Applied Energy最新文献_第5页

A trustworthy reinforcement learning framework for autonomous control of a large-scale complex heating system: Simulation and field implementation 用于大规模复杂供热系统自主控制的可信强化学习框架：模拟和实地实施

IF 10.1 1区工程技术 Q1 ENERGY & FUELS

Applied Energy

Pub Date : 2024-11-08 DOI: 10.1016/j.apenergy.2024.124815

Amirreza Heidari , Luc Girardin , Cédric Dorsaz , François Maréchal

Traditional control approaches heavily rely on hard-coded expert knowledge, complicating the development of optimal control solutions as system complexity increases. Deep Reinforcement Learning (DRL) offers a self-learning control solution, proving advantageous in scenarios where crafting expert-based solutions becomes intricate. This study investigates the potential of DRL for supervisory control in a unique and complex heating system within a large-scale university building. The DRL framework aims to minimize energy costs while ensuring occupant comfort. However, the trial-and-error learning approach of DRL raises concerns about the trustworthiness of executed actions, hindering practical implementation. To address this, the study incorporates action masking, enabling the integration of hard constraints into DRL to enhance user trust. Maskable Proximal Policy Optimization (MPPO) is evaluated alongside standard Proximal Policy Optimization (PPO) and Soft Actor–Critic (SAC). Simulation results reveal that MPPO achieves comparable energy savings (8% relative to the baseline control) with fewer comfort violations than other methods. Therefore, it is selected among the candidate algorithms and experimentally implemented in the university building over one week. Experimental findings demonstrate that MPPO reduces energy costs while maintaining occupant comfort, resulting in a 36% saving compared to a historical day with similar weather conditions. These results underscore the proactive decision-making capability of DRL, establishing its viability for autonomous control in complex energy systems.

传统的控制方法严重依赖于硬编码的专家知识，随着系统复杂性的增加，优化控制解决方案的开发也变得更加复杂。深度强化学习（DRL）提供了一种自学习控制解决方案，在基于专家的解决方案变得复杂的情况下证明了其优势。本研究调查了 DRL 在大型大学建筑内独特而复杂的供热系统中用于监督控制的潜力。DRL 框架旨在最大限度地降低能源成本，同时确保居住者的舒适度。然而，DRL 的试错学习方法引起了人们对所执行操作的可信度的担忧，从而阻碍了实际应用。为解决这一问题，本研究采用了行动掩码技术，将硬约束整合到 DRL 中，以增强用户信任度。可屏蔽近端策略优化（MPPO）与标准近端策略优化（PPO）和软行为批判（SAC）一起进行了评估。仿真结果表明，与其他方法相比，MPPO 实现了相当的节能效果（相对于基线控制为 8%），且违反舒适度的情况较少。因此，我们在候选算法中选择了 MPPO，并在大学建筑中进行了为期一周的实验。实验结果表明，MPPO 降低了能源成本，同时保持了居住舒适度，与天气条件相似的历史天数相比，节省了 36%。这些结果凸显了 DRL 的前瞻性决策能力，确立了其在复杂能源系统中进行自主控制的可行性。

{"title":"A trustworthy reinforcement learning framework for autonomous control of a large-scale complex heating system: Simulation and field implementation","authors":"Amirreza Heidari , Luc Girardin , Cédric Dorsaz , François Maréchal","doi":"10.1016/j.apenergy.2024.124815","DOIUrl":"10.1016/j.apenergy.2024.124815","url":null,"abstract":"<div><div>Traditional control approaches heavily rely on hard-coded expert knowledge, complicating the development of optimal control solutions as system complexity increases. Deep Reinforcement Learning (DRL) offers a self-learning control solution, proving advantageous in scenarios where crafting expert-based solutions becomes intricate. This study investigates the potential of DRL for supervisory control in a unique and complex heating system within a large-scale university building. The DRL framework aims to minimize energy costs while ensuring occupant comfort. However, the trial-and-error learning approach of DRL raises concerns about the trustworthiness of executed actions, hindering practical implementation. To address this, the study incorporates action masking, enabling the integration of hard constraints into DRL to enhance user trust. Maskable Proximal Policy Optimization (MPPO) is evaluated alongside standard Proximal Policy Optimization (PPO) and Soft Actor–Critic (SAC). Simulation results reveal that MPPO achieves comparable energy savings (8% relative to the baseline control) with fewer comfort violations than other methods. Therefore, it is selected among the candidate algorithms and experimentally implemented in the university building over one week. Experimental findings demonstrate that MPPO reduces energy costs while maintaining occupant comfort, resulting in a 36% saving compared to a historical day with similar weather conditions. These results underscore the proactive decision-making capability of DRL, establishing its viability for autonomous control in complex energy systems.</div></div>","PeriodicalId":246,"journal":{"name":"Applied Energy","volume":"378 ","pages":"Article 124815"},"PeriodicalIF":10.1,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662519","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A privacy-preserving heterogeneous federated learning framework with class imbalance learning for electricity theft detection 用于窃电检测的具有类不平衡学习功能的隐私保护异构联合学习框架

