Transmission network topology control offers cheap flexibility to system operators for mitigating grid congestion. However, finding the optimal sequence of topology actions remains a challenge due to the large number of possible actions. Although reinforcement learning (RL) approaches have attracted interest for long-term planning in large combinatorial action spaces, they encounter challenges such as training stability, sample efficiency, and unforeseen consequences of RL actions. Addressing these challenges, this paper proposes a hybrid curriculum-trained RL and Monte Carlo tree search (MCTS) approach to determine sequential topological actions for mitigating grid congestion. The curriculum-based approach stabilizes training by first pre-training a policy network through supervised imitation learning, followed by RL training. The policy network guides the MCTS to simulate promising future trajectories, mitigating unforeseen consequences and identifying long-term strategies to improve grid security. Moreover, the MCTS-verified actions are used for RL training, enhancing sample efficiency and training time. A distance factor is added to the MCTS, which improves convergence by prioritizing actions closer to congestion. Numerical results on the IEEE 118-bus system show that the proposed hybrid approach improves the timesteps survived by 30% compared to a standard RL approach, and by 5% compared to a brute-force baseline. Additionally, the inclusion of the distance factor increases the timesteps survived by 15%. These results highlight the potential of the proposed method for real-world applications of using sequential topological actions to effectively relieve grid congestion.

@en

Ground fault detection in inverter-based microgrid (IBM) systems is challenging, particularly in a real-time setting, as the fault current deviates slightly from the nominal value. This difficulty is reinforced when there are partially decoupled disturbances and modeling uncertainties. The conventional solution of installing more relays to obtain additional measurements is costly and also increases the complexity of the system. In this brief, we propose a data-assisted diagnosis scheme based on an optimization-based fault detection filter with the output current as the only measurement. Modeling the microgrid dynamics and the diagnosis filter, we formulate the filter design as a quadratic programming (QP) problem that accounts for decoupling partial disturbances, robustness to nondecoupled disturbances and modeling uncertainties by training with data, and ensuring fault sensitivity simultaneously. To ease the computational effort, we also provide an approximate but analytical solution to this QP. Additionally, we use classical statistical results to provide a thresholding mechanism that enjoys probabilistic false-alarm guarantees. Finally, we implement the IBM system with Simulink and real-time digital simulator (RTDS) to verify the effectiveness of the proposed method through simulations.

@en

Accelerated development of demand response service provision by the residential sector is crucial for reducing carbon-emissions in the power sector. Along with the infrastructure advancement, encouraging the end users to participate is crucial. End users highly value their privacy and control, and want to be included in the service design and decision-making process when creating the daily appliance operation schedules. Furthermore, unless they are financially or environmentally motivated, they are generally not prepared to sacrifice their comfort to help balance the power system. In this paper, we present an inverse-reinforcement-learning-based model that helps create the end users’ daily appliance schedules without asking them to explicitly state their needs and wishes. By using their past consumption data, the end consumers will implicitly participate in the creation of those decisions and will thus be motivated to continue participating in the provision of demand response services.

@en

The design of electricity markets may be facilitated by simulating actors’ behaviors. Recent studies model human decision-makers within markets as agents which learn strategies that maximize expected profits. This work investigates the problem of ‘non-stationarity’ in the context of market simulations, a problem with the learning-algorithms used by such studies which results in agents behaving irrationally, thus limiting the studies’ applicability to real-world strategic behavior. Isolating the source of the problem for a day-ahead electricity market, this paper proposes methods which meliorate this problem in simple test-cases, and proves requirements under which ‘centralized-training, decentralized-execution’ value-learning methods will converge to correct behavior in general. Subsequently, this paper proposes a framework for ‘adversarial market design’ that includes the market-designer as an agent. This allows the optimization of market-designs subject to possibly strategic behavior of participating firms — in turn enabling the automated selection of the optimal market from any set of markets.@en

