基于集合卡尔曼滤波的海洋土孔隙率预测研究  

Prediction of Marine Soil Porosity Based on Ensemble Kalman Filters

作　　者：汪明元 张国 潘孙珏徐 陶袁钦 WANG Mingyuan;ZHANG Guo;PAN Sunjuexu;TAO Yuanqin(Powerchina Huadong Engineering Co.,Ltd.,Hangzhou 311122,China;Shenzhen Shengye Construction Technology Co.,Ltd.,Shenzhen 518026,China;College of Civil Engineering,Zhejiang University of Technology,Hangzhou 310014,China)

机构地区：[1]中国电建集团华东勘测设计研究院有限公司,杭州311122 [2]深圳市盛业建筑科技集团有限公司,广东深圳518026 [3]浙江工业大学土木工程学院,杭州310014

出　　处：《工业建筑》2023年第6期37-42,共6页Industrial Construction

摘　　要：正确评估海洋土的物理力学性质是保障海洋工程安全的关键。基于集合卡尔曼滤波结合波阻抗测量数据和波阻抗-孔隙率转换式,提出了一种海洋土孔隙率的概率预测及不确定性量化的方法。它可同时考虑转换模型和状态转移的不确定性,提供孔隙率沿深度的取值及其不确定性。首先基于先验信息生成海洋土孔隙率估计的初始集合;然后通过由多传感器岩心记录仪取样测量的波阻抗数据和概率转换模型,对海洋土孔隙率进行预测和更新;最后分析转换模型误差、初始集合和观测数据量对孔隙率估计的影响,通过工程实例的验证,表明该方法可有效地估计海洋土孔隙率随深度的空间分布,并量化不确定性。To estimate parameters of physical and mechanical properties of marine soil correctly plays a significant role in ensuring the safety of ocean engineering.A probabilistic prediction method was proposed,based on the Ensemble Kalman Filters and combined with the measured date of acoustic impectance and acoustic impedance-porosity transfer model,to predict the porosity of marine soil and quantify the relevant uncertainty of the transfer model and transfer states.First,an initial set representing the primary porosity estimation was generated based on the prior information.Then,the soil porosity along the depth was predicted and updated by combining the measurements of acoustic impedance obtained by multi-sensor core loggers and the probabilistic transfer model.Finally,the influence of errors of the transfer model,the initial set and the numbers of observation data on the prediction results were analyzed.An example was given to illustrate and verify the proposed method.The result indicated that the proposed method could effectively predict the distribution of soil porosity along the depth and reasonably quantify the relevant uncertainty.

关 键 词：概率分析 集合卡尔曼滤波 海洋土 孔隙率 波阻抗 

分 类 号：P75[天文地球—海洋科学]