IF 10.1 1区工程技术 Q1 ENERGY & FUELS

Applied Energy

Pub Date : 2024-11-08 DOI: 10.1016/j.apenergy.2024.124789

Hanguan Wen , Xiufeng Liu , Bo Lei , Ming Yang , Xu Cheng , Zhe Chen

Electricity theft is a critical issue in smart grids, leading to significant financial losses for utilities and compromising the stability and reliability of the power system. Existing centralized methods for electricity theft detection raise privacy and security concerns due to the need for sharing sensitive customer data. To address these challenges, we propose HeteroFL, a novel heterogeneous federated learning framework for privacy-preserving electricity theft detection in smart grids. HeteroFL enables retailers to collaboratively train a global model without sharing their private data, while accounting for the class imbalance problem prevalent in electricity theft datasets. We introduce a data partitioning and aggregation scheme that assigns different weights to classes, ensuring a balanced contribution and representation of each class in the global model. In addition, our framework leverages the CKKS homomorphic encryption scheme to perform secure computations on encrypted parameters and employs a CNN-LSTM model to capture the spatial and temporal dependencies in electricity consumption patterns. We evaluate HeteroFL using a real-world smart grid dataset and demonstrate its effectiveness and efficiency in detecting energy theft. Furthermore, we analyze the robustness and perform ablation studies to validate the framework’s stability and identify the contributions of its key components. Although the impact of approximation errors introduced by the CKKS scheme on the CNN-LSTM model’s performance requires further investigation, our framework presents a promising solution for privacy-preserving and accurate electricity theft detection in smart grids using heterogeneous federated learning.

窃电是智能电网中的一个关键问题，会给电力公司造成重大经济损失，并损害电力系统的稳定性和可靠性。由于需要共享敏感的客户数据，现有的集中式窃电检测方法引发了隐私和安全问题。为了应对这些挑战，我们提出了 HeteroFL，这是一种新型异构联合学习框架，用于智能电网中保护隐私的窃电检测。HeteroFL 使零售商能够在不共享其隐私数据的情况下协作训练一个全局模型，同时考虑到窃电数据集中普遍存在的类不平衡问题。我们引入了一种数据分区和聚合方案，为类别分配不同的权重，确保全局模型中每个类别的贡献和代表性均衡。此外，我们的框架还利用 CKKS 同态加密方案对加密参数执行安全计算，并采用 CNN-LSTM 模型捕捉电力消费模式的空间和时间依赖性。我们使用真实世界的智能电网数据集对 HeteroFL 进行了评估，并证明了它在检测能源盗窃方面的有效性和效率。此外，我们还分析了鲁棒性并进行了消融研究，以验证该框架的稳定性并确定其关键组件的贡献。虽然 CKKS 方案引入的近似误差对 CNN-LSTM 模型性能的影响还需要进一步研究，但我们的框架为利用异构联合学习在智能电网中进行隐私保护和准确的窃电检测提供了一个前景广阔的解决方案。

{"title":"A privacy-preserving heterogeneous federated learning framework with class imbalance learning for electricity theft detection","authors":"Hanguan Wen , Xiufeng Liu , Bo Lei , Ming Yang , Xu Cheng , Zhe Chen","doi":"10.1016/j.apenergy.2024.124789","DOIUrl":"10.1016/j.apenergy.2024.124789","url":null,"abstract":"<div><div>Electricity theft is a critical issue in smart grids, leading to significant financial losses for utilities and compromising the stability and reliability of the power system. Existing centralized methods for electricity theft detection raise privacy and security concerns due to the need for sharing sensitive customer data. To address these challenges, we propose HeteroFL, a novel heterogeneous federated learning framework for privacy-preserving electricity theft detection in smart grids. HeteroFL enables retailers to collaboratively train a global model without sharing their private data, while accounting for the class imbalance problem prevalent in electricity theft datasets. We introduce a data partitioning and aggregation scheme that assigns different weights to classes, ensuring a balanced contribution and representation of each class in the global model. In addition, our framework leverages the CKKS homomorphic encryption scheme to perform secure computations on encrypted parameters and employs a CNN-LSTM model to capture the spatial and temporal dependencies in electricity consumption patterns. We evaluate HeteroFL using a real-world smart grid dataset and demonstrate its effectiveness and efficiency in detecting energy theft. Furthermore, we analyze the robustness and perform ablation studies to validate the framework’s stability and identify the contributions of its key components. Although the impact of approximation errors introduced by the CKKS scheme on the CNN-LSTM model’s performance requires further investigation, our framework presents a promising solution for privacy-preserving and accurate electricity theft detection in smart grids using heterogeneous federated learning.</div></div>","PeriodicalId":246,"journal":{"name":"Applied Energy","volume":"378 ","pages":"Article 124789"},"PeriodicalIF":10.1,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662419","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Predicting U.S. federal fleet electric vehicle charging patterns using internal combustion engine vehicle fueling transaction statistics 利用内燃机汽车加油交易统计数据预测美国联邦车队电动汽车充电模式

