基于马尔科夫链修正的RBFNN模型在碾压混凝土重力坝变形预报中的应用

Application of Improved Radial Basis Function Neural Network Prediction Model with Markov Chain to Deformation Prediction of RCC Gravity Dam

针对径向基神经网络(RBFNN)预报模型的不足,提出了一种基于马尔科夫链修正的RBFNN预报模型,以RBFNN模型的预测结果为基准,利用马尔科夫链进行误差修正,进而显著提高模型的预报精度。以某碾压混凝土重力坝的变形监测为例,建立大坝变形预报模型,并将其结果与单一的RBFNN模型的预报结果做了对比,结果表明,基于马尔科夫链修正的RBFNN预报模型精度更高,结果更符合实际。

An improved radial basis function neural network （RBFNN） prediction model with Markov chain is put forward for overcoming the shortcomings of RBFNN prediction model. Generally, the proposed model can improve the accuracy remarkably by employing the Markov chain to revise the results obtained by RBFNN model. Both the proposed model and the RBFNN model are applied to form a deformation prediction model of a gravity dam. Compared with the RBFNN model, the results show that the proposed model performs better in prediction accuracy and it accords with the actual situation.