Energy and AI最新文献_第10页

Neural-accelerated numerical model for packed bed latent heat storage system 填料床潜热蓄热系统的神经加速数值模型

IF 9.6 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Energy and AI

Pub Date : 2025-08-27 DOI: 10.1016/j.egyai.2025.100602

Dessie Tadele Embiale , Shri Balaji Padmanabhan , Mohamed Tahar Mabrouk , Stéphane Grieu , Bruno Lacarrière

Developing accurate and computationally efficient dynamic models for packed-bed latent-heat storages (PBLHS) is crucial for reliably predicting their performance across different operating scenarios and enabling their use in planning and real-time control. In this study, a novel neural-accelerated numerical model for PBLHS is proposed by coupling a neural network (NN) into a coarsely discretized equations of the Continuous-solid Phase (CP) model. The embedded NN predicts the surface temperature of the phase change material (PCM) given the fluid temperature and enthalpy of the PCM as inputs, which the CP model fails to capture. This allows the neural-accelerated model to replicate the accuracy of a high-fidelity and computationally expensive model namely Concentric Dispersion (CD) model. An innovative data generation process to generate training data for NN involving both CD and CP model is proposed. Two versions of neural-accelerated model are proposed, one with conventional NN and another using NN with a custom activation function. Both versions demonstrate an excellent accuracy, achieving MSE as low as 0.117 °C,

R^{2}

values closer to 0.995 and error percentage below 0.394

%

compared to the highly accurate CD model. As for computational efficiency, the proposed models achieved 342 times and 764 times acceleration respectively. The gain in more acceleration for the later version of the proposed model is achieved through the use of a compact architecture that benefits from the custom activation function, while also enhancing model explainability. These results highlight the model’s suitability for scenarios demanding both high accuracy and computational efficiency.

为填料床潜热储热系统（PBLHS）建立准确且计算高效的动态模型对于可靠地预测其在不同操作场景下的性能，并使其能够在规划和实时控制中使用至关重要。本文通过将神经网络（NN）与粗离散的连续固相（CP）模型方程相结合，提出了一种新的PBLHS神经加速数值模型。嵌入式神经网络以相变材料（PCM）的流体温度和焓为输入，预测相变材料（PCM）的表面温度，这是CP模型无法捕获的。这使得神经加速模型能够复制高保真度和计算昂贵的模型即同心色散（CD）模型的准确性。提出了一种创新的数据生成过程，用于同时生成CD模型和CP模型的神经网络训练数据。提出了两种版本的神经加速模型，一种是使用传统的神经网络，另一种是使用带有自定义激活函数的神经网络。与高精度的CD模型相比，这两个版本都表现出优异的精度，MSE低至0.117°C， R2值接近0.995，错误率低于0.394%。在计算效率方面，所提模型分别实现了342倍和764倍的加速。通过使用从自定义激活函数中受益的紧凑体系结构，同时还增强了模型的可解释性，为所建议模型的后续版本获得了更多的加速。这些结果突出了该模型对高精度和计算效率要求高的场景的适用性。

{"title":"Neural-accelerated numerical model for packed bed latent heat storage system","authors":"Dessie Tadele Embiale , Shri Balaji Padmanabhan , Mohamed Tahar Mabrouk , Stéphane Grieu , Bruno Lacarrière","doi":"10.1016/j.egyai.2025.100602","DOIUrl":"10.1016/j.egyai.2025.100602","url":null,"abstract":"<div><div>Developing accurate and computationally efficient dynamic models for packed-bed latent-heat storages (PBLHS) is crucial for reliably predicting their performance across different operating scenarios and enabling their use in planning and real-time control. In this study, a novel neural-accelerated numerical model for PBLHS is proposed by coupling a neural network (NN) into a coarsely discretized equations of the Continuous-solid Phase (CP) model. The embedded NN predicts the surface temperature of the phase change material (PCM) given the fluid temperature and enthalpy of the PCM as inputs, which the CP model fails to capture. This allows the neural-accelerated model to replicate the accuracy of a high-fidelity and computationally expensive model namely Concentric Dispersion (CD) model. An innovative data generation process to generate training data for NN involving both CD and CP model is proposed. Two versions of neural-accelerated model are proposed, one with conventional NN and another using NN with a custom activation function. Both versions demonstrate an excellent accuracy, achieving MSE as low as 0.117 °C, <span><math><msup><mrow><mi>R</mi></mrow><mn>2</mn></msup></math></span> values closer to 0.995 and error percentage below 0.394<span><math><mo>%</mo></math></span> compared to the highly accurate CD model. As for computational efficiency, the proposed models achieved 342 times and 764 times acceleration respectively. The gain in more acceleration for the later version of the proposed model is achieved through the use of a compact architecture that benefits from the custom activation function, while also enhancing model explainability. These results highlight the model’s suitability for scenarios demanding both high accuracy and computational efficiency.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"22 ","pages":"Article 100602"},"PeriodicalIF":9.6,"publicationDate":"2025-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144921292","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A hybrid monthly electricity demand forecasting model combining an Hodrick-Prescott filter, recurrent neural networks, and autoregressive integrated moving average 结合Hodrick-Prescott滤波、递归神经网络和自回归综合移动平均的混合月电力需求预测模型