IF 10.1 1区工程技术 Q1 ENERGY & FUELS

Applied Energy

Pub Date : 2024-11-08 DOI: 10.1016/j.apenergy.2024.124778

Karen Ficenec, Mark Singer, Cabell Hodge

Utilizing fueling transactions from internal combustion engine vehicles (ICEVs), the authors estimated how frequently midday public charging would be required for U.S. federal fleet battery electric vehicles (BEVs). Fueling transaction summary statistics are more widely available than trip-level telematics data, making this methodology more accessible and transferable to other researchers and fleet managers considering BEV replacements. For example, readers can easily apply a linear model using only the count of back-to-back fueling events at gas stations over 57 straight-line miles apart to predict days exceeding range. This linear regression predicted binned days exceeding 250 miles at 80 % accuracy on a hold-out test set from the same fleet as the training data and 66 % accuracy on a new fleet displaying different driving behaviors. The authors additionally provide linear equations for days exceeding 200 and 300 miles as alternative range estimates to account for differences in BEV range and temperature impacts. Beyond the single-feature linear models which readers can apply, the authors tuned and trained other machine learning models on a variety of fueling transaction statistics including consecutive transaction distances, transaction distance from garage, estimated miles traveled from fuel economy and fuel quantity, and transaction periodicity. Utilizing a subset of 1678 light-duty federal fleet vehicles which contained daily vehicle miles traveled (VMT) in addition to fueling statistics, the authors determined which fueling transaction statistics were most relevant in predicting driving days exceeding 250 miles (an approximation of BEV rated driving range). In support of the U.S. federal fleet transition to zero-emission vehicles (ZEVs), the authors used these statistics and machine learning models to predict the frequency of BEV midday charging. After training models on the subset with VMT, the authors predicted days exceeding rated range for 112,902 light-duty vehicles operating in similar circumstances in the federal fleet using a Support Vector Regressor (SVR). They then used the projections as part of the ZEV Planning and Charging (ZPAC) tool to identify optimal candidates for BEVs for the federal fleet. An anonymized version of ZPAC is included in the supplementary materials.

作者利用内燃机汽车（ICEV）的加油交易，估算了美国联邦车队电池电动汽车（BEV）中午需要公共充电的频率。加油交易汇总统计数据比行程级别的远程信息处理数据更容易获得，因此这种方法更容易被其他研究人员和考虑更换 BEV 的车队经理使用。例如，读者只需使用相距超过 57 英里直线距离的加油站背靠背加油事件的计数，就能轻松应用线性模型来预测超出续航里程的天数。这种线性回归预测超过 250 英里的分档天数的准确率为 80%，测试集来自与训练数据相同的车队，准确率为 66%，测试集来自显示不同驾驶行为的新车队。作者还提供了超过 200 英里和 300 英里天数的线性方程，作为替代续航里程估计值，以考虑 BEV 续航里程和温度影响的差异。除了读者可以应用的单一特征线性模型外，作者还根据各种加油交易统计数据调整和训练了其他机器学习模型，包括连续交易距离、从车库出发的交易距离、根据燃油经济性和燃油量估算的行驶里程以及交易周期。作者利用 1678 辆轻型联邦车队车辆的子集（除加油统计数据外，还包含每日车辆行驶里程 (VMT)），确定了哪些加油交易统计数据与预测行驶里程超过 250 英里（BEV 额定行驶里程的近似值）的驾驶天数最为相关。为支持美国联邦政府车队向零排放车辆（ZEV）过渡，作者利用这些统计数据和机器学习模型来预测 BEV 中午充电的频率。在对具有 VMT 的子集进行模型训练后，作者使用支持向量回归器 (SVR) 预测了联邦车队中 112,902 辆在类似情况下运行的轻型车辆超出额定范围的天数。然后，他们将预测结果作为 ZEV 规划和充电 (ZPAC) 工具的一部分，为联邦车队确定最佳的 BEV 候选车型。ZPAC 的匿名版本包含在补充材料中。

{"title":"Predicting U.S. federal fleet electric vehicle charging patterns using internal combustion engine vehicle fueling transaction statistics","authors":"Karen Ficenec, Mark Singer, Cabell Hodge","doi":"10.1016/j.apenergy.2024.124778","DOIUrl":"10.1016/j.apenergy.2024.124778","url":null,"abstract":"<div><div>Utilizing fueling transactions from internal combustion engine vehicles (ICEVs), the authors estimated how frequently midday public charging would be required for U.S. federal fleet battery electric vehicles (BEVs). Fueling transaction summary statistics are more widely available than trip-level telematics data, making this methodology more accessible and transferable to other researchers and fleet managers considering BEV replacements. For example, readers can easily apply a linear model using only the count of back-to-back fueling events at gas stations over 57 straight-line miles apart to predict days exceeding range. This linear regression predicted binned days exceeding 250 miles at 80 % accuracy on a hold-out test set from the same fleet as the training data and 66 % accuracy on a new fleet displaying different driving behaviors. The authors additionally provide linear equations for days exceeding 200 and 300 miles as alternative range estimates to account for differences in BEV range and temperature impacts. Beyond the single-feature linear models which readers can apply, the authors tuned and trained other machine learning models on a variety of fueling transaction statistics including consecutive transaction distances, transaction distance from garage, estimated miles traveled from fuel economy and fuel quantity, and transaction periodicity. Utilizing a subset of 1678 light-duty federal fleet vehicles which contained daily vehicle miles traveled (VMT) in addition to fueling statistics, the authors determined which fueling transaction statistics were most relevant in predicting driving days exceeding 250 miles (an approximation of BEV rated driving range). In support of the U.S. federal fleet transition to zero-emission vehicles (ZEVs), the authors used these statistics and machine learning models to predict the frequency of BEV midday charging. After training models on the subset with VMT, the authors predicted days exceeding rated range for 112,902 light-duty vehicles operating in similar circumstances in the federal fleet using a Support Vector Regressor (SVR). They then used the projections as part of the ZEV Planning and Charging (ZPAC) tool to identify optimal candidates for BEVs for the federal fleet. An anonymized version of ZPAC is included in the supplementary materials.</div></div>","PeriodicalId":246,"journal":{"name":"Applied Energy","volume":"378 ","pages":"Article 124778"},"PeriodicalIF":10.1,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662420","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Path signature-based life prognostics of Li-ion battery using pulse test data 利用脉冲测试数据进行基于路径特征的锂离子电池寿命预测

