Milling Surface Roughness Prediction Based on Physics-Informed Machine Learning

Sensors, Год журнала: 2023, Номер 23(10), С. 4969 - 4969

Опубликована: Май 22, 2023

Surface roughness is a key indicator of the quality mechanical products, which can precisely portray fatigue strength, wear resistance, surface hardness and other properties products. The convergence current machine-learning-based prediction methods to local minima may lead poor model generalization or results that violate existing physical laws. Therefore, this paper combined knowledge with deep learning propose physics-informed method (PIDL) for milling predictions under constraints This introduced in input phase training learning. Data augmentation was performed on limited experimental data by constructing mechanism models tolerable accuracy prior training. In training, physically guided loss function constructed guide process knowledge. Considering excellent feature extraction capability convolutional neural networks (CNNs) gated recurrent units (GRUs) spatial temporal scales, CNN-GRU adopted as main predictions. Meanwhile, bi-directional unit multi-headed self-attentive were enhance correlation. paper, experiments conducted open-source datasets S45C GAMHE 5.0. comparison state-of-the-art methods, proposed has highest both datasets, mean absolute percentage error test set reduced 3.029% average compared best method. Physical-model-guided machine be future pathway evolution.

Язык: Английский

---

A State of the Art on Surface Texture Creation Modelling Methods in Machining

Archives of Computational Methods in Engineering, Год журнала: 2025, Номер unknown

Опубликована: Фев. 6, 2025

Язык: Английский

---

Research on the Characteristics of Deformation and Axial Force Changes During Drilling of Thin-Walled AF1410 High-Strength Steel

Applied Sciences, Год журнала: 2025, Номер 15(7), С. 3481 - 3481

Опубликована: Март 22, 2025

Axial force and deformation during drilling significantly impact the hole quality of thin-walled high-strength steel components. This study analyzed process AF1410 steel, focusing on axial force, deformation, drill cap formation, exit edge characteristics. The effects cutting speed (12.6–37.7 m/min) feed rate (0.01–0.1 mm/r) were also examined. Initially, plate undergoes elastic, outward bulging deformation. driven by elastic resistance, rises from 114.9 N to 322.1 as increases 0.025 mm/r 0.1 mm/r, with minimal influence speed. As progresses, slowly. Near exit, plastic occurs beneath bit, causing material yield form a cap. results in sharp rise maximum values increasing 314.2 525.3 at higher speeds 840.1 rates. formation characteristics directly affect defects, larger thickness width leading more pronounced burrs.

Язык: Английский

---

Mechanisms of strength–plasticity enhancement and stress-induced phase transition in a medium-carbon low-alloy steel

Journal of Iron and Steel Research International, Год журнала: 2024, Номер 31(9), С. 2255 - 2270

Опубликована: Фев. 4, 2024

Язык: Английский

---

Modeling and simulation of Surface Topography in secondary milling with ellipsoid end milling cutter

International Journal on Interactive Design and Manufacturing (IJIDeM), Год журнала: 2023, Номер 18(7), С. 5043 - 5056

Опубликована: Авг. 28, 2023

Язык: Английский

---

A Literature Review on Prediction Methods for Forced Responses and Associated Surface Form/Location Errors in Milling

Journal of Vibration Engineering & Technologies, Год журнала: 2023, Номер 12(4), С. 5905 - 5934

Опубликована: Дек. 18, 2023

Язык: Английский

---

Surface topography prediction and reliability analysis of ball end milling

Measurement, Год журнала: 2024, Номер unknown, С. 115770 - 115770

Опубликована: Сен. 1, 2024

Язык: Английский

---

Tool wear evolution and its influence on cutting performance during milling ultra-high-strength steel using different cooling conditions

The International Journal of Advanced Manufacturing Technology, Год журнала: 2024, Номер unknown

Опубликована: Окт. 1, 2024

Язык: Английский

---