传统的混凝土拱坝位移预测模型主要关注水压、温度、时效等因素与拱坝位移之间的关系,未对拱坝位移数据中所包含的信息进行充分挖掘.为此,采用 Seasonal and Trend decomposition using Loess算法(STL)将拱坝位移原始数据分解为趋势序列、周期序列及残差分量.在此基础上,采用鲸鱼优化算法(WOA)结合随机森林算法(RF)对三个分量进行预测,并使用 Holt-Winters 算法充分考虑趋势序列中的趋势信息对趋势序列的预测结果进行修正.最后将修正后的趋势序列预测结果和周期序列、残差分量预测结果相加,得出拱坝位移最终预测结果.工程实例表明,基于 STL-Holt-WOA-RF的拱坝位移预测模型能够显著提高预测的准确性和稳定性,为拱坝位移预测提供了新的思路和方法.
Abstract
The traditional concrete arch dam displacement prediction model only considers the influence of water pres-sure,temperature,and aging on the displacement,but fails to explore the information contained in the arch dam displace-ment data.To address this issue,this study utilized the Seasonal and Trend decomposition using Loess algorithm(STL)to decompose the original arch dam displacement data into trend sequences,seasonal sequences,and residual sequences.Based on this decomposition,the Whale Optimization Algorithm(WOA)in conjunction with Random Forest algorithm(RF)was used to predict the three components,and the Holt-Winters algorithm was adopted to adequately consider the trend information within the trend sequences for correction the prediction results.Finally,the corrected trend sequences prediction results,the seasonal sequences and residual sequences prediction results was superposed to obtain the arch dam displacement prediction result.The engineering example shows that the arch dam displacement prediction model based on STL-Holt-WOA-RF significantly improves the accuracy and stability of predictions,providing new insights for arch dam displacement prediction.
维普
万方数据