Content area
Impact localization is a critical area of research for structural health monitoring (SHM) systems and novel vibration-based smart building systems. For SHM systems, impact localization can play a key role in the identification and localization of damage in aerospace and naval structures, which can increase the safety and efficiency of inspections and reduce downtime and cost. Additionally, impact localization is a vital component of novel vibration-based smart building systems for applications such as: occupant monitoring, gait analysis, and fall detection. Due to the wide range of applications, many impact localization algorithms have been developed. While previous works have made improvements to impact localization methodologies, few algorithms have been developed that account for both path- and frequency-dependent wave propagation properties that occur in complex structures. Ignoring these properties can cause large localization errors. This work addresses some of the current gaps in the literature by developing, validating, and evaluating two novel localization methods, namely the frequency-based ∆T (F∆T) and output-to-output frequency response function (OO-FRF) mapping methods, which account for both path- and frequency-dependent wave propagation properties.
The F∆T and OO-FRF mapping methods are first theoretically developed and then numerically validated and evaluated for 1D and 2D impact localization using simulated impacts on finite element (FE) beam and plate models. The proposed methods are compared to the ∆T mapping method from the literature as a baseline to validate the methods. Next, the proposed methods are experimentally evaluated on beam and plate structures instrumented with accelerometers to determine their performance on real structures for both 1D and 2D localization. Finally, the Ashraf Islam Engineering Building (AIEB) local structural dynamics monitoring (L-SDM) system is validated for human building interactions that occur as impact-like events, and an impact localization study is performed in a section of the AIEB L-SDM system to test the robustness of the proposed localization methods in a realistic environment. The results demonstrate that the F∆T and OO-FRF mapping methods outperform the ∆T mapping method for 1D and 2D localization in the numerical and experimental beam and plate testbeds and are able to account for path- and frequency-dependent wave propagation properties. Additionally, the F∆T mapping method outperforms the ∆T mapping method in the AIEB L-SDM system, with an average localization error of 0.203 m (7.99 in), which demonstrates the potential for the F∆T mapping method to improve footstep localization in L-SDM systems.