Monitoring the health of remote critical infrastructure poses significant challenges due to limited accessibility and harsh operational environments. Inspection drones are ubiquitous assets that enhance the reliability of critical infrastructures through improved accessibility. However, due to the harsh operation environment, it is crucial to monitor their health to ensure successful inspection operations. The battery is a key component that determines the reliability of the inspection drones and, with an appropriate health management approach, contributes to reliable and robust inspections. This paper introduces a novel hybrid probabilistic approach for predicting the end-of-discharge (EOD) voltage of lithium polymer (Li-Po) batteries in inspection drones. The proposed approach integrates Monte Carlo (MC) dropout based Convolutional Neural Networks (CNN) with electrochemistry-based battery discharge model. This integration employs an error-correction configuration that combines electrochemistry-based EOD prediction with probabilistic error correction using CNN with MC dropout. The approach is designed to infer aleatoric and epistemic uncertainty, facilitating robust battery discharge predictions through uncertainty-aware predictions. The proposed approach is empirically evaluated using a dataset comprising EOD voltage measurements under varying load conditions. The dataset, obtained from real inspection drones during offshore wind turbine inspections, underscores the practical applicability of the proposed approach. Comparative analysis with various probabilistic methods, including Quantile Linear Regression, Quantile Regression Forest, and Quantile Gradient Boosting, demonstrates a 14.8% improvement in probabilistic accuracy compared to the best-performing method. Additionally, the estimation of different uncertainties enhances the diagnosis of battery health states, contributing to more reliable inspection operations and highlighting the practical value of the work.

Transformers are crucial for reliable and efficient power system operations, particularly in supporting the integration of renewable energy . Effective monitoring of transformer health is critical to maintain grid stability and performance. Thermal insulation ageing is a key transformer failure mode, which is generally tracked by monitoring the hotspot temperature (HST). However, HST measurement is complex, costly, and often estimated from indirect measurements. Existing HST models focus on space-agnostic thermal models, providing worst-case HST estimates. This article introduces a spatio-temporal model for transformer winding temperature and ageing estimation, which leverages physics-based partial differential equations (PDEs) with data-driven Neural Networks (NN) in a Physics Informed Neural Networks (PINNs) configuration to improve prediction accuracy and acquire spatio-temporal resolution. The computational accuracy of the PINN model is improved through the implementation of the Residual-Based Attention (PINN-RBA) scheme that accelerates the PINN model convergence. The PINN-RBA model is benchmarked against self-adaptive attention schemes and classical vanilla PINN configurations. For the first time, PINN based oil temperature predictions are used to estimate spatio-temporal transformer winding temperature values, validated through PDE numerical solution and fiber optic sensor measurements. Furthermore, the spatio-temporal transformer ageing model is inferred, which supports transformer health management decision-making. Results are validated with a distribution transformer operating on a floating photovoltaic power plant.

Semiconductor devices, especially MOSFETs (Metal–oxide–semiconductor field-effect transistor), are crucial in power electronics, but their reliability is affected by ageing processes influenced by cycling and temperature. The primary ageing mechanism in discrete semiconductors and power modules is the bond wire lift-off, caused by crack growth due to thermal fatigue . The process is empirically characterized by exponential growth and an abrupt end of life, making long-term ageing forecasts challenging. This research presents a comprehensive comparative assessment of different forecasting methods for MOSFET failure forecasting applications. Classical tracking, statistical forecasting and Neural Network (NN) based forecasting models are implemented along with novel Temporal Fusion Transformers (TFTs). A comprehensive comparison is performed assessing their MOSFET ageing forecasting ability for different forecasting horizons. For short-term predictions, all algorithms result in acceptable results, with the best results produced by classical NN forecasting models at the expense of higher computations. For long-term forecasting, only the TFT is able to produce valid outcomes owing to the ability to integrate covariates from the expected future conditions. Additionally, TFT attention points identify key ageing turning points, which indicate new failure modes or accelerated ageing phases.

