Fires are a frequent threat for society. Structures are commonly designed to avoid collapse in the heating phase of the fire, allowing for safe evacuation. Structural stability in the decay phase is not considered. This is a problem for fire fighter safety, and from the perspective of sustainability and resilience (buildings are not designed to survive a fire). The idea of designing buildings for decay phase safety and post-fire usability has been gaining ground, but is hindered by a series of fundamental knowledge gaps: (i) The fire decay phase is not well-described by current models; (ii) There is a lack of data on material properties for the decay phase; (iii) Structural models are validated mainly for the heating phase; and (iv) Significant uncertainties exist in these fields, and current engineering methods are not adapted to take this into account. These fundamental knowledge gaps are tackled in the current proposal. In a first phase the state-of-the-art is gathered, resulting in publicly available databases. The second phase consists of detailed fire dynamics modelling, and an experimental test campaign for the characterization of the thermal and mechanical behavior in the decay phase, focusing on concrete. Innovative bench-scale tests are used, and test results are made public. Advanced structural models are developed and uncertainties quantified. The third phase consists of the development of a novel safety-based framework for structures exposed to realistic fires.