论文部分内容阅读
采用数值模拟对正交实验设计的4个影响因素16种组合进行模拟,并且构建相应BP神经网络预测方法,同时对覆岩移动进行预测并对数值模拟结果进行对比分析,研究结果表明:正交实验可大大降低实验次数,提高工作效率,第12组实验对煤层开采上覆岩层移动的影响最大,在正交实验设计的数值模拟结果的基础上,运用神经网络对煤层开采引起的覆岩移动预测结果误差小,具有较高的可信度。
Numerical simulation was used to simulate 16 kinds of combinations of 4 influencing factors in orthogonal experimental design, and corresponding BP neural network prediction methods were constructed. Meanwhile, the movement of overlying strata was predicted and the numerical simulation results were compared. The results show that: Experiments can greatly reduce the number of experiments and improve work efficiency. The experiment of group 12 has the most influence on the overlying strata movement of coal seam mining. Based on the numerical simulation results of orthogonal experimental design, neural network is used to influence the movement of overlying strata caused by coal seam mining The prediction error is small, with high credibility.