Keywords: Cybersecurity, Fault detection and isolation, Building supervision
Abstract: When a cyber-attacker knows the model of a plant, covert attacks consist in corrupting the control signals transmitted by the controller to cause damage on the plant while remaining undetectable from any standard model-based detection and isolation schemes. The strategy in this paper to make covert attacks visible to the defender is to introduce additional delays in the control loop unknown to the attacker. Then, a method based on the design of directional residuals generator is used to detect and isolate the attacks. An illustrative example from building automation is given to show the efficiency of the proposed method.

Keywords: Cybersecurity, Networked control system
Abstract: Cyber attacks have become a critical threat to the cyber-physical systems (CPSs), where physical components are interconnected through networks. Most existing studies on stealthy attacks focus on CPS with linear plants. This paper investigates a kind of stealthy attacks called mirror attacks on CPS with nonlinear plants. The idea is to replace the true sensor output signals with the simulated outputs computed based on the adversarial nonlinear plant model and the control input signals calculated by the controller. It is demonstrated that, with the help of joint state and parameter estimation, a mirror attack can manipulate the states of nonlinear plants while remaining stealthy, even with an inaccurate plant model. To illustrate the threat of mirror attacks, experiments have been carried out on the well-established nonlinear three-tank system.

Keywords: Cybersecurity, Networked control system, Risk analysis
Abstract: This work studies the security of multi-sensor centralized remote estimation under replay attacks launched on partial wireless channels. We consider an adversary with limited attack resources and only partial knowledge of the system dynamics, who aims to compromise sensor measurements by replaying previously observed data. To evaluate both the stealthiness of the attack and its impact on estimation performance, we adopt two key metrics: the Kullback-Leibler (K-L) and the estimation error covariance. Our analysis reveals that the system’s resilience to such attacks fundamentally depends on the stability of the state matrix. Specifically, when the system is stable, both the K-L divergence and the estimation error covariance converge to steady-state values characterized by two coupled Lyapunov equations. In contrast, if the state matrix is unstable, these metrics grow unbounded over time under replay attacks. Finally, a simulation example is provided to validate the theoretical results and demonstrate the practical implications of partial-channel replay attacks in remote estimation systems.

Keywords: Fault detection and isolation, Cybersecurity, Model-based methods
Abstract: This article proposes a novel framework for designing deceptive fault injection attacks against Cyber-Physical Systems, with a focus on misleading modern remote monitoring systems into diagnosing non-existent faults. Such attacks can induce unnecessary maintenance actions or inappropriate operational responses. We formulate the attack design as a finite-horizon optimization problem that computes optimal injection control inputs and sensor measurements for a given set of compromised communication channels. Multiple attack scenarios, ranging from full-channel compromise to resource-limited cases, are evaluated through simulation. The results show that full-channel access enables perfect deception, while carefully chosen partial compromises can still produce substantial monitoring errors. The proposed framework offers a systematic tool for assessing monitoring-system vulnerabilities and supports the development of more resilient anomaly detection and fault diagnosis strategies.

Keywords: Fault tolerant control / fault recovery, Networked control system, Model-based methods
Abstract: This paper presents an integrated approach to fault estimation (FE) and fault-tolerant control (FTC) for uncertain networked control systems subject to actuator faults, external disturbances, and lossy communication networks experiencing random packet dropouts in both output and control signal transmission channels. The proposed FTC framework consists of: 1) an augmented state observer for simultaneous state and fault estimation, and 2) a controller that leverages these estimates to achieve robust stabilization. The design is formulated as an observer-based robust control problem for a stochastic system and is solved via a single-step linear matrix inequality (LMI) approach. Simulation results demonstrate that the method achieves accurate estimation, fault compensation, and robust system stabilization over lossy networks.

Keywords: Data driven methods, Fault detection and isolation
Abstract: In this paper, a data-driven Fault Detection strategy is proposed for detecting faults in generator speed sensors of floating offshore wind turbines using Machine Learning (ML) techniques. To generate reliable datasets in both healthy and faulty conditions, a seven-turbine offshore wind farm is set up using the FOWLTY MATLAB/Simulink benchmark framework. A systematic approach is applied for signal-based feature extraction, followed by dimensionality reduction through feature selection using LightGBM. Two classifiers—Support Vector Machines (SVM) and Decision Trees (DT)—are trained and evaluated on the resulting datasets. The results show that both methods achieve acceptable accuracy in fault classification, but the SVM demonstrates better performance compared to the DT. These findings confirm that the proposed methods can accurately detect generator speed sensor faults in offshore wind turbines. The results confirm that the proposed methods can accurately detect generator speed sensor faults in offshore wind turbines.

