Article
Computer Science, Information Systems
Xiuting Gu, Zhu Tianqing, Jie Li, Tao Zhang, Wei Ren, Kim-Kwang Raymond Choo
Summary: This paper introduces a federated learning training model that balances privacy, accuracy, and model fairness using differential privacy (DP). It discusses the fairness and privacy effect of local DP and global DP in federated learning and proposes a fair and privacy quantification mechanism. The experiments demonstrate the positive effect of DP on fairness.
COMPUTERS & SECURITY
(2022)
Article
Green & Sustainable Science & Technology
Sunil Kumar Mohapatra, Sushruta Mishra, Hrudaya Kumar Tripathy, Ahmed Alkhayyat
Summary: Energy consumption analysis plays a crucial role in building energy management for commercial building infrastructures. This study proposes a computationally optimized data-driven model that utilizes advanced data analytics with minimum computing resources to estimate energy consumption sustainably. The results demonstrate that the random forest regressor model performs well for commercial building energy consumption, with a high level of accuracy and low error rates.
SUSTAINABLE ENERGY TECHNOLOGIES AND ASSESSMENTS
(2022)
Article
Chemistry, Analytical
Vincent J. L. Gan, Han Luo, Yi Tan, Min Deng, H. L. Kwok
Summary: This paper presents a new computational framework based on BIM and machine learning models to analyze the impact of natural ventilation on indoor thermal comfort, using CFD modeling and neural network analysis. The study finds that while natural ventilation can save energy consumption, it may not fully meet all thermal comfort criteria, so seasonal performance variations should be considered.
Article
Computer Science, Information Systems
Xianyao You, Ximeng Liu, Xuanwei Lin, Jianping Cai, Shaoquan Chen
Summary: Centralized learning faces constraints in data mapping and security, while federated learning with a distributed architecture addresses these issues by training locally and protecting data privacy. However, fairness becomes a concern in real-world applications of federated learning.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Review
Computer Science, Information Systems
Alex Gittens, Bulent Yener, Moti Yung
Summary: Model accuracy is a traditional metric in machine learning, but privacy, fairness, and robustness are becoming increasingly important. Trustworthy ML requires balancing these four aspects, and the defenses introduced for each aspect may have trade-offs. This paper reviews the state of the art in Trustworthy ML research and advocates for the use of causal modeling to achieve a unified approach to Trustworthy ML.
Article
Construction & Building Technology
Ying Sun, Benjamin C. M. Fung, Fariborz Haghighat
Summary: In recent years, the development and application of data-driven building models have been a hot research topic due to the massive data collected from buildings. This study proposes a sequentially balanced sampling (SBS) technique to address the issues of data volume variation and fairness. The performance of SBS is compared with four existing pre-processing techniques, showing comparable performance in accuracy and fairness improvement.
ENERGY AND BUILDINGS
(2022)
Article
Construction & Building Technology
Antonio Liguori, Romana Markovic, Thi Thu Ha Dam, Jerome Frisch, Christoph van Treeck, Francesco Causone
Summary: This study presents a data-driven approach to fill missing data in building operation, training three different autoencoder neural networks to reconstruct short-term indoor environment data time-series. The models outperform classic numerical approaches, resulting in reconstructing the corresponding variables with average RMSEs of 0.42 degrees C, 1.30 % and 78.41 ppm.
BUILDING AND ENVIRONMENT
(2021)
Review
Construction & Building Technology
Huiheng Liu, Jinrui Liang, Yanchen Liu, Huijun Wu
Summary: Building energy consumption prediction has a significant impact on energy control, design optimization, retrofit evaluation, energy price guidance, and prevention and control of COVID-19 in buildings, ensuring energy efficiency and carbon neutrality. Through reviewing 116 research papers, this study discusses feasible techniques for data-driven building energy prediction across time scales, building levels, and energy consumption types in the context of influencing factors. The review reveals that outdoor dry-bulb temperature is a crucial factor affecting building energy consumption, while data preprocessing, energy consumption feature extraction, and hyperparameter optimization enable accurate prediction.
Article
Construction & Building Technology
Andrew Sonta, Thomas R. Dougherty, Rishee K. Jain
Summary: Occupant behavioral dynamics play a key role in building energy performance, with layout optimization using clustering and genetic algorithms showing potential for reducing energy consumption. High diversity in occupant schedules is found to positively correlate with energy consumption of highly controllable lighting systems. The study demonstrates the benefits of utilizing low-cost dynamic design in building layouts to reduce energy usage and reach sustainable energy goals in the built environment.
