Abstract
Extreme loading conditions due to natural hazards such as windstorms, earthquakes, and floods can lead to the failure of both structural and non-structural elements and, subsequently, damage to the entire structural system. The significant economic repercussions of such events call for the revisiting of engineering approaches to resilience assessment towards the examination of the functional integrity of civil infrastructure. This assessment requires the development of accurate yet computationally efficient frameworks to model the failure of systems comprising structural and non-structural components. To this end, we leverage a Potential-of-Mean-Force (PMF) approach to Lattice Element Method (LEM), a class of discrete methods demonstrated to be particularly advantageous for simulating failure and fracture, to capture the mechanical response of structural systems in both linear and nonlinear regimes. The premise of the proposed framework is to discretize the system into a set of particles that interact with each other through prescribed potential functions, which represent the mechanical properties of different types of members. Lending itself to damage assessment due to its discrete nature, our PMF-based LEM transcends the limitations of continuum mechanics approaches and enables the incorporation of a range of effective interaction potentials to simulate the linear and nonlinear behavior of structural components. Since the determination of elastic behavior is a precursor to the failure analysis of structural components, harmonic potentials are initially adopted to model the linear response of various structural members, e.g., beams, columns, roofs, and walls. The calibration procedure for such potentials is thus carried out via a handshake with continuum mechanics theories, e.g., the Timoshenko beam theory and Kirchhoff-Love plate theory. This calibration is then carried out for non-harmonic potentials by adopting section properties that encapsulate the nonlinear stress-strain responses of the materials, e.g., nonlinear moment-curvature relations. Upon the calibration of non-harmonic potentials, ductile failure of the structural members is modeled by breaking bonds between particles according to an energy-based failure criterion. Finally, the utility and accuracy of the proposed framework are demonstrated through its application in (i) both quasi-static linear and nonlinear simulations of large-scale buildings under different loading conditions and (ii) the simulation of progressive structural failure due to the propagation of local structural damage.