神经网络在砖混结构健康监测中的应用

Neural network in the application of brick-concrete structure health monitoring

以砖混结构房屋为研究对象,利用具有时变适应能力的Elman动态回归神经网络,建立预测模型,通过对东南角光纤水准仪、位移传感器和梁左端应变传感器监测数据进行模拟预测,得出监测值、预测值曲线图和相对误差。

With the brick-concrete structure house as the research object. The prediction model is established by using Elman dynamic recurrent neural network which has time-varying adaptation ability. It obtains the monitoring value,forecast curve and the relative error through simulating and forecasting the monitoring data,which were collected from the strain sensor in left side of the beam and the optical fiber level and displace-ment sensors in the southeast corner of the house.