Keywords: Fault detection and isolation, Model-based methods, Power plants / energy transport
Abstract: This paper presents a model-based fault detection approach for a floating offshore wind farm consisting of seven wind turbines, using a standardized Simulink benchmark to facilitate reproducible research. A 1-dimensional momentum theory model is employed for each wind turbine, balancing fidelity and tractability for fault detection while avoiding proprietary data by using parametric approximations for power and thrust coefficients. Given the inherent uncertainty in model parameters, the study investigates the practical limits of detectability achievable with such simplified models while maintaining robustness against significant parametric variability. To rigorously account for uncertainty, the residual covariance matrix of the so-called mathcal{H}_i/mathcal{H}_{infty} 'unified solution' fault detection filter is derived through first-order sensitivity analysis. The resulting generalized likelihood ratio test demonstrates that while faults in low-uncertainty subsystems (e.g., pitch and torque actuators) can be reliably detected, subtle faults in variables related to the aerodynamics (e.g., in generator speed or power) remain challenging due to the dominant influence of model uncertainty. The results highlight the trade-offs between model simplicity and fault detectability, offering insights into the best achievable performance under realistic uncertainty constraints.

Keywords: Health monitoring, Fault detection and isolation, Data driven methods
Abstract: This paper introduces an interpretable ensemble machine learning approach specifically designed for fault detection in floating offshore wind turbines. The methodology integrates advanced statistical features extracted from residual signals with complementary machine learning models, enhancing the identification of subtle fault-induced deviations typical in offshore environments. Validated using a realistic offshore wind farm simulation benchmark, the proposed method demonstrated clear advantages over traditional threshold-based techniques and single-model approaches. The practical interpretability of the method is demonstrated through analysis of feature relevance, aiding effective fault diagnosis. Although tested primarily on specific sensor faults, the modular nature of the methodology supports its generalisation and highlights its potential suitability for broader fault detection scenarios and real-time applications.

Keywords: Fault detection and isolation, Fault tolerant control / fault recovery, Power plants / energy transport
Abstract: This paper presents a benchmark for fault detection and isolation (FDI) in floating offshore wind farms, addressing the lack of standardized evaluation frameworks for this emerging technology. Developed using the FOWLTY simulator, the benchmark models a wind farm with seven floating turbines based on the NREL 5MW reference turbine and DeepCWind platform. It incorporates ten diverse wind scenarios (5-23 m/s) and realistic fault conditions, including sensor and actuator faults with variable severity and timing. The benchmark provides sensor measurements with injected noise, actuator reference signals, and evaluation metrics such as false alarm rate (FAR), missed detection rate (MDR), and correct isolation rate. By offering a modular Simulink environment and customizable datasets, this work enables reproducible comparisons of FDI methods while highlighting the unique dynamics of floating offshore systems. The benchmark is publicly available to support research in fault detection and fault-tolerant control for offshore wind energy.

Keywords: Power plants / energy transport, Fault detection and isolation, Data driven methods
Abstract: This paper presents a hybrid fault diagnosis approach for wind turbines that integrates structural analysis through Analytical Redundancy Relations (ARRs) with data- driven modeling using Gaussian Process Regression (GPR). The proposed method leverages the physical structure of the system to define input-output dependencies and trains GPR estimators on fault-free operational data to predict key sub- system outputs. Residuals are computed by comparing sensor measurements with GPR predictions, and faults are detected using a combination of interval-based thresholds and Cumula- tive Sum (CUSUM) control charts. The proposed approach is validated on a simulated 5-MW wind turbine benchmark model under realistic operating conditions. Various fault scenarios are injected in the pitch actuator, drivetrain, and generator subsystems. Results demonstrate the fault diagnosis accuracy, robustness, and early detection capability across diverse fault types.

Keywords: Decision making, Fault tolerant control / fault recovery, Model-based methods
Abstract: Critical Infrastructure Systems (CIS) are large-scale networks that provide essential services to individual citizens and the society at large. Their intra- and inter-dependencies can cause cascading effects in the presence of natural disasters, hardware failures and cyber threats. Therefore, ensuring CIS resilience is a crucial aspect that is becoming more challenging due to the increasingly interconnected components and interactions. This work investigates how agency that emerges from foundation machine learning models can be employed to address some of these challenges and enhance CIS resilience. We propose a framework for autonomous agents based on foundation models to be embedded at different levels of CIS operations for dealing with monitoring, control and management and safeguard CIS operations. We elaborate on three distinct instantiations of foundation model based agency in CIS and discuss opportunities and challenges that arise for their introduction to large-scale and complex systems.

