What does this research mean for the field?

A physics-informed cascaded-learning framework accurately predicts clad geometry, thermal disturbance, and morphological quality in laser cladding repair of carburized gear steels, enabling effective process optimization with small sample sizes. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

To establish a physics-informed cascaded-learning framework for predicting geometry, thermal boundary depth, and clad quality during the laser cladding repair of carburized gear steels.

September 8, 2026MaterialsOpen Access

Physics-Informed Cascaded Learning for Predicting and Optimizing Geometry and Quality in Laser Cladding Repair of Carburized Gear Steel

View Full Paper

Ask AI

Bookmark

Authors

YXYingjie XuZhejiang University of Technology PZPeng ZhengShanghai University ZLZhongming LiuMaterials Technology (United Kingdom)

Discussion

Member takes

Overview

Computational modeling study demonstrates accurate geometry and quality prediction in carburized gear steel laser repair, indicating viable optimization with small sample sizes.

Key Points

To establish a physics-informed cascaded-learning framework for predicting geometry, thermal boundary depth, and clad quality during the laser cladding repair of carburized gear steels.
Deposited 15 experimental tracks of NHT.22.A01 iron-based powder onto carburized-quenched 18CrNiMo7-6 steel across varied laser powers, scanning speeds, and powder feed rates.
Implemented a cascaded machine learning framework using leave-one-out out-of-fold width predictions as inputs for height and depth models to eliminate target leakage.
Conducted bi-objective optimization of an auxiliary continuous quality index and high-hardness layer depth coupled with local sensitivity-resolved process mapping.
Leave-one-out R2 values for track width, height, and Ac1-boundary depth ranged from 0.934 to 0.968, yielding a maximum relative error of 4.3% at an unseen boundary condition.
The empirical heat-affected zone depth to track width scaling coefficient was 0.211, falling within the theoretical Rosenthal range of 0.15 to 0.25.
Bi-objective optimization identified a trade-off between cladding quality and high-hardness layer depth, with outcomes validated by an independent batch replicate.

Cite This Study

Xu et al. (2026) studied this question.

synapsesocial.com/papers/6a9fd79b58e84d0ff5b4670d https://doi.org/10.3390/ma19173793

View Full Paper

Ask AI

Bookmark

September 8, 2026MaterialsOpen Access

Physics-Informed Cascaded Learning for Predicting and Optimizing Geometry and Quality in Laser Cladding Repair of Carburized Gear Steel

View Full Paper

Ask AI

Bookmark

Authors

YXYingjie XuZhejiang University of Technology PZPeng ZhengShanghai University ZLZhongming LiuMaterials Technology (United Kingdom)

Discussion

Member takes

Overview

Computational modeling study demonstrates accurate geometry and quality prediction in carburized gear steel laser repair, indicating viable optimization with small sample sizes.

Key Points

To establish a physics-informed cascaded-learning framework for predicting geometry, thermal boundary depth, and clad quality during the laser cladding repair of carburized gear steels.
Deposited 15 experimental tracks of NHT.22.A01 iron-based powder onto carburized-quenched 18CrNiMo7-6 steel across varied laser powers, scanning speeds, and powder feed rates.
Implemented a cascaded machine learning framework using leave-one-out out-of-fold width predictions as inputs for height and depth models to eliminate target leakage.
Conducted bi-objective optimization of an auxiliary continuous quality index and high-hardness layer depth coupled with local sensitivity-resolved process mapping.
Leave-one-out R2 values for track width, height, and Ac1-boundary depth ranged from 0.934 to 0.968, yielding a maximum relative error of 4.3% at an unseen boundary condition.
The empirical heat-affected zone depth to track width scaling coefficient was 0.211, falling within the theoretical Rosenthal range of 0.15 to 0.25.
Bi-objective optimization identified a trade-off between cladding quality and high-hardness layer depth, with outcomes validated by an independent batch replicate.

Cite This Study

Xu et al. (2026) studied this question.

synapsesocial.com/papers/6a9fd79b58e84d0ff5b4670d https://doi.org/10.3390/ma19173793

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark