Energy and AI最新文献

{"title":"Enhancing PV feed-in power forecasting through federated learning with differential privacy using LSTM and GRU","authors":"Pascal Riedel ,&nbsp;Kaouther Belkilani ,&nbsp;Manfred Reichert ,&nbsp;Gerd Heilscher ,&nbsp;Reinhold von Schwerin","doi":"10.1016/j.egyai.2024.100452","DOIUrl":"10.1016/j.egyai.2024.100452","url":null,"abstract":"<div><div>Given the inherent fluctuation of photovoltaic (PV) generation, accurately forecasting solar power output and grid feed-in is crucial for optimizing grid operations. Data-driven methods facilitate efficient supply and demand management in smart grids, but predicting solar power remains challenging due to weather dependence and data privacy restrictions. Traditional deep learning (DL) approaches require access to centralized training data, leading to security and privacy risks. To navigate these challenges, this study utilizes federated learning (FL) to forecast feed-in power for the low-voltage grid. We propose a bottom-up, privacy-preserving prediction method using differential privacy (DP) to enhance data privacy for energy analytics on the customer side. This study aims at proving the viability of an enhanced FL approach by employing three years of meter data from three residential PV systems installed in a southern city of Germany, incorporating irradiance weather data for accurate PV power generation predictions. For the experiments, the DL models long short-term memory (LSTM) and gated recurrent unit (GRU) are federated and integrated with DP. Consequently, federated LSTM and GRU models are compared with centralized and local baseline models using rolling 5-fold cross-validation to evaluate their respective performances. By leveraging advanced FL algorithms such as FedYogi and FedAdam, we propose a method that not only predicts sequential energy data with high accuracy, achieving an <span><math><msup><mrow><mi>R</mi></mrow><mrow><mn>2</mn></mrow></msup></math></span> of 97.68%, but also adheres to stringent privacy standards, offering a scalable solution for the challenges of smart grids analytics, thus clearly showing that the proposed approach is promising and worth being pursued further.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"18 ","pages":"Article 100452"},"PeriodicalIF":9.6,"publicationDate":"2024-11-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142723726","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Real-world validation of safe reinforcement learning, model predictive control and decision tree-based home energy management systems","authors":"Julian Ruddick ,&nbsp;Glenn Ceusters ,&nbsp;Gilles Van Kriekinge ,&nbsp;Evgenii Genov ,&nbsp;Cedric De Cauwer ,&nbsp;Thierry Coosemans ,&nbsp;Maarten Messagie","doi":"10.1016/j.egyai.2024.100448","DOIUrl":"10.1016/j.egyai.2024.100448","url":null,"abstract":"<div><div>Recent advancements in machine learning based energy management approaches, specifically reinforcement learning with a safety layer (<span>OptLayerPolicy</span>) and a metaheuristic algorithm generating a decision tree control policy (<span>TreeC</span>), have shown promise. However, their effectiveness has only been demonstrated in computer simulations. This paper presents the real-world validation of these methods, comparing them against model predictive control and simple rule-based control benchmarks. The experiments were conducted on the electrical installation of four reproductions of residential houses, each with its own battery, photovoltaic, and dynamic load system emulating a non-controllable electrical load and a controllable electric vehicle charger. The results show that the simple rules, <span>TreeC</span>, and model predictive control-based methods achieved similar costs, with a difference of only 0.6%. The reinforcement learning based method, still in its training phase, obtained a cost 25.5% higher to the other methods. Additional simulations show that the costs can be further reduced by using a more representative training dataset for <span>TreeC</span> and addressing errors in the model predictive control implementation caused by its reliance on accurate data from various sources. The <span>OptLayerPolicy</span> safety layer allows safe online training of a reinforcement learning agent in the real world, given an accurate constraint function formulation. The proposed safety layer method remains error-prone; nonetheless, it has been found beneficial for all investigated methods. The <span>TreeC</span> method, which does require building a realistic simulation for training, exhibits the safest operational performance, exceeding the grid limit by only 27.1 Wh compared to 593.9 Wh for reinforcement learning.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"18 ","pages":"Article 100448"},"PeriodicalIF":9.6,"publicationDate":"2024-11-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142723725","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Decentralized coordination of distributed energy resources through local energy markets and deep reinforcement learning","authors":"Daniel C. May ,&nbsp;Matthew Taylor ,&nbsp;Petr Musilek","doi":"10.1016/j.egyai.2024.100446","DOIUrl":"10.1016/j.egyai.2024.100446","url":null,"abstract":"<div><div>As the energy landscape evolves towards sustainability, the accelerating integration of distributed energy resources poses challenges to the operability and reliability of the electricity grid. One significant aspect of this issue is the notable increase in net load variability at the grid edge.