IF 9.6 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Energy and AI

Pub Date : 2025-08-27 DOI: 10.1016/j.egyai.2025.100600

Zhenyu Su, Juan Zhang, Zhehan Yang, Leihao Ma

The coexistence of growth trends and seasonal fluctuations in monthly electricity demand presents significant forecasting challenges. Therefore, this study proposes a univariate time series forecasting approach that applies the Hodrick-Prescott (HP) filter to decompose the demand series into trend and seasonal components. Autoregressive integrated moving average (ARIMA) is used to forecast the trend, while recurrent neural networks (RNNs) handle the periodic component. The final prediction is obtained by combining the forecasts of both components. The model’s predictive performance is evaluated using Guangzhou’s total electricity consumption data. Compared to traditional methods such as Holt-Winters, Seasonal ARIMA, and error-trend-seasonal (ETS), the proposed HP_RNN_ARIMA hybrid model reduces mean absolute percentage error (MAPE), root mean square error (RMSE), and mean absolute error (MAE) by approximately 9.70 % to 35.66 %, 14.18 % to 35.06 %, and 20.01 % to 41.92 %, respectively. Compared to standalone neural networks such as backpropagation (BP), RNNs, and long short-term memory (LSTM), the proposed model lowers MAPE, RMSE, and MAE by approximately 9.05 % to 44.02 %, 20.88 % to 51.74 %, and 29.53 % to 56.23 %, respectively. Against other hybrid models, it reduces these metrics by 3.60 % to 33.39 %, 4.27 % to 36.67 %, and 4.43 % to 44.87 %. It also achieves the highest Willmott’s index (WI) and Legates and McCabe’s index (LMI) scores, reflecting superior model fit. Moreover, applying the HP filter for decomposition and modeling each component individually significantly improves forecasting accuracy.

每月电力需求的增长趋势和季节性波动并存，这对预测提出了重大挑战。因此，本研究提出了一种单变量时间序列预测方法，该方法采用Hodrick-Prescott （HP）滤波器将需求序列分解为趋势分量和季节分量。自回归积分移动平均（ARIMA）用于预测趋势，而循环神经网络（rnn）处理周期分量。最后的预测是将两个分量的预测结合起来得到的。利用广州市总用电量数据对模型的预测性能进行了评价。与传统的Holt-Winters、Seasonal ARIMA和error-trend- Seasonal （ETS）方法相比，HP_RNN_ARIMA混合模型将平均绝对百分比误差（MAPE）、均方根误差（RMSE）和平均绝对误差（MAE）分别降低了约9.70% ~ 35.66%、14.18% ~ 35.06%和20.01% ~ 41.92%。与反向传播（BP）、rnn和长短期记忆（LSTM）等独立神经网络相比，该模型将MAPE、RMSE和MAE分别降低了约9.05%至44.02%、20.88%至51.74%和29.53%至56.23%。与其他混合动力车型相比，它将这些指标降低了3.60%至33.39%，4.27%至36.67%，4.43%至44.87%。它还达到了最高的威尔莫特指数（WI）和Legates和McCabe指数（LMI）得分，反映了卓越的模型拟合。此外，应用HP滤波器对每个组件进行分解和单独建模显著提高了预测精度。

{"title":"A hybrid monthly electricity demand forecasting model combining an Hodrick-Prescott filter, recurrent neural networks, and autoregressive integrated moving average","authors":"Zhenyu Su, Juan Zhang, Zhehan Yang, Leihao Ma","doi":"10.1016/j.egyai.2025.100600","DOIUrl":"10.1016/j.egyai.2025.100600","url":null,"abstract":"<div><div>The coexistence of growth trends and seasonal fluctuations in monthly electricity demand presents significant forecasting challenges. Therefore, this study proposes a univariate time series forecasting approach that applies the Hodrick-Prescott (HP) filter to decompose the demand series into trend and seasonal components. Autoregressive integrated moving average (ARIMA) is used to forecast the trend, while recurrent neural networks (RNNs) handle the periodic component. The final prediction is obtained by combining the forecasts of both components. The model’s predictive performance is evaluated using Guangzhou’s total electricity consumption data. Compared to traditional methods such as Holt-Winters, Seasonal ARIMA, and error-trend-seasonal (ETS), the proposed HP_RNN_ARIMA hybrid model reduces mean absolute percentage error (MAPE), root mean square error (RMSE), and mean absolute error (MAE) by approximately 9.70 % to 35.66 %, 14.18 % to 35.06 %, and 20.01 % to 41.92 %, respectively. Compared to standalone neural networks such as backpropagation (BP), RNNs, and long short-term memory (LSTM), the proposed model lowers MAPE, RMSE, and MAE by approximately 9.05 % to 44.02 %, 20.88 % to 51.74 %, and 29.53 % to 56.23 %, respectively. Against other hybrid models, it reduces these metrics by 3.60 % to 33.39 %, 4.27 % to 36.67 %, and 4.43 % to 44.87 %. It also achieves the highest Willmott’s index (WI) and Legates and McCabe’s index (LMI) scores, reflecting superior model fit. Moreover, applying the HP filter for decomposition and modeling each component individually significantly improves forecasting accuracy.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"22 ","pages":"Article 100600"},"PeriodicalIF":9.6,"publicationDate":"2025-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144919974","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Intelligent coalbed methane drainage optimization: A deep reinforcement learning-driven life-cycle strategy 智能煤层气抽放优化：深度强化学习驱动的生命周期策略

