Related references
Note: Only part of the references are listed.
Article
Mechanics
Jan-Hendrik Bastek et al.
Summary: Numerical modeling of thin shell structures is challenging, and various finite element method (FEM) formulations have been proposed to tackle this. This study proposes a Physics-Informed Neural Network (PINN) that utilizes machine learning to predict the small-strain response of curved shells. The PINN performs well in identifying the solution field in benchmark tests when the equations are presented in their weak form, but may fail to do so when using the strong form.
EUROPEAN JOURNAL OF MECHANICS A-SOLIDS
(2023)
Review
Engineering, Electrical & Electronic
Bin Huang et al.
Summary: The advances of deep learning techniques have brought new opportunities to power systems. However, there are challenges in applying deep learning in power systems, such as the requirement for high-quality training data, production of physically inconsistent solutions, and low interpretability. Physics-informed neural networks (PINNs) can address these concerns by integrating physics rules into deep learning methodology. This survey provides a systematic overview of PINN in power systems, summarizing different paradigms and investigating their applications and relevant research.
IEEE TRANSACTIONS ON POWER SYSTEMS
(2023)
Article
Computer Science, Interdisciplinary Applications
Shaoyu Zhao et al.
Summary: This paper proposes a data-driven modeling approach to accurately predict the mechanical properties of graphene-reinforced nanocomposites with vacancy defects. The developed micromechanics models based on molecular dynamics (MD) databases and genetic programming (GP) algorithm provide efficient predictions of thermo-elastic properties with high coefficients of determination. The models are also applied in analyzing the structural behaviors of defective graphene reinforced composite beams.
ENGINEERING WITH COMPUTERS
(2023)
Article
Mechanics
Chuang Liu et al.
Summary: In this paper, a deep operator network (DeepONet) is proposed to bridge the gap between the material design space and mechanical behaviors in the development of high-performance composite materials. The network efficiently predicts the mechanical response such as stress or strain directly from material makeup, and achieves good accuracy even with limited training data. Moreover, it can predict the mechanical response of complex materials regardless of geometry, constitutive relations, and boundary conditions. Combined with optimization algorithms, the network serves as an efficient tool for solving inverse design problems of composite materials.
INTERNATIONAL JOURNAL OF APPLIED MECHANICS
(2023)
Article
Chemistry, Physical
Kequan Xia et al.
Summary: A self-powered bridge health monitoring system based on the elastic origami structure triboelectric nanogenerator (EO-TENG) array is proposed. The system contains the self-powered bridge health data acquisition/processing modular and the NB-IoTs communication modular. The EO-TENG array can convert low-frequency impact vibration energy into multiple frequency current output and provide stable DC voltage to external load through PMC.
Article
Chemistry, Multidisciplinary
Hamed Abdolmaleki et al.
Summary: An interfacial engineering approach using amine-functionalized graphene oxide (AGO) is proposed to enhance the electrical properties of fluoropolymers. This method enables the formation of beta-phase, enlargement of lamellae dimensions, and alignment of micro-dipoles. The resulting PVDF-TrFE films with AGO demonstrate exceptional remnant polarization and high voltage coefficient, energy density, and energy-harvesting figure of merit values, making them suitable for next-generation wearables and human-machine interfaces.
Review
Chemistry, Multidisciplinary
Dongwhi Choi et al.
Summary: Serious climate changes and energy-related environmental problems are critical issues in the world. Triboelectric nanogenerators (TENGs) have become one of the most promising mechanical energy harvesters, offering a solution to reduce carbon emissions and save the environment. Considerable advancements have been made in theory, materials, devices, systems, circuits, and applications in the TENG field, and it has reached the stage of prototype development and commercialization.
Article
Thermodynamics
Allah Rakhio Junejo et al.
Summary: This study focuses on the design of wind turbine in urban environments, using a vertical-axis wind turbine. The design of the robust control system becomes challenging due to the nonlinear aerodynamic characteristics of the rotor blades. By using the proposed controller and a low-inertia turbine, there is a substantial active power gain of approximately 25% compared to traditional control systems.
Article
Automation & Control Systems
Ehsan Haghighat et al.
Summary: Constitutive models are fundamental in modeling physical processes by connecting conservation laws with system kinematics. However, characterizing these models can be challenging, especially in nonlinear regimes. We believe that theory-based parametric elastoplastic models are still the most efficient and predictive.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Mechanics
GaoYuan He et al.
Summary: This study proposes a physics-informed neural network (MFLP-PINN) that combines a multiaxial fatigue critical plane model and a neural network for life prediction. By integrating different life prediction models into the loss function of a neural network, the prediction accuracy can be improved.
EUROPEAN JOURNAL OF MECHANICS A-SOLIDS
(2023)
Article
Mathematics, Applied
Hau T. Mai et al.
Summary: In this study, a direct physics-informed neural network (DPINN) is proposed to analyze the stability of truss structures without the use of incremental-iterative algorithms. A neural network is used to minimize a loss function based on structural instability information. The network parameters are considered as design variables, and joint coordinates are used as input while displacements and load factors are output. The training process involves predicting outputs, establishing a loss function, and using back-propagation and optimization to adjust network parameters until convergence is achieved. The proposed scheme is shown to be efficient and accurate in evaluating the stability of truss structures with geometric nonlinearity.
