Seismic Response Sensitivity to Uncertain Variables in RC Frames with Infill Walls

Seismic fragility analysis involves probabilistic assessment of the seismic performance of structures. A large amount of uncertainty is involved in the estimation of seismic fragility. The probabilistic modeling framework for performance assessment requires specifying the uncertainties in key input parameters in terms of probability distributions, sampling the distribution of the specified parameters in an iterative fashion, and propagating the effects of uncertainties through the model. The objective of this work is to identify and statistically predict the influence of uncertainty in the independent input parameters on which the seismic performance of RC buildings (with and without infill walls) depends. Random samples of the uncertain parameters are used in parametric nonlinear dynamic analyses of three variants of typical three-bay, four-story reinforced concrete (RC) frames, namely, bare frame, open ground story frame, and fully infilled frame. The relative importance of the uncertainties in different input parameters on response sensitivity is discussed using different statistical and graphical methods, such as displacement sensitivity radar charts, response sensitivity bar diagrams, tornado diagrams, Sobol′ indices, least absolute shrinkage and selection operator (LASSO) regression, and weighted pie charts. The seismic response of the bare and the open ground story frames is found to be most sensitive to the compressive strength of concrete and the column dimensions. Whereas the response of the fully infilled frames is sensitive to the infill properties, it is also observed that input variables with high uncertainty do not necessarily yield the highest sensitivity in the response.