IF 9.6 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Energy and AI

Pub Date : 2025-08-27 DOI: 10.1016/j.egyai.2025.100598

Chen Liu , Bin Gong , HaoQiang Wu , Hu Huang , Heng Zhao

Coalbed methane (CBM) production, as a significant portion of unconventional natural gas development, plays a crucial role in enhancing output and economic benefits through the optimization of its life-cycle drainage strategy. Traditional drainage strategies rely on experience and trial-and-error methods, making it difficult to adapt to complex and dynamic production environments. This study proposes a life-cycle production and drainage optimization strategy for CBM based on Deep Reinforcement Learning (DRL). Utilizing the Deep Q-Network (DQN) algorithm, this work learns and optimizes the drainage strategy during the production process, achieving intelligent decision-making for drainage operations. An auto-regressive surrogate model is introduced to build a reinforcement learning environment for gas production and drainage, based on a deep learning model (CNN-LSTM). This model substitutes the full-physics simulation model that requires high computational cost, significantly accelerating the interactive learning process between the agent and the environment in DRL. Whether to set the gas production or Net Present Value (NPV) as reward, two strategies for reinforcement learning were considered accordingly. The results concluded that the DRL drainage strategy with NPV as the reward increased the net gain by 5.83 % compared to historical data. Compared with traditional methods, this approach significantly improves the NPV and optimizes the drainage strategy. The findings demonstrate that the life-cycle drainage optimization method for CBM based on DRL is not only efficient and feasible but also provides an intelligent solution for the development of unconventional natural gas resources. The results highlight the method's strong adaptability and potential for addressing complex optimization problems in dynamic production environments.

煤层气生产作为非常规天然气开发的重要组成部分，通过优化煤层气全生命周期抽采策略，对提高煤层气产量和经济效益起着至关重要的作用。传统的排水策略依赖于经验和试错方法，难以适应复杂和动态的生产环境。本研究提出了一种基于深度强化学习（DRL）的煤层气全生命周期产排水优化策略。利用Deep Q-Network （DQN）算法，学习并优化生产过程中的排水策略，实现排水作业的智能决策。在深度学习模型（CNN-LSTM）的基础上，引入了自回归代理模型来构建天然气生产和排水的强化学习环境。该模型替代了计算成本较高的全物理仿真模型，显著加快了DRL中agent与环境之间的交互学习过程。以产气量或净现值（NPV）作为奖励，分别考虑了两种强化学习策略。结果表明，与历史数据相比，以净现值为奖励的DRL排水策略使净收益增加了5.83%。与传统方法相比，该方法显著提高了NPV，优化了排水策略。研究结果表明，基于DRL的煤层气全生命周期排水优化方法不仅高效可行，而且为非常规天然气资源开发提供了一种智能化解决方案。结果表明，该方法具有较强的适应性和解决动态生产环境中复杂优化问题的潜力。

{"title":"Intelligent coalbed methane drainage optimization: A deep reinforcement learning-driven life-cycle strategy","authors":"Chen Liu , Bin Gong , HaoQiang Wu , Hu Huang , Heng Zhao","doi":"10.1016/j.egyai.2025.100598","DOIUrl":"10.1016/j.egyai.2025.100598","url":null,"abstract":"<div><div>Coalbed methane (CBM) production, as a significant portion of unconventional natural gas development, plays a crucial role in enhancing output and economic benefits through the optimization of its life-cycle drainage strategy. Traditional drainage strategies rely on experience and trial-and-error methods, making it difficult to adapt to complex and dynamic production environments. This study proposes a life-cycle production and drainage optimization strategy for CBM based on Deep Reinforcement Learning (DRL). Utilizing the Deep Q-Network (DQN) algorithm, this work learns and optimizes the drainage strategy during the production process, achieving intelligent decision-making for drainage operations. An auto-regressive surrogate model is introduced to build a reinforcement learning environment for gas production and drainage, based on a deep learning model (CNN-LSTM). This model substitutes the full-physics simulation model that requires high computational cost, significantly accelerating the interactive learning process between the agent and the environment in DRL. Whether to set the gas production or Net Present Value (NPV) as reward, two strategies for reinforcement learning were considered accordingly. The results concluded that the DRL drainage strategy with NPV as the reward increased the net gain by 5.83 % compared to historical data. Compared with traditional methods, this approach significantly improves the NPV and optimizes the drainage strategy. The findings demonstrate that the life-cycle drainage optimization method for CBM based on DRL is not only efficient and feasible but also provides an intelligent solution for the development of unconventional natural gas resources. The results highlight the method's strong adaptability and potential for addressing complex optimization problems in dynamic production environments.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"22 ","pages":"Article 100598"},"PeriodicalIF":9.6,"publicationDate":"2025-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144919975","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

The pyxis project: A geospatial data system for emission estimation monitoring in the oil and gas industry pyxis项目：一个用于石油和天然气工业排放估计监测的地理空间数据系统

IF 9.6 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Energy and AI

Pub Date : 2025-08-27 DOI: 10.1016/j.egyai.2025.100601

Yaqi Fan , Mohammad S. Masnadi , Liang Jing , Bo Ren , Adam R. Brandt

Consistent estimation and monitoring of greenhouse gas (GHG) emissions in the Oil and Gas (O&G) industry is challenging due to inaccessible, fragmented, and unstandardized datasets. Earlier efforts in estimating such emissions required extensive manual analysis to harmonize diverse data sources on O&G operations. Also, these analyses depend on flaring and methane leakage datasets, which should ideally be updated in near real-time, challenging to integrate effectively to process models. To tackle these challenges, this study proposes a Geographic Information System (GIS)-based data platform called Pyxis for integrating and managing data input associated with GHG emissions estimates in the O&G sector. The Pyxis architecture includes a scalable geodatabase for source management and an automated data pipeline for data management using spatial indexing. This greatly reduces the manual labor traditionally needed for data matching and merging. In addition, top-down remote sensing data can be seamlessly associated with bottom-up field operations data through Pyxis, which improves data recency and spatiotemporal coverage. Here, we apply Pyxis to the O&G fields of Brazil as a case study to show how it can help generating accurate estimates of Carbon Intensity (CI) with data management among disparate and inconsistent data sources. This work highlights the potential of scaling up Pyxis globally via integrating artificial intelligence models for data extraction and ultimately becoming a valuable tool for GHG emissions monitoring and policymaking in the O&G industry.

