Article
Computer Science, Artificial Intelligence
Shahdad Ghassemzadeh, Maria Gonzalez Perdomo, Manouchehr Haghighi, Ehsan Abbasnejad
Summary: Physics-based reservoir simulation, while critical in the oil and gas industry, can be time-intensive and challenging to update in real time due to computational demands. In this study, a data-driven simulator using deep learning was developed, showing significantly faster and more accurate results for simulating multiple reservoirs compared to traditional methods.
NEURAL COMPUTING & APPLICATIONS
(2021)
Article
Water Resources
Yongen Lin, Dagang Wang, Yue Meng, Wei Sun, Jianxiu Qiu, Wei Shangguan, Jingheng Cai, Yeonjoo Kim, Yongjiu Dai
Summary: This study investigates the incorporation of bias learning components into data driven models for streamflow prediction. Experiments are conducted in the Andun river basin of China and 273 watersheds in the United States to validate the effectiveness of the mapping-bias-learning models. The results show that these models outperform mapping-learning-alone models and machine learning methods are superior to traditional statistical methods in terms of bias learning ability.
JOURNAL OF HYDROLOGY-REGIONAL STUDIES
(2023)
Article
Engineering, Industrial
Blazhe Gjorgiev, Laya Das, Seline Merkel, Martina Rohrer, Etienne Auger, Giovanni Sansavini
Summary: The combination of physics-based modeling and machine learning methods allows for effective monitoring and identification of faulty insulators in power transmission lines. By training on a large dataset generated by the model, accurate fault detection models with over 99% accuracy can be obtained.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2023)
Article
Engineering, Civil
Zhouteng Ye, Fengyan Shi, Xizeng Zhao, Zijun Hu, Matt Malej
Summary: The study developed a data-driven subgrid approach to extract and accurately reflect flow characteristics at the subgrid scale from a full grid model, using the Random Forest method for data training. Compared to stochastic-based subgrid models, this approach considers more appropriate upscaling factors representing anisotropic subgrid effects.
COASTAL ENGINEERING
(2021)
Article
Robotics
Kong Yao Chee, Tom Z. Jiahao, M. Ani Hsieh
Summary: This letter presents a method to enhance the dynamic models used in model predictive control (MPC) for quadrotor control using deep learning. By integrating a first-principle model and a neural network, the hybrid model can accurately predict the quadrotor dynamics and demonstrates improved performance in closed-loop control.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2022)
Article
Mechanics
Alec J. Linot, Michael D. Graham
Summary: In fluid dynamics, using a data-driven manifold dynamics modelling method, we can accurately describe the state of a fluid system in a low-dimensional coordinate system and build models to capture key characteristics and important dynamics.
JOURNAL OF FLUID MECHANICS
(2023)
Article
Engineering, Aerospace
Yanjun Chen, Shengye Wang, Wei Liu
Summary: This paper builds upon innovative ideas in data-driven turbulence modeling and reconstructs two models for turbulence transition prediction. The results demonstrate improved accuracy and generalization abilities compared to previous models, highlighting the potential of machine learning as a supplementary approach in turbulence transition modeling.
Article
Engineering, Civil
Xie Lian, Xiaolong Hu, Jiang Bian, Liangsheng Shi, Lin Lin, Yuanlai Cui
Summary: This study improves the accuracy of streamflow estimation by replacing the classical evapotranspiration (ET) module with a data-driven submodel in two hydrological models. The results show that streamflow is more sensitive to ET under the saturation excess runoff mechanism. The ETRF-XAJ model outperforms the XAJ model, highlighting the potential of hybrid models for improved streamflow simulation.
JOURNAL OF HYDROLOGY
(2023)
Review
Energy & Fuels
Hong Cai, Wenmin Qin, Lunche Wang, Bo Hu, Ming Zhang
Summary: Accurate hourly clear-sky irradiance estimation is crucial for solar technologies' cost competitiveness and supply-demand balance. This study evaluated multiple models in China, finding that Machine Learning models generally outperformed Clear-sky Irradiance models. Several models showed better performance and can guide the selection of models for hourly CSI estimation in different climatic zones in China.
Article
Economics
Heng-Guo Zhang, Tingting Cao, Houxuan Li, Tiantian Xu
Summary: This paper measures news-driven information friction in China's carbon market, finding that enterprises with different attributes face varying levels of information friction. Higher policy information friction leads to weaker impacts of effective information shocks and greater reductions in policy effects.