The transition to green energy is reshaping the energy landscape, marked by increased integration of renewables, distributed resources, and the electrification of other energy sectors. These changes challenge grid security, particularly regarding the N-1 security criterion, a crucial factor in preventing blackouts. This necessitates studying the security constrained optimal power flow (SCOPF) problem with multiple line outages (N-k). Conventional methods exhibit poor scalability as k increases. This paper proposes a constraint-driven machine learning (ML) approach using line outage distribution factors (LODF). The method shows promise in its ability to scale effectively to N-k contingencies. Key contributions include a deterministic approach for N-k security and a probabilistic security assessment. Case studies on the IEEE 39-bus and the IEEE-118 bus systems show the approach’s effectiveness in identifying violating post-contingency cases, with up to 173x speedups and close to optimal dispatch costs.@en

This letter studies the problem of coordinating aggregators in the power system to provide fast frequency response as dynamic ancillary services. We approach the problem from the perspective of suboptimal H _∞ control, and propose an efficient and tractable formulation. We further develop a distributed solution method for the investigated problem, which enables aggregator agents to learn their optimal provisions in an adaptive way. More precisely, we reformulate the original problem into a state-based potential game, where the agents interact with each other towards our designed Nash equilibrium. The proposed game-theoretic learning approach decouples the coupling Linear Matrix Inequality constraint, guarantees the convergence to the equilibrium which is close enough to the original optimum. The learning process is also robust to the changes in communication graphs. We demonstrate the efficacy of our proposed approach with a case study on a 3-aggregator system.

@en

Power system operators require advanced applications in the control centers to tackle increasingly variable power transfers effectively. One urgently needed application concerns estimating the feasible available aggregated flexibility from a power system network, which can be effectively deployed to mitigate issues in interconnected networks. This paper proposes the TensorConvolution+ algorithm to address the above application. Unlike related literature approaches, TensorConvolution+ estimates the density of feasible flexibility combinations to reach a new operating point within the p-q flexibility area. This density can improve the decision-making of system operators for efficient and safe flexibility deployment. The proposed algorithm applies to radial and meshed networks, is adaptable to new operational conditions, and can consider scenarios with disconnected flexibility areas. Using convolutions and tensors, the algorithm efficiently aggregates the combinations of flexibility providers' adjustable power output that can occur for each flexibility area set point. Simulations on the meshed Oberrhein and radial CIGRE test networks illustrate the effectiveness of TensorConvolution+ for flexibility estimation with high numerical confidence and a minor computing effort. Additional simulations highlight how system operators can interpret the estimated density of feasible flexibility combinations for decision-making purposes, the algorithm's capability to estimate disconnected flexibility areas, and adapt to new operating conditions.

@en

This paper summarizes recent advancements on spatio-temporal data-driven and machine learning methods for static and dynamic security assessment, and their particular use cases. It is a collective effort of different research groups members of the IEEE Working Group on Big Data Analytics for Transmission Systems, to provide transmission system operators (TSOs) with innovative tools and ideas for their potential implementation. The algorithms presented here are classified as non-training and training approaches, namely spatio-temporal and machine learning based, considering as input time series from time domain simulations, and or synchrophasor data from wide-area monitoring systems. The efficacy of these algorithms is then evaluated in different IEEE benchmark models and using real system measurements from different countries.

@en

Power electronic interfaced devices progressively enable the increasing provision of flexible operational actions in distribution networks. The feasible flexibility these devices can effectively provide requires estimation and quantification so the network operators can plan operations close to real- time. Existing approaches estimating the distribution network flexibility require the full observability of the system, meaning topological and state knowledge. However, the assumption of full observability is unrealistic and represents a barrier to system operators' adaptation. This paper proposes a definition of the distribution network flexibility problem that considers the limited observability in real-time operation. A critical review and assessment of the most prominent approaches are done, based on the proposed definition. This assessment showcases the limitations and benefits of existing approaches for estimating flexibility with low observability. A case study on the CIGRE MV distribution system highlights the drawbacks brought by low observability.@en