由于难以获取、碎片化和非标准化的数据集，对石油和天然气行业温室气体（GHG）排放的一致估计和监测具有挑战性。早期估算此类排放的工作需要大量的人工分析，以协调不同的油气操作数据源。此外，这些分析依赖于燃烧和甲烷泄漏数据集，理想情况下，这些数据集应该是实时更新的，很难有效地整合到过程模型中。为了应对这些挑战，本研究提出了一个基于地理信息系统（GIS）的数据平台Pyxis，用于整合和管理与油气行业温室气体排放估算相关的数据输入。Pyxis体系结构包括用于源管理的可扩展地理数据库和用于使用空间索引进行数据管理的自动数据管道。这大大减少了传统上数据匹配和合并所需的手工劳动。此外，自上而下的遥感数据可以通过Pyxis与自下而上的野外作业数据无缝关联，提高了数据的近时性和时空覆盖度。在这里，我们将Pyxis应用于巴西的o&&g油田作为案例研究，以展示它如何通过在不同和不一致的数据源中进行数据管理来帮助生成碳强度（CI）的准确估计。通过整合人工智能模型进行数据提取，Pyxis有可能在全球范围内扩大规模，并最终成为油气行业温室气体排放监测和政策制定的宝贵工具。

{"title":"The pyxis project: A geospatial data system for emission estimation monitoring in the oil and gas industry","authors":"Yaqi Fan , Mohammad S. Masnadi , Liang Jing , Bo Ren , Adam R. Brandt","doi":"10.1016/j.egyai.2025.100601","DOIUrl":"10.1016/j.egyai.2025.100601","url":null,"abstract":"<div><div>Consistent estimation and monitoring of greenhouse gas (GHG) emissions in the Oil and Gas (O&G) industry is challenging due to inaccessible, fragmented, and unstandardized datasets. Earlier efforts in estimating such emissions required extensive manual analysis to harmonize diverse data sources on O&G operations. Also, these analyses depend on flaring and methane leakage datasets, which should ideally be updated in near real-time, challenging to integrate effectively to process models. To tackle these challenges, this study proposes a Geographic Information System (GIS)-based data platform called Pyxis for integrating and managing data input associated with GHG emissions estimates in the O&G sector. The Pyxis architecture includes a scalable geodatabase for source management and an automated data pipeline for data management using spatial indexing. This greatly reduces the manual labor traditionally needed for data matching and merging. In addition, top-down remote sensing data can be seamlessly associated with bottom-up field operations data through Pyxis, which improves data recency and spatiotemporal coverage. Here, we apply Pyxis to the O&G fields of Brazil as a case study to show how it can help generating accurate estimates of Carbon Intensity (CI) with data management among disparate and inconsistent data sources. This work highlights the potential of scaling up Pyxis globally via integrating artificial intelligence models for data extraction and ultimately becoming a valuable tool for GHG emissions monitoring and policymaking in the O&G industry.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"22 ","pages":"Article 100601"},"PeriodicalIF":9.6,"publicationDate":"2025-08-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144921144","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Intelli-Dispatch-SQL: An LLM-based agent for reliable Text-to-SQL in power dispatching Intelli-Dispatch-SQL：一个基于llm的代理，用于电力调度中可靠的Text-to-SQL

IF 9.6 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Energy and AI

Pub Date : 2025-08-26 DOI: 10.1016/j.egyai.2025.100591

Binye Ni , Xinlei Cai , Zhijun Shen , Zijie Meng , Junhua Zhao , Yuheng Cheng , Xuanang Gui

The increasing complexity of modern power systems, driven by factors such as the large-scale integration of renewable energy and the proliferation of distributed generation, has placed unprecedented demands on power dispatching operations. Ensuring grid stability and safety in this new environment requires real-time monitoring and swift, data-driven decision-making. Consequently, efficient and accurate data querying capabilities have become paramount. This study introduces Intelli-Dispatch-SQL, a novel agent-based Text-to-SQL framework that leverages the Large Language Model (LLM) to enhance the accuracy and reliability of generated SQL queries in the context of power dispatching. By integrating intent recognition and SQL validation modules, Intelli-Dispatch-SQL ensures that generated queries are not only syntactically correct but also semantically aligned with user intent and executable within the operational context. Through comprehensive experiments, including ablation studies and cross-model evaluations, we demonstrate that Intelli-Dispatch-SQL significantly outperforms existing Text-to-SQL models, achieving substantial improvements in both Exact Match (EM) and Execution Accuracy (EX). Notably, the incorporation of intent recognition and SQL validation modules is shown to be critical for performance enhancement. The framework’s effectiveness was further validated across various LLMs, confirming its robustness and applicability across diverse scenarios. Intelli-Dispatch-SQL offers a high-performance and generalizable solution for Text-to-SQL in power dispatching, paving the way for more efficient and intelligent power system management.

