基于力学机理的贝叶斯动态线性岩质边坡变形预测模型研究

Study on Prediction Model of Bayesian Dynamic Linear Rocky Slope Deformation Based on Mechanical Mechanism

考虑岩质边坡变形的力学机理,将边坡变形分解为弹塑性损伤变形、粘弹性变形、周期波动及观测误差等分量,以Bayesian动态线性模型理论为基础,建立了基于力学机理的边坡变形动态演化预测模型。模型以边坡岩体力学参数为系统控制变量,应用Bayesian推断,通过观测数据及相关信息,估计其动态演化规律,并以此预测边坡变形。通过对某水电站边坡位移监测数据分析表明:监测点处边坡变形主要为弹塑性损伤蠕变,监测初期,岩体迅速损伤,变形15天后进入加速蠕变阶段,与现场巡查结果一致。

Based on mechanism of rocky slope deformation,slope deformation is decomposed into a sum of elastic-plasticity damage deformation,viscoelastic deformation,periodic deformation and observation error.The mechanism model is proposed to dynamically forecast slope deformation based on Bayesian dynamic linear model.In the model mechanics parameters of rock mass are considered as state of the system.Applying Bayesian inference,dynamic evolution of the mechanics'parameters is evaluated through monitoring data and related information by which slope deformation is forecasted.The model has been used to analyze monitoring data of the hydropower station slope.Results indicate that the slope deformation at site is mainly elastic-plasticity damage deformation and accelerate creep half a month later with rock mass rapidly damaged at early monitoring,which proves to be in agreement with the results of site inspections.