Digital Twin prototype of a 33/11kV substation
Traditional power distribution networks within developing energy sectors frequently rely on legacy monitor-<br>ing systems or completely lack digitized telemetry, which severely limits situational awareness and delays<br>fault detection. To address these operational challenges, this study designs, prototypes, and validates a scal-<br>able cyber-physical digital twin architecture for a 33/11 kilovolt electrical substation, utilizing an active<br>utility node in Nepal as a case study. The proposed framework integrates three cohesive tiers: a low-cost,<br>distributed edge-computing hardware layer, a real-time three-dimensional human-machine interface built<br>inside a high-fidelity game engine (Unreal Engine 5), and an analytical forecasting core. By implementing<br>optimized bidirectional data routing protocols, the architecture achieves reliable, sub-second synchronization<br>between the physical equipment and the virtual testbed. This setup enables operators to monitor live power<br>fluctuations through dynamic visual changes and instantaneously actuate physical isolation switches directly<br>from the virtual environment. While advanced neural networks experienced processing limitations when<br>handling highly volatile load configurations, a non-parametric probabilistic regression model demonstrated<br>mathematical precision for multi-step ahead load trend projections. Finally,this work demonstrates that cost-<br>effective, open-source digital twins can successfully bridge the divide between legacy power infrastructure<br>and real-time interactive environments, offering an accessible and practical paradigm for grid modernization<br>in developing nations.