基于BP神经网络的多测点监测模型在大坝安全监测中的应用

Multi-pointsmonitor model on BP neural network in the application of dam safety monitoring

建立了基于BP神经网络的多测点监测模型,为了验证该监测模型的有效性和精确性,通过具体的工程计算实例,将该模型的计算结果与统计回归单测点监测模型、BP神经网络单测点监测模型的计算结果比较。对比分析表明:对于TP8测值,统计回归模型平均误差为3.8%,BP神经网络单测点监测模型平均误差为9.4%,而BP神经网络的多测点监测模型平均误差仅为1.6%。因为BP神经网络的多测点监测模型考虑了各种效应量之间的相关性,预测结果比另外两种模型预测结果好,在大坝监测预测预报中具有一定的应用价值。

This paper established a multi-pointsmonitor model on BP neural network for analysis of dam monitoring data,and contrast it with the traditional single-point regression model and single-point Model on BP neural network in a project.The results showed that the multi-pointmonitor model on BP neural network takes into account the correlation of various effects and the predicted results are better than the traditional single-point regression model and single-point model on BP neural network.So it has great value in the dam monitoring and prediction.