改进灰色神经网络模型在形变预测中的应用

Application of Improved Grey Neural Network Model in Deformation Forecasting

针对围岩形变预测精度较低的问题,将灰色预测模型与神经网络相结合,提出了一种改进灰色神经网络模型的形变预测方法。首先,对传统灰色预测模型进行改进,采用改进灰色预测模型对原始数据进行预测;然后,将得到的拟合值和对应残差分别作为神经网络样本的输入和输出,确定网络结构;最后利用训练好的网络结构进行预测,将输出的数据与改进灰色预测模型的预测值相加,得到最终的预测结果。实验结果表明,改进灰色神经网络模型在围岩形变预测中具有较高的精度。

Aiming at the low prediction accuracy in surrounding rock deformation, we proposed an improved grey neural network prediction model which combined with the grey forecasting model and neural network. Firstly, the tra- ditional grey prediction model was improved and the improved grey prediction model was used to forecast original data roughly. Secondly, the fitted values and the corresponding residuals obtained were used respectively as the input and output sample of neural network to train the network. Finally, the trained neural network was used to predict the de- formation data. The final prediction results are the added value of output and the predicted value of the improved grey forecast model. The experimental result indicates that the improved grey neural network model has higher prediction accuracy in prediction of surrounding rock deformation.