Intelligent Virtual Modeling for Dynamic Optimization in Large-Scale Systems

• Elvis Mondal

  Author

Modern intelligent control strategies for building energy management should improve operational efficiency while guaranteeing adaptability to changing environmental conditions by using legacy automation systems. The paper proposes a complete framework that combines data-driven digital twins with SCADA for real-time predictive control at scale in buildings. The proposed architecture allows setting up a bi-directional communication interface between EnergyPlus-based virtual models and industry SCADA through an OPC-based integration layer without altering heavy deployment infrastructures. We develop Gaussian Process (GP) models to predict building power demand and zone temperatures with uncertainty estimates that are used in an MPC formulation.  Different modes of operation are available in the framework. A few are simulation, controller validation, and real-time deployment. Thus, it allows a practical transition from simulated to real-world operations. Testing is done using the U.S. As shown by the Department of Energy Commercial Reference Building, the predictive performance using B2G’s estimate is quite good with normalized root mean square errors of 5.8% for power demand and 3.2% for zone temperature. The GP-MPC controller design delivers reliable tracking of the demand response, allowing for almost 18% energy on climate control, while maintaining occupant comfort under various operating conditions. As shown by these outcomes, integrating probabilistic digital twins with building automation systems is feasible and can optimize energy intelligently and practically.

• PDF

Copyright (c) 2026 International Journal of Intelligent Systems and Data Science

This work is licensed under a Creative Commons Attribution 4.0 International License.