Mobile robots play an important role in search and rescue operations in post-disaster scenarios such as earthquakes and hurricanes due to the hazardous nature of these sites. Mobile robots can use Lidar devices to scan affected infrastructure and identify the hazard level of different buildings on the disaster site. This research proposes a method for mobile robots to efficiently scan a disaster site and detect damaged regions based on the acquired 3D point clouds. Disaster scenarios such as earthquakes and explosions are first simulated on a CAD model of a nuclear power plant using rigid body physics. 3D point clouds are obtained from the simulated sites using ray-tracing from a virtual mobile robot. Finally, the damage extent is estimated by comparing the acquired point cloud to the original 3D model. Validation results showed that the method achieved 43%-100% precision and 40%-82% recall rate averaged over four scenarios. The proposed research has potential to improve the safety and efficiency of search and rescue operations on disaster sites.