Study On The Sand Source Of K742-K745 On The Ejina-Hami Railway Based On Nonlinear Mathematics

The K742-K745 section of the Ejina-Hami railway that was opened in 2015 suf-fered serious sand damage,and protective measures such as nylon net sand barriers and vegetation have not been installed on both sides.Failure to timely control would have a great impact on the safety of the railway,so exploring the roadbed sediment sources of this section of the road,and according to the results of the exploration putting forward measures to deal with the wind and sand disaster,is very urgent and realistic.In this paper,the electron microscope scan and energy spectrum analysis of the sand grains from the railway subgrade and the sand dune located upstream of the main wind direction on the railway are carried out.After the experimental data are analyzed visually and conjecture is proposed,the electron microscopic images are used to calcu-late the fractal dimension for comparative analysis;Using SPSS software to perform systematic cluster analysis based on total distance of each element content among sam-ples;Using MATLAB software established a neural network model that can be used to predict the matching of sand samples,and predicted the overall matching of sand sam-ples of sand dune and railway embankment;According to the content of the main ele-ments in the sand sample,the statistics of the matching data of dunes and roadbed sam-ples were used to establish the fuzzy transformation relationship,and the weight of each element was determined by using the coefficient of variation and analytic hierarchy process,and then the mapping relationship was selected and a fuzzy comprehensive evaluation of the similarity of sand dunes and subgrade sand samples was made.These four non-linear mathematic methods respectively consider the contours of sand,the overall distance of the sample,the prediction of the training model,and the mapping based on statistics.They respectively judge the sand source from different angles and can be mutually verified.A consistent conclusion was made:The sand from the roadbed comes from dune.Based on this conclusion,a comprehensive sand control plan combing sand-fixing and sand-preventing measures was proposed.The main research results of this paper are:(1)Energy spectrum analysis was performed on the sand samples form the dune and both sides of K742-K745 in Ejina-Hami railway.It was analyzed that the main ele-ments of the sand sample are O,Si,Na,Al,Ca,Mg,Fe,C,K,and the proportions of these 9 elements in the sand grains were all measured.According to statistics,the high-est content of elements is O,up to 45%-60%,followed by Si,20%-40%.(2)The regularity of sand grains at the dune and railway embankment is generally the same,and the fractal dimension of the box is between 1.79-1.85.The fractal dimen-sion of the sand in the two areas with the same particle size is close,and they are obvi-ously different from that in other places..(3)The sand grains at the sand dune and the railway subgrade are very integrated,and the samples are clustered with staggered clustering.(4)Fuzzy comprehensive evaluation and neural network prediction show that the elements of the sand grains at the sand dune and the railway subgrade are generally sim-ilar.(5)According to the conditions of the dune and the section of the line,sand control measures have been formulated at different locations to form a comprehensive sand control system combining sand-fixing and sand-preventing ways.