雕刻
曲面(拓扑)
激光器
材料科学
工程制图
计算机科学
光学
复合材料
工程类
几何学
数学
物理
作者
Somayeh Sohrabi,Zahra Dehghanian,Alireza Bayat,Mehdi Mohammadaghaie,Mojtaba Taghipoor,Ariana Ghorashi,Nasim Hassani,Hamid Reza Rabiee
摘要
Abstract This study presents a comparative evaluation of two predictive approaches for determining microgroove dimensions in laser machining. The first approach employs response surface methodology (RSM) regression models to predict microgroove width and depth using three input parameters: laser power (10–20 W), scanning rate (50–150 mm/s), and focus distance (6–8 mm). The second approach utilizes data‐driven machine learning (ML) and deep neural network (DNN) models, incorporating five input parameters: laser power, scanning rate, focus distance, laser pass number (1–3), and measurement location (edge and middle). A total of 350 microgrooves were analyzed, and results indicate that the DNN model achieved the highest prediction accuracy, with an R 2 value exceeding 0.94 for depth prediction and a mean absolute error of 10.96 μm on the training data. These findings demonstrate the potential of data‐driven models in improving the precision of laser machining predictions.
科研通智能强力驱动
Strongly Powered by AbleSci AI