Optical Sensing and Computing: Supercharging Cities for a Climate-Resilient Energy Future

Written by Benedict Vergara and Payman Dehghanian

The transition from a centralized electricity generation and distribution system, consisting of a few large-scale units, to a decentralized system with many small-scale units has resulted in complex power distribution challenges that cannot be efficiently managed by a single centralized authority. This is particularly true for power distribution systems in cities, which are facing increasing frequency of extreme weather events and need rapid controls to protect buildings from potential damage. Fortunately, recent advancements in optical sensing and computing technology have improved the ability to capture high-resolution data and accelerate the speed at which machine learning models can make predictions. These developments pave the way for a future power grid that can effectively manage a heavily populated network of distributed energy resources (DERs) in urban areas.


Introduction

Climate change is the driving force behind the energy transition. In addition to the need for renewable energy to reduce emissions, the growing threat of weather-related damage to critical infrastructure, networks, and services necessitates the deployment of distributed energy resources for increased reliability. The International Panel on Climate Change (IPCC) has highlighted this need [1]. Furthermore, cities are crucial actors in this transition, given projections from the International Renewable Energy Agency (IRENA) that they will need to accommodate two-thirds of the world's population in a low-carbon environment by 2050 [2].

Technological advancements are the enablers of this transition. In the United States, for example, the Bipartisan Infrastructure Law has pledged $11 billion to build a smarter and stronger electric grid, reflecting the emergence and continued development of innovative hardware and software for grid-related services [3]. These advancements offer solutions that strengthen cities' ability to withstand extreme climate events. Photonic sensing and computing are particularly promising innovations for this energy transition in smart cities of the future, as they offer higher-resolution insight into grid status and the ability to supercharge the applicability of machine learning in coordinating energy resources across the city power grid.

 

Optical Sensing: Higher-resolution Measurements for Grid Diagnostics

Sensing the status of the power grid has been essential since its inception, leading to the development of first and second-generation phasor measurement units (PMUs) that are widely used at the transmission level. However, as more Distributed Energy Resources (DERs) are integrated into the grid, standards such as IEEE 1547 require grid-connected inverters (such as batteries, solar panels, fuel cells, etc.) to monitor voltage, current, and frequency to support the distribution grid properly. Inverters that contribute to grid stability, such as voltage support, frequency regulation, and power factor correction, are becoming increasingly crucial for managing the power distribution level spanning smart cities of the future [4].

To achieve an even greater understanding of the distribution grid, next-generation sensing equipment is under development. Photonic devices, which use the Pockels Effect, offer much higher resolution and timing of the grid operating state than traditional sensors that use electrical induction. State-of-the-art solutions in the market offer technologies capable of sensing 15,000 samples per second and processing information to reveal up to the 250th harmonic, providing in-depth analysis of even the most nuanced perturbations and identifying their origin [5]. This information can help protect equipment in urban distribution systems, particularly as the severity of weather events increases.

 

Optical Computing: Fast Inference for Machine Learning

Machine learning has proven effective in power system analysis, coordination, and planning, enabling fault detection, optimal energy resource dispatch, and expansion project proposals. However, the speed at which machine learning models make predictions remains a critical concern, particularly in time-sensitive applications like correcting inverter power or activating protective relays within milliseconds or nanoseconds (Figure 1).

Figure 1 Control Technique

 Figure 1: Control Technique (Top): Approach and solving time for key decision-making (optimization) problems in power systems. Time Frame (Bottom): Each of the control techniques can span over a certain set of power grid decision making, from days to nanoseconds. Demonstrating the ability, Photonic Machine Learning has to cover regions of control that have been traditionally left to “threshold” controllers.

 

While machine learning has been used to classify grid status at a transmission level [6], its complexity increases with the volume of data used to train the model. This complexity is only further increased as the grid evolves and experiences new kind of perturbations due to DER deployment and unprecedented extreme weather events. Similarly, coordinating DERs through machine learning has been successful because optimization problems (such as the unit commitment problem and optimal power flow), suffer from scalability issues (due to their NP-hardness [7]). But, like classification problems, even these machine learning models grow in complexity as the number of DERs increase, limiting the number of units that can learn to coordinate together while providing solutions fast enough to manage the grid.

Therefore, Photonic inference (making a machine learning prediction) is a promising solution to overcome digital time limits, offering superpowered detection and reaction speeds. Photonic tensor cores are currently being tested in labs [8], and NVIDIA is designing them as a product exclusively for accelerating machine learning computations [9]. Implementing this technology can provide significant acceleration in machine learning predictions, benefitting heavily DER-populated cities experiencing extreme weather events (Figure 2).

 

Figure 2 Conceptual Architecture

 Figure 2: Conceptual architecture of a photonic replacement (right) for each component of an advanced machine learning (left) smart inverter for buildings in a city’s distribution line.

