.png){kind=icon}
.png){kind=icon}
Return
Machine Learning-Assisted Study of Bond Coat Deposition via Spot Spray in Internal Diameter APS
.png){kind=icon}
PRE
AI
B
W
Q
Q
DOI：10.1007/s11666-026-02288-1.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Internal diameter atmospheric plasma spraying (ID-APS) is an effective technique for coating confined and curved internal surfaces, especially within the narrow internal passages of aerospace, energy, and automotive components where conventional thermal spray methods are unsuitable. However, the specialized structural constraints of the internal diameter torch nozzle lead to intricate plasma-substrate interactions. Compared to external surface thermal spraying, research on ID-APS remains limited, and this inherent complexity impedes systematic modeling and quantitative characterization. In this study, an internal diameter plasma torch was employed for spot spraying to extract key bond coat spray spot properties, including deposition efficiency and the geometric profiles characterized by maximum height and full width at half maximum. Subsequently, statistical association analysis was performed to examine the relationships between the input spray parameters and the spot properties. Utilizing a dataset derived from a systematic experimental design, machine learning-based regression models were established to accurately predict these spot properties in ID-APS, and an attribution analysis was conducted to identify the most influential parameters and clarify their specific impacts on bond coat deposition behavior. This study establishes a systematic, data-driven framework for ID-APS that elucidates the relationships between spray parameters and the resulting spot properties, enables predictive modeling, and offers a generalizable methodology for analyzing governing factors in powder deposition and related coating processes.
Keywords:
attribution analysis
data-driven framework
internal diameter atmospheric plasma spraying (ID-APS)
regression models
Journal
IF:
3.3
Papers:
3.7K
Citations:
7.6K
Organization
Cited Papers