Theoretical derivation and experimental investigation of dynamic displacement reconstruction based on data fusion for beam structures

Accurately obtaining the dynamic displacement response of the beam structure is of great significance. However, it is difficult to directly measure the dynamic displacement for large structures due to the low measurement accuracy or the installation difficulty of the sensor. Therefore, it is urgent to develop an indirect measurement method for displacement based on measurable physical quantities. Since acceleration and strain contain high and low frequency displacement information respectively, this paper proposes a displacement reconstruction algorithm that can realize the data fusion of the two, which is very helpful for the research of structural health monitoring. Firstly, the stochastic subspace identification (SSI) method is adopted to calculate the strain mode, and then the displacement is derived via the mode shape superposition method. Afterwards, the strain-derived displacement and acceleration are combined by the proposed algorithm to reconstruct the dynamic displacement. Both the numerical simulation and model experiment are conducted to verify the effectiveness of the proposed algorithm. Furthermore, the influences of noise, sampling rate ratio and measurement point position are analyzed. The results show that the proposed algorithm can accurately reconstruct both high-frequency and pseudo-static displacements, and the displacement reconstructed error in the model experiment is within 5%.