Article
Engineering, Chemical
Mehrad Ansari, Heta A. Gandhi, David G. Foster, Andrew D. White
Summary: Computational fluid dynamics (CFD) analysis is widely used in chemical engineering. In this study, active learning (AL) and symbolic regression (SR) are combined to obtain symbolic equations for system variables from CFD simulations. This scalable approach is applicable for any desired number of CFD design parameters.
Article
Construction & Building Technology
Stephanie Higgins, Ted Stathopoulos
Summary: The paper focuses on utilizing artificial intelligence tools to improve urban wind energy generation by studying the impact of building shapes on turbine locations. Research shows that the artificial neural network system performs better in prediction and can be used to identify optimal turbine locations for maximizing urban wind energy production.
BUILDING AND ENVIRONMENT
(2021)
Article
Engineering, Marine
Lucia Santiago Caamano, Maria Isabel Lamas Galdo, Rodrigo Carballo, Ivan Lopez, Juan Jose Cartelle Barros, Luis Carral
Summary: This study presents a methodology that combines the design of artificial reef (AR) units and a 3D hydrodynamic circulation model of estuary, validated using a CFD model. By analyzing the vertical circulation velocity profile and using CFD tool, the circulation of nutrients around the AR unit, approximately 5 times its size, can be determined and validated through towing tank experiments.
JOURNAL OF MARINE SCIENCE AND ENGINEERING
(2022)
Article
Construction & Building Technology
Ruijun Zhang, Parham A. Mirzaei
Summary: The coupling of building energy simulation tools with computational fluid dynamics offers the ability to consider the neighborhood effect on local airflow patterns and energy demand. A novel framework of virtual dynamic BES-CFD-AI coupling was proposed in this study using artificial neural network for predictions, achieving a 2/3 reduction in computational time. A case study in Los Angeles showed satisfactory predictions with an accuracy of 0.88 for local convective heat transfer coefficients on external surfaces.
BUILDING AND ENVIRONMENT
(2021)
Article
Marine & Freshwater Biology
Luis Carral, Maria Isabel Lamas-Galdo, Ma Jesus Rodriguez-Guerreiro, Andreina Vargas, Carlos Alvarez-Feal, Ivan Lopez, Rodrigo Carballo
Summary: Hydrodynamics is crucial for the evolution of artificial reefs, which can determine the geometry of a reef group for optimal water circulation and nutrient supply.
ESTUARINE COASTAL AND SHELF SCIENCE
(2021)
Article
Chemistry, Analytical
Amiya Dash, Shuvabrata Bandopadhay, Soumya Ranjan Samal, Vladimir Poulkov
Summary: This paper proposes an automatic framework for leakage detection and consequence prediction of liquefied petroleum gas (LPG) during transportation using artificial intelligence (AI) and the internet of things (IoT). An AI model is developed to predict the probable consequences of the accident, and an IoT framework is proposed for gas leakage detection and reporting to the disaster management team. The proposed solution allows for quick detection and prediction of gas leakage, aiding in efficient disaster management.
Article
Thermodynamics
Cary A. Faulkner, Dominik S. Jankowski, John E. Castellini, Wangda Zuo, Philipp Epple, Michael D. Sohn, Ali Taleb Zadeh Kasgari, Walid Saad
Summary: The study proposes a CGAN model for predicting indoor airflow distribution and addresses the limitations of current methods, including limited output prediction. A novel feature-driven algorithm is also designed to reduce the amount of expensive training data while maintaining prediction accuracy.
BUILDING SIMULATION
(2023)
Article
Engineering, Aerospace
A. J. Torregrosa, L. M. Garcia-Cuevas, P. Quintero, A. Cremades
Summary: This article proposes a method based on neural networks for calculating the dynamic aerodynamic coefficients of a flat plate, which can greatly reduce computational cost without compromising accuracy.
AEROSPACE SCIENCE AND TECHNOLOGY
(2021)
Article
Computer Science, Artificial Intelligence
Qian Zhang, Jie Lu, Yaochu Jin
Summary: Recommender systems use artificial intelligence to provide personalized services, improve prediction accuracy, and solve data sparsity and cold start issues. This paper discusses how AI can effectively enhance technological development in recommender systems and reviews current research problems and new directions in this field. It also examines the use of various AI techniques, such as fuzzy techniques, transfer learning, genetic algorithms, neural networks, and deep learning, in improving recommender systems.
COMPLEX & INTELLIGENT SYSTEMS
(2021)
Article
Construction & Building Technology
MinHo Kim, Hyung-Jun Park
Summary: The study proposes a new approach based on artificial neural network (ANN) for predicting thermodynamic parameters in an indoor environment. The approach consists of two independently trained ANN models, where the outputs of the first model are used as inputs to the second model. The predicted velocity distribution from the first model is employed as an additional input for the second model. The proposed approach outperforms existing ANN models and provides a reasonable solution for indoor airflow prediction.
JOURNAL OF BUILDING ENGINEERING
(2023)
Article
Chemistry, Physical
Bao Jin, Jun Zhao, Guangyan Chen, Yongyong He, Yiyao Huang, Jianbin Luo
Summary: Graphene-based nanoadditives, such as Mn3O4/graphene (Mn3O4#G) nanocomposite, show significant improvement in lubrication performance when added to lithium grease. The addition of 0.03 wt% Mn3O4#G can reduce coefficient of friction and wear scar depth by 35% and 76% respectively, demonstrating stable lubrication performance even at high temperatures. Computational fluid dynamics (CFD) results indicate that the shear thinning phenomenon of the grease allows for effective dispersion of the nanoadditives in the friction contact area.
APPLIED SURFACE SCIENCE
(2021)
Article
Computer Science, Artificial Intelligence
Shohreh Sheiati, Navid Ranjbar, Jes Frellsen, Elisabeth L. Skare, Rolands Cepuritis, Stefan Jacobsen, Jon Spangenberg
Summary: This study utilizes artificial neural networks to substitute the complex analysis of cement flow conducted by a CFD model, and the results show accurate predictions for both single values and full curves after training, with a significant decrease in calculation time achieved through coupling.
NEURAL COMPUTING & APPLICATIONS
(2021)
Review
Medicine, General & Internal
Inho Kim, Kyungmin Kang, Youngjae Song, Tae-Jung Kim
Summary: With the recent success of artificial intelligence (AI) in computer vision applications, many pathologists expect AI to assist them in various digital pathology tasks. Advances in deep learning have synergized with AI, enabling image-based diagnosis in the field of digital pathology. Efforts are being made to develop AI-based tools that can save time and eliminate errors for pathologists.
Article
Computer Science, Interdisciplinary Applications
Jeff D. Eldredge
Summary: The study introduces a discrete Heaviside function to mask fields on the grid and develop operators and identities for any surface geometry. The derived equations include familiar IBM forcing terms as well as additional terms to regularize field jumps onto the grid and specify constraints on field behavior on each side of the interface, referred to as immersed layers. The method is demonstrated on various incompressible flow problems, showcasing its effectiveness in simulation.
JOURNAL OF COMPUTATIONAL PHYSICS
(2022)
Article
Energy & Fuels
Sarvapriya Singh, Siddharth Suman, Santanu Mitra, Manish Kumar
Summary: A hybrid CFD-ANN approach is used to predict the thermo-hydraulic performance of a solar air heater with rotating circular ribs. An optimized ANN model is developed based on CFD simulations, and it shows good agreement with experimental results. The optimized ANN model significantly reduces the computational time compared to CFD simulations.