Review
Ecology
Da-Yeong Lee, Dae-Seong Lee, YoonKyung Cha, Joong-Hyuk Min, Young-Seuk Park
Summary: This study reviews data-based research using models to predict the biological elements of freshwater ecosystems over the last three decades. It evaluates the ability of current models to predict changes in freshwater organisms and suggests future research directions.
ECOLOGICAL INFORMATICS
(2023)
Review
Chemistry, Multidisciplinary
Ady Suwardi, FuKe Wang, Kun Xue, Ming-Yong Han, Peili Teo, Pei Wang, Shijie Wang, Ye Liu, Enyi Ye, Zibiao Li, Xian Jun Loh
Summary: Biomaterials research has historically been hindered by long development periods, but the application of machine learning in materials science has greatly accelerated progress. The combination of machine learning with high-throughput theoretical predictions and experiments has shifted the traditional trial and error paradigm to a data-driven paradigm, which is driving the discovery and application of biomaterials.
ADVANCED MATERIALS
(2022)
Article
Environmental Sciences
Huimin Bai, Zhiqiang Gong, Guiquan Sun, Li Li
Summary: This study develops an analysis model using SVM method to identify the variations in vegetation coverage based on meteorological elements data. The performance of the SVM model is evaluated and compared with other models, showing better results in simulating the changes in vegetation coverage related to meteorological elements.
Article
Engineering, Industrial
Ugur Kupper, Andreas Klink, Thomas Bergs
Summary: In order to digitalize the wire EDM process, data-driven models are utilized to assess its performance, which is challenging due to the large volume of data and the stochastic nature of the process. This study employs an FPGA system to measure and process electrical parameters and classify temporally and spatially resolved single discharges as normal or abnormal. Supervised machine learning methods, such as artificial neural networks, are utilized to train models with different datasets, enabling the prediction of machined geometrical accuracy and cutting speed based on recorded process data.
CIRP ANNALS-MANUFACTURING TECHNOLOGY
(2023)
Article
Water Resources
Yanan Chen, Donghui Li, Qiankun Zhao, Ximing Cai
Summary: This study presents a generic data-driven reservoir operation model (GDROM) that can accurately simulate the operation and release of reservoirs. GDROM, using a few input variables and employing hidden Markov-decision tree and classification and regression tree algorithms, exhibits good interpretability and performance, making it applicable to reservoirs in different regions.
ADVANCES IN WATER RESOURCES
(2022)
Article
Engineering, Electrical & Electronic
Pinar Satilmis, Thomas Bashford-Rogers, Alan Chalmers, Kurt Debattista
SIGNAL PROCESSING-IMAGE COMMUNICATION
(2020)
Article
Computer Science, Software Engineering
Murat Ture, Mustafa Ege Ciklabakkal, Aykut Erdem, Erkut Erdem, Pinar Satilmis, Ahmet Oguz Akyuz
Summary: This paper addresses the problem of controllably updating the position of the sun using a single image, achieving realistic and real-time changes through an algorithm that allows for both fine- and large-scale adjustments, including simulating the sun position behind clouds. The results demonstrate the effectiveness of leveraging a precomputed atmospheric scattering algorithm for these changes to be realistic and in real-time.
COMPUTER GRAPHICS FORUM
(2021)
Article
Computer Science, Software Engineering
Pinar Satilmis, Demetris Marnerides, Kurt Debattista, Thomas Bashford-Rogers
Summary: Current appearance models for the sky lack the representation of clouds, which are essential for realistic sky applications. This study proposes an alternative approach that synthesizes clouds using a data-driven representation. The synthesized clouds can be artistically controlled and rapidly synthesized for lighting virtual environments.
IEEE COMPUTER GRAPHICS AND APPLICATIONS
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Yitong Sun, Hanchun Wang, Pinar Satilmis, Narges Pourshahrokhi, Carlo Harvey, Ali Asadipour
Summary: Virtual reality (VR) headsets provide realistic simulated environments but can over-stimulate the human eye, especially the Non-Image-Forming (NIF) visual system. To address this, it is important to predict the emitted spectrum of the VR headset and evaluate light stimulation during the virtual environment construction phase. We propose a framework that predicts the spectrum of VR scenes by importing an optical profile of the VR headset, converting it into Five Photoreceptors Radiation Efficacy (FPRE) maps and Melanopic Equivalent Daylight Illuminance (M-EDI) values for visual prediction of detailed stimulation in virtual scenes.
2023 IEEE CONFERENCE ON VIRTUAL REALITY AND 3D USER INTERFACES ABSTRACTS AND WORKSHOPS, VRW
(2023)
