Scholarlink

<p>In this paper, the poly lactic acid (PLA)/carbon fiber (CF) specimen was fabricated using fused deposition modeling (FDM) technique. The effects of various operation parameters (printing speed, nozzle temperature, layer thickness, and platform temperature) on the tensile strength of FDM-PLA/CF specimen were researched, and the three-layered backpropagation (BP) neural network model was built for tensile strength prediction. Furthermore, the BP neural network model was optimized by genetic algorithm (GA) and the GA-BP neural network model was established. Compared with BP neural network model, the average error between the prediction values and experimental results of the GA-BP neural network model was lower, which was only 0.84%. Meanwhile, the regression coefficient of BP and GA-BP neural network model was 0.86475 and 0.97146, respectively. In addition, the evaluation indexes of GA-BP neural network model were better than those of BP neural network model, which indicated it had high prediction accuracy. Therefore, the GA-BP neural network could provide process optimization and high precision prediction for the mechanical properties of FDM-PLA/CF specimens.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Tensile Strength Assessment of FDM-PLA/CF Specimens Fabricated Using Orthogonal Experimental Design and Artificial Neural Network Approach

  • 期刊论文

  • 发表日期：

    2025年8月21日

查看全文
  • Mengyu Cao,
  • Dehao Tian,
  • Xu Zhang,
  • Chaoyu Li,
  • Xue Guo,
  • Baojin Wang

摘要

In this paper, the poly lactic acid (PLA)/carbon fiber (CF) specimen was fabricated using fused deposition modeling (FDM) technique. The effects of various operation parameters (printing speed, nozzle temperature, layer thickness, and platform temperature) on the tensile strength of FDM-PLA/CF specimen were researched, and the three-layered backpropagation (BP) neural network model was built for tensile strength prediction. Furthermore, the BP neural network model was optimized by genetic algorithm (GA) and the GA-BP neural network model was established. Compared with BP neural network model, the average error between the prediction values and experimental results of the GA-BP neural network model was lower, which was only 0.84%. Meanwhile, the regression coefficient of BP and GA-BP neural network model was 0.86475 and 0.97146, respectively. In addition, the evaluation indexes of GA-BP neural network model were better than those of BP neural network model, which indicated it had high prediction accuracy. Therefore, the GA-BP neural network could provide process optimization and high precision prediction for the mechanical properties of FDM-PLA/CF specimens.