Parametric Performance Identification And Deformation Monitoring Method Of Concrete Dam Based On Optimized Statistic Model

Water conservancy project has functions of flood control,water storage,irrigation,power generation,shipping and so on.It is an important engineering facility for regulating and optimizing the spatial and temporal distribution of water resources.As one of the basic structures of water conservancy,the dam plays a key role in retaining water and discharging flood.While the dam construction has achieved great success in China,the safety of its service is becoming more and more important.The time-varying evolution process of dam construction materials and structural properties is bound to intensify under the condition of solid-liquid-gas coupling for a long time.Based on the measured effect size of the dam,different monitoring models and evaluation indexes are established by statistics or finite element method,which can effectively perceive the operation form of the dam and ensure the safety of the dam in service.Aiming at the shortcomings of the existing deformation monitoring statistical model establishment and the identification of dam operation performance parameters,this paper comprehensively applies mathematical statistics and finite element simulation methods to carry out research on the optimization of deformation monitoring statistical model,the rapid partition inversion of mechanical parameters and the mixed monitoring model of operation performance,so as to realize the long-term accurate prediction of dam deformation effect size and the evolution law analysis of service behavior.In order to provide theoretical basis and technical support for the healthy service of concrete dam.The main research contents are as follows:(1)Considering the shortcomings of existing deformation monitoring statistical models in quantifying the lag effect of environmental temperature,Chi-square distribution was introduced as the temperature influence weight function in the early stage,and a statistical deformation monitoring model was established,which took weighted temperature and water temperature measurements as the explanatory variables of temperature deformation and took the lag effect of dam surface temperature into account.In order to make up for the shortcoming that the model is difficult to apply in the case of less water temperature data,the method of adjusting the water temperature related parameters of the model is given.In addition,in order to improve the long term prediction accuracy of the model,a modeling method based on genetic optimization of long term memory network algorithm was proposed.(2)Aiming at the problems of large computation amount,low efficiency and poor accuracy when equivalent inversion of mechanical parameters of dam-foundation system using conventional inversion framework,sample data set is generated by improved genetic Latin hypercube sampling algorithm to train the proxy model of multi-zone finite element model of dam-foundation system,and it is applied to multiobjective mechanical parameter inversion based on Bayesian optimization.The rapid identification of elastic modulus of concrete dam is realized.(3)Based on the spatio-temporal monitoring data of the dam,the mixed monitoring model of the operation performance of the concrete dam is constructed based on the improvement of the temperature component factor and the inversion results of regional mechanical parameters.The dam deformation monitoring and early warning indicators are formulated by the comprehensive use of the confidence interval estimation method,the typical small probability method and the structural analysis method,so as to provide feedback information for the adjustment of the dam reinforcement measures or operation management plan.