Study On Experimental Simulation And Modeling Of Slope Erosion Process Of Different Soils

The process of soil erosion and sediment yield on slope is very complicated and it has many influential factors. The controlled study on indoor and field soil erosion and sediment yield on slope under rainfall conditions is greatly helpful to understand the situation and process of soil erosion on field slope by analyzing the change rules of soil erosion and sediment yield on indoor slope, and to know the internal mechanism and the development process of soil erosion and sediment yield on slope from the dynamic perspective, which could provide critical support for the establishment of the slope soil erosion prediction model and have important application prospect and practical significance on the contol and harness of soil erosion and the related disasters. With the objective of revealing the mechanism of the process of water erosion and sediment yield on slope, the spatial and temporal variation rules of soil erosion and sediment yield on slope with three different soil types, the dynamic response process of sediment yield and the factors of water erosion were systematically studied by using the method of artificial simulated rainfall experiments with different rainfall intensities in the paper and through the theories such as regression analysis, gray correlation, the theories of fuzzy similarity, and the differences and similarities of the different experimental soils were discussed, and a new slope rainfall erosion prediction model was premiliarly studied and established too. The The main results in the paper were as follows:Firstly, the development rules of rainfall erosion and sediment yield on different soil type slopes were explained. In the early stage of rainfall, the runoff generation, sediment yields on all the different soil type slopes increased along with the increasing of experiment duration, then a abrup point of them would appear. After that, runoff generation, sediment yield would reach a relative steady state. But the appearing time of abrup point and the reaching stable time of runoff generation, sediment yield of different soils were very different. In addition, with the increase of rainfall intensity of the simulated rainfall, the appearing time of abrup point and the reaching stable time of runoff generation and sediment yield of different soils would arrive earlier and earlier. Both of the appearing time of abrup point and the reaching stable time of sediment yield on slope would lag about 1～2min behind runoff generation.Secondly, the differences among the soil erosion and sediment yield on different soil slopes were analyzed. The runoff generation, sediment concentration and sediment yield on different soil slopes would increase along with the increasing of slope gradient, rainfall intensitym, and the variations of the above mentioned study objects on undisturbed soil slope were far greater than those on disturbed soil slope. Under the same rainfall intensity and slope gradient, the average sediment concentration and sediment yield on different soil slope followed the following order:fly ash> Mangshan mountain soil> undisturbed soil. The quantity of generated rill, and average rill length, width, depth on slope would increase along with the rainfall intensity increasing, while the shape and pattern of soil erosion on indoor slope was different from those on field slope, which indicated that there were greater difficulty in using the development process of indoor slope pattern to simulate that of field slope pattern. Under the same other conditions, runoff generation, sediment concentration and sediment yield of the soil with smaller bulk density were relatively greater.Thirdly, the variation rules of water erosion factors on different soil slopes were verified. Under the same slope gradient, average flow shear stress, velocity, stream power and unit stream power would increase with rainfall intensity increasing. Under the same rainfall intensity, the above mentioned water erosion factors would increase with the slope gradient increasing. With the increase of rainfall duration, there would arise a surge point for flow shear stress, velocity and stream power on slope. Then the above mentioned water erosion factors would reach relatively stable status respectively, but the reaching stable times of them would be greatly different.Fourthly, the correlation relationship between slope water erosion factors and sediment yield on slope were established. Through gray correlation and multiple regression analysis methods, the relationship between sediment yield and water erosion factors were analyzed. The results indicated that, in terms of the impact of a single factor on sediment yield on slope, the influence of rainfall intensity on sediment yield would be the greatest, and that of stream power would be secondary. The analysis of the multi-factor integrated impact on sediment yield on slope showed that the effect of the integration of rainfall intensity and slope gradient on sediment yield on slope is the greatest. And the relationship between sediment yield and its main influencing factors on the slopes of three different types of soil were established in the paper independently. Fifthly, the indoor and outdoor fuzzy similarity degree were analyzed. The analysis of the results on the fuzzy close-degree calculation of the process of rainfall erosion and sediment yield on indoor and field slopes indicated that the simulation effect on Mangshan loess soil slope is better than that on fly ash slope, and the difference between the process of rainfall erosion and sediment yield on indoor Mangshan loess with larger bulk density and that on field slope was the smallest amone three different experimental soils. Thus, the experimental data of Mangshan loess could be used to establish slope rainfall erosion and sediment yield prediction model.Finally, a slope rainfall erosion and sediment yield model for a single rainfall event was established. According to statistical learning theory, a slope rainfall erosion and sediment yield prediction model SVR was obtained based on the studied influence factors in the paper, and a BP neural network model of slope rainfall erosion and sediment yield for a single rainfall event was also established. a comparision of the predicted results of slope rainfall erosion and sediment yield for a single rainfall event based on the established two models of prediction indicated that both of the predicted results of the two models were better, while the predicted effect of SVR model was much better than that of BP neural network model.