摘要:采用自适应的变分模态分解(adaptive variational modal decomposition,简称AVMD)与核极限学习机(kernel extreme learning machine,简称KELM)联合的方法对水闸在泄流过程中的监测信号进行振动预测分析,用以辅助决策和及时预警。首先,基于互信息准则确定AVMD的分解模态数,克服变分模态分解(variatronal modal decomposition,简称VMD)盲目选取分解参数的缺点,利用AVMD把水闸振动信号分解成K个固态模量(intrinsic mode function,简称IMF);其次,通过KELM对各IMF分量分别进行预测;最后,将各测点对应的IMFs预测结果相加作为最终的预测值。结合某水闸在自由泄流工况下的振动数据,分别采用AVMD-KELM和KELM模型、支持向量机(support vector machine,简称SVM)模型对其振动趋势进行预测,并将预测结果进行对比分析。结果表明,AVMD-KELM模型得到的预测结果与实测值更加接近,计算速度更快,精度更高,且误差较小,该方法可有效预测水闸结构的振动趋势。
注:因版权方要求,不能公开全文,如需全文,请咨询杂志社
热门期刊服务
Earthquake Engineering and Engineering Vibration Journal of Integrative Agriculture Advances in Atmospheric Sciences Chemical Research in Chinese Universities Journal of Integrative Plant Biology Chinese Journal of Integrative Medicine Journal of Central South University Journal of Zhejiang University Science A Applied Mathematics:A Journal of Chinese Universities