【摘要】 在综合分析地表下沉影响因素的基础上,采用不同的BP-ANN算法建立了地表下沉的预计模型,运用某矿区的实测数据,对各模型进行了训练和性能测试,并对各网络模型进行了对比分析。结果表明,采用BP-ANN模型预计地表下沉的结果是合理的,该方法减少了预计中的人为因素,将复杂问题简单化,预计结果准确可靠,具有一定的应用价值。更多还原

【Abstract】 Based on comprehensive analysis of the factors influencing surface subsidence,a prediction model for surface subsidence was established by different methods of back propagation artificial neural network(BP-ANN).A large amount of data obtained at observation stations in a mining field was used for training and performance test of BP-ANN models,and comparison and analysis on these models were made. The results indicated that it is much more reasonable to use BP-ANN model to predict the mining subsidence.The method limited artificial factors during the prediction,simplified the complex problems and made the solution more reasonable and creditable.So,it has a certain application value.更多还原