IF 10.1 1区工程技术 Q1 ENERGY & FUELS

Applied Energy

Pub Date : 2024-11-08 DOI: 10.1016/j.apenergy.2024.124820

Rasheed Ibraheem , Philipp Dechent , Gonçalo dos Reis

Common models predicting the End of Life (EOL) and Remaining Useful Life (RUL) of Li-ion cells make use of long cycling data samples. This is a bottleneck when predictions are needed for decision-making but no historical data is available. A machine learning model to predict the EOL and RUL of Li-ion cells using only data contained in a single Hybrid Pulse Power Characterization (HPPC) test is proposed. The model ignores the cell’s prior cycling usage and is validated across nine different datasets each with its cathode chemistry. A model able to classify cells on whether they have passed EOL given an HPPC test is also developed. The underpinning data-centric modelling concept for feature generation is the notion of ‘path signature’ which is combined with an explainable tree-based machine learning model and an in-depth study of the models is provided. Model validation across different SOC ranges shows that data collected from the HPPC test across a 20% SOC window suffices for effective prediction. The EOL and RUL models achieve 85 and 91 cycles MAE respectively while the classification model has an accuracy of 94% on the test data. Code for data processing and modelling is publicly available.

预测锂离子电池寿命（EOL）和剩余使用寿命（RUL）的常见模型需要使用长周期数据样本。当需要进行决策预测但又没有历史数据时，这是一个瓶颈。本文提出了一种机器学习模型，仅使用单次混合脉冲功率表征（HPPC）测试中的数据来预测锂离子电池的 EOL 和 RUL。该模型忽略了电池之前的循环使用情况，并在九个不同的数据集上进行了验证，每个数据集都有其阴极化学成分。此外，还开发了一种模型，能够根据 HPPC 测试对电池是否通过 EOL 进行分类。以数据为中心的特征生成建模概念是 "路径签名 "概念，它与基于树的可解释机器学习模型相结合，并对模型进行了深入研究。不同 SOC 范围的模型验证表明，从 HPPC 测试中收集的 20% SOC 窗口数据足以进行有效预测。EOL 和 RUL 模型的 MAE 分别为 85 和 91 个周期，而分类模型在测试数据上的准确率为 94%。数据处理和建模代码已公开发布。

{"title":"Path signature-based life prognostics of Li-ion battery using pulse test data","authors":"Rasheed Ibraheem , Philipp Dechent , Gonçalo dos Reis","doi":"10.1016/j.apenergy.2024.124820","DOIUrl":"10.1016/j.apenergy.2024.124820","url":null,"abstract":"<div><div>Common models predicting the End of Life (EOL) and Remaining Useful Life (RUL) of Li-ion cells make use of long cycling data samples. This is a bottleneck when predictions are needed for decision-making but no historical data is available. A machine learning model to predict the EOL and RUL of Li-ion cells using only data contained in a single Hybrid Pulse Power Characterization (HPPC) test is proposed. The model ignores the cell’s prior cycling usage and is validated across nine different datasets each with its cathode chemistry. A model able to classify cells on whether they have passed EOL given an HPPC test is also developed. The underpinning data-centric modelling concept for feature generation is the notion of ‘path signature’ which is combined with an explainable tree-based machine learning model and an in-depth study of the models is provided. Model validation across different SOC ranges shows that data collected from the HPPC test across a 20% SOC window suffices for effective prediction. The EOL and RUL models achieve 85 and 91 cycles MAE respectively while the classification model has an accuracy of 94% on the test data. Code for data processing and modelling is publicly available.</div></div>","PeriodicalId":246,"journal":{"name":"Applied Energy","volume":"378 ","pages":"Article 124820"},"PeriodicalIF":10.1,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662480","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Experimental energy performance assessment of a smart controlled water-flow glazing adaptive facade in heating demand conditions 供热需求条件下智能可控水流玻璃自适应外墙的实验能效评估

IF 10.1 1区工程技术 Q1 ENERGY & FUELS

Applied Energy

Pub Date : 2024-11-08 DOI: 10.1016/j.apenergy.2024.124787

Luis J. Claros-Marfil , Vicente Zetola Vargas , J. Francisco Padial , Benito Lauret

Water-flow glazing (WFG) adaptive facades can significantly enhance the energy efficiency of glazed buildings. Although the energy potential of WFG has already been studied, there remains a research gap between this potential and the control strategies needed. This study evaluates the performance of a smart controller designed to manage active WFG adaptive facades by executing programmed algorithms. These algorithms consider both internal and external ambient conditions in two reduced-scale test cells. The influence of the intelligent controller on the internal cell ambient, the stored water temperature, and the heating energy consumption is examined. The results indicate that the smart control enables the active WFG adaptive facade to reduce the indoor temperature of the test cells during solar radiation hours by absorbing a portion of the internal ambient energy. The stored energy can be used later within a specific time delay, thereby reducing heating energy consumption.

