Study On Prediction And Simulation Of Dam Foundation Curtain Grouting Based On Data-Driven

Curtain grouting,as one of the most commonly used means of dam foundation seepage prevention,has been widely used in water conservancy projects.Because of the curtain grouting with high concealment,the actual development law of rock fracture is often more complex,making the actual completed unit injection quantity,the grouting of permeable rate and the diffusion in slurry in bedrock difficult to forecast.Just a note to the unit after the intake and irrigation water rate of prediction research,failed to fully consider the geological conditions,construction conditions and comprehensive factors such as slurry properties and impact,and due to its by the crack law of development and construction conditions of uncertainty,often has strong nonlinear,the prediction of reason and difficult to achieve ideal effect.On the other hand,the numerical simulation study of curtain grouting,because of the curtain grouting rock fracture development,is unable to understand the sampling method of fracture joint development of sampling will inevitably produce errors,so the curtain grouting of the actual development condition and the quantities is difficult to accurately model.Therefore,based on the actual engineering data to carry out the curtain grouting unit injection quantity and after filling and permeable rate of prediction research,exploring new methods of numerical simulation of curtain grouting,the scientific appraisal grouting quality and improving the filling simulation accuracy has vital significance.In view of the above problem,this paper uses the machine learning method,spatial interpolation theory and numerical value equivalent method and obtains the following results:(1)According to the actual data of curtain grouting and its influence factor between the nonlinear problem,based on the gating cycle unit,bay leaf,optimization algorithm and the improved adaptive noise complete set of empirical mode decomposition,the ICEEMDAN algorithm is adopted to decompose the original data,and the Bayesian optimization algorithm is applied to the GRU-helped superparameter optimization,build the intelligent prediction model of curtain grouting,effectively improve the prediction accuracy of curtain grouting unit injection quantity,and can provide guidance for practical grouting quality control.(2)For the curtain grouting inspection hole before the lack of irrigation data,when the irrigation water rate is difficult to predict,detailed analysis in the Inverse Distance Weighted interpolation and Kriging interpolation,improved Shepherd interpolation,Theissen polygon interpolation and minimum curvature interpolation on the basis of the principle,advantages and disadvantages of establishing inspection hole filling and the combination of the former data interpolation model,in a unified inspection hole irrigation before and after irrigation data.Then,after the irrigation water rate based on the analysis of the influence factors,curtain grouting and the combination interpolation model wisdom can predict model fusion,after filling and permeable rate prediction model is established,can provide strong support for the evaluation of grouting quality.(3)The drilling test section of the water pressure test is based on the finite element model,establishing a permeable rate and permeability coefficient of the equivalent method.Numerical simulation provides reliable material parameters for the curtain grouting.At the same time,to build the finite element model based on the equivalent permeability coefficient of curtain grouting and on the basis of Darcy’s law and the law of Fick diffusion,the diffusion model of fluid flow in porous media is deduced to solve the problems of the slurry in the rock mass diffusion,which provides an accurate simulation of the curtain grouting process tool,as the curtain grouting grout diffusion and unit injection quantity simulation provides a new way of thinking.(4)Based on waste valleys pumping storage power station curtain grouting project,the application study on the prediction and simulation of curtain grouting method,results show that ICEEMDAN decomposition algorithm can effectively reduce the raw data of the nonlinear;Combination interpolation model has good effect,and has good generalization ability;ICEEMDAN-BO-GRU helped build by intelligent prediction model for the curtain filling slurry unit injection quantity and after filling and permeable rate prediction with high accuracy;Curtain grouting construction simulation method based on the equivalent permeability coefficient,the simulation results can better formation of grout curtain,grout diffusion law is consistent with the existing research results.The curtain grouting help actual prediction and simulation method of grouting engineering project quality prediction and simulation analysis can be used as an important support of grouting engineering practice.