Application Research Of Combination Forecasting And Fuzzy Comprehensive Evaluation In Subgrade Settlement On High Speed Railway

Throughout the entire life cycle of high speed rail construction and operation management,the risk assessment and settlement observation and evaluation of subgrade engineering are the prerequisites and core content to ensure its high smoothness and stability.With the optimization of China’s high speed railway network and the update of high speed rail precision engineering measurement technology,the control over the settlement and deformation of high speed railway foundations has become increasingly strict.This theses first combs out the main points of the settlement deformation measurement system,settlement evaluation work and early warning mechanism of the high speed railway from the standardization level.Based on the subject of "technical service for subgrade settlement assessment" of China Railway Shanghai Bureau Group Co.,Ltd.,according to the characteristics of high speed railway subgrade construction under construction and operation,the fundamental content and process of settlement deformation observation over the entire life cycle are summarized.Taking practical engineering case for an instance,this theses conducts an in-depth study of the high speed railway subgrade from the aspects of safety evaluation and settlement prediction:1.The parallel combination prediction model is applied to the high speed railway foundation during construction period.First of all,three methods are selected for singleterm prediction.Then a variable weight combination model is established,and weights are assigned by countdown and inverse square error method.Taking advantage of the non-linear mapping advantages of the GRNN,the single model prediction value is taken as the input of the GRNN.The target of the neural network is a weight matrix,and the network is trained.Finally,the final settlement’s weight is predicted by using the trained neural network to obtain the final settlement of the section in the example.2.The Series combined forecasting model is applied to the high speed railway foundation during the operation period.Taking the subgrade of K465+762-K466+119section of HS high speed rail TZ station for an example,the severe subsidence of this section is analyzed from three aspects,including engineering geology and implementation demands.Based on the theory in Chapter 3,the GM(1,1)is optimized from the two aspects of background value and weight matrix.The IGM(1,1)-BP residual correction series combination model is established for this subgrade.The prediction has finally achieved excellent results.In terms of safety evaluation(Chapter 5),this theses uses an improved fuzzy comprehensive evaluation method to study a certain high speed railway foundation project under construction.It selects a total of 15 factors in 4 categories to establish an evaluation system.The idea of combination weighting based on game theory is introduced,combined with the advantages of improved AHP and improved entropy weight method to reduce the weight deviation,achieve weight optimization,and then makes a comprehensive evaluation of the construction risk of the project.