基于IFOA-GNN的土石坝施工期位移预测模型

有效预测土石坝施工期的位移对大坝安全运行具有重要意义。针对基于智能学习的大坝位移预测模型存在的不足,提出了一种基于改进果蝇算法(IFOA)与灰色神经网络(GNN)相结合的土石坝位移预测模型,即通过改进的果蝇算法迭代调整灰色神经网络的权值和阈值,得到全局最佳的初始化参数,将其结果输入灰色神经网络,并将其应用于某土石坝施工期位移预测。结果表明,该预测模型预测精度高、预测结果稳定,且精度及稳定性均优于灰色神经网络模型(GNN)和果蝇算法与灰色神经网络(FOA-GNN)预测模型。

Displacement Prediction Model of Earth Dam during Construction Period Based on IFOA-GNN

The valid displacement prediction plays an important role in the safe operation of the dam during construction period. Aiming at the shortcomings of dam displacement prediction based intelligent learning, a prediction model based grey neural network (GNN) and improved fruit optimization algorithm (IFOA) was proposed to increase its accuracy. Firstly, the global initialization parameters of the GNN is optimized by adjusting initial weights and threshold values with IFOA and the results is used as the input parameters of the GNN. Then the GNN model based improved fruit optimization algorithm (IFOA-GNN) is built to predict displacement of earth dam during construction period. Research results show that the IFOA-GNN model is good in precision and stability and is superior to GNN and FOA-GNN.