ENERGY AND BUILDINGS
(2021)
Article
Construction & Building Technology
Jiyoung Seo, Anseop Choi, Minki Sung
Summary: A personalized luminous environment recommendation system was developed by analyzing personal lifelog data, which showed approximately 92% accuracy in recommending customized lighting environments based on task type, fatigue level, and emotion class.
BUILDING AND ENVIRONMENT
(2021)
Article
Construction & Building Technology
Michael Kent, Thomas Parkinson, Jungsoo Kim, Stefano Schiavon
Summary: This study utilized a new analytical approach to investigate workspace satisfaction, identifying privacy and amount of space, cleanliness and maintenance as the two main factors influencing occupant satisfaction in office buildings.
BUILDING AND ENVIRONMENT
(2021)
Article
Thermodynamics
Yanmin Wang, Zhiwei Li, Junjie Liu, Mingzhe Pei, Yan Zhao, Xuan Lu
Summary: This study proposed a novel method to calculate the indoor characteristic temperature of the heat substation and applied it to three actual district heating systems. The results demonstrated that the proposed method can accurately evaluate the comprehensive characteristics of indoor temperature data.
Article
Computer Science, Artificial Intelligence
Tao Zhang, Tianqing Zhu, Kun Gao, Wanlei Zhou, Philip S. Yu
Summary: As deep learning models mature, finding the ideal tradeoff between accuracy, fairness, and privacy is critical. Privacy and fairness can affect the accuracy of models, so balancing these needs is important. By implementing differentially private stochastic gradient descent (DP-SGD) in deep neural network models, privacy and fairness can be indirectly managed. The number of training epochs plays a central role in striking a balance between accuracy, fairness, and privacy. Based on this observation, two early stopping criteria are designed to help analysts achieve their ideal tradeoff.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Construction & Building Technology
Pandarasamy Arjunan, Kameshwar Poolla, Clayton Miller
Summary: This paper presents the design and implementation of BEEM, a data-driven energy use benchmarking system for buildings in Singapore. The system establishes peer groups for comparison using a public energy disclosure data set and utilizes an ensemble tree algorithm to accurately model building energy use and identify influential factors. Compared to baseline linear regression models used in the previous energy efficiency labeling program in Singapore and other recent models, our models significantly reduce prediction error. Using the prototype implementation of BEEM, we benchmarked and compared the energy performance of office, hotel, and retail buildings.
ENERGY AND BUILDINGS
(2022)
Article
Engineering, Electrical & Electronic
Roohollah Amiri, Srinivas Yerramalli, Taesang Yoo, Mohammed Hirzallah, Marwen Zorgui, Rajat Prakash, Xiaoxia Zhang
Summary: This paper proposes a novel sensing solution for representing an RF-environment and addresses practical challenges and wireless propagation phenomena. It utilizes offline data collection and an iterative process to locate virtual anchors and trains machine learning models to predict the dominant multipath components of the wireless channel. These models are used to improve positioning accuracy in challenging indoor environments through multipath assisted positioning.
IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS
(2023)
Article
Thermodynamics
Zhengxuan Liu, Zhun (Jerry) Yu, Tingting Yang, Mohamed El Mankibi, Letizia Roccamena, Ying Sun, Pengcheng Sun, Shuisheng Li, Guoqiang Zhang
ENERGY CONVERSION AND MANAGEMENT
(2019)
Review
Construction & Building Technology
Ying Sun, Fariborz Haghighat, Benjamin C. M. Fung
ENERGY AND BUILDINGS
(2020)
Article
Construction & Building Technology
Jianing (Tom) Luo, Mahmood Mastani Joybari, Karthik Panchabikesan, Ying Sun, Fariborz Haghighat, Alain Moreau, Miguel Robichaud
SUSTAINABLE CITIES AND SOCIETY
(2020)
Article
Construction & Building Technology
Ying Sun, Karthik Panchabikesan, Fariborz Haghighat, Jianing (Tom) Luo, Alain Moreau, Miguel Robichaud
Summary: This study developed two advanced controllers to achieve peak shifting and heating cost-saving by controlling electrically heated floor and heat extraction system. The results showed that the developed controllers effectively shift energy consumption from the peak period and decrease heating cost by around 33%.
