【摘要】 针对目前尾矿坝溃坝事故高发率,即尾矿坝失稳现象严重的问题。由于尾矿坝监测信息特征存在模糊、随机、不完全等不确定性,为提取准确参数,提出了一种可能性特征参数与神经网络法相结合的失稳特征提取方法,以提取失稳特征,即描述尾矿坝动态变化的特征参数。首先在对来自传感器的数据进行可能性处理的基础上,引入新的特征参数可能性变化率、可能性歪度、可能性峭度,然后利用神经网络法研究特征参数对风险状态变化的灵敏程度,从中可提取坝体的失稳特征。最后通过实例说明:使用改进方法提取的失稳特征计算坝体稳定性运行状态得出的结果更切合实际、更具有合理性。更多还原

【Abstract】 This paper came up with a new instability extraction method based on neural network combined with possibility feature parameters to extract instability feature which describes the tailing dam dynamic changes.First,on the basis of the possibility processing of data from sensor,we introduced the possibility rate,possibility skewness and possibility kurtosis of the new characteristic parameters,then we extracted feature parameters through researching the sensitivity of feature parameters with the dam risk status changing by neural network method.Experiments prove that the result is more practical and more reasonable to calculate dam stability using instability feature from the new method.更多还原