Experimental validation of automated OMA and mode tracking for structural health monitoring of transmission towers

Abstract

This article presents a cost-effective method to monitor the structural health of transmission towers, a critical yet aging infrastructure that plays an important role in the overall reliability of the electrical grid. The method is validated experimentally on a real-world transmission tower which was subjected to several (exaggerated) damage scenarios. The proposed monitoring strategy relies on four accelerometers installed on the four faces of the rectangular base of the transmission tower. The collected vibration data is processed using a classic operational modal analysis (OMA)-based structural health monitoring scheme, comprising; automated OMA, tracking, data normalization, and decision-making. The proposed algorithm processes the four faces independently to maximize the likelihood of detecting (local) damage near the sensors in the quasi-symmetric structure. Furthermore, with widespread deployment in mind, the current article introduces a semi-automated tracking algorithm using “Density-based spatial clustering of applications with noise.” Environmental effects were removed using principal component analysis, eliminating the need for additional (environmental) sensors. Finally, Q and T² statistics were used to assess damage on each face of the structure using all tracked modes. The experimental results of this study demonstrate that this workflow can effectively track a large number of modes; in the current study, 10 modes per face of the structure, and from them detect and to some level localize the majority of structural damage-inducing events, such as the removal of a bolt or a bar.