Investigation on surface morphology model of Si₃N₄ ceramics for rotary ultrasonic grinding machining based on the neural network

详细信息    查看全文

• 作者：Juntao Jing^a ; Pingfa Feng^a ; Shiliang Wei^b ; ^{weishiliang@hrbeu.edu.cn} ; ^{weishi1264326@126.com} ; Hong Zhao^c
• 关键词：Surface morphology ; Rotary ultrasonic grinding machining ; Back propagation neural network ; Improved algorithm
• 刊名：Applied Surface Science
• 出版年：2017
• 出版时间：28 February 2017
• 年：2017
• 卷：396
• 期：Complete
• 页码：85-94
• 全文大小：1546 K
• 卷排序：396

文摘

The BP neural network model between the surface morphology and process parameters has been established in the paper. An improved BP algorithm based on the Powell method is proposed to calculate the weights and thresholds. By comparing with the LM algorithm, the improved BP algorithm based on the Powell method is higher in convergence rate and prediction precision.