Keywords: Water treatment, Cybersecurity, Communication networks
Abstract: Networked control systems that operate urban drinking water networks are susceptible to cyber-attacks while still being required to satisfy service and safety constraints. We address resilient trajectory tracking for the aggregate Barcelona network and present a dual-layer predictive control architecture. During attack-free conditions, a nominal Model Predictive Controller (MPC) regulates the plant and buffers admissible inputs. A set-membership consistency test monitors the state's evolution through families of one-step controllable sets; any violation prompts a switch to an ADMM-based Encrypted MPC solved in the cloud via homomorphic encryption, thereby preserving data confidentiality and closed-loop stability. We establish feasibility and bounded tracking errors for arbitrary switches between the two layers.

Keywords: Cybersecurity, Networked control system, Fault detection and isolation
Abstract: This paper shows experimental validation of resilient homomorphic encryption (RHE) scheme implemented on a three-tank system. The RHE scheme has been recently proposed to improve confidentiality of signals in control systems and achieve the resilience of control systems to cyber attacks on the ciphertexts. With the help of the RHE scheme, the entire computation in the controller is carried out in ciphertexts. Moreover, a two-layer attack detection system has been developed to indicate whether there is an attack and whether the attack has influenced the control system behavior. The experimental results demonstrate the feasibility of the RHE scheme in practical systems and verify that both the controller parameters and the signals communicated over the network are encrypted and the detection system is able to detect and determine the attack.

Keywords: Networked control system, Cybersecurity, Power plants / energy transport
Abstract: This work presents an event-based control approach applied to buck-type DC-DC converters, considering the presence of stochastic deceptive cyber-attacks on control and measurement signals. A co-design strategy based on Linear Matrix Inequalities (LMIs) is adopted to co-design the state-feedback controller and the event-triggering mechanism (ETM). The objective is to ensure the asymptotic stability of the closed-loop system, even under communication constraints and cyber threats. Simulation results demonstrate the effectiveness of the proposed approach in reducing transmissions, enhancing resilience against attacks, and improving network usage efficiency compared to traditional methods reported in the literature.

Keywords: Design for reliability and safety, Fault tolerant control / fault recovery, Power plants / energy transport
Abstract: Microgrids are fundamental for future smart grids as they enhance energy resilience due to their ability to disconnect from the main grid and operate in islanded mode. This work proposes an innovative microgrid energy management scheme that is resilient to potential grid faults. Towards this direction, an optimization problem is formulated for the proactive scheduling of the microgrid operation in grid-connected mode to handle potential utility grid faults. Moreover, a second optimization problem is formulated for the outage management of the microgrid operation in islanded mode to minimize the load shedding by determining the power set-points of the energy storage system and renewable energy sources. However, solving the resulting non-convex optimization problem poses significant challenges. To reduce the computational complexity, a convex relaxation energy storage model is used, eliminating the complementarity constraints; however, it yields infeasible solutions when simultaneous charging and discharging occur. To address this issue, the considered problems are reformulated to penalise the charging power using a penalty term in the objective function and an iterative algorithm based on the bisection method is proposed to find the minimum value of the penalty term that yields feasible solutions. Simulation results indicate the proposed scheme can effectively manage microgrid operations during both grid-connected and islanded modes, ensuring economic operation while minimizing load shedding during utility grid faults.

Keywords: Power plants / energy transport, Fault detection and isolation, Data driven methods
Abstract: Hydraulic pitch actuators in offshore wind turbines are prone to wear that eventually may develop into failure of blade pitch control. This would leave the turbine non-operational, causing significant loss in revenue while the defect remains. This paper proposes a technique for detecting early signs of increased friction within the hydraulic actuator, as an indicator of wear. The method utilizes a temporal convolutional network to reconstruct essential signals, which are fed into a modified least squares algorithm to estimate the Coulomb friction coefficient and continuously track friction magnitude. The robustness and accuracy of the approach is validated by high-fidelity simulations.

Keywords: Fault-forecasting methods, Data driven methods, Decision making
Abstract: This study introduces a Transformer encoder-based normality model for the early detection of gearbox failures in wind turbines (WTs), utilizing Supervisory Control and Data Acquisition (SCADA) operational data for predictive analysis. The proposed methodology follows a pre-training and fine-tuning approach, enhancing computational efficiency, reducing training time, and ensuring scalability across large wind farms. Gearbox failures are identified through a fault prognosis indicator that quantifies deviations from normal behavior, enabling precise detection of early failure patterns. The results show that the proposed methodology can predict gearbox failures several months in advance, providing valuable lead time for preventive maintenance. By combining accurate early detection with low computational cost, this approach strengthens predictive maintenance strategies and improves operational reliability.

