Research On High-precision Deformation Monitoring And Prediction Methods For Loess Landslide

Landslide disasters occur frequently in loess areas of China,seriously damage the safety of life and property of local people,affect production,life and social stability,and cause irreversible damage to the ecological environment and natural resources.In order to reduce the loss caused by the loess landslide disaster,it is necessary to monitor,analyze and predict the safety and stability of the landslide.Because a single monitoring method is difficult to provide continuous and stable deformation data,and the prediction model of loess landslide deformation and its applicability also need to be further studied.Therefore,based on the in-depth study of the theory and method of loess landslide monitoring and prediction,taking Miaodian loess landslide as an example,a variety of high-precision deformation monitoring techniques and related data processing methods are applied to the deformation monitoring of loess landslides.The applicability of various landslide deformation prediction models is analyzed and evaluated.The main research contents and results of this paper are as follows:(1)Three high-precision surface deformation monitoring technologies,including GNSS,geo-robot and wireless sensor monitoring,were applied to the deformation monitoring of loess landslide.The corresponding monitoring technical scheme was designed and the deformation monitoring data of Miaodian loess landslide was obtained.(2)The accuracy and reliability of GNSS,geo-robot and wireless sensor monitoring technology applied to the deformation monitoring of loess landslide were studied,and the measured data were analyzed and compared.The results show that the average mean square error of the geo-robot in horizontal and vertical directions are ±4.0mm and ±2.4mm respectively.The short-baseline static relative positioning results of BDS and GPS are within ±1.6mm.The average GNSS real-time dynamic monitoring results N,E,U and the average difference of the geo-robot are ± 2.5,± 2.0,± 2.8(mm).The monitoring accuracy of the displacement meter is in the millimeter range.The monitoring results of the three technical means are basically consistent in magnitude and deformation trend.(3)The effects of meteorological factors such as temperature and pressure on the positioning accuracy of the geo-robot are studied.The results show that the influence of meteorological factors on the accuracy of the range is up to the centimeter level,which is proportional to the side length.After correcting the meteorological error by difference,the average accuracy of the distance measurement of the two stations increased respectively by 52.6% and 78.2%,this verified the effectiveness of the method of differential correction of meteorological error.The atmospheric temperature has the most significant influence on the ranging results,followed by the air pressure,and the relative humidity is the weakest.(4)The applicability of the prediction model of loess landslide deformation is studied.A variety of common landslide deformation prediction models are introduced.Combined with classical weights and optimal weight methods,combined forecasting models are established and relevant deformation prediction experiments are carried out.The results show that the exponential smoothing model has poor applicability to the prediction of the deformation sequence of this landslide.The unbiased grey Verhulst model and the grey BP model are more suitable for short-term prediction.The prediction accuracy of the combined prediction model is much higher than that of the single prediction model,and the applicability is better.