水流玻璃（WFG）自适应外墙可显著提高玻璃建筑的能效。尽管人们已经对 WFG 的能源潜力进行了研究，但在这一潜力与所需的控制策略之间仍存在研究空白。本研究评估了智能控制器的性能，该控制器旨在通过执行编程算法来管理主动式 WFG 自适应外墙。这些算法考虑了两个缩小比例测试单元的内部和外部环境条件。研究了智能控制器对室内环境、储水温度和加热能耗的影响。结果表明，智能控制使主动式 WFG 自适应外墙能够在太阳辐射时间内通过吸收部分内部环境能量来降低测试单元的室内温度。储存的能量可在特定的时间延迟内使用，从而减少供暖能耗。

引用次数: 0

Development of rooftop photovoltaic models to support urban building energy modeling 开发屋顶光伏模型，支持城市建筑能源建模

IF 10.1 1区工程技术 Q1 ENERGY & FUELS

Applied Energy

Pub Date : 2024-11-08 DOI: 10.1016/j.apenergy.2024.124811

Zhiyuan Wang , Jingjing Yang , Guangchen Li , Chengjin Wu , Rongpeng Zhang , Yixing Chen

Developing the rooftop photovoltaic (PV) system was beneficial to generate electricity and reduce carbon emissions in buildings. This paper presented the rooftop PV modeling method to support urban building energy modeling (UBEM) using the prototype UBEM method and the building-by-building UBEM method. The PV modeling method was developed, which was capable for buildings with rectangular flat rooftops, pitched rooftops, and arbitrary-shape flat rooftops. The main layout configuration parameters of the rooftop PV can be customized, including the PV dimension, tilt angle, azimuth angle, number of stacked rows, and the interrow spacing of panels. A district in Changsha, China, was selected as the case study, where basic building information was collected, including the building type, building footprint, year built, and the number of stories. The results showed that the PV models can be successfully added to all 5717 buildings with arbitrary-shape flat rooftops through manual inspection. When the interrow spacing was larger than 1 m, with the decrease of interrow spacing, the power generation increased because of the larger PV installation area, even if the self-shading impact increased. The largest PV power generation was 110.81 kWh/m² and 94.00 kWh/m² per roof area in Changsha when using the prototype UBEM method and the building-by-building UBEM method. The power generation using the building-by-building UBEM method was 15.17 % less than using the prototype UBEM method because the power generation due to shading from surrounding buildings decreased by 5.57 %, and the PV installation area decreased by 10.00 %.

开发屋顶光伏（PV）系统有利于发电和减少建筑物的碳排放。本文介绍了屋顶光伏建模方法，利用原型 UBEM 方法和逐栋建筑 UBEM 方法支持城市建筑能源建模（UBEM）。所开发的光伏建模方法适用于矩形平屋顶、坡屋顶和任意形状平屋顶的建筑。屋顶光伏的主要布局配置参数可以自定义，包括光伏尺寸、倾斜角、方位角、叠加行数和电池板的行间距。研究选择了中国长沙的一个地区作为案例，收集了建筑的基本信息，包括建筑类型、建筑占地面积、建造年份和层数。结果表明，通过人工检测，5717 栋任意形状的平屋顶建筑都能成功添加光伏模型。当间距大于 1 米时，随着间距的减小，光伏安装面积增大，即使自遮阳影响增加，发电量也会增加。在长沙，采用原型 UBEM 法和逐栋 UBEM 法时，每个屋顶面积的最大光伏发电量分别为 110.81 kWh/m2 和 94.00 kWh/m2。采用逐栋建筑物 UBEM 方法的发电量比采用原型 UBEM 方法的发电量减少了 15.17%，原因是周围建筑物遮挡造成的发电量减少了 5.57%，光伏安装面积减少了 10.00%。

{"title":"Development of rooftop photovoltaic models to support urban building energy modeling","authors":"Zhiyuan Wang , Jingjing Yang , Guangchen Li , Chengjin Wu , Rongpeng Zhang , Yixing Chen","doi":"10.1016/j.apenergy.2024.124811","DOIUrl":"10.1016/j.apenergy.2024.124811","url":null,"abstract":"<div><div>Developing the rooftop photovoltaic (PV) system was beneficial to generate electricity and reduce carbon emissions in buildings. This paper presented the rooftop PV modeling method to support urban building energy modeling (UBEM) using the prototype UBEM method and the building-by-building UBEM method. The PV modeling method was developed, which was capable for buildings with rectangular flat rooftops, pitched rooftops, and arbitrary-shape flat rooftops. The main layout configuration parameters of the rooftop PV can be customized, including the PV dimension, tilt angle, azimuth angle, number of stacked rows, and the interrow spacing of panels. A district in Changsha, China, was selected as the case study, where basic building information was collected, including the building type, building footprint, year built, and the number of stories. The results showed that the PV models can be successfully added to all 5717 buildings with arbitrary-shape flat rooftops through manual inspection. When the interrow spacing was larger than 1 m, with the decrease of interrow spacing, the power generation increased because of the larger PV installation area, even if the self-shading impact increased. The largest PV power generation was 110.81 kWh/m<sup>2</sup> and 94.00 kWh/m<sup>2</sup> per roof area in Changsha when using the prototype UBEM method and the building-by-building UBEM method. The power generation using the building-by-building UBEM method was 15.17 % less than using the prototype UBEM method because the power generation due to shading from surrounding buildings decreased by 5.57 %, and the PV installation area decreased by 10.00 %.</div></div>","PeriodicalId":246,"journal":{"name":"Applied Energy","volume":"378 ","pages":"Article 124811"},"PeriodicalIF":10.1,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662479","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Online real-time robust framework for non-intrusive load monitoring in constrained edge devices 用于受限边缘设备非侵入式负载监控的在线实时稳健框架

IF 10.1 1区工程技术 Q1 ENERGY & FUELS

Applied Energy

Pub Date : 2024-11-08 DOI: 10.1016/j.apenergy.2024.124814

L.E. Garcia-Marrero , E. Monmasson , G. Petrone

Real-time information on detailed power consumption can motivate users to make informed decisions to reduce their energy bills. In that sense, Non-Intrusive Load Monitoring (NILM) emerges as a cost-effective technique to achieve the previously mentioned benefits. This paper presents an online real-time robust NILM framework that only requires the aggregated active power, operates by updating the appliance’s state probabilities sequentially, and uses this information to predict the power consumption of each monitored appliance. The framework primarily focuses on the seamless integration and practical deployment of a real-time NILM algorithm, operating at frequencies around 1 Hz, on constrained edge devices. Starting with detecting edges and the base load in real-time, the appliance’s state probabilities are updated considering the possible presence of unknown loads. The power consumption of each appliance is then estimated by employing a modified Population-Based Incremental Learning algorithm (PBIL). Experiments on two publicly available datasets against state-of-the-art methods demonstrated its accuracy and robustness in the presence of unknown appliances. The real-time capabilities of the framework were verified through integration in a Home Automation framework running in a constrained edge device.

