Article
Energy & Fuels
Jingang Qu, Thibault Faney, Jean-Charles de Hemptinne, Soleiman Yousef, Patrick Gallinari
Summary: This paper introduces a fast and parallel framework, PTFlash, which uses vectorized algorithms and neural networks to accelerate phase equilibrium calculations, greatly reducing computation time while maintaining high precision.
Article
Mechanics
Peitong Li, Huibo Wang, Xiaoyu Li, Liejin Guo, Hui Jin
Summary: This paper aims to find a convenient method to calculate the friction coefficient on a flat plate in supercritical water (SCW) laminar flow using a pseudo-critical incoming state. Direct numerical simulation (DNS) is used to study the velocity profile characteristics in the SCW boundary layer and apply them to derive a semi-analytical formula for plate friction in the SCW fluid field. The method of obtaining a dimensionless parameter G(mu)* by DNS is given, and the dependence between G(mu)* and boundary conditions is derived by numerical experiments. The accuracy of this method is proved by comparing the results with DNS results.
Article
Computer Science, Artificial Intelligence
Mingjun Qu, Yonghuai Wang, Honghe Li, Jinzhu Yang, Chunyan Ma
Summary: This study proposes a method to automatically identify cLBBB with SF using echocardiography. The method utilizes a linear attention cascaded network (LACNet) to extract spatial and temporal features and enriches input data diversity by adding left ventricle area-time curve. The results show the possibility of using echocardiography for the automatic diagnosis of cLBBB with SF.
MEDICAL IMAGE ANALYSIS
(2022)
Article
Biochemistry & Molecular Biology
Kaixuan Diao, Jing Chen, Tao Wu, Xuan Wang, Guangshuai Wang, Xiaoqin Sun, Xiangyu Zhao, Chenxu Wu, Jinyu Wang, Huizi Yao, Casimiro Gerarduzzi, Xue-Song Liu
Summary: Seq2Neo is a pipeline that predicts the immunogenicity of neoantigens by providing a solution for neoepitope feature prediction using raw sequencing data. It supports different types of genome DNA alterations and includes a CNN-based model that shows improved performance in immunogenicity prediction compared to currently available tools.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
(2022)
Article
Chemistry, Physical
Jaime Jaramillo-Gutierrez, J. L. Lopez-Picon, Jose Torres-Arenas
Summary: The Thermodynamic Geometry (TG) of Mie fluids in the subcritical and supercritical region is studied using a third order thermodynamic perturbation theory equation of state (EOS). The R-crossing method of TG is applied to reproduce the coexistence curves related to Mie fluids. The behavior of the R-Widom line is analyzed and found to be dependent on the stiffness and range potential for the Mie fluids. A correspondence states principle is found for the R-Widom line, and a Clausius-Clapeyron-type relation near the critical point in the supercritical region is fulfilled.
JOURNAL OF MOLECULAR LIQUIDS
(2022)
Article
Anatomy & Morphology
Bart R. Thomson, Louise Francoise Martin, Paul L. Schmidle, Hannah Schlierbach, Anne Schaenzer, Henning Richter
Summary: This study developed an automated pipeline for selecting nerve fibers and calculating the g-ratio in optical microscopy. The custom UNet model demonstrated high agreement with manual delineation and good reliability. The proposed method provides a reliable measurement approach for studying nerve health and function.
FRONTIERS IN NEUROANATOMY
(2023)
Article
Neurosciences
Jennifer Faber, David Kuegler, Emad Bahrami, Lea-Sophie Heinz, Dagmar Timmann, Thomas M. Ernst, Katerina Deike-Hofmann, Thomas Klockgether, Bart van de Warrenburg, Judith van Gaalen, Kathrin Reetz, Sandro Romanzetti, Gulin Oz, James M. Joers, Jorn Diedrichsen, Martin Reuter
Summary: The study introduces CerebNet, a deep learning method for the segmentation of the cerebellum. CerebNet demonstrates high accuracy, test-retest reliability, and sensitivity to disease effects, making it an efficient and validated solution for analyzing cerebellum sub-structure volumes.
Article
Chemistry, Multidisciplinary
Qihui Hu, Nan Zhang, Yuxing Li, Wuchang Wang, Jianlu Zhu, Jiyu Gong
Summary: A new model combining a decompression wave prediction model and an improved BTC model was developed to study the arrest toughness in supercritical CO2 pipeline fracture process. Comparing the decompression wave velocity and fracture propagation velocity, it was found that in some working conditions, the minimum wall thickness calculated based on strength design alone may not meet the requirements for ductile fracture arrest.