FINITE ELEMENTS IN ANALYSIS AND DESIGN
(2023)
Article
Engineering, Electrical & Electronic
Marco Baldan et al.
Summary: Physics-informed neural networks (PINNs) have been successfully used for solving direct problems in electromagnetism (EM). However, when dealing with inverse problems or parameter studies, PINNs usually require additional training for each new set of system parameters. To address this issue, we propose a hypernetwork (HNN) that takes system parameters as input and outputs the network weights of a PINN, enabling the fast solution of inverse problems. This approach is demonstrated in a coil optimal design task in magnetostatics.
IEEE TRANSACTIONS ON MAGNETICS
(2023)
Article
Engineering, Electrical & Electronic
Ketian Ye et al.
Summary: This paper proposes a physics-informed sparse Gaussian process (SGP) for probabilistic stability assessment of large-scale power systems in the presence of uncertain dynamic PVs and loads. The SGP is developed with a stochastic variational inference technique to address scalability issues. Data preprocessing, including sample clustering and constructing separate SGPs, is introduced to handle the coexistence of stable and unstable cases. The proposed SGP achieves improved computational efficiency and accuracy with limited data.
IEEE TRANSACTIONS ON POWER SYSTEMS
(2023)
Article
Engineering, Industrial
Clayton Cooper et al.
Summary: This research explores a data-driven predictive model with Shapley additive explanation (SHAP) to improve the accuracy and interpretability of predicting mechanical properties in directed energy deposition (DED) processes. The results demonstrate that by interpreting input features and reducing model complexity, both the accuracy and effectiveness of predictive models for Inconel 718 (IN718) tensile strength can be improved.
JOURNAL OF MATERIALS PROCESSING TECHNOLOGY
(2023)
Review
Chemistry, Physical
Lin-Chuan Zhao et al.
Summary: The Artificial Intelligence of Things (AIoT) connects everything with intelligence, and the increasing energy consumption by electronic devices creates a demand for efficient power supply. Energy harvesting technology has emerged as a promising solution for zero-emission power supply for AIoT. However, there are challenges such as poor electrical output, weak environmental adaptability, and low reliability. Mechanical intelligent energy harvesting, which relies on the system's ability to identify external excitation or its own state, shows great potential in overcoming these challenges. This article discusses the definition, design methodology, and key research directions of mechanical intelligent energy harvesting, highlighting its potential to revolutionize energy harvesting technology and enable new applications.
ADVANCED ENERGY MATERIALS
(2023)
Article
Chemistry, Physical
Haobin Wang et al.
Summary: Triboelectrification between metals, polymers, and semiconductors was investigated, revealing a universal mechanism involving both electrostatic induction and dynamic junction modulation. The outputs of triboelectrification exhibited four current/voltage peaks, two of which were attributed to electrostatic induction effects and determined by surface state densities and electron affinity differences between the materials. The other two peaks were associated with dynamic junction modulation effects and determined by charging exchange between surface states and depletion regions. While metals and polymers showed no modulation of depletion regions, only the peaks induced by electrostatic induction were observable. Semiconductors, on the other hand, exhibited additional peaks induced by junction modulation when in contact with any material type. These findings provide valuable insights for developing high-performance triboelectric nanogenerators.
ADVANCED ENERGY MATERIALS
(2023)
Article
Chemistry, Physical
Mohammad M. Rastegardoost et al.
Summary: Since the introduction of triboelectric nanogenerators (TENGs), various porous materials and structural designs have been explored to improve the TENGs output performance and expand their functionalities. Porous materials like hydrogels, aerogels, foams, and fibrous media have shown to be effective in enhancing the performance and versatility of TENGs. Additionally, structural designs in the form of textiles and yarns provide larger pores for air gap implementation, enabling the fabrication of all-in-one TENGs.
Article
Chemistry, Multidisciplinary
Chun-Teh Chen et al.
Summary: Elastography is a medical imaging technique that measures tissue elasticity using ultrasound signals. Current methods suffer from low resolution and reliance on material incompressibility assumption. A new physics-informed deep-learning method is proposed, which reconstructs the Young's modulus field based on axial displacement alone, and removes the incompressibility assumption. The method provides accurate elasticity reconstruction and mitigates errors caused by material stiffness.
Article
Chemistry, Physical
Yue Liu et al.
Summary: Generative Artificial Intelligence (GAI) is receiving increasing attention in the materials community due to its excellent capability of generating required contents. This review comprehensively examines the development and methodology of GAI, and analyzes the pros and cons of various generative models. The applications of task-specific generative models in materials science, such as materials inverse design and data augmentation, are discussed. The study also explores the potential applications of general GAI in generating multiple materials content, solving differential equations, and querying materials FAQs using ChatGPT as an example. Additionally, six challenges in using GAI in materials science are identified and corresponding solutions are provided. This work paves the way for effective and explainable materials data generation and analysis approaches to accelerate materials research and development.
JOURNAL OF MATERIOMICS
(2023)
Article
Engineering, Multidisciplinary
Yu Diao et al.