</div><div>Transactive energy, implemented through local energy markets, has recently garnered attention as a promising solution to address the grid challenges in the form of decentralized, indirect demand response on a community level. Model-free control approaches, such as deep reinforcement learning (DRL), show promise for the decentralized automation of participation within this context. Existing studies at the intersection of transactive energy and model-free control primarily focus on socioeconomic and self-consumption metrics, overlooking the crucial goal of reducing community-level net load variability.</div><div>This study addresses this gap by training a set of deep reinforcement learning agents to automate end-user participation in an economy-driven, autonomous local energy market (ALEX). In this setting, agents do not share information and only prioritize individual bill optimization. The study unveils a clear correlation between bill reduction and reduced net load variability. The impact on net load variability is assessed over various time horizons using metrics such as ramping rate, daily and monthly load factor, as well as daily average and total peak export and import on an open-source dataset.</div><div>To examine the performance of the proposed DRL method, its agents are benchmarked against a near-optimal dynamic programming method, using a no-control scenario as the baseline. The dynamic programming benchmark reduces average daily import, export, and peak demand by 22.05%, 83.92%, and 24.09%, respectively. The RL agents demonstrate comparable or superior performance, with improvements of 21.93%, 84.46%, and 27.02% on these metrics. This demonstrates that DRL can be effectively employed for such tasks, as they are inherently scalable with near-optimal performance in decentralized grid management.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"18 ","pages":"Article 100446"},"PeriodicalIF":9.6,"publicationDate":"2024-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142701131","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Photovoltaic power forecasting: A Transformer based framework","authors":"Gabriele Piantadosi ,&nbsp;Sofia Dutto ,&nbsp;Antonio Galli ,&nbsp;Saverio De Vito ,&nbsp;Carlo Sansone ,&nbsp;Girolamo Di Francia","doi":"10.1016/j.egyai.2024.100444","DOIUrl":"10.1016/j.egyai.2024.100444","url":null,"abstract":"<div><div>The accurate prediction of photovoltaic (PV) energy production is a crucial task to optimise the integration of solar energy into the power grid and maximise the benefit of renewable source trading in the energy market. This paper systematically and quantitatively analyses the literature by comparing different machine learning techniques and the impact of different meteorological forecast providers. The methodology consists of an irradiance model coupled with a meteorological provider; this combination removes the constraint of a local irradiance measurement. The result is a Transformer Neural Network architecture, trained and tested using OpenMeteo data, whose performance is superior to other combinations, providing a MAE of 1.22 kW (0.95%), and a MAPE of 2.21%. The implications of our study suggest that adopting a comprehensive approach, integrating local weather data, modelled irradiance, and PV plant configuration data, can significantly improve the accuracy of PV power forecasting, thus contributing to more effective technological and economic integration.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"18 ","pages":"Article 100444"},"PeriodicalIF":9.6,"publicationDate":"2024-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142723724","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A self-growth convolution network for thermal and mechanical fault detection with very limited engine data","authors":"Gou Xin ,&nbsp;Zhu Xiaolong ,&nbsp;Wang Xinwei ,&nbsp;Wang Hui ,&nbsp;Zhang Junhong ,&nbsp;Lin Jiewei","doi":"10.1016/j.egyai.2024.100449","DOIUrl":"10.1016/j.egyai.2024.100449","url":null,"abstract":"<div><div>Severe faults occur infrequently but are critical for the prognostics and health management (PHM) of power machinery. Due to the scarcity of fault data, diagnostic models are always facing a very limited data problem. Basic convolutional neural networks require a large number of samples to train, and widely used data augmentation methods are influenced by data quality, which can exacerbate overfitting. To address this issue, a self-growth convolution network (SGNet) is proposed to make the deep learning process a self-growing scheme in both depth and width dimensions. The direct similarity measurement is utilized to supervise the depth-growth in the layer-by-layer training process. The feature redundancy metric is employed to control the width expansion. The self-growth scheme is proposed to disrupt the coadaptation between layers and that between kernels in order to mitigate the overfitting issue of small-sample cases. The SGNet is verified and implemented in the PHM of a heavy-duty diesel engine. It exhibits remarkable diagnostic capabilities in extremely sample-limited scenarios. With only three training samples per faulty type, the recognition rates of SGNet for the misfire fault and the gear tooth fracture fault are 88.44% and 98.11%, respectively. Further, the feature contrast, the information transmission, the noise resistance, and the frequency domain activation heat of SGNet are discussed by the ablation experiment in detail. The results indicate a novel path to solve the data-limitation problem in the PHM of important power machinery.