In modern microgrids, battery energy storage systems are increasingly required to respond rapidly to power dispatch commands while operating under complex and uncertain conditions. Simply tracking power references without considering battery health can accelerate degradation, increase fault risks, and ultimately compromise system reliability. As battery systems become larger and more critical, health-aware control is no longer optional—it is essential.

At ADAC Lab, we focus on the development of an intelligent Battery Agent that bridges high-level power coordination and low-level electrochemical dynamics. Among them, one core research direction is: it converts external power demands into optimized input currents that not only satisfy operational requirements, but also explicitly account for battery health and safety constraints.

This function is built upon a physics-informed battery model that captures both electrical behavior and degradation mechanisms. By explicitly modeling the negative electrode potential and the SEI side-reaction current as state variables, the Battery Agent gains situational awareness of internal health dynamics that are invisible to conventional equivalent-circuit-based controllers. This enables the agent to make informed control decisions that balance performance and longevity in real time.

By leveraging health-aware optimization and predictive control, the Battery Agent enables intelligent and cooperative energy management across distributed systems. In particular, this function plays a critical role in the following applications:

• Health-Aware Dispatch: Converts power commands into input currents that minimize harmful operating conditions, reducing excessive overpotential and suppressing accelerated SEI growth.
• Fault Risk Reduction: By constraining internal electrochemical states, the controller lowers the probability of incipient faults and improves operational safety under high-load or fast-response scenarios.
• Resilient Microgrids: Enables battery systems to participate in coordinated energy management while preserving long-term availability, supporting resilient operation during disturbances or emergency conditions.
• Scalable Agent-Based Control: Provides a unified interface between system-level coordination and cell-level health dynamics, facilitating seamless integration into multi-agent microgrid and virtual power plant architectures.

The Battery Agent transforms battery control from a purely power-tracking task into a health-aware decision-making process, forming a key building block for intelligent, resilient, and sustainable energy systems.

Lithium-ion battery performance and long-term reliability are strongly influenced by incipient degradation mechanisms that emerge and gradually evolve during operation. These faults, such as solid electrolyte interphase (SEI) growth, metal dendrite growth, and particle cracking within electrode materials—often evolve silently before manifesting as measurable capacity loss, impedance growth, or safety risks. Accurately capturing these early-stage degradation processes is essential for understanding failure pathways and improving battery durability.

Our research focuses on developing fuzzy logic–based modeling methodologies for representing battery incipient faults in an interpretable and computationally efficient manner. Traditional physics-based models, while accurate under idealized conditions, are often computationally expensive and rely on detailed material data, limiting their scalability and adaptability. Fuzzy logic offers a complementary approach by enabling the structured translation of qualitative degradation knowledge into quantitative surrogate models through transparent, rule-based representations.

As a representative case study, we investigate mechanical degradation arising from particle cracking in lithium-ion battery anodes. A fuzzy inference system is constructed to model maximum crack-induced damage, with systematic refinement of membership function granularity used to balance model structure, representational fidelity, and interpretability. This methodology-driven approach demonstrates how fuzzy systems can capture key degradation trends while remaining analytically transparent, providing a foundation for extending the framework to other incipient fault mechanisms in lithium-ion batteries.

After disasters, restoring power to critical facilities like hospitals is an urgent priority. However, disasters often severely damage both power and communication infrastructure, making restoration efforts particularly challenging when communication resources are limited. Therefore, how to optimize the communication network of a microgrid to achieve high-efficiency, and resilience is a problem worth investigating.

To address the problem, we have applied edge centrality metrics to identify critical links as well as nonessential links. We have shown that Connectivity Rank Index (CRI) is an effective centrality metric to measure the importance of communication links for distributed structure. Based on this, we also proposed a CRI-based network optimization strategy by pruning. In addition, we defined a variant of CRI, called Local-CRI, to estimate the impact of link failures on hierarchical distributed structure.

In our future work, we will explore the integration of edge centrality metrics from complex network filed and physical constraints from power system. The strategic application of edge centrality metrics can be a practical and quantitative tool for us to design more efficient and resilient communication networks for microgrids in disaster relief.

Publications:

[1] H. Tian, A. Joshi and M. -Y. Chow, “Impact of Communication Link Failures on Distributed Energy Management in Disaster Relief Microgrids,” IECON 2025 – 51st Annual Conference of the IEEE Industrial Electronics Society, Madrid, Spain, 2025, pp. 1-6.

[2] H. Tian, A. Joshi, M.-Y. Chow, J. Zhou, and G. Chen, “Microgrid communication network optimization during disaster relief: A connectivity rank index based approach,” in 2025 IEEE International Conference on Industrial Technology (ICIT), 2025, pp. 1 – 6.

The increasing integration of distributed energy resources (DERs)—such as renewable generation, energy storage systems, and responsive loads—has introduced significant variability and complexity into modern power systems. As the scale of controllable devices grows, traditional manual configuration and static management frameworks impose a heavy cognitive burden on grid operators, making it increasingly difficult to manage the system efficiently and respond to dynamic changes in real-time.

Building on this foundation, our current work focuses on the novel paradigm of Intent-Based Energy Management System. This approach bridges the semantic gap between human operational goals and machine-level control:

1. Intent Translation & Mapping: The system automatically parses abstract Operator Intents into executable control policies for physical and virtual resources.
2. Closed-Loop Verification: By establishing a feedback mechanism between the infrastructure and the intent translator, the system ensures continuous alignment with operational goals.

