Experimental Study On Special Yield Of Sandy Pebble Media

This paper took Yellow River natural sandy pebble medium as the research object, has carried on specific yield series test of river bed native-state sandy pebble medium, sandy pebble medium at same size and sandy pebble medium not at the same size, analyzed change characteristics and influence factors of sandy pebble medium specific yield at different grading; Explored the relationship between specific yield and the characteristic parameters of particle size of sandy pebble medium; The linear model, nonlinear model and the BP neural network model of specific yield is established. In this paper, the research results show that:Nonuniformity coefficient of Yellow River beach natural accumulation sandy pebble medium is156,≤1mm size group which accounting for21.52%of the total medium not only fill the pore of particles, also participated in the particle skeleton build of sand pebble medium, has great influence on specific yield of the Yellow River sand pebble medium, causes specific yield only0.04. Specific yield of the sandy pebble medium Out of≥90mm and≤1mm increases significantly than natural gradation sandy pebble medium, up to0.16, increased4times.The specific yield of sandy pebble medium with same diameter increases as the average diameter of sandy pebble medium diameter groups increase. When the average diameter≤10mm, the specific yield increases fast with the average diameter increases; When the average diameter≥10mm, the specific yield increases slowly with the average diameter increases; The relationship between specific yield and average particle size conforms to logarithmic relationship.The grading of sand gravel media has a significant impact on specific yield. The grain size characteristics which affect the specific yield of sand gravel media is nonuniform coefficient Cu, curvature coefficient Cc,effective grain size d10and constrained diameter. When sand gravel media d10change between0.03～1.00cm, d30change between0.20～ 0.50cm, d60change between1.66～5.00cm, Cc change between0.245～30.00, Cu change between2.638～133.33, the minimum specific yield is0.047, the maximum specific yield is0.294, The maximum is the minimum6times.When the nonuniform coefficient is certain, the specific yield increases with the curvature coefficient increases.A single grain characteristic indicator can not show the quantitative relation of specific yield under different grading, curvature coefficient Cc and nonuniform coefficient Cu have point group relation with specific yield, they have close quadratic polynomial correlativity. With the increase of nonuniform coefficient, the specific yield decrease first and increase then. In certain range, the specific yield decrease as the nonuniform coefficient increase, when the nonuniform coefficient is greater than a certain value, the specific yield increase as the nonuniform coefficient increase. The correlativity between nonuniform coefficient and specific yield is better than correlativity between curvature coefficient and specific yield, and apply conveniences. Forecasting specific yield by nonuniform coefficient Cu is feasible, and has high-precision.In sandy pebble media, the specific yield increases with the large and middle particle diameter increase. When the small diameter particle increases, the specific yield may increase or decrease. To the small diameter particle which form the minimum porous of sandy pebble medium, the quantity and diameter decide the compactedness of large particles porous together, it lead to uncertainty of specific yield caused by small particle diameter changed.The unit weight of sand gravel media has obvious effect on specific yield. To saturated sandy pebble medium which has same grading, the unit weight is greater, the porosity of sandy pebble medium is smaller and the specific yield is smaller too.Linear model, nonlinear model and BP neural network model which predicated specific yield by the grain size characteristics indicators is feasible. After the comparison of each model input variable and error, it is confirmed that BP neural network model is the best forecasting model to forecast the specific yield of saturated sandy pebble medium.This paper tested and analysed deeply to forecast specific yield by sandy pebble medium grain size characteristic indicators, and has important theoretical significance and practical value to forecast specific yield by sandy gravel medium grain size characteristic indicators.