Arnaud Zinflou

Transformer top-oil temperature prediction is essential for reliable thermal assessment and lifetime management. Although the IEC 60076-7 loading guide model is widely adopted for this purpose, its thermal parameters are generally assumed to remain constant regardless of environmental conditions: an assumption that has received limited attention. This paper investigates the influence of ambient temperature on the IEC top-oil model parameters and proposes an adaptive ambient-dependent parameterization framework. A segmented optimization methodology is applied to field measurements from a 66 MVA, 225/26.4 kV power transformer to identify the optimal rated top-oil temperature rise and thermal exponent x over different ambient temperatures. The results reveal an approximately linear dependence of the optimized parameters on ambient temperature, suggesting that the thermal dynamics represented by the IEC model evolve with environmental conditions. Incorporating this dependency improves prediction accuracy, achieving a mean absolute error (MAE) of 2.03 °C and a root mean squared error (RMSE) of 2.56 °C on the test dataset, outperforming both the conventional IEC model with fixed parameters and a previously published adaptive approach. These findings offer a practical extension of the IEC model with applications in transformer monitoring, dynamic loading, and digital-twin-based asset management.

The global adoption of electric vehicles (EVs) has experienced rapid growth in recent years. This has resulted in higher EV charging demands, placing additional pressure and challenges on existing smart grid infrastructure. Accurate and reliable short-term electric vehicle load predictions are essential to optimize smart grid operations, prevent overload of the grid and transformers, and ensure that electricity distribution aligns with actual demand. Although machine learning methods have found wide use in electric vehicle charging load forecasting, they struggle to capture complex feature patterns and accurately quantify uncertainties. In contrast, generative AI has shown great promise in various real-world applications due to its ability to model intricate data distributions and provide robust probabilistic forecasts. In this work, we leverage denoising diffusion probabilistic models, a type of generative model, to address the challenges of probabilistic electric vehicle load forecasting. Specifically, we propose an ensemble of conditional diffusion models as a unified framework to improve forecasting accuracy. We evaluate our approach on several real-world datasets across multiple forecasting horizons. The results demonstrate that the proposed method consistently improves forecasting accuracy compared with frequently used forecasting methods.

Building energy management is essential for achieving carbon reduction goals, improving occupant comfort and reducing energy costs. Coordinated building energy management faces critical challenges in exploiting spatial–temporal dependencies while ensuring operational safety across multibuilding systems. Current multibuilding energy systems face three key challenges: insufficient spatial–temporal information exploitation, lack of rigorous safety guarantees, and system complexity. This article proposes spatial–temporal-enhanced safe multiagent coordination (STEMS), a novel safety-constrained multiagent reinforcement learning (RL) framework for coordinated building energy management. STEMS integrates two core components: 1) a spatial–temporal graph representation learning framework using graph convolutional network (GCN)–Transformer fusion architecture to capture inter-building relationships and temporal patterns, and 2) a safety-constrained multiagent RL algorithm incorporating control barrier functions (CBFs) to provide mathematical safety guarantees. Extensive experiments on real-world building datasets demonstrate STEMS’s strong performance over existing methods, showing that STEMS achieves 21% cost reduction and 18% emission reduction over traditional methods, 65% safety violation reduction compared with learning-based baselines, and maintains the lowest discomfort rate. The framework also demonstrates strong robustness during extreme weather conditions and maintains effectiveness across different building types.

Time series data permeates our daily existence and has been recognized as of significant importance for many sectors, such as energy, transportation, telecommunication, and health care. Ensemble learning stands as a prevalent technique in time series forecasting, adept at handling dynamic data distributions and enhancing robustness. Nevertheless, the determination of model combination weights for computing ensemble outputs remains an unresolved issue. In this endeavour, we introduce a comprehensive reinforcement learning-based approach to tackle this challenge. Moreover, our method remains agnostic to the base models within the ensemble and exhibits versatility across a spectrum of time series forecasting tasks, thereby holding promise for application across diverse real-world scenarios. In this work, we extensively test the performance of our method on real-world datasets. Experimental results show that our proposed method achieves an average 6.2% forecasting accuracy improvement over all other single baseline methods.

