Research

I. Renewable Energy Grid Management and Control

The growing penetration of renewable energy resources poses challenges to modern power systems due to the variability of generation and the dynamic nature of demand. Addressing these challenges requires advanced methods for supply–demand balancing, voltage stability, and reliable system operation. Our lab focuses on developing AI-based approaches to optimize energy management under grid power flow constraints while incorporating demand response. We also study inverter-based Volt–Var control and AI-enabled mitigation strategies to regulate voltage and maintain power quality. Through these efforts, we aim to advance intelligent control methodologies that enhance the stability, reliability, and efficiency of renewable-integrated power systems.

III. AI-based Forecasting, Anomaly Detection and XAI Analysis

The variability of renewable generation and the uncertainty of load demand pose significant challenges to reliable power system operation, making accurate forecasting and robust anomaly detection increasingly crucial. Moreover, the black-box nature of conventional AI models limits interpretability, which can hinder operator trust and practical deployment in critical energy systems. In response, our lab conducts research on developing task-specific AI-based forecasting methods to improve the accuracy of renewable and load predictions, alongside anomaly detection strategies to support stable and resilient system operation. We also investigate explainable AI (XAI) techniques to enhance model transparency by quantifying feature contributions, thereby increasing the interpretability and trustworthiness of AI-driven decision-making in renewable-integrated power systems.

V. Nano-Grid Control

The increasing adoption of DERs, electric vehicles, and smart appliances has created a growing need for localized energy management systems. Nano-grids, operating at the scale of households, buildings, or small communities, provide a promising solution by enabling the flexible integration of renewable generation, energy storage, and controllable loads while also considering thermal comfort of end-users. Capable of functioning in both grid-connected and islanded modes, nano-grids enhance energy self-sufficiency, reliability, and resilience. The Grid AI Lab conducts research on advanced nano-grid control strategies that utilize deep learning and deep reinforcement learning to optimize energy flows among local generation, storage, and electricity and heating demand. Our work emphasizes intelligent demand response, real-time control, and seamless operation under variable renewable conditions. Thus, we aim to develop scalable control methodologies that improve the efficiency, autonomy, and sustainability of future energy systems at the nano-grid level.