Natural hazards such as windstorms, earthquakes, and floods can exert significant stresses on buildings and infrastructure that could cause failure of the structural and non-structural components and consequently overall damage to the entire building. The significance of economic repercussions of such events calls for a reevaluation of engineering approaches to resilience assessment and examination of functional integrity of civil infrastructures. To implement such an assessment, it is crucial to establish accurate yet computationally efficient frameworks for predicting the failure of building systems that incorporate both structural and non-structural elements. To this end, we employ a potential-of-mean-force (PMF) approach to lattice element method (LEM) to capture the mechanical response of structural systems. The proposed framework discretizes the system into a set of particles that interact through prescribed potential functions, to represent the mechanical properties of members. Due to its discrete nature, PMF-based LEM is well-suited for damage assessment and overcomes the limitations of continuum mechanics methods in simulating discontinuity. The framework is capable of simulating the linear and nonlinear behavior of structural components by incorporating harmonic and non-harmonic effective interaction potentials, respectively. Here we use both harmonic and non-harmonic potential and model failure by breaking bonds between particles according to an energy-based failure criterion. The calibration procedure for the harmonic potential parameters is carried out via a handshake with continuum mechanics theories, i.e. the Timoshenko beam theory for one-dimensional and Kirchhoff-Love plate theory for two-dimensional members. Such calibration is applied to non-harmonic potentials adopting section properties that encapsulate the nonlinear stress-strain responses of the materials, e.g., nonlinear moment-curvature relations. The efficiency and accuracy of the proposed method are explored through its application to quasi-static simulation of large-scale buildings subjected to various loading conditions. We then extend the analysis to the simulation of large-scale failure of structures due to the propagation of local structural damage. REFERENCES [1] E. Masoero, C. O’Shaughnessy, P. D. Gosling, B. M. Chiaia, Topology optimization using the discrete element method. part 2: Material nonlinearity, Meccanica 57 (6) (2022) 1233–1250.