中文摘要：

利用神经网络的自适应、自学习功能,以有限的实验数据为训练样本,建立了描述风速、粒径与沙粒起跳初速度分布函数之间映射关系的神经网络模型.利用此网络可以预测得到实验尚未给出的沙粒初速度分布函数中的拟合参数,可快捷、有效地弥补实验数据的不足.

英文摘要：

Based on some existing measurement data,several neural network models were established to describe the mapping relations between wind speed,particle diameter and the probability distribution of sand particles＇ lift-off velocities.The application of these neural networks may conveniently give the fitting parameters in the distributing functions of lift-off velocities.Such a predicting method can efficiently make up for the deficiency of the present measurement data.