡盾构施工参数进行优化。建立掘进速度、刀盘转速、土仓压力与超出土系数的关系模型。结果表明神经网络对样本数据进行学习和训练后得出的超出土系数与实测值基本吻合;经过神经网络结合遗传算法得到的最优土仓压力与实测静止土仓压力较为接近。基于人工神经网络和遗传算法研究的盾构施工参数可以为工程提供参考。The phenomenon of earthwork over-excavation is most likely to occur during tunnel shield construction.In order to reduce the impact of earthwork over-excavation,this paper combines neural network and genetic algo⁃rithm to optimize the construction parameters of earth pressure balance shield based on the background of shield sec⁃tion of a subway tunnel in Chengdu.The relationship between tunneling speed,cutter speed,earth pressure,and ex⁃cess soil coefficient is obtained.The results show that the excess soil coefficient obtained by neural network after learning and training the sample data is basically consistent with the measured value.The optimal soil pressure ob⁃tained by neural network combining with genetic algorithm is close to the measured static soil pressure.

关 键 词：隧道盾构施工 神经网络 土压平衡盾构 参数优化 

分 类 号：TU473.2[建筑科学—结构工程]