Energy and AI最新文献

The marine climate conditions are intricate and variable. In scenarios characterized by high proportions of wind and solar energy access, the uncertainty regarding the energy sources for island microgrid is significantly exacerbated, presenting challenges to both the economic viability and reliability of the capacity configuration for island microgrids. To address this issue, this paper proposes a distributionally robust optimization (DRO) method for island microgrids, considering extreme scenarios of wind and solar conditions. Firstly, to address the challenge of determining the probability distribution functions of wind and solar in complex island climates, a conditional generative adversarial network (CGAN) is employed to generate a scenario set for wind and solar conditions. Then, by combining k-means clustering with an extreme scenario selection method, typical scenarios and extreme scenarios are selected from the generated scenario set, forming the scenario set for the DRO model of island microgrids. On this basis, a DRO model based on multiple discrete scenarios is constructed with the objective of minimizing the sum of investment costs, operation and maintenance costs, fuel purchase costs, penalty costs of wind and solar curtailment, and penalty costs of load loss. The model is subjected to equipment operation and power balance constraints, and solved using the columns and constraints generation (CCG) algorithm. Finally, through typical examples, the effectiveness of this paper’s method in balancing the economic viability and robustness of the configuration scheme for the island microgrid, as well as reducing wind and solar curtailment and load loss, is verified.

Proton-exchange membrane fuel cells (PEMFCs) are critical components of zero-emission electro-hydrogen generators. Accurate performance prediction is vital to the optimal operation management and preventive maintenance of these generators. Polarization curve remains one of the most important features representing the performance of PEMFCs in terms of efficiency and durability. However, predicting the polarization curve is not trivial as PEMFCs involve complex electrochemical reactions that feature multiple nonlinear relationships between the operating variables as inputs and the voltage as outputs. Herein, we present an artificial-intelligence-based approach for predicting the PEMFCs’ performance. In that way, we propose first an explainable solution for selecting the relevant features based on kernel principal component analysis and mutual information. Then, we develop a machine learning approach based on XGBRegressor and Bayesian optimization to explore the complex features and predict the PEMFCs’ performance. The performance and the robustness of the proposed machine learning based prediction approach is tested and validated through a real industrial dataset including 10 PEMFCs. Furthermore, several comparison studies with XGBRegressor and the two popular machine learning-based methods in predicting PEMFC performance, such as artificial neural network (ANN) and support vector machine regressor (SVR) are also conducted. The obtained results show that the proposed approach is more robust and outperforms the two conventional methods and the XGBRegressor for all the considered PEMFCs. Indeed, according to the coefficient of determination criterion, the proposed model gains an improvement of 6.35%, 6.8%, and 4.8% compared with ANN, SVR, and XGBRegressor respectively.

Data-driven models for battery state estimation require extensive experimental training data, which may not be available or suitable for specific tasks like open-circuit voltage (OCV) reconstruction and subsequent state of health (SOH) estimation. This study addresses this issue by developing a transfer-learning-based OCV reconstruction model using a temporal convolutional long short-term memory (TCN-LSTM) network trained on synthetic data from an automotive nickel cobalt aluminium oxide (NCA) cell generated through a mechanistic model approach. The data consists of voltage curves at constant temperature, C-rates between $\text{C}/30$ to $1\text{C}$, and a SOH-range from 70 % to 100 %. The model is refined via Bayesian optimization and then applied to four use cases with reduced experimental nickel manganese cobalt oxide (NMC) cell training data for higher use cases. The TL models’ performances are compared with models trained solely on experimental data, focusing on different C-rates and voltage windows. The results demonstrate that the OCV reconstruction mean absolute error (MAE) within the average battery electric vehicle (BEV) home charging window (30 % to 85 % state of charge (SOC)) is less than 22 mV for the first three use cases across all C-rates. The SOH estimated from the reconstructed OCV exhibits an mean absolute percentage error (MAPE) below 2.2 % for these cases. The study further investigates the impact of the source domain on TL by incorporating two additional synthetic datasets, a lithium iron phosphate (LFP) cell and an entirely artificial, non-existing, cell, showing that solely the shifting and scaling of gradient changes in the charging curve suffice to transfer knowledge, even between different cell chemistries. A key limitation with respect to extrapolation capability is identified and evidenced in our fourth use case, where the absence of such comprehensive data hindered the TL process.