Summary: Physics-informed neural networks (PINNs) are widely used in solid mechanics but have limitations in handling multi-material problems and expressing complex material contact models. We propose a method using PINNs and domain decomposition technology to solve multi-material problems and describe material contact relations using regular terms. Through various test cases, we demonstrate the accuracy of our approach compared to common PINN algorithms. We also introduce parameter sharing in multi-task learning to provide additional freedom in choosing sharing structures, leading to higher accuracy.
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
(2023)
Article
Materials Science, Multidisciplinary
Mohammad Abu-Mualla et al.
Summary: This paper investigates the feasibility of data-driven methods in automating the engineering design process, specifically studying inverse design of cellular mechanical metamaterials. The study aims to develop an efficient inverse design methodology that accurately generates mechanical metamaterial while ensuring the manufacturability of predicted structures. Results demonstrate high computational efficiency and prediction accuracy compared to conventional methods.
MATERIALS & DESIGN
(2023)
Article
Engineering, Multidisciplinary
Xin-Yu Guo et al.
Summary: A parameter identification framework based on physics-informed neural networks (PINNs) is developed, which incorporates physical constraints into the training process. Two information acquisition principles are proposed for training data sets and physical constraints. The framework utilizes finite element computation and uniform design to generate the minimum number of training data, and applies multivariate nonlinear regression to establish physical constraints. The proposed framework guides the training process towards a physically or mechanically consistent solution, leading to a physics-informed data-driven approach. Finally, the framework is used to identify stiffness parameters of a laboratory-scale frame model and an actual frame structure.
Article
Nanoscience & Nanotechnology
Tao Huang et al.
Summary: Microplastics, sub-millimeter-sized plastic fragments, have been extensively found in the environment as relatively new pollutants that are difficult to degrade. They have irreversible adverse effects on microorganisms, animals, plants, and human health through the food chain. However, traditional detection and identification methods face challenges due to the small size, variety, and different properties of microplastics. This study presents a method using a liquid-solid triboelectric nanogenerator (LS-TENG) combined with a deep learning model to detect and classify microplastics in liquids. Experimental results demonstrate that the type and content of microplastics in the liquid significantly affect the contact electrification between the liquid and the perfluoroethylene-propylene copolymer. The LS-TENG sensor can quantitatively detect the content of microplastics and a convolutional neural network achieves high recognition accuracy in identifying different types of microplastics labels. This has great significance in expanding the application prospect of LS-TENG and realizing the detection of microplastics in liquids.
ACS APPLIED MATERIALS & INTERFACES
(2023)
Article
Engineering, Ocean
He-Wen-Xuan Li et al.
Summary: This paper investigates the application of physics-informed neural networks (PINNs) to estimate motion and identify system parameters of a moored buoy under different sea states. The results demonstrate that PINN can accurately estimate motion and identify system parameters by choosing appropriate hyperparameters (HPs). This study finds that hyperparameter optimization can significantly reduce the relative error of identified parameters. Overall, this study highlights the applicability of PINN in modeling complex offshore structures and provides insights into selecting optimal HPs for accurate and efficient estimation of motion and system parameters.
APPLIED OCEAN RESEARCH
(2023)
Article
Multidisciplinary Sciences
Pengcheng Jiao et al.
Summary: This article summarizes the research progress of mechanical metamaterials in terms of their mechanical functionalities and multi-functionality. It also proposes data-driven approaches for design and optimization. Moreover, it discusses the emerging applications of mechanical metamaterials in the field of intelligent devices and the challenges that need to be overcome.
NATURE COMMUNICATIONS
(2023)
Review
Chemistry, Physical
Jiuxu Zhang et al.
Summary: Nanogenerators (NGs) are a promising energy solution for self-powered systems, IoT, and blue energy, and their output performance can be evaluated using standardized indicators. Machine learning (ML) has been proven applicable for multi-index evaluation and distributed data analysis. Current research on NGs and ML includes equipment optimization, information security, intelligent devices, human-machine interaction, object recognition, intelligent transportation, intelligent sound systems, and environmental protection. Advanced algorithms like PCA, RF, SVM, and deep learning show better results than traditional algorithms. The development trends and challenges of NGs and ML are discussed, indicating their future extensive applications in the research field.
Article
Chemistry, Physical
Sicong Jiang et al.
Summary: In this study, a high-throughput screening of quaternary-Heusler/MgO heterostructures was conducted for spintronic applications. By analyzing various materials descriptors, 5 promising compounds out of 27,000 quaternary Heusler compounds were identified, showing potential for designing energy-efficient perpendicular magnetic tunnel junctions.
NPJ COMPUTATIONAL MATERIALS
(2023)
Article
Automation & Control Systems
Qianyun Zhang et al.
Summary: Harnessing the power of natural evolution, a concept called evolving metamaterial (EM) is introduced to directly evolve thousands of metastructures with unknown structures and new modes of operation. By randomly creating an initial population of parent metamaterial entities and passing their genetic material to offspring through variation, reproduction, and selection, desired metamaterial configurations emerge. The proposed approach demonstrates the capability to explore both 2D and 3D mechanical metamaterial structures with specific properties such as maximum bulk modulus and minimum Poisson's ratio.
ADVANCED INTELLIGENT SYSTEMS
(2023)
Article
Engineering, Multidisciplinary
Hyogu Jeong et al.