This approach fundamentally transforms grid management by shielding operators from the complexity of underlying physical resources. It allows for intuitive, goal-oriented supervision, ensuring that the power system remains manageable, resilient, and responsive even as complexity scales.

Our research focuses on rapid power restoration after extreme events, when distribution lines are physically damaged and conventional recovery actions become ineffective. In such post-disaster conditions, the system must operate in a black-start manner using resources available within the affected area, while transportation disruptions fundamentally constrain what can be deployed and where.

We propose a paradigm shift from “power flow scheduling on a fixed topology” to active topology construction: building an ad-hoc, resilient microgrid by strategically adding links in both the physical and cyber layers. On the physical layer, mobile energy assets (e.g., electric vehicles and mobile storage) are dispatched to designated sites and, once parked and connected, serve as controllable power supply and demand resources; meanwhile, robotic or crew-assisted temporary cabling can create temporary electrical connections where fixed interconnection infrastructure is unavailable or damaged. This enables restoration even when latent tie-lines or reconfiguration alone cannot bridge severed segments.

A central goal is to make these decisions fast and scalable under tight resource limits. We therefore develop metric-guided strategies for Physical Link Addition and Communication Link Addition: selecting which electrical connections to build to recover critical loads, improve networked topology stability, and control transport costs; and selecting communication edges to enhance algebraic connectivity and accelerate distributed control convergence. Our ongoing work further integrates transportation and energy coupling in complex urban environments, alongside learning-based decision policies and hardware-in-the-loop validation.

With the rapid growth of global industry and economy, the energy area is undergoing significant transformation with a notable increase in the penetration of distributed energy resources (DERs). Microgrid technology plays a crucial role in such transition by enabling the rapid and efficient integration of DERs. However, under highly dynamic and complex operating scenarios, microgrid energy management system (EMS) still faces a critical challenge: how to properly adapt the objective functions in response to dynamically changing external conditions to achieve timely and resilient EMS responses. As a representative and highly time-sensitive microgrid EMS application, post-disaster power supply is adopted as the study context to better demonstrate the practical value and potential of this research.

To address this issues, we have been developing an adaptive objective function inference and selection (AOFIS) method for disaster relief microgrids. An objective function library is first established, consisting of a variety of objective functions specifically designed for different disaster relief conditions. A fuzzy inference system is then developed to adaptively infer and assign appropriate weight values to these objectives, reflecting their relative priorities under the prevailing environmental conditions. The highest-priority objective functions can be dynamically selected from the library and aggregated into a weighted single-objective power scheduling model together with the microgrid operating constraints, enabling the EMS to better fit the current disaster environment and power supply requirements.

We have also developed an EMS research and development (R&D) fast prototyping platform for disaster relief microgrids, aiming to provide a convenient R&D environment for rapid microgrid prototyping and proof-of-concept validation of EMS solutions in disaster response areas. The software part of the platform incorporates built-in libraries for various node types, objective functions, constraints, and EMS functional modules. In addition, the platform also includes a hardware-in-the-loop (HIL) system that enables EMS algorithm deployment, testing, monitoring, and visual demonstration.

Publications:

[1] Zhiyu Long and Mo-Yuen Chow, “An ems research and development fast prototyping platform for disaster relief,” IEEE/CAA Journal of Automatica Sinica, 2025, (accepted).

[2] Zhiyu Long and Mo-Yuen Chow, “Multi-objective energy scheduling with fuzzy logic-based weight tuning for disaster relief,” in 2025 IEEE International Conference on Industrial Electronics for Sustainable Energy Systems (IESES), 2025, (accepted).

[3] Zhiyu Long and Mo-Yuen Chow, “Objective function library and adaptive objective selection for resilient microgrids in disaster relief,” in 2026 27th IEEE International Conference on Industrial Technology (ICIT), 2026, (submitted).

[4] Zhiyu Long and Mo-Yuen Chow, “Adaptive Objective Function Inference and Selection-Based Distributed Energy Scheduling (AOFIS-DES) in Disaster Relief Microgrids,” IEEE Transactions on Industrial Informatics, 2026, (submitted).

Publications:

[1] Zhiyu Long and Mo-Yuen Chow, “An ems research and development fast prototyping platform for disaster relief,” IEEE/CAA Journal of Automatica Sinica, 2025, (accepted).

[2] Zhiyu Long and Mo-Yuen Chow, “Multi-objective energy scheduling with fuzzy logic-based weight tuning for disaster relief,” in 2025 IEEE International Conference on Industrial Electronics for Sustainable Energy Systems (IESES), 2025, (accepted).

[3] Zhiyu Long and Mo-Yuen Chow, “Objective function library and adaptive objective selection for resilient microgrids in disaster relief,” in 2026 27th IEEE International Conference on Industrial Technology (ICIT), 2026, (submitted).

[4] Zhiyu Long and Mo-Yuen Chow, “Adaptive Objective Function Inference and Selection-Based Distributed Energy Scheduling (AOFIS-DES) in Disaster Relief Microgrids,” IEEE Transactions on Industrial Informatics, 2026, (submitted).