This paper presents a regularization scheme in time domain to reconstruct displacement history using acceleration data measured from structures. For given measured acceleration, displacement reconstruction can be defined as an elliptic boundary value ...
This paper presents a regularization scheme in time domain to reconstruct displacement history using acceleration data measured from structures. For given measured acceleration, displacement reconstruction can be defined as an elliptic boundary value problem and formulated by using least squares between measured and calculated accelerations. Unfortunately, the reconstructed displacement suffers from physically unadmissible drifts due to noise in measured acceleration and uncertainties of initial conditions. Regularization technique is employed to alleviate the unadmissible drift of reconstructed displacement. To determine an optimal regularization factor, a stable factor selection scheme is proposed. The time window concept is proposed to reduce computational costs for real time displacement reconstruction. The validity of the proposed method is demonstrated in reconstructing displacements using the numerically simulated acceleration of a two span truss bridge under earthquake and experimental acceleration of a cable with forced vibration.