边坡变形的LMD-ELM多尺度预测模型研究

发布时间：2018-01-14 10:13

本文关键词：边坡变形的LMD-ELM多尺度预测模型研究　出处：《中国安全科学学报》2015年11期 　论文类型：期刊论文

【摘要】：为准确预测边坡变形,有效预防边坡灾害发生,提出构建基于局域均值分解(LMD)和极限学习机(ELM)的边坡变形多尺度预测模型。用LMD方法,将边坡变形时间序列分解为多尺度且相对平稳的随机项、周期项和趋势项。针对各项时间序列,分别构建基于ELM的预测模型。经累加各分项预测值,获得模型最终预测结果。以甘肃某边坡变形为案例,进行实证分析。结果表明:LMD-ELM模型能够充分挖掘数据内部隐含的变形规律,有效诠释多尺度变形与其诱发因素间复杂的响应关系,预测精度、运行速度和拟合泛化能力较其他模型有所提高。
[Abstract]:In order to accurately predict slope deformation and effectively prevent slope disaster, a multi-scale prediction model of slope deformation based on local mean decomposition (LMD) and extreme learning machine (ELM) is proposed. The LMD method is used. The time series of slope deformation is decomposed into multi-scale and relatively stable random term, period item and trend item. According to each time series, the prediction model based on ELM is constructed, and the prediction value of each item is accumulated. The final prediction results of the model are obtained. Taking a slope deformation in Gansu province as an example, the results show that the WLMD-ELM model can fully excavate the implicit deformation law inside the data. The complex response relationship between multi-scale deformation and its induced factors is effectively explained, and the prediction accuracy, running speed and fitting generalization ability are improved compared with other models.
【作者单位】：辽宁工程技术大学矿业学院;煤科集团沈阳研究院有限公司;
【基金】：国家自然科学基金资助(51004063) 辽宁省高等学校优秀人才支持计划(LJQ2011029)
【分类号】：TU43
【正文快照】： 0引言边坡是受多因素影响而不断发展演化的复杂非线性动力系统,在不同背景条件下,边坡变形通常表现出强烈的非线性特征[1]。边坡变形监测数据携带诱导边坡失稳及破坏因素的重要信息。因此,深度挖掘边坡变形监测数据的时变规律,研究和应用智能预测方法准确预测边坡发展趋势,对

【参考文献】