摘要
针对半周期或无周期性变化规律的裂缝,基于非线性理论建立了一个三层多输入单输出的NBP人工神经网络新模型,该模型具有复杂的非线性映射功能,通过网络的自学习能自动地建立权重分配关系。实例表明,此新模型具有较高的预报精度,在大坝安全分析中有较广阔的应用前景。
Aiming at periodical and nonperiodical crack change rule, on the basis of nonlinear theory, an NBP artificial neural network model with three layer, multiinput and single output is built. This neural network model can acquire weight by itself and modify constantly the connections between the cells. The weightassigned connections are automatically built by the artificial neural network. The result shows that the new model is very useful in research of dam crack openings data analysis.
出处
《水电能源科学》
2003年第2期16-18,共3页
Water Resources and Power
基金
国家自然科学基金重点资助项目(50139030)
关键词
大坝
人工神经网络
裂缝开度
因子分量
dam
artificial neural network
crack openings
gene weight
