Research On Strain-Based Displacement Reconstruction Method Of Floating Raft

In the floating raft vibration isolation system,the raft frame serves as the main bearing component,and its surface deformation monitoring is of great significance for the alignment of the components installed on it and the improvement of the stability of the components and the concealment of the ship.However,due to the raft foundation floating,the traditional displacement monitoring method is no longer applicable.In Fiber Bragg Grating(FBG)sensor used for floating raft deformation monitoring,the displacement value needs to be obtained indirectly through the straindisplacement reconstruction method.The method research has important engineering practical value for FBG used in the monitoring of floating raft surface deformation.Based on this goal,this thesis studies the reconstruction method of floating raft displacement based on modal conversion method and machine learning method:(1)The displacement reconstruction method of floating raft based on modal conversion theory is studied,and the displacement reconstruction process of modal conversion method based on finite element analysis and experimental testing is built.In the finite element verification of the modal conversion method,the finite element model of the floating raft with fixed ends is established,and the statically loaded strain / displacement FEA data and the modal analysis of the strain and displacement modes are extracted according to the established mode conversion theory,and the floating raft surface displacement is reconstructed.It is concluded that the existence of local modes reduces the reconstruction accuracy of the concerned parts of the raft.(2)In view of the influence of local modes on the reconstruction accuracy of floating rafts,the location of the measuring points is optimized,and it’s found that the location of the measuring points should be arranged at the junction of the top plate and the horizontal and vertical webs to avoid the place where the local mode is large,that is,the structural rigidity is small.At these points,the influence of local modals on the overall reconstruction accuracy is solved.The order selection and cut-off issues are studied,and it is concluded that the order selection only needs to include the order corresponding to the energy of the working condition.In the case of unknown working conditions,the order must be selected as much as possible,if the main orders have been included,increasing the cut-off order does not significantly improve the reconstruction accuracy.A finite element model of spring-supported floating raft at both ends is established,and the effect of foundation floating on reconstruction accuracy is studied,and it is concluded that the strain-based displacement reconstruction of the floating raft has rigid displacement and rotation.The relative curved surface is constructed based on the three corner points of the upper surface of the floating raft,which solves the absolute curved surfaces do not coincide.(3)The FBG experimental test system was built,and the key experimental technologies such as FBG sensing principle,number configuration,and modal parameter identification are studied.The ITD method is used to establish a steadystate graph to eliminate the scattered frequencies to identify a stable and reliable natural frequency of the system and a matching vibration mode.(4)Research on the reconstruction method of floating raft displacement based on machine learning.The strain and displacement data extracted from the floating raft finite element transient analysis were taken as samples to study the influence of BP algorithm selection and hidden layer node setting on the training error.Finally,the optimal BP algorithm and the number of hidden layer nodes were selected and reconstructed.Deformation of key measuring points and surface of floating raft.In summary,the modal conversion method and machine learning method studied in this thesis can reconstruct the displacement of the floating raft well according to the strain.