Seismic vulnerability evaluation of Peruvian aerial electrical substations with simulated fragility functions

This research uses a probabilistic methodology to evaluate the seismic vulnerability of aerial electrical substations in Perú. Peruvian electrical distribution substations are vital structures for which exists little research about their seismic vulnerability. This paper aims to be a contribution to estimate the seismic vulnerability of vital electrification lines. The proposed methodology is based on fragility functions which represent the overall structural behavior of 3D model frames for 27 aerial distribution electrical substations located in Cañete city. The Latin Hypercube technique, an enhanced Monte Carlo-base method, was used to generate fragility functions. This technique optimizes the sampling of structural parameters and provides at least 100 reliable samples for every level of seismic demand. In case of structural models, only the concrete compressive strength (f 'c), the maximum concrete strain (εcu) and the yield stress of the reinforcing steel (fy) were considered. For the concrete, mean values of 20 MPa and 0.40% were considered for compressive strength and maximum strain, respectively. For both cases, 15% coefficient of variation with a normal probability distribution function was assumed. However, for the steel, it was assumed an average value of 412 MPa, a 6.0% coefficient of variation and a lognormal probability distribution function. This is due to higher quality requirements during its fabrication. Axial force was shown to have a direct impact on curvature ductility of reinforced concrete cross sections. The seismic demand was defined by synthetic records that were compatible with the elastic Peruvian design spectrum, considering two soil types (S1 and S2) described in Peruvian seismic code. Acceleration records were scaled based on the peak ground acceleration on rigid soil (PGA) which went from 0.05g to 1.00g. A total of 2000 structural models were considered to account for both structural and seismic variability, for each soil type and aerial distribution electrical substation category. The structural models use frame elements with concentrated plastic flexure hinges. This way, the strength and stiffness degradation of the structure persists overtime. Nonlinear time-history analysis was carried out for each model subjected to a synthetic signal. This was performed with SAP 2000 program. These samples were run without user interaction through MATLAB scripts. The coupling between structural models and signals was performed with Monte Carlo simulation. On one hand, seismic vulnerability ratios for aerial electrical substations, constructed on S1 soil type, showed expected average probabilities of 99% and 88% for simple and double post substations respectively, considering a damage limit state as yielding. For the same substations, 1% and 3% ratios were computed for collapse limit state. On the other hand, considering S2 soil type, ratios showed average probabilities of 86% and 66% in yielding limit state, and 1% ratio was computed for collapse limit in both post type cases. These ratios were computed considering a seismic demand related to 10% of probability of exceedance in 50 years, which is a requirement in the Peruvian seismic code. These results show an unacceptable seismic performance for substations. They may require extensive repairing or complete post reposition.