 

 

Conclusion

Interdisciplinary convergence has repeatedly demonstrated its ability to enhance systems by drawing on techniques from other domains. In power engineering, this is evident in the work of chemical engineers improving battery capabilities and systems engineers devising techniques to facilitate equity in energy systems planning and operation. By adopting techniques from other fields, the power sector can continue to grow and evolve. Photonics, for example, is a promising solution that can be taken to tackle grid-related problems within the cities in particular when managing extreme event consequences.

 

References

  1. Dodman, H., “Climate change 2022: Impacts, adaptation, and vulnerability. contribution of working group
    ii to the sixth assessment report of the intergovernmental panel on climate change,” (2022). https://www.ipcc.ch/report/ar6/wg2/downloads/report/IPCC_AR6_WGII_FullReport.pdf
  2. IRENA (2018), Global Energy Transformation: A roadmap to 2050, International Renewable Energy Agency, Abu Dhabi. This report is available for download from www.irena.org/publications. For further information or to provide feedback, please contact IRENA at info@irena.org
  3. “Building a Better Grid Initiative.” Energy.gov, https://www.energy.gov/gdo/building-better-grid-initiative.
  4. IEEE Standards Association. (2018). IEEE Standard for Interconnection and Interoperability of Distributed Energy Resources with Associated Electric Power Systems Interfaces--Amendment 1. IEEE 1547-2018. https://standards.ieee.org/ieee/1527/4976/
  5. Micate Inc. (2023). “Data Driven Insight for the Modern Grid”. Retrieved from https://www.micatu.com.
  6. Wang, S., Li, L., and Dehghanian, P., “Distributed intelligence for online situational awareness in power grids,” IEEE Transactions on Power Systems 37(4), 2499–2515 (2022).
  7. Bendotti, Fouilhoux, R., “On the complexity of the unit commitment problem,” Ann Oper Res 274 , 119–130 (2019).
  8. Hu, Z., Li, S., Schwartz, R. L., Solyanik-Gorgone, M., Nouri, B. M., Miscuglio, M., Gupta, P., Dalir, H., and Sorger, V. J., “Batch processing and data streaming fourier-based convolutional neural network accelerator,” in [Emerging Topics in Artificial Intelligence (ETAI) 2022 ], 12204, 68–74, SPIE (2022).
  9. NVIDIA. (2023). NVIDIA Tensor Cores. Retrieved from: https://www.nvidia.com/en-us/data-center/tensor-cores/.

 

This article was edited by Melkior Ornik.

To view all articles in this issue, please go to April 2023 eNewsletter. For a downloadable copy, please visit the IEEE Smart Cities Resource Center.

Benedict Vergara HeadShot
Benedict Vergara received his B.Sc. and M.Sc. in Electrical Engineering at The George Washington University, both with concentrations in Energy and Power Systems. Ben worked at the engineering consulting firm Strategic Analysis, Inc. as a staff engineer aiding the development of techno-economic analyses for clean energy technologies. Ben is currently a Ph.D. candidate in Electrical Engineering applying artificial intelligence to multi-energy networks to improve their reliability, resiliency, and efficiency.

PaymanDehghanian HeadShot
Payman Dehghanian (IEEE Senior Member’20) received a B.Sc. degree in electrical engineering from the University of Tehran, Tehran, Iran, in 2009, a M.Sc. degree in electrical engineering from the Sharif University of Technology, Tehran, in 2011, and a Ph.D. degree in electrical engineering from Texas A&M University, College Station, TX, USA, in 2017. He is currently an Assistant Professor with the Department of Electrical and Computer Engineering, George Washington University, Washington, DC, USA. His research interests include power system reliability and resilience assessment, data-informed decision-making for maintenance and asset management in electrical systems, and smart electricity grid applications. Dr. Dehghanian is the recipient of the 2013 IEEE Iran Section Best M.Sc. Thesis Award in Electrical Engineering, the 2014 and 2015 IEEE Region 5 Outstanding Professional Achievement Awards, the 2015 IEEE-HKN Outstanding Young Professional Award, the 2021 Early Career Award from the Washington Academy of Sciences, the 2022 George Washington University’s Early Career Researcher Award, the 2022 IEEE Industry Applications Society (IAS) Electric Safety Committee’s Young Professional Achievement Award, and the 2022 IEEE IAS Outstanding Young Member Service Award.


IEEE Smart Cities Publications Journals and Magazines Special Issues

This web page displays the effort of IEEE Smart Cities Publications Committee in proposing and guest editing special issues for IEEE Journals and Magazines which is of interests to IEEE Smart Cities Community. Please click here to view.

Past Issues

To view archived articles, and issues, which deliver rich insight into the forces shaping the future of the smart cities. Older eNewsletter can be found here. To download full issues, visit the publications section of the IEEE Smart Cities Resource Center.