CAJ | 학술논문

变形监测分析和预警对大坝健康运行起着关键的作用，但是监测数据都不可避免地存在随机误差。卡尔曼滤波法可以有效地剔除测量数据中的噪声，然而，利用其对大坝未来的趋势位移做出预测时与实际情况吻合度不高。因此提出了马尔可夫链-卡尔曼滤波法，既能剔除测量数据噪声又可以准确的预测未来位移。利用浙江某拱坝位移实测资料，拟合并预测了该大坝的变形，验证结果表明该方法的拟合预测效果良好，可用于拱坝变形预测和安全监控。
변형감측분석화예경대대패건강운행기착관건적작용，단시감측수거도불가피면지존재수궤오차。잡이만려파법가이유효지척제측량수거중적조성，연이，이용기대대패미래적추세위이주출예측시여실제정황문합도불고。인차제출료마이가부련-잡이만려파법，기능척제측량수거조성우가이준학적예측미래위이。이용절강모공패위이실측자료，의합병예측료해대패적변형，험증결과표명해방법적의합예측효과량호，가용우공패변형예측화안전감공。
Deformation monitoring and early warning analysis of the dam play a key role in the healthy operation of dam. H ow ev-er, there are unavoidable random errors in the monitoring data. Kalman filtering met hod can eliminate the noise from t he ob-served data effect ively; however, the prediction of the displacement trend of the dam does not agree w ell w ith the actual condi-tions using this method. In this paper, Markov chain-Kalman filter method w as proposed, w hich can not only remove the random error but also predict the displacement accurately. Based on the observed displacement data of an arch dam in Zhejiang Province, the method show ed an accurate prediction of dam deformation. Therefore, Markov chain-Kalman filter met hod can be used in the prediction of arch dam deformation and security monit oring.