有关详细耗电量的实时信息可以促使用户做出明智的决定，从而减少能源费用。从这个意义上说，非侵入式负载监控（NILM）是实现上述优势的一种经济有效的技术。本文介绍了一种在线实时鲁棒性 NILM 框架，它只需要聚合有功功率，通过依次更新设备的状态概率来运行，并利用这些信息来预测每个受监控设备的功耗。该框架主要关注实时 NILM 算法的无缝集成和实际部署，在受限的边缘设备上以 1 Hz 左右的频率运行。从实时检测边缘和基本负载开始，考虑到可能存在的未知负载，对设备的状态概率进行更新。然后采用改进的基于群体的增量学习算法（PBIL）来估算每个设备的功耗。在两个公开的数据集上与最先进的方法进行了对比实验，证明了该框架在存在未知设备时的准确性和鲁棒性。通过将该框架集成到在受限边缘设备中运行的家庭自动化框架中，验证了该框架的实时能力。

{"title":"Online real-time robust framework for non-intrusive load monitoring in constrained edge devices","authors":"L.E. Garcia-Marrero , E. Monmasson , G. Petrone","doi":"10.1016/j.apenergy.2024.124814","DOIUrl":"10.1016/j.apenergy.2024.124814","url":null,"abstract":"<div><div>Real-time information on detailed power consumption can motivate users to make informed decisions to reduce their energy bills. In that sense, Non-Intrusive Load Monitoring (NILM) emerges as a cost-effective technique to achieve the previously mentioned benefits. This paper presents an online real-time robust NILM framework that only requires the aggregated active power, operates by updating the appliance’s state probabilities sequentially, and uses this information to predict the power consumption of each monitored appliance. The framework primarily focuses on the seamless integration and practical deployment of a real-time NILM algorithm, operating at frequencies around 1 Hz, on constrained edge devices. Starting with detecting edges and the base load in real-time, the appliance’s state probabilities are updated considering the possible presence of unknown loads. The power consumption of each appliance is then estimated by employing a modified Population-Based Incremental Learning algorithm (PBIL). Experiments on two publicly available datasets against state-of-the-art methods demonstrated its accuracy and robustness in the presence of unknown appliances. The real-time capabilities of the framework were verified through integration in a Home Automation framework running in a constrained edge device.</div></div>","PeriodicalId":246,"journal":{"name":"Applied Energy","volume":"378 ","pages":"Article 124814"},"PeriodicalIF":10.1,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662451","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Parameter prediction of lead-bismuth fast reactor under various accidents with recurrent neural network 用递归神经网络预测各种事故下铅铋快堆的参数

IF 10.1 1区工程技术 Q1 ENERGY & FUELS

Applied Energy

Pub Date : 2024-11-06 DOI: 10.1016/j.apenergy.2024.124790

Wenshun Duan , Kefan Zhang , Weixiang Wang , Sifan Dong , Rui Pan , Chong Qin , Hongli Chen

Advanced nuclear reactor plays an important role in the sustainable development of green energy, and lead-cooled fast reactors are one of the most promising types. To further improve the safety of lead‑bismuth fast reactors, it is necessary to predict the key parameters and their changing trends under various working conditions quickly and accurately. The prediction method based on the neural network can achieve this goal. In this paper, by using the data of lead‑bismuth reactor NCLFR-Oil under four types of typical accidents, the generalized accident prediction model of lead‑bismuth fast reactor is established with the neural network. First, by comparing the performance differences between the prediction models based on six different neural networks, the gated recurrent neural network with the addition of attention mechanism (AT_GRU) performs the best. Then, a prediction model is established based on the AT_GRU coupled grey wolf optimization algorithm (GWO), and the parameter prediction analysis is carried out for 160 cases of four types of accidents. The results show that the prediction results of the four kinds of accidents are good, even the MAPE, RMSE and R² of the accidents with relatively poor performance can reach 0.165 %, 1.334 °C and 0.9980, respectively. Whether it is a single-type accident model or a general model, the average prediction time of a single case is between 0.014 and 0.035 s, which can be said that the model has realized real-time prediction. Since this paper is not about the prediction of a single working condition, the prediction model obtained is more generalized and has more practical significance.