With the increasing penetration of renewable power sources such as wind and solar, accurate short-term, nowcasting renewable power prediction is becoming increasingly important. This paper investigates the multi-modal (MM) learning and end-to-end (E2E) learning for nowcasting renewable power as an intermediate to energy management systems. MM combines features from all-sky imagery and meteorological sensor data as two modalities to predict renewable power generation that otherwise could not be combined effectively. The combined, predicted values are then input to a differentiable optimal power flow (OPF) formulation simulating the energy management. For the first time, MM is combined with E2E training of the model that minimises the expected total system cost. The case study tests the proposed methodology on the real sky and meteorological data from the Netherlands. In our study, the proposed MM- E2E model reduced system cost by 30% compared to uni-modal baselines.@en

The reader of the chapter will be able to connect techniques from machine learning (ML) and digital twins (DTs) to gain insights for monitoring and control of (dynamic) security for electrical power systems. DTs are validated and verified high-fidelity (hf) models providing high simulation accuracy. DTs can be used for simulation of the supervised process of system operation and are therefore able to provide synthetic studied data, where measurement data are scarce. However, for some real-time applications in monitoring and control, such high-fidelity simulation models are not appropriate due to the corresponding computational barrier. There, ML aims to create an application-specific, low-fidelity (lf) approximation of the digital twin. Such trained lf models are used in real-time applications where computational time is scarce and lf information is sufficient. The conceptual intersection of hf and lf models has been little explored and becomes increasingly complex. This chapter aims to provide a conceptual overview of how such hf and lf models can be combined. This chapter is split into two parts where the first part is to introduce ML, lf models, and digital twins, hf models, for power systems analysis, and the second chapter is to use these two types of models to form purpose-driven surrogate lf models, illustrated on the example of dynamic security assessment (DSA). In the first part, the concepts for using DTs as hf models for online power system studies and their corresponding tuning of model parameters are introduced. Subsequently, ML i.e., lf models, are introduced and their corresponding training frameworks.

@en

Due to the increasing system stability issues caused by the technological revolutions of power system equipment, the assessment of the dynamic security of the systems for changing operating conditions (OCs) is nowadays crucial. To address the computational time problem of conventional dynamic security assessment tools, many machine learning (ML) approaches have been proposed and well-studied in this context. However, these learned models only rely on data, and thus miss resourceful information offered by the physical system. To this end, this paper focuses on combining the power system dynamical model together with the conventional ML. Going beyond the classic Physics Informed Neural Networks (PINNs), this paper proposes Selected Physics Informed Neural Networks (SPINNs) to predict the system dynamics for varying OCs. A two-level structure of feed-forward NNs is proposed, where the first NN predicts the generator bus rotor angles (system states) and the second NN learns to adapt to varying OCs. We show a case study on an IEEE-9 bus system that considering selected physics in model training reduces the amount of needed training data. Moreover, the trained model effectively predicted long-term dynamics that were beyond the time scale of the collected training dataset (extrapolation).

@en

Machine learning (ML) for real-time security assessment requires a diverse training database to be accurate for scenarios beyond historical records. Generating diverse operating conditions is highly relevant for the uncertain future of emerging power systems that are completely different to historical power systems. In response, for the first time, this work proposes a novel split-based sequential sampling approach based on optimisation that generates more diverse operation scenarios for training ML models than state-of-the-art approaches. This work also proposes a volume-based coverage metric, the convex hull volume (CHV), to quantify the quality of samplers based on the coverage of generated data. This metric accounts for the distribution of samples across multidimensional space to measure coverage within the physical network limits. Studies on IEEE test cases with 6, 68 and 118 buses demonstrate the efficiency of the approach. Samples generated using the proposed split-based sampling cover 37.5% more volume than random sampling in the IEEE 68-bus system. The proposed CHV metric can assess the quality of generated samples (standard deviation of 0.74) better than a distance-based coverage metric which outputs the same value (standard deviation of <0.001) for very different data distributions in the IEEE 68-bus system. As we demonstrate, the proposed split-based sampling is relevant as a pre-step for training ML models for critical tasks such as security assessment.