In recent years, large language models have achieved breakthroughs on a wide range of benchmarks in natural language processing and continue to increase in performance. Recently, the advances of large language models have raised interest outside the natural language processing community and could have a large impact on daily life. In this paper, we pose the question: How will large language models and other foundation models shape the future product development process? We provide the reader with an overview of the subject by summarizing both recent advances in natural language processing and the use of information technology in the engineering design process. We argue that discourse should be regarded as the core of engineering design processes, and therefore should be represented in a digital artifact. On this basis, we describe how foundation models such as large language models could contribute to the design discourse by automating parts thereof that involve creativity and reasoning, and were previously reserved for humans. We describe how simulations, experiments, topology optimizations, and other process steps can be integrated into a machine-actionable, discourse-centric design process. As an example, we present a design discourse on the optimization of wind turbine blades. Finally, we outline the future research that will be necessary for the implementation of the conceptualized framework.

The digital transformation process of power systems towards smart grids is resulting in improved reliability, efficiency and situational awareness at the expense of increased cybersecurity vulnerabilities. Given the availability of large volumes of smart grid data, machine learning-based methods are considered an effective way to improve cybersecurity posture. Despite the unquestionable merits of machine learning approaches for cybersecurity enhancement, they represent a component of the cyberattack surface that is vulnerable, in particular, to adversarial attacks. In this paper, we examine the robustness of autoencoder-based cyberattack detection systems in smart grids to adversarial attacks. A novel iterative-based method is first proposed to craft adversarial attack samples. Then, it is demonstrated that an attacker with white-box access to the autoencoder-based cyberattack detection systems can successfully craft evasive samples using the proposed method. The results indicate that naive initial adversarial seeds cannot be employed to craft successful adversarial attacks shedding insight on the complexity of designing adversarial attacks against autoencoder-based cyberattack detection systems in smart grids.

With the advancement of artificial intelligence, the dominance of deep learning (DL) models over ordinary machine learning (ML) algorithms has become a reality in recent years due to its capability of handling complex pattern recognition without manual feature pre-definition. With the growing demands for power savings, building energy loss reduction could benefit from DL techniques. For buildings/rooms with the varying number of occupants, heating, ventilation, and air conditioning (HVAC) systems are often found in operations without much necessity. To reduce the building’s energy loss, accurate occupancy detection/prediction (ODP) results could be used to control the proper operations of HVACs. However, ODP is a challenging issue due to multiple reasons, such as improper selection/deployment of sensors, inefficient learning algorithms for pattern recognition, varying room conditions, etc. To overcome the above challenges, we propose a DL-based framework, i.e., Deep Weighted Fusion Learning (DWFL), to detect and predict occupancy counts with optimal multi-sensor fusion structure. DWFL fuses the extracted features from multiple types of sensors with the priority/weight assignment to each sensor. Such weight assignment considers different room conditions and the pros/cons of each type of sensor. To evaluate DWFL model in terms of occupancy prediction accuracy, we have set up an experimental testbed with low-cost cameras, carbon dioxide (${\mathrm{CO}}_2$), and passive infrared (PIR) sensors. Among the recently proposed occupancy detection models, DeepFusion utilized deep learning model on heterogeneous sensor data and achieved 88% accuracy in occupancy count estimation (Xue et al., 2019). Another deep learning-based model MI-PIR achieved 91% accuracy on raw analog data from PIR sensors (Andrews et al., 2020). Our research outcome is 94%. Therefore, the experiment results show that our DWFL scheme outperforms the state-of-the-art ODP methods by 3%.

