Gaussian Kernel Methods for Seismic Fragility and Risk Assessment of Mid-Rise Buildings

Seismic fragility functions can be evaluated using the cloud analysis method with linear regression which makes three fundamental assumptions about relation between structural response and seismic intensity: log-linear median relationship, constant standard deviation, Gaussian distributed errors. While is a popular method, degree to these individual compounded affect risk of mid-rise buildings needs systematically studied. This paper conducts such study considering building archetypes that make up bulk stock: RC moment frame, steel wood shear wall. kernel methods are employed capture data-driven variations in deviation distributions residuals intensity level. With reference kernels approach, it found while may not lower damage states, this conclusion does hold for higher states (such as Complete state). In addition, effects on evaluated. For predicting demand hazard, impact computed larger values. However, loss hazard downtime decision variable, considered adequate all practical purposes.