Article
Biochemical Research Methods
Ina Bang, Sang-Mok Lee, Seojoung Park, Joon Young Park, Linh Khanh Nong, Ye Gao, Bernhard O. Palsson, Donghyuk Kim
Summary: DEOCSU is a machine learning-based ChIP-exo peak calling suite that accurately predicts DNA-binding protein binding sites and demonstrates high accuracy and reliability across different types of datasets.
BRIEFINGS IN BIOINFORMATICS
(2023)
Article
Chemistry, Physical
Lucien Roach, Gian-Marco Rignanese, Arnaud Erriguible, Cyril Aymonier
Summary: Machine learning has been increasingly implemented as a predictive tool in chemical and physical sciences, offering a computational data-driven approach to accelerate scientific discovery. Although well established in other fields, it is still in its early stages in supercritical fluids research, but is expected to accelerate significantly in the future.
JOURNAL OF SUPERCRITICAL FLUIDS
(2023)
Article
Robotics
Shengchao Yan, Tim Welschehold, Daniel Buescher, Wolfram Burgard
Summary: This letter proposes a novel approach using deep reinforcement learning to optimize traffic flow at intersections in mixed traffic situations. The method allows connected autonomous vehicles to yield to other vehicles to improve traffic flow at unsignalized intersections.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2022)
Article
Ergonomics
Franco Basso, Raill Pezoa, Mauricio Varas, Matias Villalobos
Summary: This article introduces a new image-inspired data architecture and accident-prediction model using Convolutional Neural Networks and Deep Convolutional Generative Adversarial Networks. Computational experiments show that this model outperforms traditional prediction methodologies and has greater applicability in real-life scenarios.
ACCIDENT ANALYSIS AND PREVENTION
(2021)
Article
Multidisciplinary Sciences
Jayasuriya Senthilvelan, Neema Jamshidi
Summary: This study developed a automated workflow called PADLLS that utilizes Deep Convolutional Neural Networks (DCNNs) to achieve fully automated liver segmentation with higher accuracy compared to existing models.
SCIENTIFIC REPORTS
(2022)
Article
Chemistry, Analytical
Suan Lee, Byeonghak Kim
Summary: In this study, a machine learning-based method for water leakage detection was proposed and achieved a high accuracy of 99.79% using data collected from leak detection sensors. This method effectively reduces the time for leak detection and response, minimizing water waste and economic losses.
Article
Plant Sciences
L. Minh Dang, Muhammad Nadeem, Tan N. Nguyen, Han Yong Park, O. New Lee, Hyoung-Kyu Song, Hyeonjoon Moon
Summary: This research introduces a novel approach to automate the measurement of pumpkin biophysical properties using high-resolution images. The experimental results reveal that this method is faster and more efficient compared to conventional techniques for monitoring pumpkin properties.
Article
Biochemistry & Molecular Biology
Amgad Salama, Adel Alyan, Mohamed El Amin, Shuyu Sun, Tao Zhang, Mohamed Zoubeik
Summary: This work studied the effects of deteriorating affinity-related properties of membranes on their rejection capacity due to leaching and erosion through computational fluid dynamics (CFD). The study showed that a decrease in the contact angle can lead to the permeation of droplets that would otherwise be rejected, as demonstrated in the CFD analysis of droplet behavior in a crossflow field.
Article
Energy & Fuels
Siyuan Chen, Yu Li, Tao Zhang, Xingyu Zhu, Shuyu Sun, Xin Gao
Summary: Researchers are actively looking for alternatives like Helium-3 on the Moon due to energy crisis and environmental impacts of fossil fuels. They developed a deep learning method to identify multiple lunar features simultaneously for potential energy source discovery, which showed effectiveness in comprehensive experiments on three datasets.
Review
Energy & Fuels
Tao Zhang, Yiteng Li, Yin Chen, Xiaoyu Feng, Xingyu Zhu, Zhangxing Chen, Jun Yao, Yongchun Zheng, Jianchao Cai, Hongqing Song, Shuyu Sun
Summary: This paper critically reviews space energy resources, focusing on space solar power stations and energy resources on Mars, particularly solar energy, geothermal energy, and wind energy. It also discusses the technical gaps in practical application of SSPS and the potential of Helium-3 as a resource for nuclear fusion.
Article
Energy & Fuels
Tao Zhang, Shuyu Sun
Summary: In this study, a Thermodynamics-Informed Neural Network is developed for the first time to provide a fast, accurate, and robust method for calculating phase equilibrium properties for unconventional reservoirs. The trained model performs well in phase stability tests and phase splitting calculations under a wide range of reservoir conditions, enabling further multi-component multi-phase flow simulations with a strong thermodynamic basis.
