arXiv subjects

Veit Hagenmeyer

Publications and source records attributed to Veit Hagenmeyer.

At least 37 records · Page 2Linked to original sources

arXiv2026

CritBench: A Framework for Evaluating Cybersecurity Capabilities of Large Language Models in IEC 61850 Digital Substation Environments

The advancement of Large Language Models (LLMs) has raised concerns regarding their dual-use potential in cybersecurity. Existing evaluation frameworks overwhelmingly focus on Information Technology (IT) environments, failing to capture the constraints, and specialized protocols of Operational Technology (OT). To address this gap, we introduce CritBench, a novel framework designed to evaluate the cybersecurity capabilities of LLM agents within IEC 61850 Digital Substation environments. We assess five state-of-the-art models, including OpenAI's GPT-5 suite and open-weight models, across a corpus of 81 domain-specific tasks spanning static configuration analysis, network traffic reconnaissance, and live virtual machine interaction. To facilitate industrial protocol interaction, we develop a domain-specific tool scaffold. Our empirical results show that agents reliably execute static structured-file analysis and single-tool network enumeration, but their performance degrades on dynamic tasks. Despite demonstrating explicit, internalized knowledge of the IEC 61850 standards terminology, current models struggle with the persistent sequential reasoning and state tracking required to manipulate live systems without specialized tools. Equipping agents with our domain-specific tool scaffold significantly mitigates this operational bottleneck. Code and evaluation scripts are available at: https://github.com/GKeppler/CritBench

cs.CR↗

arXiv2026

Stochastic Model Predictive Control based on Mixed Random Variables for Economic Energy Management

Optimal scheduling of batteries has significant potential to reduce electricity costs and to enhance grid resilience. However, effective battery scheduling must account for both physical constraints as well as uncertainties in consumption and generation of renewable energy sources. Instead of optimizing fixed battery power setpoints, we propose an approach that optimizes battery power intervals, allowing the optimization to explicitly account for uncertain consumption and generation as well as how the battery system should respond to them within its physical limits. Our method is based on mixed random variables, represented as mixtures of discrete and continuous probability distributions. Building on this representation, we develop an analytical stochastic formulation for minimizing electricity costs in a residential setting with load, photovoltaics, and battery storage. We demonstrate its effectiveness across real-world data from 15 residential buildings over five consecutive months. Compared with deterministic and probabilistic benchmark controllers, the proposed interval-based optimization achieves the lowest costs. These results show that mixed random variables are a practical and promising tool for decision-making under uncertainty.

math.OC↗

arXiv2026

Knowledge Distillation for Efficient Transformer-Based Reinforcement Learning in Hardware-Constrained Energy Management Systems

Transformer-based reinforcement learning has emerged as a strong candidate for sequential control in residential energy management. In particular, the Decision Transformer can learn effective battery dispatch policies from historical data, thereby increasing photovoltaic self-consumption and reducing electricity costs. However, transformer models are typically too computationally demanding for deployment on resource-constrained residential controllers, where memory and latency constraints are critical. This paper investigates knowledge distillation to transfer the decision-making behaviour of high-capacity Decision Transformer policies to compact models that are more suitable for embedded deployment. Using the Ausgrid dataset, we train teacher models in an offline sequence-based Decision Transformer framework on heterogeneous multi-building data. We then distil smaller student models by matching the teachers' actions, thereby preserving control quality while reducing model size. Across a broad set of teacher-student configurations, distillation largely preserves control performance and even yields small improvements of up to 1%, while reducing the parameter count by up to 96%, the inference memory by up to 90%, and the inference time by up to 63%. Beyond these compression effects, comparable cost improvements are also observed when distilling into a student model of identical architectural capacity. Overall, our results show that knowledge distillation makes Decision Transformer control more applicable for residential energy management on resource-limited hardware.

cs.LG↗

arXiv2026

RTS-ABAC: Real-Time Server-Aided Attribute-Based Authorization & Access Control for Substation Automation Systems

