| Abstrakt: |
Fault identification in a structure using the vibration characteristics is a quiet smart technique for health monitoring of the structure. In this research, the radial basis function neural network (RBFNN) method has been implemented for damage detection in the rotating circular cross-section cantilever type rotor immersed in the altered viscous medium environment as an inverse method. The proposed radial basis function neural network (RBFNN) Controller has five inputs and four outputs parameters. The database for neural network has been developed from the experimental, theoretical and FEM analysis. Theoretical expression has been developed to estimate the influence of the position and size of damage on the vibration characteristics. For damage cantilever rotor with extra mass at the free end, the vibration signature is computed in both transversal directions using the influence coefficient approaches. The effect of fluid forces is obtained by using the Navier–Stokes equation. Local stiffness has been calculated to employ the Strain energy release rate at the crack region of the rotor shaft. The finite element software ANSYS is used for numerical analysis on cracked and non-cracked cantilever rotor shaft to estimate the vibration signature such as natural frequencies and amplitude of Vibration. To determine the robustness of the proposed neural network technique, the computed results from the radial basis function neural network (RBFNN) controller are validated with the experimental, theoretical, and FEM analyses results. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink