基于VMD与优化LightGBM的混凝土拱坝变形预测

Deformation Prediction of Concrete Arch Dam Based on VMD and Optimized LightGBM

变形是反映混凝土拱坝安全运行状态的重要指标,因此变形预测模型的研究对拱坝结构健康评价具有重要意义。为充分挖掘拱坝变形监测数据的有效信息,提高监控模型的预测精度,提出一种基于变分模态分解与优化LightGBM的混凝土拱坝变形预测模型。首先,采用VMD将变形实测数据分解为多个模态分量;其次,引入改进灰狼算法与LightGBM相结合建立混凝土拱坝变形预测模型;随后,对模态分量进行单独建模和预测,最后叠加以得到最终的预测结果。工程实例分析表明,通过有效地分解重构,构建的变形预测模型具有较高的预测精度和泛化性能。

Deformation is an important index to reflect the safe operation state of concrete arch dams.Therefore,the study of deformation prediction model is of great significance to the structural health evaluation of arch dams.In order to fully excavate the effective information of arch dam deformation monitoring data and improve the prediction accuracy of monitoring model,a deformation prediction model of concrete arch dam based on variational mode decomposition and optimization LightGBM is proposed.Firstly,the measured deformation data is decomposed into multiple modal components by using VMD.Secondly,the improved grey wolf algorithm combined with LightGBM is introduced to build the deformation prediction model of concrete arch dam.Then,the modal components are modeled and predicted respectively,and the final prediction result is obtained by superposition.The analysis of engineering examples shows that the deformation prediction model constructed by effective decomposition and reconstruction has high prediction accuracy and generalization performance.