先进核反应堆在绿色能源的可持续发展中发挥着重要作用，而铅冷快堆是最具发展前景的类型之一。为了进一步提高铅铋快堆的安全性，有必要快速准确地预测各种工况下的关键参数及其变化趋势。基于神经网络的预测方法可以实现这一目标。本文利用铅铋堆 NCLFR-Oil 在四种典型事故下的数据，利用神经网络建立了铅铋快堆的广义事故预测模型。首先，通过比较基于六种不同神经网络的预测模型之间的性能差异，发现添加注意机制的门控递归神经网络（AT_GRU）性能最佳。然后，建立了基于 AT_GRU 耦合灰狼优化算法（GWO）的预测模型，并对四类事故的 160 个案例进行了参数预测分析。结果表明，四种事故的预测结果均较好，甚至性能相对较差的事故的 MAPE、RMSE 和 R2 分别可达 0.165 %、1.334 ℃ 和 0.9980。无论是单类事故模型还是通用模型，单例事故的平均预测时间都在 0.014 至 0.035 s 之间，可以说该模型已经实现了实时预测。由于本文不是针对单一工况的预测，因此得到的预测模型更具有普适性，更有实际意义。

{"title":"Parameter prediction of lead-bismuth fast reactor under various accidents with recurrent neural network","authors":"Wenshun Duan , Kefan Zhang , Weixiang Wang , Sifan Dong , Rui Pan , Chong Qin , Hongli Chen","doi":"10.1016/j.apenergy.2024.124790","DOIUrl":"10.1016/j.apenergy.2024.124790","url":null,"abstract":"<div><div>Advanced nuclear reactor plays an important role in the sustainable development of green energy, and lead-cooled fast reactors are one of the most promising types. To further improve the safety of lead‑bismuth fast reactors, it is necessary to predict the key parameters and their changing trends under various working conditions quickly and accurately. The prediction method based on the neural network can achieve this goal. In this paper, by using the data of lead‑bismuth reactor NCLFR-Oil under four types of typical accidents, the generalized accident prediction model of lead‑bismuth fast reactor is established with the neural network. First, by comparing the performance differences between the prediction models based on six different neural networks, the gated recurrent neural network with the addition of attention mechanism (AT_GRU) performs the best. Then, a prediction model is established based on the AT_GRU coupled grey wolf optimization algorithm (GWO), and the parameter prediction analysis is carried out for 160 cases of four types of accidents. The results show that the prediction results of the four kinds of accidents are good, even the MAPE, RMSE and R<sup>2</sup> of the accidents with relatively poor performance can reach 0.165 %, 1.334 °C and 0.9980, respectively. Whether it is a single-type accident model or a general model, the average prediction time of a single case is between 0.014 and 0.035 s, which can be said that the model has realized real-time prediction. Since this paper is not about the prediction of a single working condition, the prediction model obtained is more generalized and has more practical significance.</div></div>","PeriodicalId":246,"journal":{"name":"Applied Energy","volume":"378 ","pages":"Article 124790"},"PeriodicalIF":10.1,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142593872","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Technological, economic, and emission analysis of the oxy-combustion process 全氧燃烧工艺的技术、经济和排放分析

IF 10.1 1区工程技术 Q1 ENERGY & FUELS

Applied Energy

Pub Date : 2024-11-06 DOI: 10.1016/j.apenergy.2024.124821

Brenda Raho , Marcello Giangreco , Gianpiero Colangelo , Marco Milanese , Arturo de Risi

Abstract

The high concentration of polluting emissions, and in particular of CO₂ in the atmosphere, determines the greenhouse effect, therefore it is necessary to reduce its quantity as much as possible. For this reason, a strong commitment is underway to obtain effective technological improvements and to study adequate operational measures. One measure among these may be the oxy-combustion process.

Many researchers have studied this process, its characteristics, and operating conditions but what is not known in the literature is the economic feasibility of a plant employing this technology and its environmental impact. There are few plants powered by oxy-combustion and many of these are still pilot plants, for this reason using Retscreen it was possible to evaluate and optimize the technical and financial feasibility of an oxy-fuel cogeneration plant for a university campus in such a way as to demonstrate the cost-effectiveness and lower environmental impact that an oxy-fuel system causes compared to a traditional system. It was evaluated the return on investment for the cogeneration plant as the economic parameters varied: in almost all cases analyzed the investment turned out to be convenient and the minimum calculated payback time was 2.5 years.

With this software, it was also possible to determine the environmental impact of this technology which corresponds to a reduction of approximately 3700 tons/year of carbon dioxide compared to a traditional type of system. This work will encourage the investors and corporate sector to embrace this alternative technology for decreasing polluting emissions from the process.

摘要高浓度的污染排放物，特别是大气中的二氧化碳，决定了温室效应，因此有必要尽可能地减少其数量。为此，我们正在大力进行有效的技术改进，并研究适当的操作措施。许多研究人员都对全氧燃烧工艺、其特点和操作条件进行了研究，但文献中并不清楚采用这种技术的工厂的经济可行性及其对环境的影响。因此，使用 Retscreen 可以评估和优化大学校园全氧燃烧热电联产厂的技术和财务可行性，从而证明全氧燃烧系统与传统系统相比具有成本效益，对环境的影响更小。随着经济参数的变化，对热电联产厂的投资回报率进行了评估：在几乎所有分析的情况下，投资都很方便，计算出的最短投资回收期为 2.5 年。通过该软件，还可以确定该技术对环境的影响，与传统系统相比，每年可减少约 3700 吨二氧化碳。这项工作将鼓励投资者和企业部门采用这种替代技术，以减少生产过程中的污染排放。

