Machine Learning for Scalable and Optimal Load Shedding Under Power System Contingency
2405.05521
0
0
🏷️
Abstract
Prompt and effective corrective actions in response to unexpected contingencies are crucial for improving power system resilience and preventing cascading blackouts. The optimal load shedding (OLS) accounting for network limits has the potential to address the diverse system-wide impacts of contingency scenarios as compared to traditional local schemes. However, due to the fast cascading propagation of initial contingencies, real-time OLS solutions are challenging to attain in large systems with high computation and communication needs. In this paper, we propose a decentralized design that leverages offline training of a neural network (NN) model for individual load centers to autonomously construct the OLS solutions from locally available measurements. Our learning-for-OLS approach can greatly reduce the computation and communication needs during online emergency responses, thus preventing the cascading propagation of contingencies for enhanced power grid resilience. Numerical studies on both the IEEE 118-bus system and a synthetic Texas 2000-bus system have demonstrated the efficiency and effectiveness of our scalable OLS learning design for timely power system emergency operations.
Create account to get full access
Overview
- Proposes a decentralized approach to optimal load shedding (OLS) for power system resilience
- Uses a neural network (NN) model trained offline to autonomously construct OLS solutions from local measurements
- Aims to reduce computation and communication needs during online emergency responses
Plain English Explanation
Our power grids need to be resilient and able to quickly respond to unexpected problems to prevent widespread blackouts. Optimal load shedding (OLS) is a technique that can help address the diverse impacts of unexpected events on the power system. However, real-time OLS solutions are challenging for large power systems because they require a lot of computing power and communication.
This paper proposes a new decentralized approach that uses a neural network (NN) model. The NN is trained offline to learn how to construct OLS solutions using only local measurements available at each load center. This can greatly reduce the computing and communication needed during an actual emergency, helping prevent cascading blackouts and improve power grid resilience.
The researchers tested their approach on both a medium-sized power system (IEEE 118-bus) and a larger, synthetic system (Texas 2000-bus). The results show their scalable OLS learning design can provide timely and effective emergency responses.
Technical Explanation
The paper presents a decentralized design for optimal load shedding (OLS) that leverages offline training of a neural network (NN) model. The goal is to enable individual load centers to autonomously construct OLS solutions from locally available measurements, rather than requiring centralized, real-time optimization that is computationally intensive and communication-heavy.
The proposed learning-for-OLS approach first trains the NN model offline using data from a full AC-OPF solver. The trained NN at each load center can then quickly determine the optimal load shedding actions during online emergencies based only on its local measurements, without the need for extensive coordination.
Numerical experiments on both the IEEE 118-bus system and a larger, synthetic Texas 2000-bus system demonstrate the scalability and effectiveness of this decentralized OLS learning design. Compared to traditional local load shedding schemes, the NN-based approach can significantly reduce computation and communication requirements while still effectively mitigating the cascading propagation of contingencies.
Critical Analysis
The paper provides a promising approach to improving power system resilience through a decentralized optimal load shedding (OLS) strategy. By leveraging offline trained neural networks, the solution can overcome the real-time computation and communication challenges of traditional centralized OLS methods.
However, the paper does not fully address the potential challenges of training and deploying these NN models in practice. For example, it is unclear how the NN models would be kept up-to-date as the power system evolves over time, or how their performance would be validated and monitored in a live grid environment. Online learning strategies could be an area for future research to address such practical deployment issues.
Additionally, while the numerical results demonstrate the scalability of the approach, more real-world testing and validation would be valuable to ensure the NN-based OLS can be reliably and robustly deployed. Weather-related power outage prediction is another area where similar decentralized ML techniques could prove useful for enhancing grid resilience.
Conclusion
This paper presents a novel decentralized approach to optimal load shedding (OLS) for power system resilience, leveraging offline trained neural networks to enable autonomous emergency response at individual load centers. By reducing the real-time computation and communication requirements, the proposed learning-for-OLS design can help prevent the cascading propagation of contingencies and improve overall grid resilience.
The scalable and effective performance demonstrated through numerical studies on both medium and large-scale power systems is promising. Further research is needed to address practical deployment challenges, such as online model adaptation and probabilistic forecasting of power system imbalances. Nonetheless, this work represents an important step towards enhancing the resilience of our critical power infrastructure through advanced machine learning techniques.
This summary was produced with help from an AI and may contain inaccuracies - check out the links to read the original source documents!
