{"title":"Coordinating renewable microgrids for reliable reserve services: a distributionally robust chance-constrained game","authors":"Yifu Ding, Siyuan Wang, B. Hobbs","doi":"10.1145/3575813.3597342","DOIUrl":"https://doi.org/10.1145/3575813.3597342","url":null,"abstract":"Networked microgrids aggregate distributed energy resources (DERs) and flexible loads to reach the minimum capacity for market participation and provide reserve services for the main grid. However, due to uncertain renewable generations such as solar power, microgrids might under-deliver reserve services and breach the day-ahead contracts in the real-time market. If multiple microgrids fail to deliver services as promised, this could lead to a severe grid contingency. This paper designs a distributionally robust chance-constrained (DRCC) market game simulating risk-aware bidding and system-wide reserve policy. Leveraging historical error samples, the reserve bidding strategy of each microgrid is formulated into a two-stage Wasserstein-metrics distribution robust optimization (DRO) model. A CC regulates the under-delivered reserve allowance of all microgrids’ reserve contracts in a non-cooperative game. Case studies are performed using the CAISO data. The proposed game is simulated under the reserve policy at different risk-aware levels and numbers of players to model the market behaviors and DER adoption. Results show the under-delivered reserve can be effectively regulated while securing the profit of microgrids in this game framework.","PeriodicalId":359352,"journal":{"name":"Proceedings of the 14th ACM International Conference on Future Energy Systems","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114650864","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Saidur Rahman, Javier Sales-Ortiz, Omid Ardakanian
{"title":"Making a Virtual Power Plant out of Privately Owned Electric Vehicles: From Contract Design to Scheduling","authors":"Saidur Rahman, Javier Sales-Ortiz, Omid Ardakanian","doi":"10.1145/3575813.3597353","DOIUrl":"https://doi.org/10.1145/3575813.3597353","url":null,"abstract":"With the rollout of bidirectional chargers, electric vehicle (EV) battery packs can be used in lieu of utility-scale energy storage systems to support the grid. These batteries, if aggregated and coordinated at scale, will act as a virtual power plant (VPP) that could offer flexibility and other services to the grid. To realize this vision, EV owners must be incentivized to let their battery be discharged before it is charged to the desired level. In this paper, we use contract theory to design incentive-compatible, fixed-term contracts between the VPP and EV owners. Each contract defines the maximum amount of energy that can be discharged from an EV battery and exported to the grid over a certain period of time, and the compensation paid to the EV owner upon successful execution of the contract, for reducing the cycle life of their battery. We then propose an algorithm for the optimal operation of this VPP that participates in day-ahead and balancing markets. This algorithm maximizes the expected VPP profit by taking advantage of the accepted contracts that are still valid, while honoring day-ahead commitments and fulfilling the charging demand of each EV by its deadline. We show through simulation that by offering a menu of fixed-term contracts to EVs that arrive at the charging station, trading energy and scheduling EV charging according to the proposed algorithm, the VPP profitability increases by up to 12.2%, while allowing EVs to partially offset the cost of charging their battery.","PeriodicalId":359352,"journal":{"name":"Proceedings of the 14th ACM International Conference on Future Energy Systems","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126437107","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Modeling Participation of Storage Units in Electricity Markets using Multi-Agent Deep Reinforcement Learning","authors":"Nick Harder, A. Weidlich, P. Staudt","doi":"10.1145/3575813.3597351","DOIUrl":"https://doi.org/10.1145/3575813.3597351","url":null,"abstract":"Modeling electricity markets realistically plays a crucial role for understanding complex emerging market dynamics and guiding policy making. In systems with a high share of variable renewable generation, accurately modeling the behavior of storage units can be particularly challenging, as their bidding strategies depend on expected electricity prices. While there exist a wide variety of electricity market models, they typically rely on rule-based bidding strategies or optimization approaches, which may not be sufficient to represent competitive and strategic behavior on the market. In this paper, we present a multi-agent deep reinforcement learning modeling framework that allows representing competitive and strategic