Critical energy infrastructures increasingly rely on information and communication technology for monitoring and control, which leads to new challenges with regard to cybersecurity. Recent advancements in this domain, including attribute-based access control (ABAC), have not been sufficiently addressed by established standards such as IEC 61850 and IEC 62351. To address this issue, we propose a novel real-time server-aided attribute-based authorization and access control for time-critical applications called RTS-ABAC. We tailor RTS-ABAC to the strict timing constraints inherent to the protocols employed in substation automation systems (SAS). We extend the concept of conventional ABAC by introducing real-time attributes and time-dependent policy evaluation and enforcement. To safeguard the authenticity, integrity, and non-repudiation of SAS communication and protect an SAS against domain-typical adversarial attacks, RTS-ABAC employs mandatory authentication, authorization, and access control for any type of SAS communication using a bump-in-the-wire (BITW) approach. To evaluate RTS-ABAC, we conduct a testbed-based performance analysis and a laboratory-based demonstration of applicability. We demonstrate the applicability using intelligent electronic devices, merging units, and I/O boxes communicating via the GOOSE and SV protocol. The results show that RTS-ABAC is able to secure low-latency communication between SAS devices, as up to 99.82 % of exchanged packets achieve a round-trip time below 6 ms. Moreover, the results of the evaluation indicate that RTS-ABAC is a viable solution to enhance the cybersecurity not only in a newly constructed SAS but also via retrofitting of existing substations.

cs.CR↗

arXiv2026

JanusBM: A Dual-Fidelity Multi-Zone White-Box Building Modeling Framework

Accurate building energy models are crucial for analyzing sector-coupled energy systems, where buildings interact with electrified heating, energy storage, and advanced control across various scenarios. High-fidelity (HiFi) white-box models that resolve hydronic distribution and emitter dynamics can capture short-term transients, yet their numerical stiffness and computational burden limit long-term simulations and large-scale scenario exploration. Conversely, reduced-order low-fidelity (LoFi) representations enable rapid annual assessments but may fail to capture the hydronic- and control-induced dynamics that govern transient and peak behavior. This paper proposes a dual-fidelity, multi-zone white-box building modeling framework, which is called JanusBM, built on a novel topology-driven modeling tool RoomFlex6D, coupling a HiFi hydronic model and a LoFi ideal-load surrogate that removes explicit hydronic states in Modelica. To ensure applicability and physical consistency across time scales, we introduce a two-stage hybrid validation and calibration pipeline that uses complementary data: the IEA EBC Annex 60 benchmark for energy-scale validation and time-series measurements from real-world experimental buildings for hydronic dynamics-scale calibration. Results show that the generated LoFi models achieve a high degree of consistency with Annex 60 benchmark on the energy scale, and the proposed calibration workflow robustly improves loop-level return water temperature transients and zone-level temperature dynamics. Moreover, the LoFi model achieves orders-of-magnitude faster simulations suited to annual energy analyses, whereas the HiFi model becomes necessary when the required heat differs from the actual delivered heat due to distribution and control limitations, especially in transient and peak-oriented assessments.

eess.SY↗

arXiv2026

Improving Spatial Allocation for Energy System Coupling with Graph Neural Networks

In energy system analysis, coupling models with mismatched spatial resolutions is a significant challenge. A common solution is assigning weights to high-resolution geographic units for aggregation, but traditional models are limited by using only a single geospatial attribute. This paper presents an innovative method employing a self-supervised Heterogeneous Graph Neural Network to address this issue. This method models high-resolution geographic units as graph nodes, integrating various geographical features to generate physically meaningful weights for each grid point. These weights enhance the conventional Voronoi-based allocation method, allowing it to go beyond simply geographic proximity by incorporating essential geographic information.In addition, the self-supervised learning paradigm overcomes the lack of accurate ground-truth data. Experimental results demonstrate that applying weights generated by this method to cluster-based Voronoi Diagrams significantly enhances scalability, accuracy, and physical plausibility, while increasing precision compared to traditional methods.

cs.LG↗

arXiv2026

A Multi-Agent System for Building-Age Cohort Mapping to Support Urban Energy Planning