Multivariate Time Series (MTS) forecasting has a wide range of applications in both industry and academia. Recent advances in Spatial-Temporal Graph Neural Network (STGNN) have achieved great progress in modelling spatialtemporal correlations. Limited by computational complexity, most STGNNs for MTS forecasting focus primarily on short-term and local spatial-temporal dependencies. Although some recent methods attempt to incorporate univariate history into modeling, they still overlook crucial long-term spatial-temporal similarities and correlations across MTS, which are essential for accurate forecasting. To fill this gap, we propose a framework called the Long-term Multivariate History Representation (LMHR) Enhanced STGNN for MTS forecasting. Specifically, a Long-term History Encoder (LHEncoder) is adopted to effectively encode the long-term history into segment-level contextual representations and reduce point-level noise. A non-parametric Hierarchical Representation Retriever (HRetriever) is designed to include the spatial information in the long-term spatial-temporal dependency modelling and pick out the most valuable representations with no additional training. A Transformer-based Aggregator (TAggregator) selectively fuses the sparsely retrieved contextual representations based on the ranking positional embedding efficiently. Experimental results demonstrate that LMHR outperforms typical STGNNs by 10.72% on the average prediction horizons and state-of-the-art methods by 4.12% on several real-world datasets. Additionally, it consistently improves prediction accuracy by 9.8% on the top 10% of rapidly changing patterns across the datasets.

The power transformer is a critical component in any transmission and distribution grid. This vital machine faces new thermal stresses arising from challenges related to energy transition along with the ever-increasing load. Understanding and predicting transformer thermal behavior is fundamental to optimizing operation and maintenance, and consequently ensuring the system’s reliability. Transformer thermal modeling (TTM) has garnered significant attention among engineers and researchers. Various approaches to TTM exist, including physical, semi-physical, physical-based numerical, and artificial intelligence (AI)-based models, with the latter being relatively unexplored in the literature. This contribution presents a comprehensive review of AI-based applications for transformer thermal modeling, examining commonly used techniques, inputs, and outputs. Perspectives in the field are discussed, with a focus on gray-box and adaptive models. The impacts of AI-based models in developing digital transformer twins are also explored. Prominent models in TTM include artificial neural networks and fuzzy systems, with support vector regression also featuring among the techniques utilized. Load and ambient temperature are primary inputs in top-oil temperature predictions, while top-oil temperature is crucial for hot-spot temperature predictions. Incorporating historical data is increasingly common in both cases. This review serves as a guide for researchers interested in TTM and highlights perspectives for future developments. AI-based applications offer powerful tools for TTM and, despite present challenges, hold significant potential for transformation in the field.

Hacking and false data injection from adversaries can threaten power grids' everyday operations and cause significant economic loss. Anomaly detection in power grids aims to detect and discriminate anomalies caused by cyber attacks against the power system, which is essential for keeping power grids working correctly and efficiently. Different methods have been applied for anomaly detection, such as statistical methods and machine learning-based methods. Usually, machine learning-based methods need to model the normal data distribution. In this work, we propose a novel anomaly detection method by modeling the data distribution of normal samples via multiple encoders and decoders. Specifically, the proposed method maps input samples into a latent space and then reconstructs output samples from latent vectors. The extra encoder finally maps reconstructed samples to latent representations. During the training phase, we optimize parameters by minimizing the reconstruction loss and encoding loss. Training samples are re-weighted to focus more on missed correlations between features of normal data. Furthermore, we employ the long short-term memory model as encoders and decoders to test its effectiveness. We also investigate a meta-learning-based framework for hyper-parameter tuning of our approach. Experiment results on network intrusion and power system datasets demonstrate the effectiveness of our proposed method, where our models consistently outperform all baselines.