在可再生能源大规模并网和分布式发电普及等因素的推动下，现代电力系统日益复杂化，对电力调度业务提出了前所未有的要求。在这种新环境下，确保电网的稳定和安全需要实时监控和快速的、数据驱动的决策。因此，高效和准确的数据查询功能变得至关重要。本研究介绍了一种新的基于代理的文本到SQL框架，它利用大语言模型（LLM）来提高电力调度环境中生成的SQL查询的准确性和可靠性。通过集成意图识别和SQL验证模块，Intelli-Dispatch-SQL确保生成的查询不仅在语法上正确，而且在语义上与用户意图一致，并且在操作上下文中可执行。通过综合实验，包括烧消研究和跨模型评估，我们证明了Intelli-Dispatch-SQL显著优于现有的Text-to-SQL模型，在精确匹配（EM）和执行精度（EX）方面都取得了实质性的改进。值得注意的是，意图识别和SQL验证模块的结合对于性能增强至关重要。该框架的有效性在不同的法学硕士中得到进一步验证，证实了其在不同场景中的鲁棒性和适用性。Intelli-Dispatch-SQL为电力调度中的文本到sql提供了一种高性能和通用的解决方案，为更高效和智能的电力系统管理铺平了道路。

{"title":"Intelli-Dispatch-SQL: An LLM-based agent for reliable Text-to-SQL in power dispatching","authors":"Binye Ni , Xinlei Cai , Zhijun Shen , Zijie Meng , Junhua Zhao , Yuheng Cheng , Xuanang Gui","doi":"10.1016/j.egyai.2025.100591","DOIUrl":"10.1016/j.egyai.2025.100591","url":null,"abstract":"<div><div>The increasing complexity of modern power systems, driven by factors such as the large-scale integration of renewable energy and the proliferation of distributed generation, has placed unprecedented demands on power dispatching operations. Ensuring grid stability and safety in this new environment requires real-time monitoring and swift, data-driven decision-making. Consequently, efficient and accurate data querying capabilities have become paramount. This study introduces Intelli-Dispatch-SQL, a novel agent-based Text-to-SQL framework that leverages the Large Language Model (LLM) to enhance the accuracy and reliability of generated SQL queries in the context of power dispatching. By integrating intent recognition and SQL validation modules, Intelli-Dispatch-SQL ensures that generated queries are not only syntactically correct but also semantically aligned with user intent and executable within the operational context. Through comprehensive experiments, including ablation studies and cross-model evaluations, we demonstrate that Intelli-Dispatch-SQL significantly outperforms existing Text-to-SQL models, achieving substantial improvements in both Exact Match (EM) and Execution Accuracy (EX). Notably, the incorporation of intent recognition and SQL validation modules is shown to be critical for performance enhancement. The framework’s effectiveness was further validated across various LLMs, confirming its robustness and applicability across diverse scenarios. Intelli-Dispatch-SQL offers a high-performance and generalizable solution for Text-to-SQL in power dispatching, paving the way for more efficient and intelligent power system management.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"22 ","pages":"Article 100591"},"PeriodicalIF":9.6,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144908072","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Enhancing wind speed prediction in wind farms through ordinal classification 通过顺序分类加强风电场风速预测

IF 9.6 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Energy and AI

Pub Date : 2025-08-25 DOI: 10.1016/j.egyai.2025.100596

A.M. Gómez-Orellana , M. Vega-Bayo , D. Guijo-Rubio , J. Pérez-Aracil , V.M. Vargas , P.A. Gutiérrez , L. Prieto-Godino , S. Salcedo-Sanz , C. Hervás-Martínez

This paper presents and evaluates two novel ordinal classification methods for wind speed prediction, considering three prediction time-horizons: 1h, 4h, and 8h. To address the problem, wind speed values are discretised into four classes, critical for wind farm management. Each class represents essential information for wind farm production, ranging from very low wind speeds to extreme wind speed events and the corresponding production conditions, facilitating operational decisions for wind farm operators. Ordinal classifiers are more suitable than nominal methods to tackle this problem. The study’s primary objective is to compare recently proposed ordinal classifiers for addressing the challenges of wind speed prediction with a focus on extreme wind conditions, which are responsible for many turbine shutdowns. Hourly wind speed measurements from a Spanish wind farm and predictor variables from the European Centre for Medium-Range Weather Forecasts Reanalysis v5 (ERA5 Reanalysis) model are used. The proposed methods include an Artificial Neural Network (ANN) model implementing the Cumulative Link Model as an ordinal output function (

{MLP-CLM}^{O}

), which emphasises overall performance, and an ANN model optimised using a soft labelling technique based on triangular distributions (

{MLP-T}^{O}

), which excels at handling extreme class performance. The results demonstrate the superiority of both approaches over other nominal and ordinal methods across performance metrics that account for the unbalanced nature and ordinality of the data.

{MLP-CLM}^{O}

excels in overall and ordinal performance, while

{MLP-T}^{O}

demonstrates superior handling of the extreme class predictions.