Keywords: Power plants / energy transport
Abstract: This paper presents novel findings on the application of Individual Pitch Control (IPC) for multi-megawatt, two-bladed downwind turbines, as demonstrated in a Goldwind turbine design. The results show that IPC effectively reduces fatigue loads by 10-20% and extreme loads by 5-10%. This is achieved through optimization strategies that utilize rotor lift and activation techniques to balance load reduction with the lifetimes of the main and blade bearings. Furthermore, the study explores a passive yaw configuration, where yaw control is achieved via IPC (yaw-by-IPC), demonstrating a significant 45% reduction in tower top/yaw torsional moment loads while maintaining load increases on other components below 5%. This approach supports a more cost-efficient tower design, enabling the use of a lattice tower structure.

Keywords: Data driven methods, Power plants / energy transport, Decision making
Abstract: This paper presents an enhanced Deep Deterministic Policy Gradient (DDPG) controller for collective pitch control of the DTU 10-MW OO-Star Floating Wind Turbine. The DDPG agent is pre-trained via imitation learning from a Proportional Integral (PI) controller to improve training efficiency and stability. Training is conducted using a reduced-order linear state-space model of the DTU 10 MW turbine on the OO-Star semi-submersible platform, and the controller is evaluated in FAST (Fatigue, Aerodynamics, Structures, and Turbulence) simulations under Design Load Cases (DLCs) 12 and 13, both are on region 3. Compared to the baseline ROSCO controller, the DDPG agent achieves improved rotor speed regulation and platform pitch damping, with a trade-off of increased blade pitch actuation. The results demonstrate the viability of imitation-enhanced reinforcement learning for floating wind turbine control under realistic conditions.

Keywords: Fault detection and isolation, Data driven methods
Abstract: Model-based fault diagnosis is the most powerful supervision approach when a model of the dynamic system under study is available or identifiable. When the underly- ing system presents complex nonlinearities or time varying components so that it is not feasible to obtain an accurate mathematical description of the system, knowledge-based fault diagnosis approaches are a viable alternative. However, these techniques require classification algorithms to understand if residuals are associated with a fault condition or not. This work presents a novel classification approach based on the concept of quadratic entropy. The proposed methodology is validated on the LiU-ICE benchmark, which requires to detect and isolate faults on a spark ignite engine used in the automotive industry. Results are compared with the current best solution of the benchmark which employs open-set classification techniques.

Keywords: Fault detection and isolation, Fault tolerant control / fault recovery, Data driven methods
Abstract: This paper proposes a novel framework for active fault diagnosis and parameter estimation in linear systems operating in closed-loop, subject to unknown but bounded faults. The approach integrates set-membership identification with a cost function designed to accelerate fault identification. Informative excitation is achieved by minimizing the size of the parameter uncertainty set, which is approximated using ellipsoidal outer bounds. Combining this formulation with a scheduling parameter enables a transition back to nominal control as confidence in the model estimates increases. Unlike many existing methods, the proposed approach does not rely on predefined fault models. Instead, it only requires known bounds on parameter deviations and additive disturbances. Robust constraint satisfaction is guaranteed through a tube-based model predictive control scheme. Simulation results demonstrate that the method achieves faster fault detection and identification compared to passive strategies and adaptive ones based on persistent excitation constraints.

Keywords: Fault detection and isolation, Fault tolerant control / fault recovery, Health monitoring
Abstract: This paper proposes a non-invasive method for the detection of parameter variation in DC-DC buck power converters. It is computationally efficient and only requires the usual current and voltage sensors. The method is presented in several steps. First, a generic buck converter model is presented. Then, the general design of a sliding mode observer is described as well as its method of fault reconstruction. This sliding mode observer is then adapted to detect parameter changes of the buck converter model. Finally, simulations are performed to validate the observer design.

Keywords: Model-based methods, Fault detection and isolation, Transportation systems
Abstract: Detecting anomalies in maritime vessels routes can anticipate potential collisions and prevent illicit activities. Common methods in the literature employ Automatic Identification System (AIS) data to build models of normal vessel behavior. The predictions of the models are then compared with actual data to detect discrepancies in vessel motion patterns. Data-driven or geometrical approaches can be used to build such models, requiring however a large set of historical data and lacking in interpretation. In this paper, we propose a Model Predictive Control (MPC)-based anomaly detection framework that aims at overcoming the aforementioned issues. Results on publicly available real AIS data with artificially simulated anomalies show the benefits of the proposed physics-based MPC approach when compared with a geometrical-based approach.

