Improving accuracy of estimating building capacity curves from acceleration data using SDOF analysis

Displacement response is critical data for post-earthquake fast building assessment. However, directly measuring building displacement response remains a great challenge in practice. Therefore, it is practically common to estimate displacement response from acceleration response using double integration. Unfortunately, the low-frequency component of displacement obtained by double integrating acceleration often contains noise that is indistinguishable from low-frequency displacement components. Consequently, the maximum displacement estimated from acceleration is commonly underestimated in comparison to its true value due to the removal of the low-frequency components. This can potentially lead to an underestimation of post-earthquake building damage state, especially when a building undergoes significant nonlinear deformation. This study develops a framework to improve the accuracy of the maximum displacement obtained from floor acceleration data by fusing it with the low-frequency displacement component estimated from an equivalent SDOF analysis for both reinforced concrete (RC) and steel structures. Procedures for constructing the equivalent SDOF model of a building and a procedure for extracting the low-frequency component from the analysis displacement were developed. The proposed method was verified using a diverse range of case studies from numerical simulation and experimental studies under different seismic records for both RC and steel structures. The results showed that the proposed method was effective with a range of seismic characteristics and damage levels. A significant reduction in maximum displacement errors was observed for cases with significant nonlinear deformation, which may contribute to a more accurate post-earthquake building assessment.