Development of a Mechanically informed Machine Learning framework to guide risk assessment of existing Pre-1970s RC Buildings

The Italian building stock is ill-prepared for the next seismic event. Particularly vulnerable are rein- forced concrete buildings constructed before the 1970s, which make up a considerable share of the country's built environment. Assessing the seismic performance of these structures requires extensive data acquisition and in-situ inspections, which can be time-consuming. To address this challenge, this paper proposes a framework to efficiently guide analysts in an “incremental” data collection process by leveraging a Machine Learning model informed by a mechanical/analytical simplified approach, i.e. the Simple Lateral Mechanism Analysis (SLaMA). One case study demonstrates the framework's work- flow and effectiveness.