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"18 ","pages":"Article 100449"},"PeriodicalIF":9.6,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142701189","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Federated learning and non-federated learning based power forecasting of photovoltaic/wind power energy systems: A systematic review","authors":"Ferial ElRobrini ,&nbsp;Syed Muhammad Salman Bukhari ,&nbsp;Muhammad Hamza Zafar ,&nbsp;Nedaa Al-Tawalbeh ,&nbsp;Naureen Akhtar ,&nbsp;Filippo Sanfilippo","doi":"10.1016/j.egyai.2024.100438","DOIUrl":"10.1016/j.egyai.2024.100438","url":null,"abstract":"<div><div>Renewable energy sources, particularly photovoltaic and wind power, are essential in meeting global energy demands while minimising environmental impact. Accurate photovoltaic (PV) and wind power (WP) forecasting is crucial for effective grid management and sustainable energy integration. However, traditional forecasting methods encounter challenges such as data privacy, centralised processing, and data sharing, particularly with dispersed data sources. This review paper thoroughly examines the necessity of forecasting models, methodologies, and data integrity, with a keen eye on the evolving landscape of Federated Learning (FL) in PV and WP forecasting. Commencing with an introduction highlighting the significance of forecasting models in optimising renewable energy resource utilisation, the paper delves into various forecasting techniques and emphasises the critical need for data integrity and security. A comprehensive overview of non-Federated Learning-based PV and WP forecasting is presented based on high-quality journals, followed by in-depth discussions on specific non-Federated Learning approaches for each power source. The paper subsequently introduces FL and its variants, including Horizontal, Vertical, Transfer, Cross-Device, and Cross-Silo FL, highlighting the crucial role of encryption mechanisms and addressing associated challenges. Furthermore, drawing on extensive investigations of numerous pertinent articles, the paper outlines the innovative horizon of FL-based PV and wind power forecasting, offering insights into FL-based methodologies and concluding with observations drawn from this frontier.</div><div>This review synthesises critical knowledge about PV and WP forecasting, leveraging the emerging paradigm of FL. Ultimately, this work contributes to the advancement of renewable energy integration and the optimisation of power grid management sustainably and securely.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"18 ","pages":"Article 100438"},"PeriodicalIF":9.6,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142701129","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-objective decoupling control of thermal management system for PEM fuel cell","authors":"Jun-Hong Chen ,&nbsp;Pu He ,&nbsp;Ze-Hong He ,&nbsp;Jia-Le Song ,&nbsp;Xian-Hao Liu ,&nbsp;Yun-Tian Xiao ,&nbsp;Ming-Yang Wang ,&nbsp;Lu-Zheng Yang ,&nbsp;Yu-Tong Mu ,&nbsp;Wen-Quan Tao","doi":"10.1016/j.egyai.2024.100447","DOIUrl":"10.1016/j.egyai.2024.100447","url":null,"abstract":"<div><div>Operating temperature is an important factor that affects the efficiency, durability, and safety of proton exchange membrane fuel cells (PEMFC). Thus, a thermal management system is necessary for controlling the appropriate temperature. In this paper, a novel thermal management system based on two-stage utilization of cooling air is first established, whose core characteristic is utilizing the temperature difference between the cooling air leaving the main radiator and the auxiliary radiator. The novel thermal management system can reduce the parasitic power of the fan by 59.27 % and improve the temperature control effect to a certain extent. The traditional feedforward decoupling control based on system identification is first adopted to control the temperature and surpasses dual-PID on all the 5 indexes, which are Integral Absolute Error Criterion (IAE) of temperature difference, IAE of inlet coolant temperature, parasitic power of fan, average overshoot of temperature difference and average overshoot of inlet coolant temperature. The multi-objective decoupling control based on multi-objective optimization is then proposed to further improve the temperature control effect on the basis of traditional feedforward decoupling control. The above 5 indexes are chosen as the optimization objectives, the decoupling coefficients are chosen as the decision variables, and the Pareto set is obtained by NSGAⅡ and NSGAⅢ. The results show that the proposed multi-objective decoupling control has the main advantages as follows: (1) It can provide comprehensive optimization options for different design preferences; (2) It can significantly optimize a certain objective while other objectives are not too extreme; (3) It has the ability to surpass traditional feedforward decoupling control on all the 5 indexes; (4) It does not rely on the system identification.