Summary: Physics-Informed Neural Networks (PINNs) have gained attention in the field of topology optimization. This paper proposes a novel framework, CPINNTO, which integrates two distinct PINNs to achieve complete machine-learning-based topology optimization. The research findings indicate that CPINNTO can achieve optimal topologies without labeled data nor FEA, and it demonstrates stability and favorable compliance values in various topology optimization applications.
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
(2023)
Article
Materials Science, Multidisciplinary
Shaoyu Zhao et al.
Summary: In this paper, a highly efficient micromechanical modeling approach based on molecular dynamics simulation and genetic programming algorithm is developed to predict the mechanical properties of graphene origami metallic metamaterials. The experimental results show that the model has efficient and accurate predictive capabilities, which are essential for the analysis and design of functionally graded metal metamaterial composite structures.
Article
Chemistry, Multidisciplinary
Jingzhe Sun et al.
Summary: This study utilizes chitosan blends to enhance triboelectric performance, filling the gap in improving chitosan triboelectric output. The enhanced chitosan film in TENG shows higher output voltage and multiple applications in wearable devices.
ADVANCED FUNCTIONAL MATERIALS
(2022)
Review
Engineering, Industrial
Shenghan Guo et al.
Summary: Machine learning has proven to be an effective alternative to physical models in quality prediction and process optimization of metal additive manufacturing. However, the interpretability of machine learning outcomes within the complex thermodynamics of additive manufacturing has been a challenge. Physics-informed machine learning (PIML) addresses this challenge by integrating data-driven methods with physical domain knowledge.
JOURNAL OF MANUFACTURING SYSTEMS
(2022)
Article
Materials Science, Multidisciplinary
Chenchen Cai et al.
Summary: The triboelectric properties of materials in liquid energy harvesting and emerging applications are important. Chemical functionalization can control the triboelectric properties, improving the utilization of liquid energy resources and reducing electrostatic hazards. This review systematically summarizes the latest research progress in molecular modification to control triboelectric properties through chemical functionalization. It discusses the mechanism of contact electrification between liquid and solid materials, the influence of solid surface charge density, wettability, and liquid properties on contact electrification, and highlights the progress in improving the hydrophobicity of solid materials, surface charge density, and triboelectric properties of liquid materials through chemical functionalization. The significance of enhanced liquid-solid contact electrification in energy harvesting, self-powered sensors, and metal corrosion protection is emphasized.
Article
Chemistry, Physical
Han Wu et al.
Summary: This work presents a self-powered sensor network based on TENG technology for aeolian vibration monitoring in overhead transmission lines. The study enhances the vibration sensing capability by adjusting structural parameters and weight, and successfully applies it to practical warning systems and vibration distribution monitoring systems.
ADVANCED ENERGY MATERIALS
(2022)
Article
Chemistry, Physical
Jian Yu et al.
Summary: The article discusses a potential method for real-time monitoring of sediment particles parameters using PLDD-TENG combined with deep learning, achieving high identifying accuracy with a CNN deep learning model. This provides a new approach for monitoring suspended sediment particles.
Article
Materials Science, Multidisciplinary
Jian Jiang et al.
Summary: This paper combines molecular dynamic simulation and machine learning techniques to predict the mechanical properties of graphene reinforced aluminium nanocomposites and modify the Halpin-Tsai model. The study considers the effects of graphene's volume fraction, alignment angle, chirality, and environment temperature on the properties. Machine learning models are developed and optimized using the simulation data to estimate the Young's modulus and ultimate tensile strength. The modified Halpin-Tsai model provides significantly improved accuracy for determining the Young's modulus of graphene-reinforced aluminium nanocomposite structures.
MATERIALS & DESIGN
(2022)
Article
Chemistry, Multidisciplinary
Chao Ye et al.
Summary: In this study, an IntelliSense yarn was developed to sense and identify different materials in real-time. By integrating the yarn into an IntelliSense system using Internet of Things techniques, it can be used to recognize and control various electronic and electrical systems, showing great potential in wearable energy supply, IntelliSense fabrics, and human-machine interactions.
Article
Chemistry, Multidisciplinary
Kaveh Barri et al.
Summary: This paper discusses the unmet need for smart medical implants and proposes the concept of self-aware implants to create a new generation of multifunctional metamaterial implantable devices. The integration of nano energy harvesting and mechanical metamaterial design paradigms enables these implants to respond to their environment, empower themselves, and self-monitor their condition.
ADVANCED FUNCTIONAL MATERIALS
(2022)
Article
Engineering, Multidisciplinary
Jeremy Yu et al.
Summary: The article introduces a new method called gradient-enhanced physics-informed neural networks (gPINNs) to improve the accuracy of PINNs. gPINNs leverage the gradient information of the PDE residual to enhance the loss function. Experimental results demonstrate that gPINNs outperform traditional PINN with fewer training points.
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
(2022)
Article
Engineering, Multidisciplinary
Somdatta Goswami et al.
Summary: This study proposes a physics-informed variational formulation of DeepONet (V-DeepONet) for brittle fracture analysis. By incorporating physics laws and labeled data in training, V-DeepONet accurately predicts key quantities in brittle fracture and has wide application potential.
