摘要
用养护规范中17个评价指标作为输入层网络神经元,把桥梁损伤等级参数作为输出层神经元,建立了桥梁评估3层BP神经网络模型。选用湖北省110座旧桥的评估数据作为训练样本,后10个作为测试样本,经过2068次迭代运算的网络训练,得到了误差满足精度要求的收敛网络。将待评估的桥梁参数输入训练好的网络,得到评估桥梁的技术状态等级。
With 17 evaluation index of the bridge maintenance standard as input layers neurons,the bridge damage level parameters as the output layer neurons,the paper established a 3 layer BP neural network model of bridge assessment.The paper choused the evaluation data of 110 old bridges in Hubei province as the training samples,10 of them for test samples.Through 2068 iterations computing by network training,the paper obtained a convergence network,whose error meets the demand of precision.Input the bridge parameters to be assessed to the trained network,the level of bridge technical state can be got.
出处
《交通科技》
2011年第5期41-44,共4页
Transportation Science & Technology
关键词
桥梁评估
BP神经网络
桥梁技术状态
bridge assessment
BP neural network
technical state of bridges
