Meet Our Energizer - Eli Shirazi

Meet our Energizer! Eli Shirazi

Eli (Elham) Shirazi, originally from Iran, holds a PhD in Electrical Engineering with a focus on smart grid control. She joined the University of Twente in 2021 after completing her post-doctoral research at the Department of Electrical Engineering at Katholieke Universiteit Leuven, Belgium. She has also worked at the Energy Department of IMEC in Belgium. Her multidisciplinary research focuses on the modelling and control of energy systems.

Today, Eli is an Assistant Professor in the Advanced Manufacturing, Sustainable Products, and Energy Systems group within the Department of Design, Production, and Management at the Faculty of Engineering Technology. Earlier this year, she joined the Advisory Board for 4TU.Energy.

Modelling and control of Energy Systems

The energy sector is a major contributor to greenhouse gas emissions, and reducing these emissions is essential to mitigate the impacts of global warming. By shifting to renewable energy sources such as solar and wind, we can significantly reduce our carbon footprint and promote a cleaner, and more sustainable environment. “In a nutshell, my research focuses on modelling and control of energy systems, which plays a key role in making energy transition feasible, as the current grid infrastructure was not designed to handle such high variability and penetration of renewable energy sources.”

Eli continues, “My work aims to make our energy systems smarter and more efficient. Modern energy systems handle more diverse generation and load types with different dynamics and characteristics compared to traditional systems. In my work, I model these components to optimize the flows within the energy system.”

In addition to new energy sources, consumers’ behaviours are changing too, imposing new constraints on system operation. Consumers are becoming “prosumers”, meaning they not only consume energy from the system but also produce energy and feed it back into the system. Prosumers’ interactions with the system make its operation more challenging.

Eli gives an example, “In the residential sector, significant solar energy is generated by building-integrated/applied photovoltaics (PV) during noon, which is not used locally and is therefore injected into the grid. This can create congestion and strain the distribution network. Conversely, high energy extraction from the grid in the evening, when everyone returns home, also poses a problem. This timing imbalance between peak demand and renewable electricity production is called the 'duck curve', which places significant strain on the grid and leads to new challenges such as negative energy prices”.

“I also find the new challenges that come with modernizing energy systems interesting. The adoption of electric vehicles (EV) and photovoltaic systems (PV) is outpacing changes on the user side, make operation of energy systems challenging. One solution is to develop control frameworks to manage these complexities and uncertainties.”

Engineering a Grey Box

“What I am looking for in my work, at the core of it, are patterns, to better understand non-dispatchable* generation and load profiles within the energy system” says Eli. According to her, two key issues challenge the design of control models: uncertainties and complexities.

Uncertainties in energy system control arise when optimizing the system's operation. For example, consider charging your EV using solar energy. Unlike fossil-fuel based resources, solar energy generation is uncertain. Eli explains, “We can use AI-based forecasting to predict available solar energy, which indicates the best time to charge the EV. On the other hand, we don’t know the exact arrival and departure times of the EV at the charging station, as well as EV energy demand and energy prices. Forecasting can reduce the uncertainties involved with controlling the energy system, reduce curtailing renewable power generation, lower the energy costs, and make better use of the grid.”

“In my research, I try to address the complexities in energy system operation raised by inflexible and non-dispatchable energy resources, consumer behaviour and various load types with deep reinforcement learning. This involves using different agents in a model that learn from their environment, make observations, take decisions, and perform actions.” Eli elaborates, “This can occur on two levels: demand-side or generation-side. Demand-side management might suggest not charging your EV now but in five hours. Generation-side management could recommend curtailing PV generation or storing surplus power in batteries instead of delivering it to the grid.”

As Eli explains her work, it becomes clear that her research is multidisciplinary. She uses a combination of the mathematical optimization and AI-based approaches to design a control model and find optimal solutions for controlling energy systems. Eli notes, “A data-driven approach often appears as a black box, especially in forecasting. In contrast, a mathematical approach models the system considering its physics, which is referred to as the white box modelling. My goal is to design something in between—a grey box model. I combine AI and mathematical modelling to create a ‘physics-aware AI’ for energy system control.”

Energizing Collaborations: Tackling the challenges of strong partnerships

“Beyond the inherent complexity and uncertainties in scientific research, I experience challenges in accessing public-private partnerships. Collaborations between academia and industry are crucial because it combines practical, real-world challenges with research and theoretical knowledge. Industry benefits from access to academic expertise, which can drive product development and improve processes. Meanwhile, academia gains insight into current industry needs, ensuring that research is relevant and applicable, and that students are better prepared for the workforce. This partnership accelerates the translation of research into marketable solutions, driving economic growth and addressing societal challenges more effectively. However, aligning the interests of industry and academia can be challenging, especially when applying for funding,” says Eli.

“We need agreements at higher institutional levels to seek partnerships with larger companies, benefiting researchers a few levels down the hierarchy. Speaking of which, I was surprised by the robust hierarchy in Dutch academia, as I expected a flatter social structure similar to broader Dutch society. Adapting to this hierarchy has been a learning experience.”

Eli continues, “We train students for industry through Challenge-Based Learning (CBL) and Project-Based Learning (PBL). While these educational collaborations are well-established, research project collaborations are more challenging. The financial structures and expectations between industry and academic institutions might be a significant barrier.”

To tackle these difficulties, Eli focuses on growing her network and raising awareness. She takes a holistic approach, extending this from her approach to science to how she builds her professional relationships. “I have been elected to the faculty council, where we advocate for inclusiveness and engage in round table discussions with the faculty board. Participating in these discussions is crucial for addressing challenges in building strong partnerships.”

4TU.Energy stimulates collaborative approaches in the Energy Transition

Eli observes, “There are many current initiatives and developments in the energy transition, but major challenges still need to be addressed. Significant challenges exist for researchers, industries, and policymakers. As a scientist, I am aware of the importance of policy and regulation. Think of the situation where one person has solar panels on their rooftop, while their neighbour does not. But the person with the solar panels is unable to deliver or sell electricity to their next-door neighbour. This is something that can be addressed via regulations, to alleviate pressure on the energy system. At this moment you cannot trade electricity peer-to-peer. There is definitely room for improvement, and it takes a multidisciplinary approach to solve these challenges.

Moreover, climate change and energy transition are closely interrelated, adding a sense of urgency. One of the most important strategies is the electrification in various industries, which requires substantial infrastructure expansion, that is time-consuming and resource-intensive. Regarding the energy transition, we must think far ahead, even generations into the future, to achieve true sustainability in energy systems. Despite the hurdles, I think we are on the right track!

What I truly appreciate about 4TU.Energy is its stimulation of multidisciplinary research and activities. This includes diverse topics such as energy system integration, social aspects of the energy transition, user behaviour, human capital and policy and regulation. I strongly advocate for holistic solutions to these challenges, and 4TU.Energy’s approach to fostering partnerships aligns with my view. The 4TU.Energy Community Day, for instance, is an excellent initiative. Furthermore, 4TU.Energy offers workshops like the PhD course, where candidates can learn, connect, and explore other disciplines. Finally, 4TU.Energy is committed to promoting collaborative and educational initiatives. I look forward to continuing to contribute to the 4TU.Energy community.” 

More about Eli

* Dispatchable generation refers to electricity generation sources that can be controlled and adjusted to meet demand as needed, where non-dispatchable generation refers to electricity generation sources that cannot be controlled or adjusted to meet demand in real-time. Non-dispatchable generation typically produce power based on availability rather than on-demand, a prime example is solar photovoltaics.