Model uncertainty applied to the failure analysis of transmission towers
R.R. Roman,L.F.F. Miguel,F. Alminhana
DOI: https://doi.org/10.1016/j.engfailanal.2024.108023
IF: 4
2024-02-08
Engineering Failure Analysis
Abstract:The structural design of self-supporting transmission line towers (TLTs) typically employs first-order elastic analysis to streamline practical engineering. Accordingly, due to their serial industrial production, prototype tests are often conducted to validate the adopted simplifications. Alternatively, developing robust and reliable nonlinear structural models to simulate progressive collapse could reduce the need for costly and time-consuming prototype test programs. In addition, a more comprehensive failure description is achieved by treating the mechanical model parameters as random variables. This approach leads to describing the load factor associated with tower collapse through probability density functions instead of relying on a single deterministic value. Although significant progress has been made in modeling the progressive collapse of lattice TLTs using both deterministic and stochastic approaches, critical gaps in this field remain unexplored. For instance, the existing literature has not yet addressed the modeling of stiffness degradation caused by residual stresses in tower angle sections, even in a deterministic approach. Furthermore, the few studies employing state-of-the-art structural models (e.g., including the bolt slippage effect) while simultaneously considering uncertainties have been limited to assessing numerical models without full-scale experimental validations. Hence, the present paper presents a novel structural model uncertainty assessment for the failure analysis of lattice TLTs. Originality arises from including the stiffness degradation caused by residual stresses in a numerical simulation that addresses all the remaining modeling complexities influencing tower collapse. Yet, the numerical framework is validated using three pre-existing prototype tests of TLTs available to the authors. For this purpose, up to eight distinct structural models are constructed, progressively incorporating different modeling complexities. This methodology allows for an examination of the impact of each modeling aspect on tower collapse simulation. The results indicate that numerical predictions align with prototype test results only when all modeling features are incorporated. In this scenario, the corresponding experimental load factors fall within the range determined by the uncertainty analysis. Finally, the proposed failure analysis tool is readily adaptable to assess other pertinent real-world engineering applications, including the cascading collapse of complete transmission line segment models under wind or seismic loads.
engineering, mechanical,materials science, characterization & testing