长壁工作面顶板的神经网络控制

Roof control in longwall faces with neural networks

建立了长壁工作面直接顶分类与支护方法选择、老顶分级与支护强度确定的人工神经网络（ＢＰ网络）模型，并通过网络自适应学习与匹配联想，得出了工作面支护强度（方法）与顶板级（类）别相对应的结果。同时，通过领域相互作用算法与ＢＰ网络配合，预报了工作面顶板来压。网络试验表明。

Artificial neural network methods are applied to two district aspects of strata control in coals. First is the roof classification and support capacity determination by means of adaptive learning and associating in BP(Block Propagation) algorithm.Second is predication of likely future weighting of main roof through the neighborhood interaction operation and the neural networks.Theory outputs are in a good agreement with the in-situ results.