DC microgrids (MGs) are cyber-physical systems (CPSs) prone to cyber attacks which could disrupt the normal operation of DC MGs. Accurate estimation of the attack vector is crucial to recover correct signals from compromised measurements for safe DC MG operation, while it has not been effectively achieved by existing methods and the accuracy is challenged by unmodeled uncertainties in practical power electronic converters. This paper proposes a digital twin (DT)-based cyber attack detection and mitigation scheme for DC MGs. First, the lightweight radial basis function neural network (RBFNN) is adopted to compensate for the mismatch between the ideal model and the real system for accurate converter modeling. Second, a composite descriptor observer-based local DT is designed to achieve accurate estimations of attack signals and correct observations of converter states. In addition, a global DT is developed at the system level to accurately estimate and eliminate cyber attacks in the secondary control. As a result, the proposed method can mitigate attacks by replacing the corrupted signals with estimated true values provided by DT, leading to accurate and stable operation of the system. Finally, simulation and experimental results are given to validate the effectiveness of the proposed method.

The information security of IEC 61850-compliant substations is a growing concern for researchers and industry practitioners. IEC 62351, developed to address such concerns, recommends the use of intrusion detection systems (IDSs) as a defense, prompting extensive research on their development, particularly in data-driven approaches. Data-driven IDSs rely on high-quality and comprehensive training data. However, capturing complete datasets for each unique substation at scale is challenging due to the diverse and dynamic operating states between substations. Transfer learning (TL) has been shown to improve model performance in data-scarce environments; however, to the best of our knowledge, no prior work has formulated its use in the context of knowledge transfer between IEC 61850 substations. To address this gap, we propose cross-substation transfer learning (XSTL), a strategy that leverages knowledge transfer between substations that share the same protocol stack but differ in architecture. We demonstrate the value of XSTL using two publicly available datasets collected from substations with contrasting architectures, and show that XSTL can improve IDS performance compared to training IDSs in an isolated manner. Using data from a generic object-oriented substation event (GOOSE) flooding attack, we show that IDS performance is significantly improved in cross-domain tests (using data from two different substations) compared with baseline tests (using data from one substation), with statistical analyses confirming the significance of the improvement. These findings indicate that XSTL can reduce reliance on large datasets, thereby enabling more practical and scalable IDS development across substations where collecting diverse training data is challenging.

Power systems are evolving in response to the growing integration of variable renewable resources. Considering this, balancing reserves are becoming more important for maintaining the continuous balance between total production and demand in the power system. However, the allocation of these reserves comes at a cost, given their inherent link to the energy traded in the day-ahead electricity market. This study examines the implications of different reserve allocation strategies using a modeling framework that incorporates dispatchable hydropower and intermittent wind power alongside demand to evaluate the costs of securing sufficient reserves in a highly renewable energy system. The analysis indicates that aligning reserve allocation more closely with the conditions expected at the time of operation results in improved outcomes, leading to better utilization of stored water in hydropower reservoirs and lower operational costs. Therefore, considering the increasing integration of variable renewable energy sources, the findings highlight the need to investigate more effective approaches for reserve allocation that are better aligned with anticipated system conditions.

With the growing electrification and integration of renewables, network operators face unprecedented challenges. Coordinated control of Flexible AC Transmission Systems (FACTS) setpoints using real-time optimization techniques has been proposed to substantially improve voltage and power flow control. However, optimizing the setpoints of several FACTS devices is rarely done in practice. In part, this can be derived from the challenges with model-based methods. As alternative control methods, data-driven methods based on reinforcement learning (RL) have shown great promise. However, RL has its own challenges that include data and safety during learning. Motivated by the increasing collection of data, we study an RL-based optimization of FACTS setpoints and how datasets can be leveraged for pre-training to improve safety. We demonstrate on the IEEE 14-bus and IEEE 57-bus systems that an offline to online RL algorithm can significantly reduce voltage deviations and constraint violations. The performance is compared against an RL agent learning from scratch and the original control policy that generated the dataset. Moreover, our analysis shows that dataset coverage and the amount of pre-training updates affect the performance considerably. Finally, to identify the gap to an optimal policy, the proposed approach is benchmarked against an optimal controller with perfect information.

