Font Size: a A A

Remote Monitoring And Early Warning Model Of Frozen Soil In Dam Areas

Posted on:2020-04-24Degree:MasterType:Thesis
Country:ChinaCandidate:X J ZhangFull Text:PDF
GTID:2392330575990615Subject:Internet of Things works
Abstract/Summary:PDF Full Text Request
The area of frozen soil accounts for about 70% of the mainland area in the world.The area of frozen soil accounts for about 53% of the country's land area in China.Frozen soil is a soil medium that is extremely sensitive to temperature.Its freezing and thawing effect is affected by the temperature change during the seasons.The volume change of frozen soil is unavoidable.Under the freezing and thawing of frozen soil,the rock in the frozen soil area is destroyed,causing the building to suffer from deformation and collapse.In particular,water conservancy projects such as dams and dams will cause immeasurable losses if they are damaged.In the construction of our country,hydraulic structures such as dams are an important part.If it affected by the freezing and thawing of frozen soil,the foundation of hydraulic structures will change due to the expansion or settlement of frozen soil,resulting in buildings being damaged.Deformation damage and even more serious collapse,damage to buildings will seriously affect its service life,and it will cost a lot of money to repair.At present,the research on frozen soil is mostly roadbed frozen soil,but the research on frozen soil in dam is relatively rare.The existing research has mostly stayed on the data of artificially collected frozen soil,and most of them only collect the temperature data of frozen soil separately.The change of temperature is used to study the freezing and thawing characteristics of frozen soil.However,the artificial collection of frozen soil data is cumbersome,time-consuming and laborintensive,and cannot be analyzed and prevented in time.This paper aims to prevent the deformation and damage of dams caused by changes in frozen soil,and proposes a study on real-time monitoring and early warning models of frozen soil in dams.The study is divided into two stages: real-time monitoring and early warning models.At the first stage,the influencing factors of early warning model of frozen soil in dams were frozen soil temperature,frozen soil displacement,frozen soil type and building construction time according to analysis the freezing and thawing characteristics of frozen soil and the influencing factors of dam deformation.After determining the impact factor,the frozen soil temperature and displacement data are selected as the real-time monitoring parameters,the Pt100 temperature sensor and JM seam gauge were selected as the measurement tool by consulting the data and surveying the geological environment of the monitoring area,4G is used as a data transmission network to design a communication protocol to implement data interaction between the data collection end and the monitoring terminal.At the second stage,using BP neural network to train and predict the temperature and displacement data of frozen soil,in this process,in the face of the slow convergence of BP neural network,the additional momentum term and learning rate adaptive design method are selected to improve the convergence speed of BP neural network and better to predict frozen soil data.Finally,the early warning model was established by using fuzzy inference technology,and the early warning function of dam deformation was realized.The frozen soil monitoring and early warning model was studied in this paper solves the problems of cumbersome manual collection,untimely observation,and danger prevention.It provides data support for studying the freezing and thawing characteristics of frozen soil,and it has a certain promotion effect on the study of frozen soil in dams.The realization of real-time monitoring and early warning function provides technical support for avoiding the occurrence of construction engineering hazards in the frozen soil area.
Keywords/Search Tags:Frozen soil in dam areas, sensor, BP neural network, fuzzy reasoning, early warning model
PDF Full Text Request
Related items