Transformers are critical assets for the reliable operation of the power grid. Transformers may fail in service if monitoring models do not identify degraded conditions in time. Dissolved gas analysis (DGA) focuses on the examination of dissolved gasses in transformer oil to diagnose the state of a transformer. Fusion of black-box (BB) classifiers, also known as an ensemble of diagnostics models, have been used to improve the accuracy of diagnostics models across many fields. When independent classifiers diagnose the same fault, this method can increase the veracity of the diagnostics. However, if these methods give conflicting results, it is not always clear which model is most accurate due to their BB nature. In this context, the use of white-box (WB) models can help resolve conflicted samples effectively by incorporating uncertainty information and improve the classification accuracy. This paper presents an uncertainty-aware fusion method to combine BB and WB diagnostics methods. The effectiveness of the proposed approach is validated using two publicly available DGA datasets.

In engineering practice, multiple repair actions are considered carefully by designers, and their success or failure defines further control actions and the evolution of the system state. Such treatment is not fully supported by the current state-of-the-art in dependability analysis. We propose a novel approach for explicit modelling and analysis of repairable systems, and describe an implementation, which builds on HiP-HOPS, a method and tool for model-based synthesis of dependability evaluation models. HiP-HOPS is augmented with Pandora, a temporal logic for the qualitative analysis of Temporal Fault Trees (TFTs), and capabilities for quantitative dependability analysis via Stochastic Activity Networks (SAN). Dependability prediction is achieved via explicit modelling of local failure and repair events in a system model and then by: (i) propagation of local effects through the model and synthesis of repair-aware TFTs for the system, (ii) qualitative analysis of TFTs that respects both failure and repair logic and (iii) quantification of dependability via translation of repair-aware TFTs into SAN. The approach provides insight into the effects of multiple and alternative failure and repair scenarios, and can thus be useful in reconfigurable systems that typically employ software to utilise functional redundancies in a variety of ways.

Prognostics predictions estimate the remaining useful life (RUL) of assets. This information enables the implementation of condition-based maintenance strategies by scheduling intervention when failure is imminent. Circuit breakers (CBs) are key assets for the correct operation of the power network, fulfilling both a protection and a network reconfiguration role. Certain breakers will perform switching on a deterministic schedule, while operating stochastically in response to network faults. Both types of operation increase wear on the main contact, with high fault currents leading to more rapid aging. This paper presents a hybrid approach for prognostics of CBs, which integrates deterministic and stochastic operation through piecewise deterministic Markov processes. The main contributions of this paper are: 1) the integration of hybrid prognostics models with dynamic reliability concepts for a more accurate RUL forecasting and 2) the uncertain failure threshold modeling to integrate and propagate uncertain failure evaluation levels in the prognostics estimation process. Results show the effect of dynamic operation conditions on prognostics predictions and confirm the potential for its use within a condition-based maintenance strategy.

The use of average data for dependability assessments results in an outdated system-level dependability estimation, which can lead to incorrect design decisions. With increasing availability of online data, there is room to improve traditional dependability assessment techniques. Namely, prognostics is an emerging field, which provides asset-specific failure information that can be reused to improve the system-level failure estimation. This paper presents a framework for prognostics-updated dynamic dependability assessment. The dynamic behavior comes from runtime updated information, asset interdependencies, and time-dependent system behavior. A case study from the power generation industry is analyzed, and results confirm the validity of the approach for improved near real-time unavailability estimations.

Camera networks are an important component of modern complex systems, be it for surveillance, human/machine interaction or healthcare. Having smart cameras that can, by themselves, perform part of the data processing improves scalability both in processing and network resources. In this paper, we present the HYBRID algorithm for multiple person tracking intended for implementation on a smart camera platform, along with the development methodology to implement said algorithm in an FPGA-based smart camera. The HYBRID strategy outperforms the well-known Markov Chain Monte Carlo based particle filter (MCMC-PF) in terms of (i) parallelization capabilities as the MCMC-PF sequentially processes the particles, and (ii) tracking performances (i.e., robustness and precision).