behavior of energy storage units. This framework can be executed in large-scale electricity market models, thus facilitating market design analyses. We show that the proposed approach performs very well when compared with widely used modeling approaches, and its computational efficiency makes its use in energy market modeling practical.","PeriodicalId":359352,"journal":{"name":"Proceedings of the 14th ACM International Conference on Future Energy Systems","volume":"317 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115647204","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"FDA-HeatFlex: Scalable Privacy-Preserving Temperature and Flexibility Prediction for Heat Pumps using Federated Domain Adaptation","authors":"Subina Khanal, N. Ho, T. Pedersen","doi":"10.1145/3575813.3595194","DOIUrl":"https://doi.org/10.1145/3575813.3595194","url":null,"abstract":"Heat pumps are a significant source of flexibility in energy systems since they can be operated flexibly, e.g., turned up when electricity is green (low CO2) or cheap, and turned down when electricity is expensive or mainly from fossil sources. However, the indoor temperature has to be kept within a user-specified comfort interval, e.g., 20-24° C, for residents to accept this flexible operation. To estimate the available flexibility, we need to know how the indoor temperature changes depending on the heat pump input power and outdoor temperature. Machine learning (ML) models can learn this given enough historical data, typically at least one year, to account for seasonal variations. However, for new buildings and/or newly retrofitted heat pumps, there is no or little data and users may be reluctant to share such sensitive data. To estimate the heat pump flexibility of such buildings, we propose FDA-HeatFlex (Federated Domain Adaptation Heat Pump Flexibility) framework where we transfer the knowledge from the source domain (a known building) to multiple target domains (new buildings) to accurately predict the indoor temperature of new buildings and derive their flexibility, making the prediction scale easily to many new buildings. Particularly, we leverage the idea of parameter-based transfer learning and adaptive boosting (AdaBoost) techniques for indoor temperature prediction to address the data shift problem, i.e., the discrepancy of data distributions between buildings, and employ the idea of federated learning to address the privacy concerns raised by data sharing between source and target domains. We conduct an extensive experimental evaluation on widely used real-world heat pump datasets which shows that our FDA-HeatFlex outperforms the state-of-the-art training approaches for indoor temperature prediction, and the state-of-the-art baseline for flexibility prediction with and improvement (on average), respectively.","PeriodicalId":359352,"journal":{"name":"Proceedings of the 14th ACM International Conference on Future Energy Systems","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115855705","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Graph neural networks for fault diagnosis of geographically nearby photovoltaic systems","authors":"Jonas Van Gompel, D. Spina, Chris Develder","doi":"10.1145/3575813.3595200","DOIUrl":"https://doi.org/10.1145/3575813.3595200","url":null,"abstract":"Faults in photovoltaic (PV) systems significantly reduce their efficiency and can pose safety risks. Nevertheless, most residential PV systems are not actively monitored, because existing methods often require expensive sensors, which are only cost-effective for large PV systems. Therefore, we propose a graph neural network (GNN) to monitor a group of nearby PV systems without relying on dedicated sensors. Instead, the GNN compares 24 h of current and voltage measurements obtained from the inverters. Four GNN variants are experimentally compared using simulated data of six different PV systems in Colorado. Results show that all GNN variants outperform a state-of-the-art PV fault diagnosis method based on gradient boosted trees. Moreover, some GNN variants can even generalize to PV systems which were not in the training data, enabling monitoring of new PV systems without retraining.","PeriodicalId":359352,"journal":{"name":"Proceedings of the 14th ACM International Conference on Future Energy Systems","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116593443","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Kaleb Phipps, Stefan Meisenbacher, Benedikt Heidrich, Marian Turowski, R. Mikut, V. Hagenmeyer
