Simulating the fragmentation and demolition of structures requires handling complex interactions between large amounts of debris as they separate and collide. To efficiently address these contact-rich scenarios, we adopt a Position-Based Dynamics (PBD) approach based on the Parallel Particles (P²) Method, a framework specifically designed for large-scale simulations and well suited to parallel computation. Although the P2 method was originally designed for systems of independent particles, many applications require the representation of rigid bodies. To address this, we model rigid bodies as clusters of particles connected through bilateral constraints. These can include constant-distance and alignment constraints, which together preserve the relative configuration of particles and enforce rigid behaviour. For example, as shown in Fig. 1, bricks can be modelled using three particles: two constant-distance constraints between particles 1-2 and 2-3, and one alignment constraint between the vectors 1to2 and 2to3. This configuration allows us to simulate the collapse of a wall. As a preprocessing step, the user has to define the particle groups that form each rigid body and has to specify the constraints that bind them. The simulation then proceeds as in standard P2 method: contact detection, constraint generation, position correction (including both contact and bilateral constraints), and position updates. To maximize parallelism, particles belonging to the same rigid body can be processed together within a single computational thread. This allows internal constraints to be solved efficiently. The proposed method is general and applicable to a wide range of scenarios involving interacting rigid elements. One potential application is the simulation of structure fragmentation and debris handling in demolition environments. These simulations can also be used for training purposes—for example, to help personnel or equipment operators better understand how structures behave and break apart in realistic scenarios.