首页|基于MK-LSTM算法的盾构掘进参数相关性分析及结构变形预测

基于MK-LSTM算法的盾构掘进参数相关性分析及结构变形预测

扫码查看
提出了MIC-K-median-LSTM(MK-LSTM)算法,用于对盾构掘进过程进行参数相关性分析和结构变形预测。首先,运用改进的MIC(MK)算法对涉及盾构掘进过程中的各参数与结构变形进行相关性分析;然后,在得到相关系数的基础上提出输入参数的修正方法;最后,通过LSTM模型对不同维度输入参数的预测效果进行分析,确定合理的输入参数维度。结果表明:盾构参数对既有结构变形的影响大于土体参数;MK算法可以有效降低计算复杂度和减小噪声对数据的影响,基于参数相关系数的数据前处理方法有利于提高模型的预测精度;MK-LSTM可以有效预测结构随时间的变形规律,考虑数据维度对预测精度的提升效果和计算效率的影响,进行实际工程预测时可以根据参数相关性大小进行维度删减。
Correlation analysis of shield driving parameters and structural deformation prediction based on MK-LSTM algorithm
MIC-K-median-LSTM(MK-LSTM)algorithm was proposed to analyze the correlation of parameters and predict the structural deformation.Firstly,the improved MIC algorithm is developed to analyze the correlation between the different input parameters and structural deformation,then to preprocess the input parameters based on their correlation coefficients.The prediction accuracy and efficiency using different dimensions of input parameters are analyzed through the LSTM model and the optimal input parameter dimensions are selected.The results show that:The influence of the shield parameters on the existing structural deformation is larger than soil parameters;The MK algorithm can effectively reduce the computational complexity and the impact of noise in raw data and the data pre-process is beneficial to improve prediction accuracy;MK-LSTM algorithm can effectively predict the deformation law of the structure over time,considering the effect of the data dimension on the improvement of the prediction accuracy and the influence of the calculation efficiency,dimension pruning can be adopted in the actual engineering based on the parameter correlation.

shield tunnellingmachine learningparameter dimensionparameter correlationdeformation

陈城、史培新、贾鹏蛟、董曼曼

展开 >

苏州大学 轨道交通学院,江苏 苏州 215006

东北大学 资源与土木工程学院,沈阳 110819

常熟理工学院,江苏 常熟 215506

盾构隧道 机器学习 参数维度 参数相关性 变形

国家自然科学基金面上项目中国博士后科学基金项目

522784052021M702400

2024

吉林大学学报(工学版)
吉林大学

吉林大学学报(工学版)

CSTPCD北大核心
影响因子:0.792
ISSN:1671-5497
年,卷(期):2024.54(6)
  • 10