基于MA-RF的混凝土坝变形预测模型及其应用

Deformation Prediction Model of Concrete Dam Based on MA-RF

针对传统随机森林算法在参数寻优及性能验证方法方面的不足,提出了一种基于蜉蝣优化算法和随机森林(MA-RF)的混凝土坝变形预测模型。以某混凝土重力坝变形监测为例进行建模分析,发现在参数寻优方面,MA算法的寻优精度明显优于经验法和PSO算法,且收敛速度更快;在预测性能方面,相比EM-RF、PSO-RF、LSTM、SVM,MA-RF模型的预测精度更高、稳定性更强,为高精度预测大坝变形提供了一种新方法。

Aiming at the shortcomings of traditional random forest algorithm in parameter optimization and verification methods of performance,a concrete dam deformation prediction model based on mayfly optimization algorithm and random forest(MA-RF)was proposed.Taking the deformation monitoring of a certain concrete gravity dam as an example for modeling analysis,it is found that in terms of parameter optimization,the accuracy of MA is obviously better than that of empirical method and PSO algorithm,and the convergence speed is faster;In terms of prediction performance,compared with EM-RF,PSO-RF,LSTM and SVM,MA-RF model has higher prediction accuracy and stronger stability,which provides a new method for high-precision prediction of dam deformation.