论文部分内容阅读
为准确预测工作面导水裂隙带发育高度,在总结顶板导水裂隙带高度预测方法和理论的基础上,结合大量实际资料,分析归纳出采深、煤层倾角、煤层厚度、煤层硬度、岩层结构、顶板岩石单轴抗压强度、开采厚度和采空区斜长是影响导水裂隙带高度的主要因素。根据全国典型案例,建立了基于BP神经网络的导水裂隙带高度预测模型,确定了BP神经网络所需的输入样本和测试样本,运用Matlab软件对网络进行了训练,得到了优化的网络模型,并利用建立的模型预测了焦作煤田赵固一矿11011工作面导水裂隙带高度。通过与实测结果对比,证明基于BP神经网络建立的导水裂隙带高度预测模型的计算结果比规程提供的公式计算的结果更接近实际。
In order to accurately predict the development height of hydraulic fractured zone in the working face, based on the summary of the method and theory of predicting the height of hydraulic conductivity fractured zone in the roof, combining with a large amount of actual data, the mining depth, coal seam inclination angle, coal seam thickness, coal seam hardness, , Uniaxial compressive strength of roof rock, mining thickness and gob-side miter length are the main factors affecting the height of water-conducting fractured zone. According to the typical case of the whole country, the forecasting model of water-conducted fractured zone height was established based on BP neural network, the input samples and test samples needed for BP neural network were determined, the network was trained by Matlab software, the optimized network model was obtained, The model was used to predict the height of water-carrying fracture zone in 11011 face of Zhaoguanyi Mine in Jiaozuo coalfield. Compared with the measured results, it is proved that the prediction model of the height of water-bearing fractured zone based on BP neural network is more realistic than the formula provided by the procedure.