From PINNs to PIKANs: recent advances in physics-informed machine learning

Machine learning for computational science and engineering, Journal Year: 2025, Volume and Issue: 1(1)

Published: March 11, 2025

Language: Английский

---

Application of physics-informed neural networks in fault diagnosis and fault-tolerant control design for electric vehicles: A review

Measurement, Journal Year: 2025, Volume and Issue: unknown, P. 116728 - 116728

Published: Jan. 1, 2025

Language: Английский

---

Surface Wettability Modeling and Predicting via Artificial Neural Networks

Materials, Journal Year: 2025, Volume and Issue: 18(1), P. 191 - 191

Published: Jan. 5, 2025

Surface wettability, defined by the contact angle, describes ability of a liquid to spread over, absorb or adhere solid surface. wetting analysis is important in many applications, such as lubrication, heat transfer, painting and wherever liquids interact with surfaces. The behavior on surfaces depends mainly texture chemical properties Therefore, these studies show possibility modeling surface wettability adjusting parameters texturing process. prediction angle describing was performed using artificial neural networks. In order select most effective model, activation functions neurons, number hidden layers network training algorithms were changed. model presented capable predicting an efficiency coefficient determination R2 between real predicted angles over 0.9.

Language: Английский

---

A Quasilinear Algorithm for Computing Higher-Order Derivatives of Deep Feed-Forward Neural Networks

Published: Jan. 10, 2025

The use of neural networks for solving differential equations is practically difficult due to the exponentially increasing runtime autodifferentiation when computing high-order derivatives. We propose \(n\)-TANGENTPROP , natural extension TANGENTPROP formalism[1] arbitrarily many computes exact derivative \({d^{n}/d}x^{n}f{(x)}\) in quasilinear, instead exponential time, a densely connected, feed-forward network \(f\) with smooth, parameter-free activation function. validate our algorithm empirically across range depths, widths, and number demonstrate that method particularly beneficial context physics-informed where allows significantly faster training times than previous methods has favorable scaling respect both model size loss-function complexity as measured by required code this paper can be found at https://github.com/kyrochi/n_tangentprop. [https://github.com/kyrochi/n_tangentprop]

Language: Английский

---

Dynamic modeling of a novel catamaran robotic system actuated by bio-inspired propulsion devices using physics-informed neural networks

Ocean Engineering, Journal Year: 2025, Volume and Issue: 326, P. 120806 - 120806

Published: March 10, 2025

Language: Английский

---

Exploiting Artificial Neural Networks for the State of Charge Estimation in EV/HV Battery Systems: A Review

Batteries, Journal Year: 2025, Volume and Issue: 11(3), P. 107 - 107

Published: March 13, 2025

Artificial Neural Networks (ANNs) improve battery management in electric vehicles (EVs) by enhancing the safety, durability, and reliability of electrochemical batteries, particularly through improvements State Charge (SOC) estimation. EV batteries operate under demanding conditions, which can affect performance and, extreme cases, lead to critical failures such as thermal runaway—an exothermic chain reaction that may result overheating, fires, even explosions. Addressing these risks requires advanced diagnostic strategies, machine learning presents a powerful solution due its ability adapt across multiple facets management. The versatility ML enables application material discovery, model development, quality control, real-time monitoring, charge optimization, fault detection, positioning it an essential technology for modern systems. Specifically, ANN models excel at detecting subtle, complex patterns reflect health performance, crucial accurate SOC effectiveness applications this domain, however, is highly dependent on selection datasets, relevant features, suitable algorithms. Advanced techniques active are being explored enhance improving models’ responsiveness diverse nuanced behavior. This compact survey consolidates recent advances estimation, analyzing current state field highlighting challenges opportunities remain. By structuring insights from extensive literature, paper aims establish ANNs foundational tool next-generation systems, ultimately supporting safer more efficient EVs robust safety protocols. Future research directions include refining dataset quality, optimizing algorithm selection, precision, thereby broadening ANNs’ role ensuring reliable vehicles.

Language: Английский

---

Harnessing dynamic graph differential operators for efficient data-driven wind prediction

GeoInformatica, Journal Year: 2025, Volume and Issue: unknown

Published: March 19, 2025

Language: Английский

---

Torque tracking position control of DLR-HIT II robotic hand using a real-time Physics-informed neural network

Applied Mathematical Modelling, Journal Year: 2025, Volume and Issue: unknown, P. 116110 - 116110

Published: March 1, 2025

Language: Английский

---

Incorporating Physical Constraints inside Neural Networks to Improve their Accuracy and Physical Reliability for Chemical Engineering Unit Operations Modeling

Computers & Chemical Engineering, Journal Year: 2025, Volume and Issue: unknown, P. 109156 - 109156

Published: April 1, 2025

Language: Английский

---

Physics-Informed Neural Networks in Polymers: A Review

Polymers, Journal Year: 2025, Volume and Issue: 17(8), P. 1108 - 1108

Published: April 19, 2025

The modeling and simulation of polymer systems present unique challenges due to their intrinsic complexity multi-scale behavior. Traditional computational methods, while effective, often struggle balance accuracy with efficiency, especially when bridging the atomistic macroscopic scales. Recently, physics-informed neural networks (PINNs) have emerged as a promising tool that integrates data-driven learning governing physical laws system. This review discusses development application PINNs in context science. It summarizes recent advances, outlines key methodologies, analyzes benefits limitations using for property prediction, structural design, process optimization. Finally, it identifies current future research directions further leverage advanced modeling.

Language: Английский

---