RBF神经网络技术预测煤矿地下水位方法的优化研究

Optimization of RBF Neural Network Technology in Forecasting Coal Mine Underground Water Level

实时预测煤矿地下水水位及其分布范围对于煤矿的安全生产至关重要。以韩城象山矿区的历史水文数据为研究依据,采用RBF神经网络技术,对预测该地区的地下水水位进行了研究。针对RBF神经网络的不足,对RBF算法的中心点数量,计算欧氏距离时各分量的权重进行了优化,导出了ARBF算法。数据对比表明,ARBF神经网络算法达到了更为理想的效果。

Real-time prediction of underground water level distribution range are critical for the safety operation of coal mines. This paper predicts the water level of Hancheng Xiangshan mining area with historical hydro-geological data by using RBF neural network technology. To overcome the weakness of RBF algorithm, this paper introduces ARBF algorithm, which optimizes the weight of each component when calculating Euclidean distance. ARBF delivers better result comparing to RBF