@en

Implementing accurate Distribution System State Estimation (DSSE) faces several challenges, among which the lack of observability and the high density of the distribution system. While data-driven alternatives based on Machine Learning models could be a choice, they suffer in DSSE because of the lack of labeled data. In fact, measurements in the distribution system are often noisy, corrupted, and unavailable. To address these issues, we propose the Deep Statistical Solver for Distribution System State Estimation (DSS^2), a deep learning model based on graph neural networks (GNNs) that accounts for the network structure of the distribution system and the governing power flow equations of the problem. DSS^2 is based on GNN and leverages hypergraphs to model the network as a graph into the deep-learning algorithm and to represent the heterogeneous components of the distribution systems. A weakly supervised learning approach is put forth to train the DSS^2: by enforcing the GNN output into the power flow equations, we force the DSS^2 to respect the physics of the distribution system. This strategy enables learning from noisy measurements and alleviates the need for ideal labeled data. Extensive experiments with case studies on the IEEE 14-bus, 70-bus, and 179-bus networks showed the DSS^2 outperforms the conventional Weighted Least Squares algorithm in accuracy, convergence, and computational time while being more robust to noisy, erroneous, and missing measurements. The DSS^2 achieves a competing, yet lower, performance compared with the supervised models that rely on the unrealistic assumption of having all the true labels.

@en

This paper presents a novel, unified approach for generating high-quality datasets for training machine-learned models for real-time security assessment in power systems. Synthetic data generation methods that extrapolate beyond historical data can be inefficient in generating feasible and rare operating conditions (OCs). The proposed approach balances the trade-off between historically relevant OCs and rare but feasible OCs. Unlike conventional methods that rely on historical records or generic sampling, our approach results in datasets that generalise well beyond similar distributions. The proposed approach is validated through experiments on the IEEE 118-bus system, where a decision tree model trained on data generated using our approach achieved 97% accuracy in predicting the security label of rare OCs, outperforming baseline approaches by 41% and 20%. This work is crucial for deploying reliable machine-learned models for real-time security assessment in power systems undergoing decarbonisation and integrating renewable energy sources.

@en

With recent telemetric advancements, the real-time availability of power grid measurements has opened challenging opportunities for the design of advanced protection and control schemes. Artificial neural networks (ANN) are promising approaches for detecting and classifying disturbance events from measurement data. Numerous offline ANN-based classification algorithms were proposed in the past, which increased the interest for their real-world deployment. However, these algorithms are inadequate due to their conventional offline training procedures, model updating, and large backend computing requirements. Besides, most ANN-based algorithms require disturbance event samples to be collectively available during training. This availability may be uncommon in practice as disturbance events are rare, non-deterministic, and uncertain. Hence, an online training procedure where the model processes the events on-the-fly is required. However, ANNs may also suffer from catastrophic forgetting where the model may unintentionally unlearn an occurred disturbance under the learning of new event types; this means ANN may not detect very similar disturbances of the same type in the future. In this paper, we propose Dynamic Incremental Learning (IL) method for ANN models, which is updated in real-time when a new disturbance is detected. Our proposed method adopts a Replay-based IL strategy for designing long-term IL, balancing the accuracy with catastrophic forgetting of disturbance events. The method is designed in a way to learn efficiently for incoming disturbance data with minimized training time and the highest classification accuracy eliminating catastrophic forgetting. The results describe the methodology’s performance regarding classification accuracy, training time, and storage memory. The findings demonstrate that the Dynamic IL method is promising for efficient learning and event classification.@en