There is limited application of closed-loop control using model-based approaches in wide area monitoring, protection, and control. Challenges that impede model-based approaches include engineering complexity, convergence issues, and model errors. Specifically, considering the rapid growth of distributed generation and renewables in the grid, maintaining an updated model without model errors is challenging. As an alternative to model-based approaches, data-driven control architectures based on reinforcement learning (RL) have shown great promise. In this work, we confront safety concerns with data-driven approaches by studying safe RL to improve voltage and power flow control. For both a model-free RL agent and a model-based RL agent, the accumulated constraint violation is investigated in a case study on the IEEE 14-bus and IEEE 57-bus systems. To evaluate performance, agents are compared against a model-based approach subject to errors. Our findings suggest that RL could be considered for optimizing voltage and current setpoints in systems when topological model errors are present.

This paper introduces FDILAB (False Data Injection LABoratory), an open-source educational tool tailored for power system cybersecurity and machine learning experiments. It offers an intuitive interface, accommodating users of various computer skill levels and bridging the gap between power systems and data science expertise. In addition to a brief explanation of the application architecture, we showcase a set of FDILAB's capabilities by using it to address an anomaly detection problem relevant to power system cybersecurity and machine learning. The demonstration includes data generation, developing and training a well-known machine learning model, and testing it within the FDILAB environment While room for improvement remains, we now consider the application suitable for classroom use and plan to include it in a future course for master's students in power systems engineering.

Decentralized energy generation, often in the form of industrial Combined Heat and Power (CHP) units, meets a significant part of the global energy demand. The power production from these units has to be forecast, both individually and collectively, to ensure balance and avoid congestion events on the power grid. However, despite its importance, forecasting CHP generation remains an under-studied problem. With increasing proliferation of renewable energy sources, large forecast errors for CHP generation are rapidly taking center-stage as well. In this paper, we propose a global, ensemble-based machine learning (ML) method to improve short-term forecasting accuracy. We demonstrate its efficacy using data from one hundred largest (industrial) CHP units in Belgium, showing a forecast error reduction of up to 27% compared to the baseline method. This can greatly reduce balancing needs and costs, as well as congestion issues. Our results also highlight several nuances that must be kept in mind by grid operators and market players alike, including fitting global models that leverage data from many CHPs which can lead to better forecast accuracy but limit interpretability of the results, and the fact that there is no single best forecast model for all CHPs.

KTH har lämnat in ett remissvar på utredningen ”Promemoria Finansiering och riskdelning vid investeringar i ny kärnkraft”, Finansdepartementet, Fi 2023:F, Augusti 2024" vilket vi stöder. Vi vill dock lämna in följande komplettering. Utredningen är oerhört omfattande, och det behövs därmed mycket mer detaljerade kommentarer än de som fick utrymme i KTH:s mer övergripande svar.

Av: Cecilia Hermansson: docent och forskare i fastighetsekonomi och finans, tidigare ledamot i Finanspolitiska rådet och tidigare ordförande i Klimatpolitiska rådet. Lars Nordström: professor i kraftsystemstyrning och tillhörande informationsutbyte. Lennart Söder: professor i elektriska energisystem. KTH