Article
Engineering, Chemical
Tao Zhang, Hua Bai, Shuyu Sun
Summary: A novel deep learning method combining thermodynamics-informed neural network and compressor Boolean neural network is proposed for safety check and energy supply predictions of natural gas pipelines.
Article
Biochemistry & Molecular Biology
Tao Zhang, Chenguang Li, Shuyu Sun
Summary: The effect of temperature on the penetration velocities of water droplets in oil-water separations using a separation membrane is studied. A numerical scheme based on the phase-field model is developed to simulate the motion of droplets under different temperatures. It is found that higher temperature leads to faster droplet penetration, and a larger density difference in the oil-water system aids in the separation process.
Article
Energy & Fuels
Tao Zhang, Hua Bai, Shuyu Sun
Summary: This paper proposes an intelligent pipeline dispatch technique using deep learning methods to predict the change in energy supply of natural gas. Practical operation data is used for training and validation, allowing for accelerated predictions and control planning for compressor operations.
FRONTIERS IN ENERGY RESEARCH
(2022)
Article
Biochemistry & Molecular Biology
Jihong Shi, Liang Gong, Tao Zhang, Shuyu Sun
Summary: The study designed and assembled an experimental module for electrodialysis desalination, investigating its ion removal efficiency and single membrane mass transfer flux. The results showed that the desalination performance of the module can be improved by adjusting operating conditions.
Article
Physics, Multidisciplinary
Jie Liu, Tao Zhang, Shuyu Sun
Summary: This paper investigates the imbibition phenomenon in porous media and its dependence on heterogeneity. The results validate that smaller pore throats and wetting surfaces are more favorable for imbibition. The study demonstrates that the SPH method can effectively solve imbibition problems, although stability remains a concern.
Article
Biochemistry & Molecular Biology
Jie Liu, Tao Zhang, Shuyu Sun
Summary: The ion transport in protein nanochannels during peritoneal dialysis was investigated using molecular dynamics (MD) simulations and the MD Monte Carlo (MDMC) algorithm. The spatial distribution of ions and their temporal properties were accurately predicted, validating the suitability of the MDMC method for handling ion transport problems in protein nanochannels.
INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
(2023)
Article
Environmental Sciences
Tao Zhang, Guangyuan Tan, Weihua Bai, Yueqiang Sun, Yuhe Wang, Xiaotian Luo, Hongqing Song, Shuyu Sun
Summary: In this paper, the possible relationship between the maximum electron density, the corresponding critical frequency, and the occurrence of earthquakes is explored. A new disturbance frequency index is proposed as a novel method to help forecast earthquakes. By using this index, earthquake occurrence can be forecasted five days ahead of detection.
Article
Electrochemistry
Xin-Chun Zhang, Nan-Nan Liu, Si-Jie Dong, Tao Zhang, Xiao-Di Yin, Tie-Jun Ci, He-Xiang Wu
Summary: This study investigates the dynamic failure mechanisms of cylindrical lithium-ion batteries (LIBs) under different impact loadings. The crushing behaviors of 18650 LIBs were experimentally analyzed through drop weight impact tests with different impactor heads. The force-electric responses of a LIB under multiple impacts were explored by changing the state of charge (SOC) of the battery, impactor types, and impact energy. The results provide guidance for the crashworthiness design and safety assessment of batteries under multiple impacts.
JOURNAL OF ELECTROCHEMICAL ENERGY CONVERSION AND STORAGE
(2023)
Review
Green & Sustainable Science & Technology
Tao Zhang, Dongxin Huo, Chengyao Wang, Zhengrong Shi
Summary: This study summarizes different methods to solve the phase transition process, discusses the advantages of employing CFD software to simulate the process, and reveals the influences of natural convection and nanoparticles on solidification/melting processes. The challenges and future developments in solution methods are also prospected.
RENEWABLE & SUSTAINABLE ENERGY REVIEWS
(2023)
Article
Energy & Fuels
Xuhai Pan, Langqing Lu, Tao Zhang, Yiming Jiang, Yunyu Li, Zhilei Wang, Min Hua, Juncheng Jiang
Summary: The suppression of hydrogen spontaneous combustion is crucial for the sustainable development of hydrogen energy. Reducing the area of the leak port has been proven to be a feasible method. This study investigates the effect of leak port areas and tube lengths on the suppression of hydrogen spontaneous combustion. The results show that reducing the leak port area increases combustible concentration while decreasing shock wave intensity and the possibility of hydrogen self-ignition. Furthermore, the effect of tube length on the suppression method diminishes as the leak area decreases.
JOURNAL OF ENERGY STORAGE
(2023)