考虑1h、4h和8h三个预测时段，提出并评价了两种新的风速排序预测方法。为了解决这个问题，风速值被离散成四个等级，这对风电场的管理至关重要。每一类都代表了风电场生产的基本信息，从极低风速到极端风速事件以及相应的生产条件，为风电场运营商的运营决策提供便利。序数分类器比名义分类器更适合解决这个问题。该研究的主要目的是比较最近提出的顺序分类器，以解决风速预测的挑战，重点是极端风条件，这是导致许多涡轮机关闭的原因。每小时风速测量来自西班牙风电场和预测变量来自欧洲中期天气预报中心再分析v5 （ERA5再分析）模型。提出的方法包括人工神经网络（ANN）模型，将累积链接模型实现为有序输出函数（MLP-CLMO），强调整体性能，以及使用基于三角分布的软标记技术优化的人工神经网络模型（MLP-TO），该模型擅长处理极端类性能。结果表明，在考虑数据的不平衡性质和有序性的性能指标上，这两种方法都优于其他标称和有序方法。MLP-CLMO在整体和顺序性能方面表现出色，而MLP-TO在极端类别预测方面表现出色。

{"title":"Enhancing wind speed prediction in wind farms through ordinal classification","authors":"A.M. Gómez-Orellana , M. Vega-Bayo , D. Guijo-Rubio , J. Pérez-Aracil , V.M. Vargas , P.A. Gutiérrez , L. Prieto-Godino , S. Salcedo-Sanz , C. Hervás-Martínez","doi":"10.1016/j.egyai.2025.100596","DOIUrl":"10.1016/j.egyai.2025.100596","url":null,"abstract":"<div><div>This paper presents and evaluates two novel ordinal classification methods for wind speed prediction, considering three prediction time-horizons: 1h, 4h, and 8h. To address the problem, wind speed values are discretised into four classes, critical for wind farm management. Each class represents essential information for wind farm production, ranging from very low wind speeds to extreme wind speed events and the corresponding production conditions, facilitating operational decisions for wind farm operators. Ordinal classifiers are more suitable than nominal methods to tackle this problem. The study’s primary objective is to compare recently proposed ordinal classifiers for addressing the challenges of wind speed prediction with a focus on extreme wind conditions, which are responsible for many turbine shutdowns. Hourly wind speed measurements from a Spanish wind farm and predictor variables from the European Centre for Medium-Range Weather Forecasts Reanalysis v5 (ERA5 Reanalysis) model are used. The proposed methods include an Artificial Neural Network (ANN) model implementing the Cumulative Link Model as an ordinal output function (<span><math><msup><mrow><mtext>MLP-CLM</mtext></mrow><mrow><mtext>O</mtext></mrow></msup></math></span>), which emphasises overall performance, and an ANN model optimised using a soft labelling technique based on triangular distributions (<span><math><msup><mrow><mtext>MLP-T</mtext></mrow><mrow><mtext>O</mtext></mrow></msup></math></span>), which excels at handling extreme class performance. The results demonstrate the superiority of both approaches over other nominal and ordinal methods across performance metrics that account for the unbalanced nature and ordinality of the data. <span><math><msup><mrow><mtext>MLP-CLM</mtext></mrow><mrow><mtext>O</mtext></mrow></msup></math></span> excels in overall and ordinal performance, while <span><math><msup><mrow><mtext>MLP-T</mtext></mrow><mrow><mtext>O</mtext></mrow></msup></math></span> demonstrates superior handling of the extreme class predictions.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"22 ","pages":"Article 100596"},"PeriodicalIF":9.6,"publicationDate":"2025-08-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144908071","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Opening the AI black-box: Symbolic regression with Kolmogorov–Arnold Networks for advanced energy applications 打开人工智能黑盒子：先进能源应用的Kolmogorov-Arnold网络符号回归

IF 9.6 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Energy and AI

Pub Date : 2025-08-23 DOI: 10.1016/j.egyai.2025.100595

Nataly R. Panczyk, Omer F. Erdem, Majdi I. Radaideh

While most modern machine learning methods offer speed and accuracy, few promise interpretability or explainability—two key features necessary for highly sensitive industries, like medicine, finance, and engineering. Using eight datasets representative of one especially sensitive industry, nuclear power, this work compares a traditional feedforward neural network (FNN) to a Kolmogorov–Arnold Network (KAN). We consider not only model performance and accuracy, but also interpretability through model architecture and explainability through a post-hoc SHapley Additive exPlanations (SHAP) analysis, a game-theory-based feature importance method. In terms of accuracy, we find KANs and FNNs comparable across all datasets when output dimensionality is limited. KANs, which transform into symbolic equations after training, yield perfectly interpretable models, while FNNs remain black-boxes. Finally, using the post-hoc explainability results from Kernel SHAP, we find that KANs learn real, physical relations from experimental data, while FNNs simply produce statistically accurate results. Overall, this analysis finds KANs a promising alternative to traditional machine learning methods, particularly in applications requiring both accuracy and comprehensibility.

虽然大多数现代机器学习方法都提供了速度和准确性，但很少有人承诺可解释性或可解释性——这是医药、金融和工程等高度敏感行业所必需的两个关键特征。利用核能这一特别敏感的行业的8个数据集，这项工作比较了传统的前馈神经网络（FNN）和Kolmogorov-Arnold网络（KAN）。我们不仅考虑了模型的性能和准确性，还考虑了模型架构的可解释性和基于博弈论的特征重要性分析方法——事后SHapley加性解释（SHAP）分析的可解释性。在准确性方面，当输出维数有限时，我们发现KANs和fnn在所有数据集上都具有可比性。经过训练后转化为符号方程的KANs可以产生完美的可解释模型，而fnn仍然是黑盒子。最后，利用Kernel SHAP的事后可解释性结果，我们发现KANs从实验数据中学习真实的物理关系，而fnn只是产生统计准确的结果。总体而言，该分析发现KANs是传统机器学习方法的有前途的替代方法，特别是在需要准确性和可理解性的应用中。