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
(2022)
Article
Chemistry, Physical
Min Jiang et al.
Summary: This research introduces an AI algorithm model based on deep neural networks for TENGs, accurately predicting the output performance of TENGs under various structures and conditions. The results were consistent with physical experimental data, providing a wider parameter range scale for researchers to analyze the data and obtain a better experimental law. Additionally, the DNN model was also used to predict the output performance of sliding mode TENG under various load conditions.
Article
Automation & Control Systems
Ahed Habib et al.
Summary: Base isolation systems have made significant progress in recent decades, and a direct design approach relying on a physics-informed neural network model has been proposed to optimize the design process. The study used a large dataset and compared the performance of the proposed model to other approaches, showing that it achieved high accuracy and provided an effective way to rapidly design quintuple friction pendulum isolators.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2022)
Article
Chemistry, Physical
Yingshuang Shang et al.
Summary: Triboelectric nanogenerator (TENG) has attracted much interest due to its efficiency, flexibility and cost-effectiveness. This study investigates the relationship between the chemical composition of triboelectric materials and their output performance in TENGs, and reveals that the structure-dependent performance is closely related to the potential energy difference between the energy levels of the positive and negative materials.
Review
Chemistry, Physical
Yang Lyu et al.
Summary: Biodegradable triboelectric nanogenerators (BD-TENGs) have gained attention in the field of green energy due to their eco-friendliness, safety, and biocompatibility. However, their relatively low output performance remains a challenge. This review summarizes the existing optimizing methods of BD-TENGs, which include altering material compositions, changing surface microstructures, and improving dielectric properties.
Article
Mathematics, Applied
Salvatore Cuomo et al.
Summary: Physics-Informed Neural Networks (PINN) are a type of neural network that incorporates model equations, such as partial differential equations, as a component. PINNs have been used to solve various types of equations, including fractional equations and stochastic partial differential equations. Current research focuses on optimizing PINN through different aspects, such as activation functions, gradient optimization techniques, neural network structures, and loss function structures. Despite the demonstrated feasibility of PINN in certain cases compared to traditional numerical techniques, there are still unresolved theoretical issues.
JOURNAL OF SCIENTIFIC COMPUTING
(2022)
Article
Multidisciplinary Sciences
Chuanfu Xin et al.
Summary: The study proposes a novel energy harvesting device, CSF-TENG, which can act as a self-powered sensor and detect its own damage. Experimental results demonstrate a high accuracy of the defect identification model for the device.
Article
Chemistry, Physical
Ziyuan Jiang et al.
Summary: A ribbon-cage-based triboelectric bearing (RTB) is proposed for the fault diagnosis of rotating machinery. The RTB utilizes existing structures of rolling bearings and adds an insulating film and electrodes to form a compact triboelectric nanogenerator. The fabricated RTB prototype demonstrates self-sensing and self-powering capabilities and is used in fault diagnosis of rotating machinery. Results show that the RTB output current can accurately diagnose faults with a classification accuracy exceeding 90%.
Review
Chemistry, Multidisciplinary
Yuan Bai et al.
Summary: Triboelectric nanogenerators (TENGs), as an emerging energy-harvesting technology, have made rapid progress in the past decade. Alongside the well-known self-powering behavior, TENGs also have the unique feature of high-voltage and low-current output, providing opportunities for the development of safe high-voltage applications.
CELL REPORTS PHYSICAL SCIENCE
(2022)
Article
Chemistry, Multidisciplinary
Bolei Deng et al.
Summary: This study utilizes mechanical metamaterials based on hinged quadrilaterals to achieve target nonlinear mechanical responses. By changing the shape of the quadrilaterals, the amount of internal rotations can be adjusted, leading to a wide range of mechanical responses. Furthermore, a neural network and evolution strategy are introduced to efficiently design structures with desired mechanical properties.
ADVANCED MATERIALS
(2022)
Article
Engineering, Electrical & Electronic
Rahul Nellikkath et al.
Summary: This paper introduces physics-informed neural networks for accurately estimating the AC-Optimal Power Flow (AC-OPF) result and providing rigorous guarantees about their performance. Experimental results show that physics-informed neural networks achieve higher accuracy and lower constraint violations compared to standard neural networks.
ELECTRIC POWER SYSTEMS RESEARCH
(2022)
Article
Engineering, Electrical & Electronic
Steven de Jongh et al.
Summary: This paper applies geometric deep learning techniques to learn approximate models for power system estimation and calculation tasks, comparing nine different graph neural network architectures. By considering the underlying graph and known physical algebraic equations, the learned models are able to estimate system states with high accuracy and can be applied to previously unseen grid topologies.
ELECTRIC POWER SYSTEMS RESEARCH
(2022)
Review
Chemistry, Multidisciplinary
Huijing Xiang et al.
Summary: This review discusses the potential of triboelectric nanogenerators (TENGs) in alleviating the energy shortage crisis and facilitating self-powered sensors. It highlights the importance of triboelectric interfaces with functionalized molecular groups and tunable surface charge densities for improving the electrical output capability of TENGs and the versatility of future electronics. Recent advances in multifunctional triboelectric sensing are briefly introduced, and future challenges and chemical perspectives in the field of TENG-based electronics are discussed.