Determining the age distribution of the urban building stock is crucial for sustainable municipal heat planning and upgrade prioritization. However, existing approaches often rely on datasets gathered via sensors or remote sensing techniques, leaving inconsistencies and gaps in data. We present a multi-agent LLM system comprising three key agents, the Zensus agent, the OSM agent, and the Monument agent, that fuse data from heterogeneous sources. A data orchestrator and harmonizer geocodes and deduplicates building imprints. Using this fused ground truth, we introduce BuildingAgeCNN, a satellite-only classifier based on a ConvNeXt backbone augmented with a Feature Pyramid Network (FPN), CoordConv spatial channels, and Squeeze-and-Excitation (SE) blocks. Under spatial cross validation, BuildingAgeCNN attains an overall accuracy of 90.69% but a modest macro-F1 of 67.25%, reflecting strong class imbalance and persistent confusions between adjacent historical cohorts. To mitigate risk for planning applications, the address-to prediction pipeline includes calibrated confidence estimates and flags low-confidence cases for manual review. This multi-agent LLM system not only assists in gathering structured data but also helps energy demand planners optimize district-heating networks and target low-carbon sustainable energy systems.

cs.CV↗

arXiv2026

Period-aware asymptotic gain with application to a periodically forced synchronization circuit

The classical asymptotic gain (AG) is a concept known from the input-to-state stability theory. Given a uniform input bound, AG estimates the asymptotic bound of the output. Sometimes, however, more information is known about the input than just a bound. In this paper we consider the case of a periodic input. Under the assumption that the system converges to a periodic solution, we introduce a new gain, called period-aware asymptotic gain (PAG), which employs periodicity to enable a sharper asymptotic estimation of the output. Since the PAG can distinguish between short-period ("high-frequency") and long-period ("low-frequency") signals, it is able to rigorously quantify such properties as bandwidth, resonant behavior, and high-frequency damping. We discuss how the PAG can be computed and illustrate it with a numerical example from the field of power electronics.

math.OC↗

arXiv2026

Hardware test and validation of the angular droop control: Analysis and experiments

We present a hardware-based validation of angular droop control for grid-forming DC/AC converters, a control strategy that establishes active power-to-angle droop. Angular droop control enables exact frequency regulation at steady state, thereby combining primary and secondary control into a single layer. We provide traceable analysis and suggest solutions to the main implementation challenges with angular droop control, specifically addressing the challenges concerning discretization and clock drift in hardware experiments. This is illustrated in two different scenarios. Experimental results from the single converter to load scenario demonstrate black start capability and power-to-angle droop behavior for two different implementation schemes. A multi-converter setup validates frequency synchronization and power-sharing properties, proving the ancillary services that angular droop control provides in the real-world experimental setup.

eess.SY↗

arXiv2026

Fast-Response Balancing Capacity of Alkaline Electrolyzers

The energy transition requires flexible technologies to maintain grid stability, and electrolyzers are playing an increasingly important role in meeting this need. While previous studies often question the dynamic capabilities of large-scale alkaline electrolyzer systems, we assess their potential to provide balancing services using real manufacturer data. Unlike common approaches, we propose the decoupling between the total electrolyzer power and a smaller fractions of power actually offered on balancing markets. Adapting an existing methodology, we analyze alkaline electrolyzer systems and extend the assessment to Germany and Europe. Our results show that large-scale electrolyzers are technically capable of delivering fast-response balancing services, with significantly lower dynamic requirements than previously assumed. The planned electrolyzers in Germany could cover the entire balancing capacity market, potentially saving around 13 % of their electricity costs, excluding energy balancing revenues. The decoupling also resolves part of the trade-off for electrolyzer manufacturers, enabling the design of less dynamic but more stable systems.

eess.SY↗

arXiv2026

HeatPrompt: Zero-Shot Vision-Language Modeling of Urban Heat Demand from Satellite Images

Accurate heat-demand maps play a crucial role in decarbonizing space heating, yet most municipalities lack detailed building-level data needed to calculate them. We introduce HeatPrompt, a zero-shot vision-language energy modeling framework that estimates annual heat demand using semantic features extracted from satellite images, basic Geographic Information System (GIS), and building-level features. We feed pretrained Large Vision Language Models (VLMs) with a domain-specific prompt to act as an energy planner and extract the visual attributes such as roof age, building density, etc, from the RGB satellite image that correspond to the thermal load. A Multi-Layer Perceptron (MLP) regressor trained on these captions shows an $R^2$ uplift of 93.7% and shrinks the mean absolute error (MAE) by 30% compared to the baseline model. Qualitative analysis shows that high-impact tokens align with high-demand zones, offering lightweight support for heat planning in data-scarce regions.

