基于MATLAB的Elman神经网络在大坝位移预测中的应用

Application of elman neural network and MATLAB to dam displacement prediction

根据MATLAB的Elman神经网络模型能够逼近任意非线性函数的特点和具有反映系统动态特性的能力,提出了一种利用Elman神经网络建立大坝位移预报模型的方法。通过实例比较Elman模型与BP模型的预测效果,结果表明,所建立Elman神经网络模型不仅收敛速度快,且其预测精度高于BP网络模型。利用MATLAB神经网络工具箱能使工作者从繁琐的编程中解脱出来,有效的提高工作效率。

Based on the fact that the Elman model of neural network tool-box in MATLAB can well approach any nonlinear continuous function and has ability to reflect dynamic features of the systems, a method of dam displacement prediction by Elman model is presented. Through the displacement prediction by Elman model and BP model, the results show that the convergence and accuracy of prediction by the Elman model is higher than that by the BP network model. Taking advantage of network tool-box in MATLAB, Analysts can free themselves from editing hard program and raise working efficiency.