基于融合型WNN的大坝裂缝开合度预测模型研究

The Appliance of Wavelet Neural Network Model in Dam Safety Monitoring

大坝的裂缝开合度监测是大坝安全监测中重要的项目之一。因此,大坝裂缝开合度预测的准确性对大坝安全监控十分重要。外界诸多因素都会对大坝裂缝开合度造成一定的影响,导致情况非常复杂。为了提高大坝裂缝开合度预测的精度,尝试将融合型WNN(小波神经网络)应用于大坝裂缝开合度预测,并将该模型应用于某混凝土大坝的裂缝开合度预测中,并与BP神经网络模型、松散型WNN模型及传统的多元回归模型预测结果进行对比。结果表明,融合型WNN用于大坝裂缝开合度预测精度更高,效果更好。

Crack openness monitoring of the dam is an important item in dam safety monitoring. Theretore, the accuracy ol the pre diction is of great importance to the safety assessment of dams. Crack openness of the dam is influenced by lots of external factors, which makes the circumstances complex. In order to improve the accuracy and reliability of the prediction of the crack openness mo- nitoring data, this paper tries to apply the Wavelet Neural Network Model which integrates the Wavelet Analysis and the Artificial Neural Networks with the prediction of dam crack openness, and the model is applied to the crack openness of an actual dam. The prediction results are compared with the results obtained from BP neural networks, the Relaxing Wavelet Neural Network Model and the conventional Multiple Regression Model, which shows that the Wavelet Neural Network Model can predict the dam crack open- ness more accurately.