Article
Chemistry, Physical
Mikkel L. Bodker et al.
Summary: This study introduces a novel approach combining statistical mechanics and machine learning to accurately predict the structure of glasses, improving both interpolation and extrapolation abilities.
NPJ COMPUTATIONAL MATERIALS
(2022)
Article
Energy & Fuels
Daniel Vazquez Pombo et al.
Summary: This paper reviews and compares the most relevant machine learning methods and statistical methods identified in the literature for solar photovoltaic (PV) power system forecasting. The authors propose a methodology that integrates a PV-performance model and physics-informed feature selection, and validate it through a case study in Denmark. The results show that the best machine learning models consistently utilize physics-informed features.
Article
Materials Science, Multidisciplinary
Z. Zhang et al.
Summary: This study presents a multi-physical field coupling model for an acoustic driven TENG, and systematically simulates the energy conversion process using the finite element method. It provides theoretical guidelines and optimization strategies for the design and optimization of TENG acoustic energy harvesting systems.
MATERIALS TODAY PHYSICS
(2022)
Article
Engineering, Manufacturing
B. V. S. S. Bharadwaja et al.
Summary: A model based on Physics-Informed Neural Networks (PINN) is used to solve the elastic deformation of heterogeneous solids and associated uncertainty quantification. The study shows that PINN can accurately capture stress jumps at material interfaces, and the introduction of surrogate features in PINN leads to better solutions. The uncertainty quantification results are in good agreement with traditional finite element methods, demonstrating the effectiveness of PINN for solving elastic deformation of heterogeneous solids.
INTEGRATING MATERIALS AND MANUFACTURING INNOVATION
(2022)
Article
Chemistry, Multidisciplinary
Da Zhao et al.
Summary: This paper proposes a quantitative standard based on the internal equivalent circuit to evaluate the effects of different factors on the outputs of TENGs. It establishes the relationship between the equivalent circuits and TENGs and explores the influence of mechanical structures and external excitations on output power and matching reactance. Experimental results show the order of influencing factors and the difference in frequency direction for TENGs.
ENERGY & ENVIRONMENTAL SCIENCE
(2022)
Article
Physics, Multidisciplinary
Lu Lu et al.
Summary: This paper proposes a multifidelity neural operator based on deep neural networks, which can reduce the demand for high-fidelity data and achieve smaller errors in solving heat transport problems. By combining with genetic algorithms and topology optimization, it enables fast solvers and inverse design for the phonon Boltzmann transport equation (BTE).
PHYSICAL REVIEW RESEARCH
(2022)
Article
Green & Sustainable Science & Technology
Chutian Wu et al.
Summary: Wind farm layout optimization is crucial in wind energy project design, with physical understanding and different mesh approaches used to determine optimal turbine positions. Staggered arrangement is more efficient in extracting energy from wind compared to aligned arrangement. Different mesh approaches show varying performance under different wind conditions, and the sunflower method generally performs well overall.
Article
Computer Science, Artificial Intelligence
Enrico Schiassi et al.
Summary: X-TFC is a physics-informed neural network method that combines the Theory of Functional Connections and Neural Networks to solve differential equation problems accurately and quickly. It achieves high accuracy with low computational time, even for large scale problems, by using a single-layer neural network trained via the Extreme Learning Machine algorithm.
Article
Energy & Fuels
Weijia Liu et al.
Summary: This study developed a hierarchy of four new physics-informed persistence models to simultaneously forecast global horizontal irradiance, direct normal irradiance, and diffuse horizontal irradiance. The new models generally outperformed simple and smart persistence models, improving forecast accuracy from 1.25 h to 6 h lead times, with forecast errors highly related to the error and temporal variability of the assumed cloud predictor. The best model for forecasting different radiative components can be explained by the relationship between solar irradiances and cloud properties.
Article
Chemistry, Physical
Jiajia Shao et al.
Summary: The distribution of electric field and energy dynamics of triboelectric nanogenerators (TENGs) are determined using 3D mathematical modeling. The stored electrical energy and output efficiency of TENGs are quantitatively calculated, and guidelines for optimization are discussed based on device parameters, geometry, and optimum conditions. Adjustments in gap distance between neighboring TENG devices improve the performance of TENG arrays, with universal design rules and holistic optimization strategies presented for the network structure of TENGs for the first time.
ADVANCED ENERGY MATERIALS
(2021)
Article
Multidisciplinary Sciences
Aditya Shekhar Nittala et al.
Summary: This research presents a computational approach for designing multi-modal electro-physiological sensors that optimize the electrode layout design, achieving an optimal balance between high signal quality and small device size.
NATURE COMMUNICATIONS
(2021)
Article
Engineering, Electrical & Electronic
Yunsik Ohm et al.
Summary: The silver-hydrogel composite exhibits a high electrical conductivity of over 350 S cm(-1) and a low Young's modulus of less than 10 kPa. It demonstrates soft compliance and deformability, making it suitable for various applications such as stingray-inspired swimmers and neuromuscular electrical stimulation electrodes.
NATURE ELECTRONICS
(2021)
Article
Chemistry, Physical
Yiqun Wang et al.