{"title":"Loss-Customised Probabilistic Energy Time Series Forecasts Using Automated Hyperparameter Optimisation","authors":"Kaleb Phipps, Stefan Meisenbacher, Benedikt Heidrich, Marian Turowski, R. Mikut, V. Hagenmeyer","doi":"10.1145/3575813.3595204","DOIUrl":"https://doi.org/10.1145/3575813.3595204","url":null,"abstract":"To mitigate climate change, renewable energy sources are increasingly integrated into the energy system. Due to their volatility and uncertainty, smart grid applications are required to deal with this uncertainty and maintain grid stability. To operate effectively and in an automated manner, each of these applications requires a quantification of the uncertainty in future demand and generation, which probabilistic forecasts can provide. Furthermore, these applications often require specific probabilistic forecast properties, such as coverage rate and sharpness. However, existing probabilistic forecasts cannot be easily customised to exhibit these required properties. Therefore, we present a novel approach that creates loss-customised probabilistic forecasts using automated hyperparameter optimisation based on custom loss metrics. We combine a deterministic base forecaster and a conditional Invertible Neural Network to include specified uncertainty in a deterministic forecast. This uncertainty is defined by automated hyperparameter optimisation based on flexible and adaptable loss metrics, enabling the generation of loss-customised probabilistic forecasts with different properties without computationally expensive retraining. We evaluate our approach on four real-world data sets and compare the generated loss-customised forecasts with three state-of-the-art probabilistic forecasting benchmarks. We show that our approach generates probabilistic forecasts that can be customised to achieve state-of-the-art performance in either Continuous Ranked Probability Score, Pinball Loss, or Coverage Rate Error, depending on the selected customised loss metric.","PeriodicalId":359352,"journal":{"name":"Proceedings of the 14th ACM International Conference on Future Energy Systems","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117084368","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Heat FlexOffers: a device-independent and scalable representation of electricity-heat flexibility.","authors":"F. Lilliu, T. Pedersen, Laurynas Siksnys","doi":"10.1145/3575813.3597347","DOIUrl":"https://doi.org/10.1145/3575813.3597347","url":null,"abstract":"The increasing relevance of Renewable Energy Sources (RES) makes energy flexibility an extremely important aspect, not only regarding electricity, but also for other energy vectors such as heat. Because of this, there is the need for a flexibility model which can i) provide a common representation of flexibility for different device types, ii) perform aggregation, optimization and disaggregation while scaling for long time horizons and many devices, iii) capture most of the available flexibility, and iv) support energy conversion between different vectors. Properties i)-iii) are addressed by FlexOffer (FO), a device-independent model that describes energy constraints in an approximate yet accurate way. This paper proposes an extension of FOs, Heat FlexOffers (HFOs), capable of modeling flexibility for different energy vectors such as heat and handling energy conversion, and therefore addressing iv) as well as i)-iii). HFOs can model the optimal power curve for heat pumps, and can provide constraints for continuous optimization problems while complying to the Smart Grid-Ready (SG-Ready) interface, which operates on discrete states. We show that HFOs are very accurate, being able to retain up to of total flexibility before aggregation and of it after aggregation. HFOs aggregation is scalable, as 2 · 106 devices can be aggregated for a 24 hours time horizon, vastly outperforming exact models as they fail to aggregate more than 500 devices.","PeriodicalId":359352,"journal":{"name":"Proceedings of the 14th ACM International Conference on Future Energy Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122067107","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Seyedali Tabaeiaghdaei, Simon Scherrer, Jonghoon Kwon, A. Perrig
{"title":"Carbon-Aware Global Routing in Path-Aware Networks","authors":"Seyedali Tabaeiaghdaei, Simon Scherrer, Jonghoon Kwon, A. Perrig","doi":"10.1145/3575813.3595192","DOIUrl":"https://doi.org/10.1145/3575813.3595192","url":null,"abstract":"The growing energy consumption of Information and Communication Technology (ICT) has raised concerns about its environmental impact. However, the carbon footprint of data transmission over the Internet has so far received relatively modest attention. This carbon footprint can be reduced by sending traffic over carbon-efficient inter-domain paths. However, challenges in estimating and disseminating carbon intensity of inter-domain paths have prevented carbon-aware path selection from becoming a reality. In this paper, we take advantage of path-aware network architectures to overcome these challenges. In particular, we design CIRo, a system for forecasting the