Critical Warning Method For Landslide Deformation And Improvement Of Grey Model For Prediction

In recent years,due to the frequent occurrence of earthquakes,coupled with increasingly intensified engineering activities and other factors,major landslide disasters have occurred from time to time,and people's lives and property have suffered serious losses.The landslide system is extremely complex.So far,people have not fully studied and understood it.Landslide prediction,warning,and forecasting are still global problems.This paper summarizes a variety of landslide warning and forecasting criteria and prediction methods,and considers that it is difficult to accurately determine the critical time at which the landslide enters into the acceleration deformation stage or the acceleration and deformation stage by using a single criterion index;in terms of landslide prediction and prediction,it is generally said that The grey GM(1,1),gray Verhulst and other grey models are suitable for short-term and impending prediction lack of rationality;to establish grey GM(1,1),gray Verhulst landslide prediction model with displacement as the parameters lacks theoretical basis,and the establishment of the model does not combine the mechanical mechanism and the movement mechanism of landslide evolution;besides,at present,there are many researches on the prediction using the displacement after the constant-velocity deformation stage,and there are few studies on the prediction and forecasting using the isokinetic deformation stage and the previous deformation information.However,the constant deformation stage and the previous deformation are also important information to reveal the laws of landslide movement.If we can use the deformation information of this stage and then realize the accurate prediction of landslide,it will promote the prediction and forecast research of landslide actively.In view of the above problems,combined with the existing multiple historical landslide monitoring data,statistical methods and matlabR2014 numerical calculation combined with research methods,and the introduction of gray system theory,shear strength Coulomb's law and coordination theory,etc.Some critical warning methods for landslides and grey-slope prediction models have been further studied.The main research contents are as follows:(1)Based on the principle of hypothesis testing,the distribution characteristics of displacement variable,displacement velocity variable,and displacement acceleration variable of eight landslide instances at constant deformation stage are analyzed first,and the characteristics of these variables and the improved displacement tangent angle at the critical point of accelerated deformation are analyzed.The distribution characteristics of displacement variable,displacement velocity variable,and displacement acceleration variable of the 19 landslide cases in the accelerated deformation stage are analyzed,as well as the characteristics of these variables and the improved displacement tangent angle at the critical acceleration acceleration deformation point.Proposed that the correlation coefficient between the displacement of the landslide body and the time is reduced,the displacement velocity variable,the displacement acceleration variable no longer obey the normal distribution,and the improved displacement tangent angle is greater than 45°,etc.As the critical point for judging the landslide entering the accelerating deformation stage from the constant velocity deformation stage Criteria comprehensive indicators;The comprehensive index of displacement acceleration variable of the landslide body no longer obeys the normal distribution and the improved displacement tangent angle is more than 85° is used as a critical deformation warning method for judging the criticality criterion of the landslide from the accelerated deformation stage to the accelerated acceleration stage.(2)According to the exponential law of accelerated deformation,the landslide is divided into:accelerated deformation non-exponent landslide,accelerated deformation near exponential landslide,and accelerated deformation high growth strong exponential landslide.Based on this classification,the original gray GM(1,1)is studied.The Applicability and Limitation of Model,Gray Verhulst Model and Gray GM(1,1)Power Model in Short-term and Impending Forecast.(3)Combining with the existing experimental results of several shear strengths and some experimental phenomena,the paper summed up the Mohr-Coulomb strength theory considering the influence of shear rate on rock shear strength,and established a grey GM(1,1)Model with rate parameter based on grey system theory,and further deduced the general form of the traditional grey GM(1,1)model—a new gray NGM(1,1,k)model for landslide forecasting;combining with examples of landslides,a new improved grey NGM(1,1,k)model is compared with the original grey GM(1,1)model;the idea of modeling with the rate as a parameter is proposed,and the original Verhulst model and the original GM(1)are proposed.1)The power model has been improved and compared with several examples of landslides.The results show that the forecasting effect of the improved model is obviously better than that of the original model.(4)The curve features of grey GM(1,1)inverse function landslide prediction model are studied.Taking Xintan landslide and Wolongsi new landslide as examples,the applicability and limitations of the model for mid-term forecast and short-term forecast are analyzed.In addition,combining several examples of landslides,the grey GM(1,1)inverse function model is compared with the GM(1,1,K)model,the Verhulst landslide prediction model with velocity as a parameter,and the GM(1,1)power landslide The prediction and forecasting results of forecasting models are compared and analyzed.The results of case studies show that the GM(1,1)inverse function model is more suitable for short-term impending prediction,and the forecasting criteria are more reasonable.(5)For the deficiency of grey GM(1,1)inverse function model in landslide forecasting,the range of its power exponent is extended,and a grey GM(1,1)inverse function is established to extend the power exponent range.The landslide prediction and forecasting model discusses the superiority of the gray GM(1,1)inverse function model that expands the power exponent range in the medium-term forecast and short-term forecast.