Sensitivities of the seismic response and fragility estimate of a transmission tower to structural and ground motion uncertainties

Abstract

Seismic fragility analysis is a powerful tool for evaluating structural seismic performance; however, this approach involves various sources of uncertainty. Electricity transmission systems are exceedingly important infrastructures that may incorporate many uncertainties; hence, this paper aims to numerically investigate the influences of various sources of uncertainty on the seismic responses and fragility estimates of transmission towers. A typical transmission tower that was previously tested at a full-scale test is selected as a case study. A tornado diagram-based sensitivity analysis is conducted to determine the relative sensitivity of the seismic responses to each uncertain parameter independently by stochastic and probability methods. The results suggest that the uncertainties in the elastic modulus, damping ratio and yield strength significantly impact the seismic responses, while uncertainties in the geometric dimensions have negligible effects on the seismic responses compared with those in the material properties. Subsequently, the Latin hypercube sampling (LHS) technique is employed to generate random samples of different uncertain parameters; these samples are then used to build uncertain models considering different levels of uncertainties. Incremental dynamic analyses (IDAs) are carried out to establish probabilistic seismic demand models (PSDMs), which are used to develop fragility curves. Comparisons among these fragility curves demonstrate that the fragility curves developed under the assumption that all structural parameters are deterministic are acceptable for concisely estimating seismic fragility; additionally, fragilities are far more sensitive to material property parameters than to geometric dimension parameters. Therefore, fragility curves can be obtained by considering only material property parameters for an exhaustive fragility estimate, which will tremendously reduce the simulation time and improve the computational efficiency.