基于神经网络的矿山地质灾害评估模型研究

Ming Geological Hazard Based on Neural Network

目前矿山工程的地质灾害评估与预测模型只依靠某单一的地质因素进行分析,无法融合造成地质灾害的众多因素,形成准确率高的全面性的地质灾害评估模型。提出一种基于神经网络的矿山地质灾害评估模型,运用神经网络强大的非线性因素融合与推断的能力建立评估模型,将海量的历史监控数据输入到神经网络中不断进行模型训练,最终形成分级别的矿山地质灾害预测模型,使用模型对测试样本检验发现,这种技术对灾害的预测准确率高达83%,具有很强的实用性。

At present the mine engineering geological hazard assessment and prediction model only rely on a single geological factors analysis, can＇ t fusion cause geological disaster, many factors forming accuracy high comprehensive geological hazard assessment model. This paper puts forward a neural network based on the mine geological hazard assessment model, using neural network strong nonlinear factors fusion and inference ability to establish evaluation model, the magnitude of the historical monitoring data input to the neural network model of ongoing training, and finally form a hierarchical other mine geological hazard prediction model, use the model to test sample inspection found that the technology to disaster prediction accuracy reaches as high as 83%, have very strong practicability.