Weather forecast models are essential for sustainable energy systems. However, forecast accuracy may not be the best metric for developing forecast models. A more or less conservative forecast may be preferred over pure accuracy. For example, forecasting accurately in times of energy-deprived situations may be more important than in times of excess wind power generation. This is not accounted for when learning a wind power forecast without system knowledge. Wind power forecasts directly impact system energy schedules. Therefore, to optimise system costs, our paper proposes a neural network structure for wind power forecasts directly considering varying energy system conditions. To train this neural network optimally, this paper models the system dispatch and their costs as an optimisation problem. Then, this paper connects the neural network with the optimisation model. In this connection, the implicit function theorem and the Karush–Kuhn–Tucker (KKT) conditions provide the system cost gradients during the training of the neurons. A case study using onshore and offshore weather data from Germany and The Netherlands showed forecast errors of system costs reduced by up to 10% with high wind capacity. Wind curtailment was reduced by more than 20% for individual cases but increased overall. Modified cases of the optimal power flow (OPF) with tighter line constraints led to a higher advantage of the proposed method. The approach led to more consistent forecast performance and reduced error variance by up to 70%.

@en

This paper presents a decentralized Multi-Agent Reinforcement Learning (MARL) approach to an incentive-based Demand Response (DR) program, which aims to maintain the capacity limits of the electricity grid and prevent grid congestion by financially incentivizing residential consumers to reduce their energy consumption. The proposed approach addresses the key challenge of coordinating heterogeneous preferences and requirements from multiple participants while preserving their privacy and minimizing financial costs for the aggregator. The participant agents use a novel Disjunctively Constrained Knapsack Problem optimization to curtail or shift the requested household appliances based on the selected demand reduction. Through case studies with electricity data from 25 households, the proposed approach effectively reduced energy consumption's Peak-to-Average ratio (PAR) by 14.48% compared to the original PAR while fully preserving participant privacy. This approach has the potential to significantly improve the efficiency and reliability of the electricity grid, making it an important con-tribution to the management of renewable energy resources and the growing electricity demand.@en

Presently, transmission system operators are tackling challenging dynamic issues in scenarios close to real-time utilizing their dynamic stability assessment tools and data acquisition devices that have in operation. These devices use different types of technology and the majority of the tools used in the control centers are tailored to each of them. Most challenges affecting the security of these systems are associated with current paradigms of management and planning such as the inclusion of deregulated markets, the interconnection with neighboring regional systems, the diversification of energy sources, and the involvement of environmental constraints. Under these conditions, the security and reliability of the transmission system can be compromised with unexpected disturbances causing violations of the security limits established and subsequently leading to the system collapse. To contribute to the finding of solutions to these problems, the members of this IEEE Task Force have worked together to gather information from eleven utilities in nine different countries in America and Europe. The main objective is to establish a baseline of these common concerns, which can be used to develop, in the next step, solutions based on machine learning and data-driven innovative algorithms. @en

Machine learning has been used in the past to construct predictors, also known as classifiers, for dynamic security assessment. Although accurate classifiers can be trained for a single topology, often they do not work for another. However, the power system topology can change frequently during operation due to maintenance and control actions. At one topological configuration, the system may have a different response to a fault than at another as the underlying distribution of power flows can be completely different. Quantifying the impact of changes in the topology on the predictive models’ performance is an important step forward to minimize inaccurate predictions and improve their reliability. In this paper, for the first time, a metric for quantifying the impact of a topology change on the accuracy of the classification model is proposed. The key novelty is to first select a subset of power flow features with a physically informed feature selection technique and subsequently compute the metric with a novel convex hull-based analysis. In addition, the approach can advise to effectively constructing new training databases that improve the accuracy of new machines trained after high-impact topology changes. Through a case study using transient stability on the IEEE 68-bus system, the use of the proposed metric in real-time operation was demonstrated. 17 high-impact topology changes were successfully detected among 42 studied topological changes. The subsequent effective construction of the training database improved the predictive accuracy by around 10%. An interesting finding is the amount of newly generated data can be reduced by up to 85% as often the generated data is the barrier for data-driven DSA. The proposed workflow significantly reduces data and trains robust classifiers against topological changes marking a fundamental step forward.

@en