Study On The Optimization Theory And Key Mechanics Problems Of High Centerline Tailings Dam

In our country, there is a serious lack of theoretical study and engineering practice of centerline embankment method(CEM), but there are many mining factories need CEM to build high dam to dump the tailings now. In order to solve the enterprises'practical problems and innovate in techniques, this paper studied the theory and methods on CEM in the strong seismic zone on the background of a planned copper mine tailing dam in Yunnan province. The results could be summarized as follows:①A lot of physical experiments have been undertaken to investigate the physical and mechanical characters of tailings included in the centerline embankment dam. The results show that the larger the grain size, the larger the internal friction angle and permeability. The sorted coarse tailing has larger permeability and is propitious to the tailing dam stability.②The dynamic triaxial test have been undertaken to investigate the dynamic characters of tailings included different coarse grain sand. The results show that the effect of coarse grain content on the vibration strain and dynamic pore pressure is quite clear; there is a positive correlation between the coarse grain content and the dynamic intensity. The dynamic strain improves with the dynamic stress under the same vibration times; the dynamic intensity improves with the consolidation pressure and there is a positive correlation between the consolidation pressure and the dynamic intensity, but the effect of the consolidation ratio K c on the dynamic intensity is not immovable. The result shows the effect of compress is distinctive at the beginning stage too, and the proposed best vibrating times is about 2 to 5 cycle times, when using vibirating compress method to improve the safety of the tailing dam.③Because there is a very different development pattern for different tailing,the pore pressure model shouldn't use the uniform and fixed function format. According to the test, the Seed's formula is modified using a power function model so that the new formula can forecast the dynamic pore-water pressure of saturated tailings material more precisely. In addition, the neural networks model was proposed to solve the dynamic pore pressure prediction in order to overcome the shortage of the normal modeling methods.④A integrated spatial visualization site selection model is proposed based on the GIS and fuzzy optimization method. On the one hand ,it could enhance the scientificity, automaticity ,efficiency and validity of the tailing dam locating using the GIS'spatial analysis and visualization function, on the other hand, by introducing the fuzzy optimization method, it could overcome the subjective experience judgment problem effectively, meet the quantification demand of random fuzzy information processing, and make the judgment more scientific.⑤A computing model of most dangerous sliding surface based on the chaos optimization search, and the result shows this search method could improve both speed and security clearly compared to the normal method, and it is fitful to search the most dangerous sliding surface.⑥The stability analysis result shows the shear strength has a significant influence on the stability, and it has a remarkable effect to improve the stability by using classified coarse tailing to construct the dam. The height of phreatic line has a significant influence on the stability too, the safety coefficient could improve about 0.06 if the height of phreatic line decreases 0.1 meter.⑦According to the dynamic response analysis of the dam under special working condition, there is a global maximum displacement at the middle and lower part of the dam slope, and there is a local maximum displacement at lower part of the deposited beach, but the displacement of the dam crest outer slope is very small all the time, and it is bigger difference between this result of centerline dam and other research results of upstream dam.⑧To overcome the difficulty of the tailing sand liquefaction prediction, a new hybrid optimization model is presented based on the complementary of BP neural network and chaos optimization algorithm. This model not only has a BP algorithm's quick local search capability, but also can converge strongly to the global optimal result by use the chaos optimization's global search character. The results show that it is an effective and feasible method to predict tailing sand liquefaction.