Papers and ongoing research from the OpenCEM project at CUHK-Shenzhen.
Introduces the OpenCEM Simulator, the first open-source digital twin explicitly designed to integrate rich, unstructured contextual information with quantitative renewable energy dynamics. Includes a unique language-rich dataset and modular component-based architecture.
A novel framework that enhances traditional MPC by integrating real-time contextual information through a Language-to-Distribution (L2D) module, translating natural language context into predictive disturbance trajectories for the MPC optimization.
The microgrid continues to operate, producing new measurement rows. Future releases will cover a wider range of operational conditions and workloads.
More sophisticated lithium-ion battery and PV generation models capturing nonlinearities and degradation effects.
Integration with reinforcement learning frameworks to provide a benchmark environment for optimization, in-context reasoning, and RL-based control.
Expanding the physical installation with additional servers to enrich load diversity and contextual information patterns.