carbon intensity of inter-domain paths and disseminating them across the Internet. We implement a proof of concept for CIRo on the codebase of the SCION path-aware Internet architecture and test it on the SCIONLab global research testbed. Further, through large-scale simulations, we demonstrate the potential of CIRo for reducing the carbon footprint of endpoints and end domains: With CIRo, half of domain pairs can reduce the carbon intensity of their inter-domain traffic by at least 47%, and 87% of end domains can reduce their carbon footprint of Internet use by at least 50%.","PeriodicalId":359352,"journal":{"name":"Proceedings of the 14th ACM International Conference on Future Energy Systems","volume":"31 5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116343213","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ruihang Wang, Zhi-Ying Cao, Xiaoxia Zhou, Yonggang Wen, Rui Tan
{"title":"Phyllis: Physics-Informed Lifelong Reinforcement Learning for Data Center Cooling Control","authors":"Ruihang Wang, Zhi-Ying Cao, Xiaoxia Zhou, Yonggang Wen, Rui Tan","doi":"10.1145/3575813.3595189","DOIUrl":"https://doi.org/10.1145/3575813.3595189","url":null,"abstract":"Deep reinforcement learning (DRL) has shown good performance in data center cooling control for improving energy efficiency. The main challenge in deploying the DRL agent to real-world data centers is how to quickly adapt the agent to the ever-changing system with thermal safety compliance. Existing approaches rely on DRL’s native fine-tuning or a learned data-driven dynamics model to assist the adaptation. However, they require long-term unsafe exploration before the agent or the model can capture a new environment. This paper proposes Phyllis, a physics-informed reinforcement learning approach to assist the DRL agent’s lifelong learning under evolving data center environment. Phyllis first identifies a transition model to capture the data hall thermodynamics in the offline stage. When the environment changes in the online stage, Phyllis assists the adaptation by i) supervising safe data collection with the identified transition model, ii) fitting power usage and residual thermal models, iii) pretraining the agent by interacting with these models, and iv) deploying the agent for further fine-tuning. Phyllis uses known physical laws to inform the transition and power models for improving the extrapolation ability to unseen states. Extensive evaluation for two simulated data centers with different system changes shows that Phyllis saves 5.7% to 13.8% energy usage compared with feedback cooling control and adapts to new environments 8x to 10x faster than fine-tuning with at most 0.74°C temperature overshoot.","PeriodicalId":359352,"journal":{"name":"Proceedings of the 14th ACM International Conference on Future Energy Systems","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125960574","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Hsin-Yu Liu, Bharathan Balaji, Rajesh K. Gupta, Dezhi Hong
{"title":"Rule-based Policy Regularization for Reinforcement Learning-based Building Control","authors":"Hsin-Yu Liu, Bharathan Balaji, Rajesh K. Gupta, Dezhi Hong","doi":"10.1145/3575813.3595202","DOIUrl":"https://doi.org/10.1145/3575813.3595202","url":null,"abstract":"Rule-based control (RBC) is widely adopted in buildings due to its stability and robustness. It resembles a behavior cloning methodology refined by human experts; however, it is incapable of adapting to distribution drifts. Reinforcement learning (RL) can adapt to changes but needs to learn from scratch in the online setting. On the other hand, the learning ability is limited in offline settings due to extrapolation errors caused by selecting out-of-distribution actions. In this paper, we explore how to incorporate RL with a rule-based control policy to combine their strengths to continuously learn a scalable and robust policy in both online and offline settings. We start with representative online and offline RL methods, TD3 and TD3+BC, respectively. Then, we develop a dynamically weighted actor loss function to selectively choose which policy for RL models to learn from at each training iteration. With extensive experiments across various weather conditions in both deterministic and stochastic scenarios, we demonstrate that our algorithm, rule-based incorporated control regularization (RUBICON), outperforms state-of-the-art methods in offline settings by and improves the baseline method by in online settings with respect to a reward consisting of thermal comfort and energy consumption in building-RL environments.","PeriodicalId":359352,"journal":{"name":"Proceedings of the 14th ACM International Conference on Future Energy Systems","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134449572","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}