论文部分内容阅读
本文以塿土和黄绵土作为实验材料,尝试使用BP神经网络方法(Back-Propagation neural network)模拟人工降雨条件下,间隔覆盖坡面的产流产沙状况。通过设置不同坡度、降雨强度、面积比,获得各种因素不同水平组合下的实测数据;以实际降雨强度、坡度、面积比、径流起始时间和初始含水率5个因子为输入变量、坡面产流量和产沙量为输出变量,利用BP神经网络模型与多元线性回归模型对数据进行模拟分析,并检验其模拟效果。研究结果表明:训练样本集平均相对误差为18.23%,预测样本集平均相对误差为5.21%;与多元线性回归模型相比,BP神经网络模型拟合精度较高,拟合效果更理想,表现出更强的预测能力。另外,比较不同土质坡面产流量与产沙量模拟效果,塿土优于黄绵土。从本研究的结果看,BP神经网络模型应用于坡面产流产沙模拟预测,省时省力,方便快捷,具有一定的应用潜力,但其实际的模拟预测能力尚需进一步探索。
In this paper, loessial and loessial soil were used as experimental materials to simulate the runoff and sediment production at the interval-covered slope under the condition of artificial rainfall using BP neural network (Back-Propagation neural network). By setting different slope, rainfall intensity and area ratio, the measured data under different levels and combinations of different factors were obtained. Based on the actual rainfall intensity, slope, area ratio, runoff start time and initial water cut, Runoff and sediment yield as output variables, the BP neural network model and multivariate linear regression model were used to simulate the data and test the simulation results. The results show that the average relative error of the training sample set is 18.23% and the average relative error of the prediction sample set is 5.21%. Compared with the multivariate linear regression model, the BP neural network model has a higher fitting accuracy and a better fitting result, More predictive ability. In addition, the simulation results of runoff yield and sediment yield of different soil slopes are better than loess soil. From the results of this study, BP neural network model is applied to simulation and prediction of runoff and sediment production on the slope surface, which saves time and labor, is convenient and fast and has some potential applications. However, its actual capability of modeling and forecasting needs to be further explored.