A rapid identification framework for post-earthquake local damage of the multi-span interconnected slender equipment in substation

Abstract

Substation equipment is interconnected through conductors to form a large-scale multi-span system. During an earthquake, individual equipment pieces may suffer local damage that is not visible. Due to the structural integrity of a system, the seismic response is extremely insensitive to the local damage. Thus, the local damage is difficult to detect with traditional time-domain or vibration-mode methods. This hinders the rapid post-earthquake recovery of substations, exacerbating the earthquake-induced losses. This study is based on a generalized transfer matrix in the frequency domain to decouple and eliminate the effects of structural integrity. By inputting the measured displacement responses of equipment during an earthquake into the transfer matrix and inversely calculating, a key indicator EG, the relative error between the theoretically calculated and the actual ground motions experienced by each equipment piece can be obtained. The EG is highly sensitive to locally damaged equipment pieces and has a monotonic relationship with the degree of damage, enabling damage localization and quantitative assessment of stiffness degradation. Based on this principle, a rapid identification framework is proposed. Applying the framework to the randomly generated damage scenarios, it is demonstrated to effectively identify various damage modes, including single-component and multi-component damage within the system. The accuracy of damage localization is 100 %, and the quantitative estimation error for the damage degree is almost within ± 5 %. This framework provides a new path for rapidly identifying post-earthquake damage in substation equipment systems.