The construction and optimization of resilient community living sphere driven by digital twins

Against the backdrop of global climate change and digital transformation in urban planning, enhancing community risk perception capabilities and developing the potential of community technology empowerment have become important issues in responding to sudden disasters and long-term risks. This study integrates digital twin technology with the concept of Community Resilience Living Sphere. A five-layer digital twin framework: "Terminal-Data-Model-Scenario-User" is established to achieve bidirectional physical-virtual coordination, balancing routine service provision with emergency response demands. Community material and perceptual resilience data were collected through mobile sensing and questionnaires. By combining Random Forest (RF) algorithms with MassMotion evacuation simulations, we validated the optimization effects of resilience factors under dual operating states. The empirical validation conducted in the Bajiao short-term resilience living sphere of Beijing's Shijingshan District demonstrated that post-optimization evacuation completion time decreased by an average of 22%, the crowd density reduction rate increased by an average of 40%, and the 15-minute evacuation efficiency improved by an average of 22%.This research advances digital twin applications in urban resilience planning, providing a replicable paradigm for mitigating systemic urban risks.