基于小波降噪的神经网络盾构泥水分离系统参数预测方法  

Parameter prediction method of shield slurry separation systemby neural network based on wavelet denoising

作　　者：周翠红[1] 周富强 刘兆赫 翟志国[2] ZHOU Cuihong;ZHOU Fuqiang;LIU Zhaohe;ZHAI Zhiguo(School of Mechanical Engineering,Beijing Institute of Petrochemical Technology,Beijing 102617,P.R.China;The 2nd Engineering Co.Ltd.of China Railway Tunnel Group,Sanhe 065201,Hebei,P.R.China)

机构地区：[1]北京石油化工学院机械工程学院,北京102617 [2]中铁隧道集团二处有限公司,河北三河065201

出　　处：《土木与环境工程学报（中英文）》2025年第1期11-17,共7页Journal of Civil and Environmental Engineering

基　　金：国家自然科学基金(51104022)。

摘　　要：泥水盾构穿越复合地层时,掘进控制参数和泥水分离系统参数往往出现大幅波动,影响施工安全和掘进效率。为提升施工过程的安全稳定性,实现异常工况预测,依托望京隧道盾构工程,针对地层状况采用筛分、双旋流、离心/压滤固液分离协同控制技术,采集盾构机掘进参数(掘进速度、刀盘转速和总推进力等)和泥水分离系统运行参数(进浆量、进浆密度和进浆黏度等),通过Cook距离离群检测和小波阈值去噪处理提升数据质量;以双旋流分离密度比值、黏度比值等12个参数为输入,排浆量、排浆密度和排浆黏度为输出,建立BP神经网络泥水分离系统参数的预测模型,并选取3个不同地层环段进行预测对比分析。预测结果表明:预测平均绝对误差均在5%以内,该预测模型在复合地层下仍具有较高的准确性。When the slurry shield passes through the composite formation,the parameters of shield tunneling control and slurry separation system generally fluctuate greatly,influencing construction safety and tunnelling efficiency.In order to improve the safety and stability of the construction process and prevent abnormal working condition prediction,based on the Wangjing Tunnel,the coordinated control technique,including solid-liquid separation screening,two-stage cyclone,and centrifugal/pressure filtration are adopted according to the stratum conditions.Shield tunneling parameters(tunnelling speed,cutter head speed,total propulsion force,etc.)and slurry separation parameters(feed quantity,feed gravity,feed viscosity,etc)are collected.Data quality can be improved through Cook distance outlier detection,wavelet threshold denoising.12 parameters are selected as inputs,such as specific gravity ratio and viscosity ratio of two-stage cyclone separation,and the output parameters are discharge volume,discharge specific gravity and discharge viscosity,A BP neural network was established to predict parameters of the slurry separation system, three different formation annuluswere selected for prediction. Results show that the average prediction errors are all within 5%, while predictionshave high accuracy under the composite formation.

关 键 词：盾构隧道 泥水分离 COOK距离 小波去噪 BP神经网络 参数预测 

分 类 号：U231.3[交通运输工程—道路与铁道工程] U455.43[建筑科学—桥梁与隧道工程]