cs.CV↗

arXiv2026

Self-Supervised Temporal Super-Resolution of Energy Data using Generative Adversarial Transformer

To bridge the temporal granularity gap in energy network design and operation based on Energy System Models, resampling of time series is required. While conventional upsampling methods are computationally efficient, they often result in significant information loss or increased noise. Advanced models such as time series generation models, Super-Resolution models and imputation models show potential, but also face fundamental challenges. The goal of time series generative models is to learn the distribution of the original data to generate high-resolution series with similar statistical characteristics. This is not entirely consistent with the definition of upsampling. Time series Super-Resolution models or imputation models can degrade the accuracy of upsampling because the input low-resolution time series are sparse and may have insufficient context. Moreover, such models usually rely on supervised learning paradigms. This presents a fundamental application paradox: their training requires the high-resolution time series that is intrinsically absent in upsampling application scenarios. To address the mentioned upsampling issue, this paper introduces a new method utilizing Generative Adversarial Transformers (GATs), which can be trained without access to any ground-truth high-resolution data. Compared with conventional interpolation methods, the introduced method can reduce the root mean square error (RMSE) of upsampling tasks by 10%, and the accuracy of a model predictive control (MPC) application scenario is improved by 13%.

cs.LG↗

arXiv2025

Explainable time-series forecasting with sampling-free SHAP for Transformers

Time-series forecasts are essential for planning and decision-making in many domains. Explainability is key to building user trust and meeting transparency requirements. Shapley Additive Explanations (SHAP) is a popular explainable AI framework, but it lacks efficient implementations for time series and often assumes feature independence when sampling counterfactuals. We introduce SHAPformer, an accurate, fast and sampling-free explainable time-series forecasting model based on the Transformer architecture. It leverages attention manipulation to make predictions based on feature subsets. SHAPformer generates explanations in under one second, several orders of magnitude faster than the SHAP Permutation Explainer. On synthetic data with ground truth explanations, SHAPformer provides explanations that are true to the data. Applied to real-world electrical load data, it achieves competitive predictive performance and delivers meaningful local and global insights, such as identifying the past load as the key predictor and revealing a distinct model behavior during the Christmas period.

cs.LG↗

arXiv2025

ComEMS4Build: Comfort-Oriented Energy Management System for Residential Buildings using Hydrogen for Seasonal Storage

Integrating flexible loads and storage systems into the residential sector contributes to the alignment of volatile renewable generation with demand. Besides batteries serving as a short-term storage solution, residential buildings can benefit from a Hydrogen (H2) storage system, allowing seasonal shifting of renewable energy. However, as the initial costs of H2 systems are high, coupling a Fuel Cell (FC) with a Heat Pump (HP) can contribute to the size reduction of the H2 system. The present study develops a Comfort-Oriented Energy Management System for Residential Buildings (ComEMS4Build) comprising Photovoltaics (PV), Battery Energy Storage System (BESS), and H2 storage, where FC and HP are envisioned as complementary technologies. The fuzzy-logic-based ComEMS4Build is designed and evaluated over a period of 12 weeks in winter for a family household building in Germany using a semi-synthetic modeling approach. The Rule-Based Control (RBC), which serves as a lower benchmark, is a scheduler designed to require minimal inputs for operation. The Model Predictive Control (MPC) is intended as a cost-optimal benchmark with an ideal forecast. The results show that ComEMS4Build, similar to MPC, does not violate the thermal comfort of occupants in 10 out of 12 weeks, while RBC has a slightly higher median discomfort of 0.68 Kh. The ComEMS4Build increases the weekly electricity costs by 12.06 EUR compared to MPC, while RBC increases the weekly costs by 30.14 EUR. The ComEMS4Build improves the Hybrid Energy Storage System (HESS) utilization and energy exchange with the main grid compared to the RBC. However, when it comes to the FC operation, the RBC has an advantage, as it reduces the toggling counts by 3.48% and working hours by 7.59% compared to MPC...