Summary: This study utilized the support vector regression algorithm and other numerical analysis methods to simulate cylindrical grating-structured TENGs and investigated the influence of structural parameters on performance, offering new insights for the optimization of TENG structures.
Review
Chemistry, Physical
Pengcheng Jiao
Summary: Piezoelectric nanogenerators and triboelectric nanogenerators are paving the way for sustainable energy harvesting, with recent advancements incorporating artificial intelligence to optimize their mechanical-to-electrical performance. This integration of AI in nanogenerators shows promising potential for future innovations in green energy solutions.
Article
Chemistry, Multidisciplinary
Zongrui Pei et al.
Summary: With the emergence of new alloys like high-entropy alloys, the traditional trial-and-error method faces challenges in alloy design; assisted by machine-learning method, microstructure images can be identified and a new neural network method proposed for inverse alloy design; this work lays the foundation for inverse alloy design based on microstructure images.
Article
Chemistry, Physical
Victor Fung et al.
Summary: The study introduces an inverse design framework utilizing invertible neural networks to map between design space and target properties for generating materials candidates with desired properties. The approach demonstrates the capability to generate novel, high fidelity, and diverse materials candidates in MoS2, and can be extended to other materials and their corresponding design spaces and target properties.
NPJ COMPUTATIONAL MATERIALS
(2021)
Article
Acoustics
Zhilu Lai et al.
Summary: This paper introduces a new approach of using Physics-informed Neural ODEs for structural identification, which enhances adaptability and flexibility through a two-level representation involving physics information and discrepancy terms, addressing complex issues in structural identification.
JOURNAL OF SOUND AND VIBRATION
(2021)
Article
Chemistry, Physical
Ruoxing Wang et al.
Summary: The study presents a systematic, data-driven learning of the process-property-performance relation in ZnO nanowires piezoelectric nanogenerators, establishing a behavioral model to reveal the connections between production parameters and output performance. The optimized ZnO nanowires piezoelectric nanogenerator was integrated with a photosensor into a self-powered sensor system, demonstrating potential for future system-level improvements.
Article
Multidisciplinary Sciences
Jianxiong Zhu et al.
Summary: The study presents a machine learning-enhanced ion mobility analyzer with a triboelectric-based ionizer, which provides effective identification and measurement of VOCs with specific design. The device is compact, easy to operate, and features real-time response and low power consumption, suitable for future IoT environmental monitoring applications.
Article
Chemistry, Multidisciplinary
Qianyun Zhang et al.
Summary: This study proposes a new concept of structural elements (TENG-SE) that can self-empower and self-monitor in civil infrastructure systems. By developing a proof-of-concept multifunctional composite rebars with TENG mechanisms and conducting empirical and theoretical studies to verify its electrical and mechanical performance, the capability of embedded structural elements to detect damage at multiscale is demonstrated.
ADVANCED FUNCTIONAL MATERIALS
(2021)
Article
Engineering, Multidisciplinary
Ehsan Haghighat et al.
Summary: This study presents the application of Physics Informed Neural Networks (PINN) in solid mechanics, improving accuracy and convergence with a multi-network model and Isogeometric Analysis. The study demonstrates the importance of honoring physics in improving robustness and highlights the potential application of PINN in sensitivity analysis and surrogate modeling.
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
(2021)
Review
Chemistry, Physical
Yuankai Zhou et al.
Summary: The sensor based on the triboelectric nanogenerator has excellent material compatibility, low cost, and high flexibility, making it a unique candidate technology for artificial intelligence. Triboelectric nanogenerators effectively provide the critical infrastructure for the new generation of sensing systems that collect information through a large number of self-powered sensors. The application of triboelectric nanogenerators in intelligent sports, security, touch control, and document management systems has attracted increasing attention in the field of artificial intelligence.
Article
Chemistry, Physical
Kaveh Barri et al.
Summary: Researchers have introduced a novel concept called self-aware composite mechanical metamaterial (SCMM) that can transform mechanical metamaterials into nanogenerators and active sensing mediums. By studying new paradigms, they have achieved contact electrification between snapping microstructures composed of topologically different triboelectric materials, leading to self-powering and self-sensing meta-tribomaterial systems.
Article
Materials Science, Multidisciplinary
Y. Du et al.
Summary: This study demonstrates a method combining physics-informed machine learning, mechanistic modeling, and experimental data to reduce common defects in additive manufacturing. By analyzing experimental data and using a physics-informed machine learning approach, a quantitative formalism for predicting defects in real-time is provided, helping to address issues in additive manufacturing processes.
APPLIED MATERIALS TODAY
(2021)
Article
Materials Science, Multidisciplinary
Zhili He et al.
Summary: A novel data-driven framework based on PINNs is proposed for direct analysis and parameter inversion of heat conduction problems, showing satisfactory accuracy and the potential to replace finite element modeling to some extent through case studies on wood and steel. The framework can effectively address both direct and inverse heat conduction problems and has broad application prospects in materials science.
MATERIALS TODAY COMMUNICATIONS
(2021)
Article
Mechanics
Zhizhou Zhang et al.
Summary: A physics-informed neural network (PINN) for digital materials analysis is introduced, trained without ground truth data using minimum energy criteria. The model reached similar accuracy as supervised ML models and prevented erroneous deformation gradients with hinge loss on the Jacobian. Parallel computing on GPU for strain energy calculation showed linear scalability with the number of nodes, laying the foundation for label-free learning in designing next-generation composites.
