摘要
如何依据矿井突水的标型组分含量来识别其水源所在的水文地质类别是一个很有意义的课题。为克服一般BP神经网络收敛慢和过学习问题,采用改进遗传算法搜索BP神经网络的拓扑结构和学习步数,建立了改进遗传神经网络模型。算例表明,该方法对地下水化学特征组分的识别结果是令人满意的。
How to identify gushing water sources of mines depending on the content of standard components of ions in the water is a important subject. In order to overcome the problems of slowly convergence and over- fitting of normal neural net work, the improved genetic algorithm was used to search the topology structure and learning step number of BP net work, then the model is constructed to identify standard components of ions of groundwater. The application of the true examples shows that the method is feasible and precise.
出处
《有色矿冶》
2006年第1期3-5,49,共4页
Non-Ferrous Mining and Metallurgy
关键词
突水
化学特征组分
改进遗传算法
BP神经网络
识别
water burst
standard components of ictus
improved genetic algorithm
BP neural network
identification