JOURNAL OF BUILDING ENGINEERING
(2021)
Article
Construction & Building Technology
Ying Sun, Benjamin C. M. Fung, Fariborz Haghighat
Summary: In recent years, the development and application of data-driven building models have been a hot research topic due to the massive data collected from buildings. This study proposes a sequentially balanced sampling (SBS) technique to address the issues of data volume variation and fairness. The performance of SBS is compared with four existing pre-processing techniques, showing comparable performance in accuracy and fairness improvement.
ENERGY AND BUILDINGS
(2022)
Article
Construction & Building Technology
Ying Sun, Benjamin C. M. Fung, Fariborz Haghighat
Summary: This study proposes four in-processing methods to improve the predictive fairness of regression models in terms of having similar predictive performance between different conditions. The results show that the mean square error constrained (MSEC) method is the most effective in improving fairness, while the mean square error penalized (MSEP) method is another good option without significantly decreasing the overall accuracy. The mean residual difference constrained (MRDC) method effectively improves the similarity of absolute mean residual difference between different conditions, while the mean residual difference penalized (MRDP) method does not affect the predictive result.
ENERGY AND BUILDINGS
(2022)
Article
Construction & Building Technology
Yapin Yang, Ying Sun, Mingsi Chen, Yuekuan Zhou, Ran Wang, Zhengxuan Liu
Summary: This paper develops a BIM-based fire safety management system platform for construction sites, which includes four subsystems: remote monitoring, fire visualization, multi-stage fire alarm, and escape route optimization. Results show that this system can provide informative and on-time fire protection measures to construction sites.
Article
Energy & Fuels
Ying Sun, Karthik Panchabikesan, Mahmood Mastani Joybari, Dave Olsthoorn, Alain Moreau, Miguel Robichaud, Fariborz Haghighat
JOURNAL OF ENERGY STORAGE
(2018)
Article
Thermodynamics
Yong Cheng, Fukai Song, Lei Fu, Saishuai Dai, Zhiming Yuan, Atilla Incecik
Summary: This paper investigates the accessibility of wave energy absorption by a dual-pontoon floating breakwater integrated with hybrid-type wave energy converters (WECs) and proposes a hydraulic-pneumatic complementary energy extraction method. The performance of the system is validated through experiments and comparative analysis.
Article
Thermodynamics
Jing Gao, Chao Wang, Zhanwu Wang, Jin Lin, Runkai Zhang, Xin Wu, Guangyin Xu, Zhenfeng Wang
Summary: This study aims to establish a new integrated method for biomass cogeneration project site selection, with a focus on the application of the model in Henan Province. By integrating Geographic Information System and Multiple Criterion Decision Making methods, the study conducts site selection in two stages, providing a theoretical reference for the construction of biomass cogeneration projects.
Article
Thermodynamics
Mert Temiz, Ibrahim Dincer
Summary: The current study presents a hybrid small modular nuclear reactor and solar-based system for sustainable communities, integrating floating and bifacial photovoltaic arrays with a small modular reactor. The system efficiently generates power, hydrogen, ammonia, freshwater, and heat for residential, agricultural, and aquaculture facilities. Thermodynamic analysis shows high energy and exergy efficiencies, as well as large-scale ammonia production meeting the needs of metropolitan areas. The hybridization of nuclear and solar technologies offers advantages of reliability, environmental friendliness, and cost efficiency compared to renewable-alone and fossil-based systems.
Editorial Material
Thermodynamics
Wojciech Stanek, Wojciech Adamczyk
Article
Thermodynamics
Desheng Xu, Yanfeng Li, Tianmei Du, Hua Zhong, Youbo Huang, Lei Li, Xiangling Duanmu
Summary: This study investigates the optimization of hybrid mechanical-natural ventilation for smoke control in complex metro stations. The results show that atrium fires are more significantly impacted by outdoor temperature variations compared to concourse/platform fires. The gathered high-temperature smoke inside the atrium can reach up to 900 K under a 5 MW train fire energy release. The findings provide crucial engineering insights into integrating weather data and adaptable ventilation protocols for smoke prevention/mitigation.
Article
Thermodynamics
Da Guo, Heping Xie, Mingzhong Gao, Jianan Li, Zhiqiang He, Ling Chen, Cong Li, Le Zhao, Dingming Wang, Yiwei Zhang, Xin Fang, Guikang Liu, Zhongya Zhou, Lin Dai
Summary: This study proposes a new in-situ pressure-preserved coring tool and elaborates its pressure-preserving mechanism. The experimental and field test results demonstrate that this tool has a high pressure-preservation capability and can maintain a stable pressure in deep wells. This study provides a theoretical framework and design standards for the development of similar technologies.
