Direction‐Aware Period Prediction for Seismic Risk Assessment and Application to an Urban Tall Building Portfolio in China

ABSTRACT Reliable estimation of the principal periods is essential for regional seismic risk assessment (RSRA) of urban building portfolios. For tall buildings, however, existing period‐prediction formulas are constrained by the lack of representative data, while building translational azimuths are neglected despite their importance for regional seismic risk. This study therefore establishes a database of 1333 tall‐building models in China, with heights ranging from 80 to 200 m, derived from real engineering projects, and develops a direction‐aware framework for period prediction and RSRA. In addition to building height, significance assessments show that seismic design intensity and lateral force‐resisting system are critical aspects governing the translational periods. An inertia‐based equivalence approach is proposed to identify the principal translational azimuths, achieving mean azimuth errors of 4.8° for plan‐regular buildings and 8.1° for plan‐irregular buildings. Based on the key parameters, including an azimuth‐dependent effective building width, period‐prediction models are developed within the proposed framework and then applied to an urban tall‐building portfolio in Guangzhou, China, through a scenario‐based seismic damage assessment. It is shown that neglecting translational azimuths can substantially underestimate seismic demands and portfolio‐level damage, with the median portfolio‐level damage ratio underestimated by up to 80% for the moderate damage state. Across the considered ground‐motion incidence angles, the ratio of the largest to the smallest median portfolio‐level damage ratio reached 1.85 for the moderate damage state, demonstrating pronounced directional sensitivity and underscoring the need to account for the combined influence of building azimuth and ground‐motion directionality in RSRA.