基于灰色BP神经网络的混凝土连续刚构桥应力预测

为了对混凝土连续刚构桥施工过程中的应力进行控制与预测，保证桥梁施工的安全，根据灰色理论和BP神经网络建立了灰色BP神经网络应力预测模型GNNM（1，1）．以贵阳花溪特大公路刚构桥施工过程实测应力为输入数据，对模型GNNM（1，1）进行训练，得到了应力预测值．该预测值与检验值的相对误差在3％以内，表明GNNM（1，1）模型的预测精度高，可以对混凝土连续刚构桥施工进行短期应力预测，同时，对大桥应力控制及安全施工具有一定的参考价值．

Prediction of Stress of Concrete Rigid Frame Bridge Based on Gray BP Neural Network

In order to predict and control stress in processing of concrete rigid frame bridge and guarantee safety of bridge construction, a kind of gray BP neural network model is constituted based on the gray theory and BP neural network. Taking the actual stresses of Huaxi Concrete Rigid Frame Bridge in processing of construction as input data, the GNNM（1,1） model is trained; and the stresses of prediction values are obtained. Relative error is less than 3 ~ between the prediction and the test values so as to show that the model GNNM（1,1） has a high prediction accuracy; it can be used to predict the recent stresses in processing of concrete rigid frame bridge construction; and it has simultaneously a certain reference value for stress control and safety construction.