PREDICTION OF DISPLACEMENT ANGLE OF FIRST FLOOR OF FRAME STRUCTURE BASED ON DEPTH LEARNING TECHNOLOGY
Based on the depth learning method,this paper first fits the moment angle curve of cast-in-place frame node,The node stiffness of the fabricated frame under the same conditions is obtained.Then a cast-in-place and fabricated frame is numerically simulated to obtain the first floor displacement angle of the two frame structures under the same seismic load.The neural network is used to simulate the two-dimensional curved surface relationship of the first floor displacement angle of the cast-in-place and fabricated frames under the same conditions.The precision of prediction results is high through finite element simulation.The network can be used to predict the first floor displacement angle of fabricated frame structure,reduce calculation cost,save time,and is more suitable for engineering practice.
NETL
NSTL
万方数据