Sammanfattning

Kärnkraft är en viktig teknologi i dagens elsystem. Utöver energi och effekt bidrar den med systemstabiliserande egenskaper. Som ett tecken på betydelsen av kraftslaget har befintliga kärnkraftverk genomgått flera moderniseringar och insatser för effektökning. Sannolikt pågår fortsatta planer på ytterligare livstidsförlängning av dagens kärnkraftverk. Utöver detta finns flera intressanta utvecklingsmöjligheter med småskaliga reaktorer, nya reaktorteknologier, återvinning av använt kärnbränsle, nya säkerhetssystem etc. Denna utveckling, d v s både livstidsförlängning av existerande verk, och utveckling av nya teknologier ställer krav på kompetensförsörjning och forskning. Sverige har goda möjligheter att ta en ledande roll i utvecklingen av kommande generationer av kärnkraftsteknologi. Syftet med den föreliggande utredningen är dock inte att stödja sådan teknisk utveckling utan istället att så snabbt som möjligt komma till beslut om storskalig investering i redan etablerad kärnkraftsteknologi. Det är av mycket stor betydelse att denna skyndsamhet inte förhindrar den nödvändiga kompetensuppbyggnaden och tekniska utvecklingen som behöver ske på området.Det är i grunden en politisk fråga om kostnaden för elkraftssystemet ska hamna på elkunderna eller på skattebetalarna. Men oavsett storleken på statens inblandning i finansieringen så är det centralt att man får ett kostnadseffektivt totalt elkraftssystem där kraftverk konkurrerar med sina förmågor, där man får betalt för vad man bidrar med och får betala för de kostnader man orsakar. Det är därför mycket välkommet att det kommer en utredning som beskriver kärnkraftens bidrag och kostnader samt metoder för att göra kärnkraften till en attraktiv möjlighet för investerare. Investeringskostnaden förtillkommande kärnkraft bedöms i utredningen vara ”80 miljoner kronor/MW” ( [1] sidan 84) vilket är högre än tidigare uppskattningar. Förutom detta grundantagande beaktas även en möjlig kostnadsöverskridande om 100 procent ( [1] sidan 181) i avsnittet om riskdelning. Det är alltså dessa kostnader som skall ställas i relation till de nyttor som den tillkommande kärnkraften kommer lämna som bidrag till elsystemet. Vi menar att utredningen inte har klarat av att visa att kostnaderna motsvarar nyttan med investeringen. Vi ställer oss därför inte bakom förslaget om finansiering och riskdelning av ny kärnkraft enligt utredningens modell. Vårt remissvar är indelat i tre delområden, påverkan på elsystemet, elmarknaden samt samhällsekonomin och de offentliga finanserna. Slutsatserna från dessa sammanfattas i bifogat dokument.

It is envisioned that the future electric grid will be underpinned by a vast number of smart inverters linking renewables at the grid edge. These inverters' dynamics are typically characterized as impedances, which are crucial for ensuring grid stability and resiliency. However, the physical implementation of these inverters may vary widely and may be kept confidential. Existing analytical impedance models require a complete and precise understanding of system parameters. They can hardly capture the complete electrical behavior when the inverters are performing complex functions. Online impedance measurements for many inverters across multiple operating points are impractical. To address these issues, we present the InvNet, a machine learning framework capable of characterizing inverter impedance patterns across a wide operation range, even with limited impedance data. Leveraging transfer learning, the InvNet can extrapolate from physics-based models to real-world ones and from one inverter to another with the same control framework but different control parameters with very limited data. This framework demonstrates machine learning as a powerful tool for modeling and analyzing black-box characteristics of grid-tied inverter systems that cannot be accurately described by traditional analytical methods, such as inverters under model-predictive control. Comprehensive evaluations were conducted to verify the effectiveness of the InvNet.

Data-driven approaches are promising to address the modeling issues of modern power electronics-based power systems, due to the black-box feature. Frequency-domain analysis has been applied to address the emerging small-signal oscillation issues caused by converter control interactions. However, the frequency-domain model of a power electronic system is linearized around a specific operating condition. It thus requires measurement or identification of frequency-domain models repeatedly at many operating points (OPs) due to the wide operation range of the power systems, which brings significant computation and data burden. This article addresses this challenge by developing a deep learning approach using multilayer feedforward neural networks (FNNs) to train the frequency-domain impedance model of power electronic systems that is continuous of OP. Distinguished from the prior neural network designs relying on trial-and-error and sufficient data size, this article proposes to design the FNN based on latent features of power electronic systems, i.e., the number of system poles and zeros. To further investigate the impacts of data quantity and quality, learning procedures from a small dataset are developed, and K-medoids clustering based on dynamic time warping is used to reveal insights into multivariable sensitivity, which helps improve the data quality. The proposed approaches for the FNN design and learning have been proven simple, effective, and optimal based on case studies on a power electronic converter, and future prospects in its industrial applications are also discussed.