Article
Thermodynamics
Aolin Lai, Qunwei Wang
Summary: This study assesses the impact of China's de-capacity policy on renewable energy development efficiency (REDE) using the Global-MSBM model and the difference-in-differences method. The findings indicate that the policy significantly enhances REDE, promoting technological advancements and marketization. Moreover, regions with stricter environmental regulations experience a higher impact.
Article
Thermodynamics
Mostafa Ghasemi, Hegazy Rezk
Summary: This study utilizes fuzzy modeling and optimization to enhance the performance of microbial fuel cells (MFCs). By simulating and analyzing experimental data sets, the ideal parameter values for increasing power density, COD elimination, and coulombic efficiency were determined. The results demonstrate that the fuzzy model and optimization methods can significantly improve the performance of MFCs.
Article
Thermodynamics
Zhang Ruan, Lianzhong Huang, Kai Wang, Ranqi Ma, Zhongyi Wang, Rui Zhang, Haoyang Zhao, Cong Wang
Summary: This paper proposes a grey box model for fuel consumption prediction of wing-diesel hybrid vessels based on feature construction. By using both parallel and series grey box modeling methods and six machine learning algorithms, twelve combinations of prediction models are established. A feature construction method based on the aerodynamic performance of the wing and the energy relationship of the hybrid system is introduced. The best combination is obtained by considering the root mean square error, and it shows improved accuracy compared to the white box model. The proposed grey box model can accurately predict the daily fuel consumption of wing-diesel hybrid vessels, contributing to operational optimization and the greenization and decarbonization of the shipping industry.
Article
Thermodynamics
Huayi Chang, Nico Heerink, Junbiao Zhang, Ke He
Summary: This study examines the interaction between off-farm employment decisions between couples and household clean energy consumption in rural China, and finds that two-paycheck households are more likely to consume clean energy. The off-farm employment of women is a key factor driving household clean energy consumption to a higher level, with wage-employed wives having a stronger influence on these decisions than self-employed ones.
Article
Thermodynamics
Hanguan Wen, Xiufeng Liu, Ming Yang, Bo Lei, Xu Cheng, Zhe Chen
Summary: Demand-side management is crucial to smart energy systems. This paper proposes a data-driven approach to understand the relationship between energy consumption patterns and household characteristics for better DSM services. The proposed method uses a clustering algorithm to generate optimal customer groups for DSM and a deep learning model for training. The model can predict the possibility of DSM membership for a given household. The results demonstrate the usefulness of weekly energy consumption data and household socio-demographic information for distinguishing consumer groups and the potential for targeted DSM strategies.
Article
Thermodynamics
Xinglan Hou, Xiuping Zhong, Shuaishuai Nie, Yafei Wang, Guigang Tu, Yingrui Ma, Kunyan Liu, Chen Chen
Summary: This study explores the feasibility of utilizing a multi-level horizontal branch well heat recovery system in the Qiabuqia geothermal field. The research systematically investigates the effects of various engineering parameters on production temperature, establishes mathematical models to describe their relationships, and evaluates the economic viability of the system. The findings demonstrate the significant economic feasibility of the multi-level branch well system.
Article
Thermodynamics
Longxin Zhang, Songtao Wang, Site Hu
Summary: This investigation reveals the influence of tip leakage flow on the modern transonic rotor and finds that the increase of tip clearance size leads to a decline in rotor performance. However, an optimal tip clearance size can extend the rotor's stall margin.
Article
Thermodynamics
Kristian Gjoka, Behzad Rismanchi, Robert H. Crawford
Summary: This paper proposes a framework for assessing the performance of 5GDHC systems and demonstrates it through a case study in a university campus in Melbourne, Australia. The results show that 5GDHC systems are a cost-effective and environmentally viable solution in mild climates, and their successful implementation in Australia can create new market opportunities and potential adoption in other countries with similar climatic conditions.
Article
Thermodynamics
Jianwei Li, Guotai Wang, Panpan Yang, Yongshuang Wen, Leian Zhang, Rujun Song, Chengwei Hou
Summary: This study proposes an orientation-adaptive electromagnetic energy harvester by introducing a rotatable bluff body, which allows for self-regulation to cater for changing wind flow direction. Experimental results show that the output power of the energy harvester can be greatly enhanced with increased rotatory inertia of the rotating bluff body, providing a promising solution for harnessing wind-induced vibration energy.