Toward digitalized microgrids with blockchain, data driven techniques, and cyber-resilience enhancement methods
Technical Outline of the Session and Topics:
Digitalization allows the modern microgrids to benefit from unprecedented flexibility and scalability. Microgrids are now experiencing a massive evolution from its conventional form of the Industrial Control System (ICS) to a cyber-physical coupled system which incorporates heterogeneous renewable energy resources and communication components. However, the microgrid digitalization also introduces more threats of cyber-attacks as more complex mechanisms involving hardware protection and cyber links are utilized to regulate microgrid performance. The bidirectional information flows between the hardware layer in microgrid agents and the network devices in the backbone communication infrastructures are vulnerable to a well-prepared and maliciously performed cyber-attack. “More advanced, more vulnerable. Trapped in the convention, wait for the failure.” The modern cyber-physical microgrids are destined to be exposed to potential cyber-attacks, and it should be provided with cyber-resilience against these cyber-attacks. Although the service ability enabled by the microgrid digitalization will be impaired, the cyber-physical microgrids could quickly recover from abnormal states and provide sufficient service ability to support the functions of microgrids. In this sense, the analysis of both the impact evaluation of cyber-attacks and the corresponding defense solution is critical to answer the challenging dilemma in this frontier field. On the other hand, many of cosmopolitans, including Singapore, have been moving one step further in terms of energy system to integrate technologies, such as reinforcement learning, blockchain, software-defined network, and federal learning.
Edge intelligence has been developed and introduced for general data communications and mobile networking. The cryptographic mechanism and the consensus protocol utilized in blockchain can theoretically safeguard the microgrid system from being injected with malicious data. The communication, network and service management fields are hungry for machine learning (ML) decision-making solutions to replace the traditional model-driven approaches to address the ever-growing complexity and heterogeneity of the modern cyber-physical microgrid. Distributed ML techniques (e.g., edge/fog computing, blockchain-based) are required to support power system operation and management with different functions. Data sharing and collaborative model training are promising ways to improve the quality of data-driven models. Federated learning aims to train collaboratively on distributed data sources without disclosing private data from each of the data sources, thus enabling privacy-preserving data sharing and collaboration. These solutions can potentially offer reliable, efficient, secure, and trustworthy collective learning for participants and service providers by deriving decisions using federated learning models for diversity of fields and applications.
This special session aims to investigate the applications of these new method and technology to support the microgrid digitalization. The session encourages the creative ideas and cutting-edge research works from both academia and industry. The topics of interest include but not are limited to the following:
- Cyber-resilience enhancement by reinforcement learning or federated learning in hierarchical microgrid control framework.
- Power system resilience by optimization in microgrid control.
- Blockchain-enabled applications in secure and privacy-preserving energy trading and dispatch in microgrids, networked microgrids.
- Cyber-attack impact evaluations in microgrids, such as impact of denial-of-service attack, false data injection attack on control performance of microgrids.
- Advanced co-simulation or hardware-in-the-loop testbed for the validation of control performance or cyber-resilience in microgrids.
- Communication-efficient ML-based data-driven methods for modern cyber-physical microgrid via federated learning.
- Vulnerability analysis, mitigation/defense strategies, and trustworthiness of data-driven methods for modern cyber-physical microgrid.