{"title":"Opening the AI black-box: Symbolic regression with Kolmogorov–Arnold Networks for advanced energy applications","authors":"Nataly R. Panczyk, Omer F. Erdem, Majdi I. Radaideh","doi":"10.1016/j.egyai.2025.100595","DOIUrl":"10.1016/j.egyai.2025.100595","url":null,"abstract":"<div><div>While most modern machine learning methods offer speed and accuracy, few promise interpretability or explainability—two key features necessary for highly sensitive industries, like medicine, finance, and engineering. Using eight datasets representative of one especially sensitive industry, nuclear power, this work compares a traditional feedforward neural network (FNN) to a Kolmogorov–Arnold Network (KAN). We consider not only model performance and accuracy, but also interpretability through model architecture and explainability through a post-hoc SHapley Additive exPlanations (SHAP) analysis, a game-theory-based feature importance method. In terms of accuracy, we find KANs and FNNs comparable across all datasets when output dimensionality is limited. KANs, which transform into symbolic equations after training, yield perfectly interpretable models, while FNNs remain black-boxes. Finally, using the post-hoc explainability results from Kernel SHAP, we find that KANs learn real, physical relations from experimental data, while FNNs simply produce statistically accurate results. Overall, this analysis finds KANs a promising alternative to traditional machine learning methods, particularly in applications requiring both accuracy <em>and</em> comprehensibility.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"22 ","pages":"Article 100595"},"PeriodicalIF":9.6,"publicationDate":"2025-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145045870","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Semi-supervised battery state of health estimation for field applications 现场应用的半监督电池健康状态估计

IF 9.6 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Energy and AI

Pub Date : 2025-08-21 DOI: 10.1016/j.egyai.2025.100575

Nejira Hadzalic , Jacob Hamar , Marco Fischer , Simon Erhard , Jan Philipp Schmidt

Battery electric vehicles are exposed to highly diverse operating conditions and driving behaviors that strongly influence degradation pathways, yet these real-world complexities are only partially captured in laboratory aging tests. This study investigates a semi-supervised learning approach for robust estimation of battery state of health, defined as the ratio of remaining to nominal capacity. The method integrates a multi-view co-training algorithm with a rule-based pseudo labeling mechanism and is developed and validated using field data from 3000 BMW i3 vehicles with battery capacity of 60 Ah, collected since 2013 across 34 countries. The available data comprises standardized full charge capacity measurements, which serve as ground truth labels. The proposed training and validation pipeline is designed to address challenges inherent in real-world data generation and is particularly advantageous during early deployment of new battery technologies, when labeled data is scarce. By incrementally incorporating newly available labeled data into both evaluation and retraining, the model adapts to heterogeneous aging patterns observed in the field. Comparative analysis demonstrates that, relative to a supervised benchmark, the proposed method reduces estimation error by 28 % under limited-label conditions and by 6 % under optimally labeled scenarios, highlighting its robustness for field applications.

纯电动汽车暴露在高度多样化的操作条件和驾驶行为中，这些条件和驾驶行为对老化路径有很大影响，但这些现实世界的复杂性仅在实验室老化测试中得到部分体现。本研究研究了一种半监督学习方法，用于稳健估计电池健康状态，定义为剩余容量与标称容量的比率。该方法将多视图协同训练算法与基于规则的伪标签机制相结合，并使用自2013年以来在34个国家收集的3000辆电池容量为60 Ah的宝马i3汽车的现场数据进行了开发和验证。可用的数据包括标准化的全充电容量测量，作为地面真实值标签。拟议的培训和验证管道旨在解决现实世界数据生成中固有的挑战，并且在新电池技术的早期部署中，当标记数据稀缺时，特别具有优势。通过逐步将新获得的标记数据纳入评估和再训练中，该模型适应了现场观察到的异构老化模式。对比分析表明，相对于监督基准，该方法在有限标签条件下减少了28%的估计误差，在最佳标记场景下减少了6%，突出了其对现场应用的鲁棒性。

{"title":"Semi-supervised battery state of health estimation for field applications","authors":"Nejira Hadzalic , Jacob Hamar , Marco Fischer , Simon Erhard , Jan Philipp Schmidt","doi":"10.1016/j.egyai.2025.100575","DOIUrl":"10.1016/j.egyai.2025.100575","url":null,"abstract":"<div><div>Battery electric vehicles are exposed to highly diverse operating conditions and driving behaviors that strongly influence degradation pathways, yet these real-world complexities are only partially captured in laboratory aging tests. This study investigates a semi-supervised learning approach for robust estimation of battery state of health, defined as the ratio of remaining to nominal capacity. The method integrates a multi-view co-training algorithm with a rule-based pseudo labeling mechanism and is developed and validated using field data from 3000 BMW i3 vehicles with battery capacity of 60Ah, collected since 2013 across 34 countries. The available data comprises standardized full charge capacity measurements, which serve as ground truth labels. The proposed training and validation pipeline is designed to address challenges inherent in real-world data generation and is particularly advantageous during early deployment of new battery technologies, when labeled data is scarce. By incrementally incorporating newly available labeled data into both evaluation and retraining, the model adapts to heterogeneous aging patterns observed in the field. Comparative analysis demonstrates that, relative to a supervised benchmark, the proposed method reduces estimation error by 28% under limited-label conditions and by 6% under optimally labeled scenarios, highlighting its robustness for field applications.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"22 ","pages":"Article 100575"},"PeriodicalIF":9.6,"publicationDate":"2025-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144908147","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Joint online identification method for topology and line parameters in distribution systems based on PLTGNN 基于PLTGNN的配电系统拓扑与线路参数联合在线辨识方法

