基于改进GWO-SVR算法的大坝变形预测模型研究

Mam Deformation Prediction Model Based on Improved GWO-SVR Algorithm

大坝的工作条件复杂、影响因子众多,极易发生形变而导致灾害发生.然而,当前常见的大坝变形预测方法由于其自身的局限又无法满足大坝变形预测的精度要求.因此,提出一种基于改进GWO-SVR算法的大坝变形预测模型,通过对灰狼算法初始种群位置、收敛因子及迭代权重三个方面进行改进,再将滑动时间窗口的思想嵌入其中,最终建立模型实现对大坝变形的高精度预测.试验结果表明,标准SVR模型、GWO-SVR模型以及改进GWO-SVR模型都能够对大坝变形进行预测,其中改进GWO-SVR大坝变形预测模型具有更优秀的预测精度与稳定性.

The working conditions of dam are complex and the impact factors are numerous. It is very easy to deform and cause disasters. However,the current common dam deformation prediction methods cannot meet the accuracy requirements of dam deformation prediction due to their own limitations. Therefore,this paper proposes a dam deformation prediction model based on the improved GWO-SVR algorithm. Through improving the initial population position,convergence factor and iterative weight of the Grey Wolf algorithm,the idea of sliding time window is embedded in it,and finally the model is established to achieve high-precision prediction of dam deformation.The test results show that the standard SVR model,GWO-SVR model and the improved GWO-SVR model can predict the dam deformation,and the improved GWO-SVR dam deformation prediction model has better prediction accuracy and stability.