THEORETICAL AND APPLIED MECHANICS LETTERS
(2021)
Review
Physics, Applied
George Em Karniadakis et al.
Summary: Physics-informed learning seamlessly integrates data and mathematical models through neural networks or kernel-based regression networks for accurate inference of realistic and high-dimensional multiphysics problems. Challenges remain in incorporating noisy data seamlessly, complex mesh generation, and addressing high-dimensional problems.
NATURE REVIEWS PHYSICS
(2021)
Review
Chemistry, Multidisciplinary
Kai Guo et al.
Summary: Artificial intelligence, especially machine learning and deep learning algorithms, is increasingly utilized in materials and mechanical engineering for predicting materials properties and designing new materials. Trained ML models offer fast exploration of design spaces, but challenges remain in data collection, preprocessing and model selection. Recent breakthroughs in ML techniques have opened up vast opportunities in overcoming mechanics problems and developing novel materials design strategies.
MATERIALS HORIZONS
(2021)
Article
Engineering, Electrical & Electronic
Xingyu Lei et al.
Summary: This study presents a data-driven approach for OPF that decomposes OPF model features into three stages to reduce learning complexity and correct learning bias. An enhanced feature attraction is achieved through a sample pre-classification strategy based on active constraint identification.
IEEE TRANSACTIONS ON POWER SYSTEMS
(2021)
Article
Multidisciplinary Sciences
Teeratorn Kadeethum et al.
Review
Engineering, Multidisciplinary
Shao JiaJia et al.
SCIENCE CHINA-TECHNOLOGICAL SCIENCES
(2020)
Article
Chemistry, Physical
Tao Jiang et al.
ADVANCED ENERGY MATERIALS
(2020)
Article
Chemistry, Physical
Mohammad Khorsand et al.
Article
Optics
Yuyao Chen et al.
Review
Chemistry, Multidisciplinary
He Zhang et al.
NANOTECHNOLOGY REVIEWS
(2020)
Review
Multidisciplinary Sciences
K. R. Sanjaya D. Gunawardhana et al.
Article
Materials Science, Characterization & Testing
Khemraj Shukla et al.
JOURNAL OF NONDESTRUCTIVE EVALUATION
(2020)
Article
Engineering, Multidisciplinary
Ruiyang Zhang et al.
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
(2020)
Article
Chemistry, Physical
Yaqian Liu et al.
Review
Chemistry, Physical
Jianjun Luo et al.
Review
Chemistry, Physical
Renyun Zhang et al.
Article
Automation & Control Systems
Renato G. Nascimento et al.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2020)
Review
Nanoscience & Nanotechnology
Sunae So et al.
Article
Computer Science, Interdisciplinary Applications
M. Raissi et al.
JOURNAL OF COMPUTATIONAL PHYSICS
(2019)
Article
Chemistry, Physical
He Zhang et al.
Review
Chemistry, Physical
R. D. I. G. Dharmasena et al.
Review
Multidisciplinary Sciences
Matthew R. Begley et al.
Review
Engineering, Mechanical
Shuaihang Pan et al.
Article
Chemistry, Physical
Mohammad Khorsand et al.
Article
Materials Science, Multidisciplinary
Bao Dong Chen et al.
Article
Chemistry, Physical
R. D. Ishara G. Dharmasena et al.
ADVANCED ENERGY MATERIALS
(2018)
Article
Chemistry, Physical
Cheng Jiang et al.
Article
Chemistry, Physical
Kyungdoc Kim et al.
NPJ COMPUTATIONAL MATERIALS
(2018)
Article
Chemistry, Physical
Wanchul Seung et al.
ADVANCED ENERGY MATERIALS
(2017)
Article
Physics, Applied
Shuaihang Pan et al.
JOURNAL OF APPLIED PHYSICS
(2017)
Article
Chemistry, Physical
Yange Feng et al.
Article
Physics, Applied
Xiangyu Chen et al.
APPLIED PHYSICS LETTERS
(2017)
Article
Chemistry, Multidisciplinary
Nuanyang Cui et al.
Article
Chemistry, Multidisciplinary
Zhi-Min Dang et al.
Article
Chemistry, Multidisciplinary
Tao Jiang et al.
ADVANCED FUNCTIONAL MATERIALS
(2015)
Article
Engineering, Electrical & Electronic
Simiao Niu et al.
IEEE TRANSACTIONS ON ELECTRON DEVICES
(2015)
Article
Chemistry, Physical
Xiaofeng Wang et al.
ADVANCED ENERGY MATERIALS
(2015)
Article
Chemistry, Physical
Simiao Niu et al.
Article
Nanoscience & Nanotechnology
Suibin Luo et al.
ACS APPLIED MATERIALS & INTERFACES
(2014)
Article
Chemistry, Multidisciplinary
Simiao Niu et al.
ADVANCED FUNCTIONAL MATERIALS
(2014)
Article
Chemistry, Multidisciplinary
Simiao Niu et al.
ADVANCED MATERIALS
(2013)
Article
Chemistry, Physical
Feng-Ru Fan et al.