The increasing drive towards eco-friendly environment motivates the generation of energy from renewable energy sources (RESs). The rising share of RESs in power generation poses potential challenges, including uncertainties in generation output, frequency fluctuations, and insufficient voltage regulation capabilities. As a solution to these challenges, energy storage systems (ESSs) play a crucial role in storing and releasing power as needed. Battery energy storage systems (BESSs) provide significant potential to maximize the energy efficiency of a distribution network and the benefits of different stakeholders. This can be achieved through optimizing placement, sizing, charge/discharge scheduling, and control, all of which contribute to enhancing the overall performance of the network. In this paper, we provide a comprehensive overview of BESS operation, optimization, and modeling in different applications, and how mathematical and artificial intelligence (AI)-based optimization techniques contribute to BESS charging and discharging scheduling. We also discuss some potential future opportunities and challenges of the BESS operation, AI in BESSs, and how emerging technologies, such as internet of things, AI, and big data impact the development of BESSs.

This paper presents a dynamic knowledge graph approach that offers a reusable, interoperable, and extensible framework for modelling power systems. Domain ontologies have been developed to support a linked data representation of infrastructure data, socio-demographic data, areal attributes like demand, and models describing power systems. The knowledge graph links the data with a hierarchical representation of administrative regions, supporting geospatial queries to retrieve information about the population within the vicinity of a power plant, the number of power plants, total generation capacity, and demand within specific areas. Computational agents were developed to operate on the knowledge graph. The agents performed tasks including data uploading, updating, retrieval, processing, model construction and scenario analysis. A derived information framework was used to track the provenance of information calculated by agents involved in each scenario. The knowledge graph was populated with data describing the UK power system. Two alternative models of the transmission grid with different levels of structural resolution were instantiated, providing the foundation for the power system simulation and optimisation tasks performed by the agents. The application of the dynamic knowledge graph was demonstrated via a case study that investigates clean energy transition trajectories based on the deployment of Small Modular Reactors in the UK.

This paper presents a knowledge graph-based approach for the dynamic control of a district heating network with integrated emission dispersion modelling. We propose an interoperable and extensible implementation to forecast the anticipated heat demand of a municipal heating network, minimise associated total generation cost based on a previously devised methodology, and couple it with dispersion simulations for induced airborne pollutants to provide automatic insights into air quality implications of various heat sourcing strategies. We create cross-domain interoperability in the nexus of energy and air quality via newly developed ontologies and semantic software agents, which can be chained together via The World Avatar dynamic knowledge graph to resemble the behaviour of complex systems. Furthermore, we integrate the City Energy Analyst into this ecosystem to provide building-level insights into energy demand and renewable generation potential to foster strategic analyses and scenario planning. Underlying calculations use building and weather data from the knowledge graph in place of inherent assumptions in the official software release, facilitating a more data-driven approach. All use cases are implemented for a mid-size town in Germany as a proof-of-concept, and a unified visualisation interface is provided, allowing for the examination of 3D buildings alongside their corresponding energy demand and supply time series, as well as emission dispersion data. With this work, we outline the potential of Semantic Web technologies to connect digital twins for holistic energy modelling in smart cities, thereby addressing the increasing complexity of interconnected energy systems.

With the rapid development of urban power grids and the large-scale integration of renewable energy, traditional power grid fault diagnosis techniques struggle to address the complexities of diagnosing faults in intricate power grid systems. Although artificial intelligence technologies offer new solutions for power grid fault diagnosis, the difficulty in acquiring labeled grid data limits the development of AI technologies in this area. In response to these challenges, this study proposes a semi-supervised learning framework with self-supervised and adaptive threshold (SAT-SSL) for fault detection and classification in power grids. Compared to other methods, our method reduces the dependence on labeling data while maintaining high recognition accuracy. First, we utilize frequency domain analysis on power grid data to filter abnormal events, then classify and label these events based on visual features, to creating a power grid dataset. Subsequently, we employ the Yule–Walker algorithm extract features from the power grid data. Then we construct a semi-supervised learning framework, incorporating self-supervised loss and dynamic threshold to enhance information extraction capabilities and adaptability across different scenarios of the model. Finally, the power grid dataset along with two benchmark datasets are used to validate the model’s functionality. The results indicate that our model achieves a low error rate across various scenarios and different amounts of labels. In power grid dataset, When retaining just 5% of the labels, the error rate is only 6.15%, which proves that this method can achieve accurate grid fault detection and classification with a limited amount of labeled data.