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"18 ","pages":"Article 100447"},"PeriodicalIF":9.6,"publicationDate":"2024-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142701130","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Imitation learning with artificial neural networks for demand response with a heuristic control approach for heat pumps","authors":"Thomas Dengiz,&nbsp;Max Kleinebrahm","doi":"10.1016/j.egyai.2024.100441","DOIUrl":"10.1016/j.egyai.2024.100441","url":null,"abstract":"<div><div>The flexibility of electrical heating devices can help address the issues arising from the growing presence of unpredictable renewable energy sources in the energy system. In particular, heat pumps offer an effective solution by employing smart control methods that adjust the heat pump’s power output in reaction to demand response signals. This paper combines imitation learning based on an artificial neural network with an intelligent control approach for heat pumps. We train the model using the output data of an optimization problem to determine the optimal operation schedule of a heat pump. The objective is to minimize the electricity cost with a time-variable electricity tariff while keeping the building temperature within acceptable boundaries. We evaluate our developed novel method, PSC-ANN, on various multi-family buildings with differing insulation levels that utilize an underfloor heating system as thermal storage. The results show that PSC-ANN outperforms a positively evaluated intelligent control approach from the literature and a conventional control approach. Further, our experiments reveal that a trained imitation learning model for a specific building is also applicable to other similar buildings without the need to train it again with new data. Our developed approach also reduces the execution time compared to optimally solving the corresponding optimization problem. PSC-ANN can be integrated into multiple buildings, enabling them to better utilize renewable energy sources by adjusting their electricity consumption in response to volatile external signals.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"18 ","pages":"Article 100441"},"PeriodicalIF":9.6,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142701188","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predicting the thermal conductivity of polymer composites with one-dimensional oriented fillers using the combination of deep learning and ensemble learning","authors":"Yinzhou Liu ,&nbsp;Weidong Zheng ,&nbsp;Haoqiang Ai ,&nbsp;Lin Cheng ,&nbsp;Ruiqiang Guo ,&nbsp;Xiaohan Song","doi":"10.1016/j.egyai.2024.100445","DOIUrl":"10.1016/j.egyai.2024.100445","url":null,"abstract":"<div><div>Polymer composites with one-dimensional (1D) oriented fillers, recognized for their high thermal conductivity (TC), are extensively utilized in cooling electronic components. However, the prediction of the TC of composites with 1D oriented fillers poses a challenge due to the significant impact of filler orientation on composite TC. In this paper, we use a strategy that combines deep learning and ensemble learning to efficiently and quickly predict the TC of composites with 1D oriented fillers. First, as a control, we used convolutional neural network (CNN) model to predict the TC of 1D carbon fiber-epoxy composite, and the R-squared (R²) on the test set reached 0.924. However, for composites consist of different matrices and fillers, the CNN model needs to be retrained, which greatly wastes computing resources. Therefore, we define a descriptor ‘Orientation degree (<em>O_d</em>)’ to quantitatively describe the spatial distribution of the 1D fillers. CNN model was used to predict this structural parameter, the accuracy R² can reach 0.950 on the test set. Using <em>O_d</em> as a feature, random forest regression (RFR) was used to predict the TC, and the accuracy R² reached 0.954 on the test set, which was higher than that of CNN control group. We further successfully extended this strategy to composites consist of different 1D fillers and matrices, and only one CNN model and one RFR model needed to be trained to achieve fast and accurate TC prediction. This strategy provides valuable insights and guidance for machine learning-based material property prediction.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"18 ","pages":"Article 100445"},"PeriodicalIF":9.6,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142659395","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A hybrid wind power prediction model based on seasonal feature decomposition and enhanced feature extraction","authors":"Weipeng Li ,&nbsp;Yuting Chong ,&nbsp;Xin Guo ,&nbsp;Jun Liu","doi":"10.1016/j.egyai.2024.100442","DOIUrl":"10.1016/j.egyai.2024.100442","url":null,"abstract":"<div><div>Efficient and accurate wind power prediction is crucial for enhancing the reliability and safety of power system. The data-driven forecasting methods are regarded as an effective solution. However, the inherent randomness and nonlinearity of wind power systems, along with the abundance of redundant information in measurement data, present challenges to forecasting methods. The integration of precise and efficient techniques for data feature decomposition and extraction is essential in conjunction with advanced data-driven forecasting models. Focus on the seasonal variation characteristics of wind energy, a hybrid wind power prediction model based on seasonal feature decomposition and enhanced feature extraction is proposed. The effectiveness and superiority of the proposed method in predictive accuracy are demonstrated through comprehensive multi-model experiment comparisons.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"18 ","pages":"Article 100442"},"PeriodicalIF":9.6,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142659396","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}