论文部分内容阅读
通过相关性分析得到了煤层瓦斯渗透率与有效应力、温度、抗压强度及瓦斯压力间的相关系数,此外,对各影响因素进行互相关分析,发现彼此间存在不同程度的相关性。依据实验数据,利用MIV算法对煤层瓦斯渗透率影响因素进行优选,得到有效应力、温度和瓦斯压力参与最终BP神经网络建模。研究最终建立了2个煤层瓦斯渗透率预测模型,模型1不做影响因素优选,模型2基于影响因素优选,对模型进行试算和误差分析,结果表明:模型2具有更好的预测稳定性和精度,能很好地反映煤层瓦斯渗透率与其影响因素间隐含的映射关系。
Through the correlation analysis, the correlation coefficient between gas permeability of coal seam and effective stress, temperature, compressive strength and gas pressure is obtained. In addition, the cross-correlation analysis of each influencing factor shows that there are different degrees of correlation between each other. Based on the experimental data, MIV algorithm was used to optimize the influencing factors of gas permeability in coal seam, and effective stress, temperature and gas pressure were obtained for final BP neural network modeling. The study finally established two prediction models of gas permeability of coal seam. Model 1 did not make the best choice of influencing factors. Based on the optimization of model 2, model 2 was tested and the error analysis was carried out. The results show that Model 2 has better predictive stability and Accuracy can well reflect the implied mapping relationship between coal seam gas permeability and its influencing factors.