Guidelines (GEM-ASV) for developing analytical seismic vulnerability functions are offered for use within the framework of the Global Earthquake Model (GEM). Emphasis is on low/mid-rise buildings and cases where the analyst has the skills and time to perform non-linear analyses. The target is for a structural engineer with a Master’s level training and the ability to create simplified non-linear structural models to be able to determine the vulnerability functions pertaining to structural response, damage, or loss for any single structure, or for a class of buildings defined by the GEM Taxonomy level 1 attributes. At the same time, sufficient flexibility is incorporated to allow full exploitation of cutting-edge methods by knowledgeable users. The basis for this effort consists of the key components of the state-of-art PEER/ATC-58 methodology for loss assessment, incorporating simplifications for reduced effort and extensions to accommodate a class of buildings rather than a single structure, and multiple damage states rather than collapse only considerations.

To inject sufficient flexibility into the guidelines and accommodate a range of different user needs and capabilities, a distinct hierarchy of complexity (and accuracy) levels has been introduced for (a) defining index buildings, (b) modelling, and (c) analysing. Sampling-wise, asset classes may be represented by random or Latin hypercube sampling in a Monte Carlo setting. For reduced-effort representations of inhomogeneous populations, simple stratified sampling is advised, where the population is partitioned into a number of appropriate subclasses, each represented by one “index” building. Homogeneous populations may be approximated using a central index building plus 2k additional high/low observations in each of k dimensions (properties) of interest. Structural representation of index buildings may be achieved via typical 2D/3D element-by-element models, simpler 2D storey-by-storey (stick) models or an equivalent SDOF system with a
user-defined capacity curve. Finally, structural analysis can be based on variants of Incremental Dynamic Analysis (IDA) or Non-linear Static Procedure (NSP) methods.

A similar structure of different level of complexity and associated accuracy is carried forward from the analysis stage into the construction of fragility curves, damage to loss function definition and vulnerability function derivation.

In all cases, the goal is obtaining useful approximations of the local storey drift and absolute acceleration response to estimate structural, non-structural, and content losses. Important sources of uncertainty are identified and propagated incorporating the epistemic uncertainty associated with simplifications adopted by the user. The end result is a set of guidelines that seamlessly fits within the GEM framework to allow the generation of vulnerability functions for any class of low/mid-rise buildings with a reasonable amount of effort by an informed engineer. Two illustrative examples are presented for the assessment of reinforced-concrete moment-resisting frames with masonry infills and unreinforced masonry structures, while a third example treating ductile steel moment-resisting frames appears in a companion document.