The nexus between transportation, the power grid, and consumer behavior is much more pronounced than ever before as the race to decarbonize intensifies. Electrification in the transportation sector has led to technology shifts and rapid deployment of electric vehicles (EVs). The potential increase in stochastic and spatially heterogeneous charging load presents a unique challenge that is not well studied, and will have significant impacts on grid operations, emissions, and system reliability if not managed effectively. Realistic scenario generators can help operators prepare, and machine learning can be leveraged to this end. In this work, we develop generative adversarial networks (GANs) to learn distributions of electric vehicle (EV) charging sessions and disentangled representations. We show that this model successfully parameterizes unlabeled temporal and power patterns and is able to generate synthetic data conditioned on these patterns. We benchmark the generation capability of this model with Gaussian Mixture Models (GMMs), and empirically show that our proposed model framework is better at capturing charging distributions and temporal dynamics.
Rob Buechler (a PhD student in my group) and I were the first to use GANs within this domain. Our paper introduced a new method that improved on existing methods for modelling EV charging. We also improved the implementation of the similarity constrained GAN or SCGAN that led to a 50X speed up over its existing implementation. I believe with more open data the capabilities of this work can be further extended to do things we imagine might not have been possible a decade ago. We are open to collaborating with entities that have data to share in exchange for co-publishing, learnings, code-sharing, and useful insights.
To enable the electrification of transportation systems, it is important to understand how technologies such as grid storage, solar photovoltaic systems, and control strategies can aid the deployment of electric vehicle charging at scale. In this work, we present EV-EcoSim, a co-simulation platform that couples electric vehicle charging, battery systems, solar photovoltaic systems, grid transformers, control strategies, and power distribution systems, to perform cost quantification and analyze the impacts of electric vehicle charging on the grid. This python-based platform can run a receding horizon control scheme for real-time operation and a one-shot control scheme for planning problems, with multi-timescale dynamics for different systems to simulate realistic scenarios. We demonstrate the utility of EV-EcoSim via a case study focused on economic evaluation of battery size to reduce electricity costs while considering impacts of fast charging on the power distribution grid. We present qualitative and quantitative evaluations on the battery size in tabulated results. The tabulated results delineate the trade-offs between candidate battery sizing solutions, providing comprehensive insights for decision-making under uncertainty. Additionally, we demonstrate the implications of the battery controller model fidelity on the system costs and show that the fidelity of the battery controller can completely change decisions made when planning an electric vehicle charging site.
Link to Project Website and Publication
Read my blog for a simple, non-technical discussion of the work.
We will be deploying a web-tool to accompany this work, to make it accessible to as many academics, practitioners, or non-technical audiences as possible. Entire software will be released 12/17/2023.
Stochastic generators are useful for estimating climate impacts on various sectors. Projecting climate risk in various sectors, e.g. energy systems, requires generators that are accurate (statistical resemblance to ground-truth), reliable (do not produce erroneous examples), and efficient. Leveraging data from the North American Land Data Assimilation System, we introduce TemperatureGAN, a Generative Adversarial Network conditioned on months, locations, and time periods, to generate 2m above ground atmospheric temperatures at an hourly resolution. We propose evaluation methods and metrics to measure the quality of generated samples. We show that TemperatureGAN produces high-fidelity examples with good spatial representation and temporal dynamics consistent with known diurnal cycles.
Preprint available here.
Coordination of distributed energy resources is critical for electricity grid management. Although nodal pricing schemes can mitigate congestion and voltage deviations, the resulting prices are not necessarily equitable. In this work, we leverage market mechanisms for DER coordination and propose a daily dynamic nodal pricing scheme that incorporates equity. We introduce a pricing "oracle," which we call the Power Distribution Authority, that sets equitable prices to manage the grid. We present two algorithms for executing this scheme and show that both methods are able to set prices that satisfy both voltage and equity constraints. Both proposed algorithms also outperform the common utility time-of-use pricing schemes by at least 45%. New market mechanisms are needed as the grid is transforming, and power system operators may leverage these methods for pricing electricity in a grid-aware, equitable fashion.
The goal of this research was to introduce equity in designing power distribution grids through the lens of dynamic pricing. To understand the overarching goal of this work, let's talk about something called "Locational Marginal Prices" or LMPs. In the electricity grid, there's a "wholesale market," where participants (usually generators) place bids for minimum price at which they can supply energy, if you're familiar with the stock market, this is very similar albeit, we cannot afford the same level of volatility and dynamism within financial markets with the technology we have today.
So, in this work we formulate a distribution management grid study as a convex optimization problem and propose two methods for producing prices that satisfy voltage fluctuation and equity objectives.
Publication here.
This paper won the best paper award at the IEEE SmartGridComm '23 Conference in Glasgow! The details of the conference and award can be found here.
Also see my presentation of this work at Stanford.