eess.SY↗

arXiv2025

Averaging favors MPC: How typical evaluation setups overstate MPC performance for residential battery scheduling

Residential prosumers with PV-battery systems increasingly manage their electricity exchange with the power grid to minimize costs. This study investigates the performance of Model Predictive Control (MPC) and Rule-Based Control (RBC) under 15/30/60 minute averaging commonly used in research, when Net Billing and battery degradation are considered. We simulate five consecutive months for 15 buildings in northern Germany, generating costs at up to 1-minute resolution while scheduling at 15/30/60 minutes. We find that time-averaged evaluations make MPC look consistently better than RBC, yet when costs are recomputed at minute-level ground-truth, the reported advantage shrinks by 69\% on average for hourly schedulers. For individual buildings, the finer evaluation can reverse conclusions, and simple RBC can achieve lower total costs than an MPC with perfect foresight. These findings caution against drawing conclusions from coarse averages and show how a fair assessment of battery scheduling approaches can be obtained.

math.OC↗

arXiv2025

A TSO-DSO Coordination Framework via Analytical Representation and Monetization of PQV-Based Distribution System Flexibility

As the role of distribution system (DS) flexibility in transmission system operator (TSO) network management becomes increasingly vital, data privacy concerns hinder seamless interoperability. The notion of the feasible operating region (FOR), defined in the PQ domain, has emerged as a promising privacy-preserving approach. However, effectively leveraging FOR in TSO operations remains challenging due to three key factors: its accurate determination in large-scale, meshed DS networks; its tractable analytical representation; and its economic valuation. In the present paper, we propose a novel AC optimal power flow (OPF)-based method to construct a three-dimensional PQV-FOR, explicitly accounting for voltage variability and diverse flexibility-providing unit (FPU) characteristics. The construction process employs a two-stage sampling strategy that combines bounding box projection and Fibonacci direction techniques to efficiently capture the FOR. We then introduce an implicit polynomial fitting approach to analytically represent the FOR. Furthermore, we derive a quadratic cost function over the PQV domain to monetize the FOR. Thus, the proposed framework enables single-round TSO-DSO coordination: the DSO provides an analytical FOR and cost model; the TSO determines operating point at the point of common coupling (PCC) within the FOR-based AC-OPF; and the DSO computes FPU dispatch by solving its local OPF, without computationally intensive disaggregation or iterative coordination. Case studies on meshed DS with up to 533 buses, integrated into TS, demonstrates the method's efficiency compared to standard AC-OPF. On average, the proposed approach yields negligible cost deviations of at most 0.058% across test cases, while reducing computation times by up to 58.11%.

eess.SY↗

arXiv2025

A Global Analysis of Cyber Threats to the Energy Sector: "Currents of Conflict" from a Geopolitical Perspective

The escalating frequency and sophistication of cyber threats increased the need for their comprehensive understanding. This paper explores the intersection of geopolitical dynamics, cyber threat intelligence analysis, and advanced detection technologies, with a focus on the energy domain. We leverage generative artificial intelligence to extract and structure information from raw cyber threat descriptions, enabling enhanced analysis. By conducting a geopolitical comparison of threat actor origins and target regions across multiple databases, we provide insights into trends within the general threat landscape. Additionally, we evaluate the effectiveness of cybersecurity tools -- with particular emphasis on learning-based techniques -- in detecting indicators of compromise for energy-targeted attacks. This analysis yields new insights, providing actionable information to researchers, policy makers, and cybersecurity professionals.

cs.CR↗

arXiv2025

Error Accumulation using Linearized Models for Aggregating Flexibility in Distribution Systems

This paper investigates flexibility aggregation approaches based on linear models. We begin by examining the theoretical foundations of linear AC power flow, two variants of so-called DC power flow, and the LinDistFlow model, along with their underlying assumptions. The discussion covers key system details, including network topology, voltage constraints, and line losses. Simulations are conducted on the KIT Campus Nord network with real demand and solar data. Results show that, in the absence of negative losses, line losses are generally underestimated by linear models. Furthermore, line losses errors tend to accumulate both at the point of common coupling (PCC) and over extended time horizons.

eess.SY↗