IF 9.6 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Energy and AI

Pub Date : 2025-08-21 DOI: 10.1016/j.egyai.2025.100593

Wei Wei , Wenwen Ji , Xu Huang , Lingxu Guo , Tao Xu , Yang Wang

Deep learning is an important method for the online identification of topologies and parameters in new power distribution systems. However, practical applications of such methods are hindered by insufficient labeled data, data noise interference, and lack of physical interpretability. To address these issues, this paper proposes a joint online identification method for topology and line parameters in power distribution system based on a Pseudo-label-trained Graph Neural Network (PLTGNN). By generating pseudo-labels through confidence-weighted historical data, the method effectively mitigates the impact of insufficient labeled data on identification accuracy. Furthermore, this paper constructs a comprehensive loss function that integrates pseudo-label learning loss, consistency regularization loss, and power deviation loss, thereby enhancing the model's physical interpretability and noise resistance. Experimental results demonstrate that the proposed method exhibits strong robustness and accuracy in the joint online identification of topology and line parameters.

深度学习是新型配电系统拓扑和参数在线识别的重要方法。然而，这些方法的实际应用受到标记数据不足、数据噪声干扰和缺乏物理可解释性的阻碍。针对这些问题，本文提出了一种基于伪标签训练图神经网络（PLTGNN）的配电系统拓扑和线路参数联合在线识别方法。该方法通过置信度加权历史数据生成伪标签，有效缓解了标注数据不足对识别精度的影响。此外，本文构建了一个综合了伪标签学习损失、一致性正则化损失和功率偏差损失的综合损失函数，从而增强了模型的物理可解释性和抗噪声性。实验结果表明，该方法对拓扑和线路参数的联合在线识别具有较强的鲁棒性和准确性。

{"title":"Joint online identification method for topology and line parameters in distribution systems based on PLTGNN","authors":"Wei Wei , Wenwen Ji , Xu Huang , Lingxu Guo , Tao Xu , Yang Wang","doi":"10.1016/j.egyai.2025.100593","DOIUrl":"10.1016/j.egyai.2025.100593","url":null,"abstract":"<div><div>Deep learning is an important method for the online identification of topologies and parameters in new power distribution systems. However, practical applications of such methods are hindered by insufficient labeled data, data noise interference, and lack of physical interpretability. To address these issues, this paper proposes a joint online identification method for topology and line parameters in power distribution system based on a Pseudo-label-trained Graph Neural Network (PLTGNN). By generating pseudo-labels through confidence-weighted historical data, the method effectively mitigates the impact of insufficient labeled data on identification accuracy. Furthermore, this paper constructs a comprehensive loss function that integrates pseudo-label learning loss, consistency regularization loss, and power deviation loss, thereby enhancing the model's physical interpretability and noise resistance. Experimental results demonstrate that the proposed method exhibits strong robustness and accuracy in the joint online identification of topology and line parameters.</div></div>","PeriodicalId":34138,"journal":{"name":"Energy and AI","volume":"22 ","pages":"Article 100593"},"PeriodicalIF":9.6,"publicationDate":"2025-08-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145266408","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Physics-embedded graph learning unlocks integrated energy system modeling 物理嵌入式图学习解锁集成能源系统建模

IF 9.6 Q1 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Energy and AI

Pub Date : 2025-08-20 DOI: 10.1016/j.egyai.2025.100597

Chongshuo Yuan , Xiaojie Lin , Wei Zhong

Integrated energy system plays a crucial role in global carbon neutrality. Accurate dynamic modeling is essential for optimizing integrated energy system, requiring concurrent modeling of network topology and multi-energy flow dynamics. Existing dynamic modeling approaches often struggle to solve dynamic characteristics with differential-algebraic coupling forms. With the rapid advancements in AI technologies, the integration of AI with energy systems has become not only a promising avenue but also a critical necessity for modeling the modern energy networks. This study innovatively integrates graph neural networks with physical principles, proposing an interpretable neural network methodology. The proposed energy-adapted graph to sequence model (EnG2S) represents a significant advancement for energy systems, pioneering the embedding of fluid dynamics theory to systematically reveal intrinsic connections between multi-energy flow dynamics and neural network characteristics. Overall, this study sets up a new paradigm for energy system modeling, broadening the boundaries of the integration between AI and energy systems.

综合能源系统在全球碳中和中发挥着至关重要的作用。准确的动态建模是优化集成能源系统的必要条件，需要同时进行网络拓扑和多能流动力学建模。现有的动态建模方法往往难以求解具有微分-代数耦合形式的动态特性。随着人工智能技术的快速发展，人工智能与能源系统的集成不仅是一个有前途的途径，而且是现代能源网络建模的关键必要条件。本研究创新性地将图神经网络与物理原理相结合，提出了一种可解释的神经网络方法。所提出的能量自适应图序列模型（EnG2S）代表了能量系统的重大进步，开创了流体动力学理论的嵌入，系统地揭示了多能流动力学与神经网络特性之间的内在联系。总体而言，本研究建立了能源系统建模的新范式，拓宽了人工智能与能源系统集成的边界。

引用次数: 0