基于混合遺傳蟻群優(yōu)化隨機森林算法的激光熔覆Ni60裂紋預(yù)測與工藝參數(shù)優(yōu)化

打開文本圖片集

中圖分類號：TG174.4；TP391.7DOI：10.3969/j.issn.1004-132X.2025.06.019 開放科學(xué)（資源服務(wù)）標(biāo)識碼（OSID）：

Prediction of Cracks and Optimization of Processing Parameters in Laser Cladding of Ni60 Based on HGA-ACO-RFA

LI Tao DENG Linhui*MO Bin SHI Feifan LIU Weiwei School of Mechanical Engineering，Dalian University of Technology，Dalian，Liaoning，116024

Abstract： To explore the complex nonlinear mapping relationship between the cracks in the cladding layer and the processing parameters during laser cladding of Ni6o，the entropy method combined with TOPSIS comprehensive evaluation method was used to comprehensively characterize and evaluate the cracks in the cladding layers. The HGA-ACO was used to optimize the hyperparameters of the RFA，and a prediction model between processing parameters and crack evaluation indicators was constructed. Finally，the genetic algorithm was used for reverse optimization of processing parameters. Results show that compared with the ACO-RFA model， HGA-ACO-RFA significantly improves prediction accuracy and evaluation indicators，and the optimal processng parameters obtained through reverse optimization may prepare almost crack free cladding layers.

Key words： laser cladding; crack;evaluation method; hybrid genetic ant colony algorithm（HGAACO）；random forest algorithm（RFA）

0 引言

激光熔覆技術(shù)以高能激光束為熱源，以惰性氣體為保護氣，將粉末、絲材等金屬或陶瓷材料與基體表面同時快速熔凝，形成一層牢固的冶金結(jié)合涂層。（剩余11019字）