摘要
建立了长壁工作面底板分类及单体液压支柱底座选型的人工神经网络（ＢＰ网络）模型，并通过网络自适应学习与匹配联想，得出了采场底板类别与单体液压支柱底座型式相对应的结果。同时，通过邻城相互作用算法与ＢＰ网络耦合，预报了采煤工作面顶板来压。网络试验表明，所得结果与实际吻合良好。
Artificial neural network methods are applied to two distinct aspects of the strata control in coal mine. The first aspect is the floor classification and prop base selection by means of adaptive learning and associating in BP (Back Propagation) algorithm. The second aspect is the prediction of the likely future weighting of main roof through hybrid algorithms of the neighbourhood interaction operation and the neural networks. And the network outputs are in a close agreement with the measured data.
关键词
神经网络
采场
底板分类
顶板来压
回采工作面
artificial neural networks
longwall stope
floor classification
adaptive learning,neighbourhood interaction
main roof weighting
prediction
