Nature frequency identification of prestressed concrete beam with RBF network

The relationship of pre-stressed concrete (PSC) beam natural frequency and pre-stressed force is difficult to described accurately with mechanical model. The past experimental data are collected. Then five unbonded post-tensioned PSC beams are designed. Frequencies and damps are collected in the dynamic experiment of five PSC beams. Radial basis function neural network is constructed to identify the natural frequencies of prestressed beam with different levels prestressing force based on previous test data and new dynamic test beam data. Then the input and output node numbers of neural network are selected and the appropriate training algorithm and expansion coefficient is determined. In order to verify that the network performance, one prestressed concrete beam test data are left to simulation test. Simulation results show that the radial basis function neural network is feasibility to recognize the frequency of PSC beams.