Keywords: Fault detection and isolation, Health monitoring, Data driven methods
Abstract: Dual-fuel engines are well known for their low environmental impacts compared to other engines. However, continuous monitoring during propulsion is crucial to ensure safe navigation and reliability. Faults in the exhaust system can lead to severe consequences, including increased toxic emissions, power loss, and high vibrations. To identify abnormalities in exhaust valve closing dead time (EVCDT)—the duration between when the closing command is sent and the actual movement of the valve, this study applies a correlation-based clustering method using K-means on EVCDT data from 12 cylinders. The developed algorithm continuously monitors the EVCDT data, distinguishing between healthy and potentially faulty cylinders. The findings underscore the effectiveness of correlation clustering-based diagnostics for predictive maintenance, supporting sustainable marine operations and relations among cylinders EVCDTS.

Keywords: Data driven methods, Power plants / energy transport
Abstract: The abrupt growth of the world population has raised global concerns about sustainable energy supply. In response, energy policies are continuously evolving to guarantee stability across the energy sector. Stakeholders in the energy supply chain are striving to maintain reliability and balance in the system. Renewable energy sources, because of their low generation cost and sustainability, are increasingly replacing fossil fuel-based power plants. However, their integration introduces significant uncertainty and variability. In this context, accurate forecasting becomes essential for effective planning in renewable-powered microgrids. Reliable forecasts enhance the availability, stability, and overall reliability of microgrids, thereby supporting a resilient and efficient energy supply chain. To address this need, a novel bi-long short-term memory attention algorithm is proposed to enhance the forecasting accuracy of key microgrid parameters that drive power-system decision making. The proposed model demonstrates superior performance over an existing electrical net-load forecasting approach that combines deep neural networks with wavelet transform on the same real-world dataset. Specifically, it achieves a 4% reduction in Mean Absolute Percentage Error and a 2% decrease in Root Mean Square Error. Furthermore, the proposed model accurately tracks the peak and trough points in photovoltaic and wind energy, which is a key factor in forecasting.

Keywords: Fault detection and isolation, Model-based methods, Decision making
Abstract: The complexity of battery packs and limited sensor coverage make accurate fault detection and isolation (FDI) challenging, especially when faults show similar symptoms. This study presents a model-based diagnosis approach for series-connected lithium-ion (li-ion) batteries, combining structural analysis with observer-based residuals. Hardware redundancy via additional voltage sensors improves isolability of sensor, short-circuit, and connection faults. Residuals are evaluated using adaptive thresholds to enhance robustness to modeling errors and noise. The proposed method enables multifault detection within a single diagnostic framework, isolating faults robustly and without added complexity.

Keywords: Model-based methods, Fault tolerant control / fault recovery, Supervisory control
Abstract: This paper studies an estimator-based switching supervisory control problem for a class of nonlinear systems using output feedback. Since the standard input-to-state stability (ISS) assumption for estimation error systems is generally difficult to verify for output-feedback design, the notion of quasi-disturbance-to-error stable (qDES) observer is applied to the supervisory control framework, which guarantees the estimation error systems ISS under state constraints. The proposed supervisory control algorithm ensures that the system output asymptotically converges to the origin in a semi-global sense with the other states remaining bounded. The fault tolerance of the proposed supervisory control has also been validated based on a redesigned qDES multi-observer using adaptive approximator. Finally, a systematic solution to the supervisory control design for a Chua’s circuit system is derived with a numerical simulation provided.

Keywords: Fault detection and isolation, Model-based methods
Abstract: This paper proposes a fault detection and isolation (FDI) scheme for the InfinityWEC prototype developed by Ocean Harvesting Technologies AB. This device is a wave energy converter (WEC), mainly composed of a float and a power take-off (PTO) system, that includes ball screw mechanisms (BSMs) and electrical generators. The proposed FDI scheme is given by a set of structured residuals, which is verified via numerical simulation in MATLAB/Simulink.

Keywords: Metal processing, Power plants / energy transport, Transportation systems
Abstract: Envelope modeling is an efficient way for obtaining large-signal amplitude and phase dynamics of fast-varying sinusoidal signals, required for e.g. phase or current/voltage magnitude regulation in resonant power converters. A significant drawback of an envelope model is its inherited system model order augmentation, resulting in high complexity of the derived plant. The latter becomes a challenging factor for controller design. This paper focuses on envelope model order reduction of capacitor powered resonant inverter feeding a time-varying series RLC load under the slow varying amplitude derivative and phase assumption, yielding a simplified low-order system plant suitable for controller design. Simulation results are presented to validate the suggested methodology.