T1: AI-driven Decarbonization for Energy Systems
Shengrong Bu, Brock University, Canada
Dawei Qiu, Imperial College, UK
Zhu Han, University of Houston, TX, USA
Decarbonization of energy systems is urgently needed to help achieve the Paris climate agreement goals. Decarbonizing our current electricity generation is currently altering the fundamental structure of system operation and planning by increasing the penetration of renewable energy, and the electrification of transportation and efficient heating facilities will make the situation even more complex by significantly increasing electricity demand. Boosting the expansion of renewable energy resources (RES) could be effective in driving the world’s energy revolution towards a low-carbon future. However, a significant increase in flexibility is needed to balance the effects of less controllable output of RES.
This tutorial will first provide an overview of various approaches for energy systems decarbonization. It will then examine two types of decarbonization approaches--electric vehicles (EVs) and peer-to-peer (P2P) energy trading--related to those challenges and explore how state-of-the-art deep reinforcement learning and big data analytic tools could be applied to meet these challenges. The tutorial will explain the multiple inter-dependent services offered by EVs, explore the challenges of EVs to make their routing and scheduling decisions in a coupled power-transportation network, and demonstrate how hierarchical and hybrid multi-agent reinforcement learning methods can be employed to address the challenges. The tutorial will also examine aspects of P2P energy trading, such as how it can be combined with multi-energy converters to further improve the flexibility of the power systems, which is critical for decarbonization in microgrids.
T2: IoT-Based Load-Altering Attacks Against Power Grids
Subhash Lakshminarayana, University of Warwick, Coventry, UK
Charalambos Konstantinou, KAUST, Thuwal, KSA
The growing integration of Internet-of-Things (IoT) high-wattage consumer appliances including electric vehicles and heating, ventilation, and air conditioning (HVAC) systems can pose a severe vulnerability to the electric grid's operations. An abrupt manipulation of power grid demand by large-scale botnet-type attacks against IoT-smart-home appliances can severely affect the balance between the power supply and demand, and lead to unsafe operation of the grid. Such load-altering attacks (LAAs) can lead to high operational costs (at the grid side), unsafe frequency excursions, and even severe frequency and voltage stability issues that can further cause generator trips and cascading failures. The information required to execute such attacks can be gathered by publicly available information, such as the charging patterns of plug-in-electric vehicles and the information on the power grid infrastructure.
The tutorial will present a comprehensive overview of the threat of LAAs in power grids. It will specifically focus on (1) the theoretical foundations for analysing a high-impact least-effort LAA targeted at high-wattage IoT-based devices, (2) the demonstration of how LAAs can lead to realistic cyber-attacks capable of identifying the most vulnerable locations and amount of load needed to be compromised in order to cause unsafe frequency fluctuations in a (low-inertia) power system, (3) application of machine learning (ML) techniques to detect and localise LAAs, (4) the presentation of simulation-based experiments to demonstrate the effects of the formulated LAA in future power systems with a high penetration of renewable energy resources. Finally, future development of LAAs in the power grid and other cyber-physical systems will be identified.
T3: Using global wireless standard-based networks to modernize the communication infrastructure used by grid operators
Larry J. Horner, Intel Principal Engineer, Senior Solution Architect
The Smart grid transformation is intersecting with a transformation in the connectivity where in some jurisdictions Private Network are being deployed with Cellular technology. Existing mesh networks and WIFI will coexist with the introduction of 4G and 5G technologies. Grid operators have started early trials and deployments, in this tutorial we will discuss some current efforts and touch on the challenges ahead based on industry feedback.
T4: Synchro-Waveforms: A New Frontier in Advanced Smart Grid Sensing and Data Analytics
Hamed Mohsenian-Rad, University of California, Riverside, CA, USA
Alireza Shahsavari, San Diego Gas and Electric, San Diego, CA, USA
Waveform measurement units (WMUs) are an emerging class of smart grid synchronized measurement technologies that provide synchronized measurements for voltage and current waveforms. Since WMUs provide synchronized waveform measurements, as opposed to synchronized phasor measurements that are provided by phasor measurement units (PMUs), the data from WMUs is much more granular than the data from PMUs. This calls for fundamentally new methodologies to analyze WMU data. In this tutorial, we cover the advancements in this field, in various areas, including sensor technologies, data collection, data analytics, and use cases. The tutorial has three parts: 1) Technology and Real-World Data; 2) Data Analytics Methods; and 3) Applications and Use Cases. The speakers are among the pioneers in installing and testing PMUs and WMUs at power distribution networks. The discussions in this Tutorial are inspired by their experiences in working with several terabytes of sychro-phasor and synchro-waveform data.