摘要
针对区域性地面沉降问题,文章运用MODFLOW软件中的水井子程序包,在地面沉降监测点处设置了虚拟水位观测井,以虚拟水井模拟的年均水位和统计的年开采量作为输入,各监测点的年沉降量作为输出,引入并构建了BP神经网络模型。对区内6个控沉点的沉降量预测结果表明,随着开采量的减少和水位的回升,地面沉降趋于缓和。同时,作者对BP神经网络模型的物理意义进行了有益的探索性研究。
In order to control land subsidence efficiently, the article has arranged the dummy water level observation well right on the monitoring point using the well program package of MODFLOW, input the annul water level simulated by the dummy well and the statistical annul produced quantity, output the annul land subsidence of each monitoring points, and established the BP Neural Network Model. From the predicting outcome of 6 sites, it has been found that the deep under - groundwater level is rising, the funnel area is decreasing, and the annul land subsidence tendency is tempered. At the same time, the author explored the physical significance of the BP neural network model.
出处
《矿产勘查》
2010年第6期569-575,共7页
Mineral Exploration
