Non-contact automated identification of earthquake-induced micro damage in substation equipment system based on local damping parameter screening with a surrogate model

Substation equipment systems may experience invisible micro-damage to local components after an earthquake, which can lead to electrical functionality failure. Such micro-damage typically has little effect on equipment stiffness or natural frequency, posing challenges to conventional damage monitoring techniques. As a result, post-earthquake inspections currently rely on manual diagnosing of each piece of equipment, making rapid recovery a challenging task. Given that micro-damage is likely to alter the damping of materials, this paper proposes a method to detect local damping variation of a system, aiding in the swift screening of potentially damaged components. The approach is based on a surrogate theoretical model developed from the motion equilibrium equations of separated components within a system and requires only measured ground motion and displacement at the top of the equipment. The method is solved in the frequency domain, effectively suppressing the influence of undamaged components and highlighting abnormalities caused by damaged parts. A key output, the indicator EG, is susceptible to local damping variation and allows for both the localization and quantification of damage. When applied to randomly generated damage scenarios, the method is shown to accurately identify different damage modes, including both single- and multiple-component damage modes. The damage localization accuracy is 100%, and the estimation error for quantifying damping variation is within ±5%. This method offers a new path for efficiently detecting minor post-earthquake damage in substation equipment systems.