{"title":"Technological, economic, and emission analysis of the oxy-combustion process","authors":"Brenda Raho , Marcello Giangreco , Gianpiero Colangelo , Marco Milanese , Arturo de Risi","doi":"10.1016/j.apenergy.2024.124821","DOIUrl":"10.1016/j.apenergy.2024.124821","url":null,"abstract":"<div><h3>Abstract</h3><div>The high concentration of polluting emissions, and in particular of CO<sub>2</sub> in the atmosphere, determines the greenhouse effect, therefore it is necessary to reduce its quantity as much as possible. For this reason, a strong commitment is underway to obtain effective technological improvements and to study adequate operational measures. One measure among these may be the oxy-combustion process.</div><div>Many researchers have studied this process, its characteristics, and operating conditions but what is not known in the literature is the economic feasibility of a plant employing this technology and its environmental impact. There are few plants powered by oxy-combustion and many of these are still pilot plants, for this reason using Retscreen it was possible to evaluate and optimize the technical and financial feasibility of an oxy-fuel cogeneration plant for a university campus in such a way as to demonstrate the cost-effectiveness and lower environmental impact that an oxy-fuel system causes compared to a traditional system. It was evaluated the return on investment for the cogeneration plant as the economic parameters varied: in almost all cases analyzed the investment turned out to be convenient and the minimum calculated payback time was 2.5 years.</div><div>With this software, it was also possible to determine the environmental impact of this technology which corresponds to a reduction of approximately 3700 tons/year of carbon dioxide compared to a traditional type of system. This work will encourage the investors and corporate sector to embrace this alternative technology for decreasing polluting emissions from the process.</div></div>","PeriodicalId":246,"journal":{"name":"Applied Energy","volume":"378 ","pages":"Article 124821"},"PeriodicalIF":10.1,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142593834","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Modeling EV dynamic wireless charging loads and constructing risk constrained operating strategy for associated distribution systems 电动汽车动态无线充电负载建模及相关配电系统风险约束运行策略构建

IF 10.1 1区工程技术 Q1 ENERGY & FUELS

Applied Energy

Pub Date : 2024-11-06 DOI: 10.1016/j.apenergy.2024.124735

Xin Cui , Liang Liang , Wei Liu , Wenqian Yin , Junhong Liu , Yunhe Hou

Dynamic wireless charging (DWC) is an emerging technology that enables the charging of electric vehicles (EVs) while they are in motion. However, previous load modeling methods have not thoroughly explored the detailed analysis of DWC load characteristics. Existing research only considers the single-node supply mode for dynamic wireless charging roads (DWCRs), and the assessment of operational risks arising from the uncertain DWC loads has not been addressed. This paper begins by conducting an equivalent circuit analysis of a typical EV DWC system with multiple segmented coils. We present a more accurate trapezoidal power model for a single EV. Subsequently, we model the aggregated EV DWC load, accounting for traffic flow and headway using Poisson and negative exponential distribution functions, respectively. In the operation process, we consider a multi-node supply mode for DWCRs. To address the inaccuracy of long-term predictions, we propose a rolling optimization model to coordinate DWC and renewables with heterogeneous uncertainties by introducing a risk metric to manage potential uncertain risks. The proposed optimization model is transformed into a mixed-integer second-order cone programming (MISOCP) problem after convex relaxation. Finally, we conduct case studies to validate the proposed methods.

动态无线充电（DWC）是一项新兴技术，可在电动汽车（EV）行驶过程中为其充电。然而，以往的负载建模方法并未深入探讨 DWC 负载特性的详细分析。现有研究只考虑了动态无线充电道路（DWCR）的单节点供电模式，而对不确定的 DWC 负载所产生的运行风险的评估尚未涉及。本文首先对具有多个分段线圈的典型电动汽车 DWC 系统进行了等效电路分析。我们为单个电动汽车提出了一个更精确的梯形功率模型。随后，我们分别使用泊松分布函数和负指数分布函数建立了电动汽车 DWC 总负载模型，并考虑了车流量和车速。在运行过程中，我们考虑了 DWCR 的多节点供电模式。针对长期预测的不准确性，我们提出了一种滚动优化模型，通过引入风险度量来管理潜在的不确定风险，从而协调具有异构不确定性的 DWC 和可再生能源。经过凸松弛后，所提出的优化模型被转化为混合整数二阶锥编程（MISOCP）问题。最后，我们进行了案例研究，以验证所提出的方法。

{"title":"Modeling EV dynamic wireless charging loads and constructing risk constrained operating strategy for associated distribution systems","authors":"Xin Cui , Liang Liang , Wei Liu , Wenqian Yin , Junhong Liu , Yunhe Hou","doi":"10.1016/j.apenergy.2024.124735","DOIUrl":"10.1016/j.apenergy.2024.124735","url":null,"abstract":"<div><div>Dynamic wireless charging (DWC) is an emerging technology that enables the charging of electric vehicles (EVs) while they are in motion. However, previous load modeling methods have not thoroughly explored the detailed analysis of DWC load characteristics. Existing research only considers the single-node supply mode for dynamic wireless charging roads (DWCRs), and the assessment of operational risks arising from the uncertain DWC loads has not been addressed. This paper begins by conducting an equivalent circuit analysis of a typical EV DWC system with multiple segmented coils. We present a more accurate trapezoidal power model for a single EV. Subsequently, we model the aggregated EV DWC load, accounting for traffic flow and headway using Poisson and negative exponential distribution functions, respectively. In the operation process, we consider a multi-node supply mode for DWCRs. To address the inaccuracy of long-term predictions, we propose a rolling optimization model to coordinate DWC and renewables with heterogeneous uncertainties by introducing a risk metric to manage potential uncertain risks. The proposed optimization model is transformed into a mixed-integer second-order cone programming (MISOCP) problem after convex relaxation. Finally, we conduct case studies to validate the proposed methods.</div></div>","PeriodicalId":246,"journal":{"name":"Applied Energy","volume":"378 ","pages":""},"PeriodicalIF":10.1,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142587016","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0