Related Papers
🏅
End-to-End Reinforcement Learning of Curative Curtailment with Partial Measurement Availability
Hinrikus Wolf, Luis Bottcher, Sarra Bouchkati, Philipp Lutat, Jens Breitung, Bastian Jung, Tina Mollemann, Viktor Todosijevi'c, Jan Schiefelbein-Lach, Oliver Pohl, Andreas Ulbig, Martin Grohe
0
0
In the course of the energy transition, the expansion of generation and consumption will change, and many of these technologies, such as PV systems, electric cars and heat pumps, will influence the power flow, especially in the distribution grids. Scalable methods that can make decisions for each grid connection are needed to enable congestion-free grid operation in the distribution grids. This paper presents a novel end-to-end approach to resolving congestion in distribution grids with deep reinforcement learning. Our architecture learns to curtail power and set appropriate reactive power to determine a non-congested and, thus, feasible grid state. State-of-the-art methods such as the optimal power flow (OPF) demand high computational costs and detailed measurements of every bus in a grid. In contrast, the presented method enables decisions under sparse information with just some buses observable in the grid. Distribution grids are generally not yet fully digitized and observable, so this method can be used for decision-making on the majority of low-voltage grids. On a real low-voltage grid the approach resolves 100% of violations in the voltage band and 98.8% of asset overloads. The results show that decisions can also be made on real grids that guarantee sufficient quality for congestion-free grid operation.
6/21/2024
Online learning for robust voltage control under uncertain grid topology
Christopher Yeh, Jing Yu, Yuanyuan Shi, Adam Wierman
0
0
Voltage control generally requires accurate information about the grid's topology in order to guarantee network stability. However, accurate topology identification is challenging for existing methods, especially as the grid is subject to increasingly frequent reconfiguration due to the adoption of renewable energy. In this work, we combine a nested convex body chasing algorithm with a robust predictive controller to achieve provably finite-time convergence to safe voltage limits in the online setting where there is uncertainty in both the network topology as well as load and generation variations. In an online fashion, our algorithm narrows down the set of possible grid models that are consistent with observations and adjusts reactive power generation accordingly to keep voltages within desired safety limits. Our approach can also incorporate existing partial knowledge of the network to improve voltage control performance. We demonstrate the effectiveness of our approach in a case study on a Southern California Edison 56-bus distribution system. Our experiments show that in practical settings, the controller is indeed able to narrow the set of consistent topologies quickly enough to make control decisions that ensure stability in both linearized and realistic non-linear models of the distribution grid.
4/1/2024
Deep Learning-Based Weather-Related Power Outage Prediction with Socio-Economic and Power Infrastructure Data
Xuesong Wang, Nina Fatehi, Caisheng Wang, Masoud H. Nazari
0
0
This paper presents a deep learning-based approach for hourly power outage probability prediction within census tracts encompassing a utility company's service territory. Two distinct deep learning models, conditional Multi-Layer Perceptron (MLP) and unconditional MLP, were developed to forecast power outage probabilities, leveraging a rich array of input features gathered from publicly available sources including weather data, weather station locations, power infrastructure maps, socio-economic and demographic statistics, and power outage records. Given a one-hour-ahead weather forecast, the models predict the power outage probability for each census tract, taking into account both the weather prediction and the location's characteristics. The deep learning models employed different loss functions to optimize prediction performance. Our experimental results underscore the significance of socio-economic factors in enhancing the accuracy of power outage predictions at the census tract level.
4/5/2024
🧠
Probabilistic forecasting of power system imbalance using neural network-based ensembles
Jonas Van Gompel, Bert Claessens, Chris Develder
0
0
Keeping the balance between electricity generation and consumption is becoming increasingly challenging and costly, mainly due to the rising share of renewables, electric vehicles and heat pumps and electrification of industrial processes. Accurate imbalance forecasts, along with reliable uncertainty estimations, enable transmission system operators (TSOs) to dispatch appropriate reserve volumes, reducing balancing costs. Further, market parties can use these probabilistic forecasts to design strategies that exploit asset flexibility to help balance the grid, generating revenue with known risks. Despite its importance, literature regarding system imbalance (SI) forecasting is limited. Further, existing methods do not focus on situations with high imbalance magnitude, which are crucial to forecast accurately for both TSOs and market parties. Hence, we propose an ensemble of C-VSNs, which are our adaptation of variable selection networks (VSNs). Each minute, our model predicts the imbalance of the current and upcoming two quarter-hours, along with uncertainty estimations on these forecasts. We evaluate our approach by forecasting the imbalance of Belgium, where high imbalance magnitude is defined as $|$SI$| > 500,$MW (occurs 1.3% of the time in Belgium). For high imbalance magnitude situations, our model outperforms the state-of-the-art by 23.4% (in terms of continuous ranked probability score (CRPS), which evaluates probabilistic forecasts), while also attaining a 6.5% improvement in overall CRPS. Similar improvements are achieved in terms of root-mean-squared error. Additionally, we developed a fine-tuning methodology to effectively include new inputs with limited history in our model. This work was performed in collaboration with Elia (the Belgian TSO) to further improve their imbalance forecasts, demonstrating the relevance of our work.
4/24/2024