摘要
随着城市地下空间发展,地下工程沉降问题应引起重视。传统沉降预测模型常存在模型预测精度不高、模型可解释性差和难以模拟沉降规律的动态变化等局限性和不足,对此提出一种基于数字孪生自演化的地下工程沉降预测方法。建立的地下工程孪生体为沉降预测模型的高精度建模提供了有效的支撑,开发的“多项式回归+总体卡尔曼滤波”自演化算法每一步都具有可解释性,同时模型参数自主校正功能使预测模型能够动态追踪沉降规律的改变。实验结果表明所提方法能够有效解决传统沉降预测模型的缺陷。
With the development of urban underground space,more and more attention has been paid to the subsidence problem arising from the operation of underground structure.However,the traditional subsidence prediction model often has some limitations and shortcomings,such as low prediction accuracy,poor model interpretability and being difficult to simulate the dynamic variation of the subsidence law.Therefore,this paper proposes a subsidence prediction method of underground structures based on digital twin self-evolution.The twin of the underground structure established by this method can provide effective support for the high-precision modeling of a subsidence prediction model,the developed“polynomial regression+total Kalman filter”self-evolution algorithm can be interpreted at every step,and the function of autonomous calibrating model parameters enables the prediction model to dynamically track the variation of the subsidence law.The experimental results show that the proposed method can effectively solve the defects of the traditional subsidence prediction models.
作者
胡辰熙
杨启亮
邢建春
秦霞
李苏亮
贾海宁
HU Chenxi;YANG Qiliang;XING Jianchun;QIN Xia;LI Suliang;JIA Haining(College of National Defense Engineering,Army Engineering University of PLA,Nanjing 210007,China)
出处
《陆军工程大学学报》
2022年第5期66-73,共8页
Journal of Army Engineering University of PLA
基金
国家自然科学基金(52178307)
江苏省自然科学基金(BK20201335)。
关键词
地下工程孪生体
沉降预测
数字孪生自演化
多项式回归模型
总体卡尔曼滤波算法
the twin of the underground structure
subsidence prediction
digital twin self-evolution
polynomial regression model
total Kalman filter algorithm
