Scholarlink

<p>CNC wire electrical discharge machines are used in industries such as the molding, aerospace, and medical device industries. However, multiple parameters need to be adjusted during wire electrical discharge machining, such as the external parameters of mechanical workpieces and discharge circuit parameters, which affect the machining speed and geometric accuracy. In this study, the issues of multiple factors and the number of levels are addressed. Experiments on the machining parameters were designed using the uniform design method, which is combined with a backpropagation neural network (BPNN) for prediction and with the genetic algorithm (GA) for optimization. The objective functions are the machining speed and geometric accuracy functions. The maximum machining speeds that correspond to electrode wire diameters of 0.2&#xa0;mm, 0.25&#xa0;mm, and 0.30&#xa0;mm are 125 mm<sup>2</sup>/min, 144 mm<sup>2</sup>/min, and 168 mm<sup>2</sup>/min, respectively. The BPNN prediction and GA optimization system has a mean absolute percentage error (MAPE) of less than 20%. In the five-factor test, the uniform design reduces the number of experimental runs by 80% relative to the orthogonal design. Therefore, this study demonstrates highly satisfactory prediction and optimization results.</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

Application of uniform design and the genetic algorithm in parameter design for wire electrical discharge machining

  • 期刊论文

  • 发表日期：

    2025年11月7日

查看全文
  • Shao-Hsien Chen,
  • Wun-Syuan Huamg

摘要

CNC wire electrical discharge machines are used in industries such as the molding, aerospace, and medical device industries. However, multiple parameters need to be adjusted during wire electrical discharge machining, such as the external parameters of mechanical workpieces and discharge circuit parameters, which affect the machining speed and geometric accuracy. In this study, the issues of multiple factors and the number of levels are addressed. Experiments on the machining parameters were designed using the uniform design method, which is combined with a backpropagation neural network (BPNN) for prediction and with the genetic algorithm (GA) for optimization. The objective functions are the machining speed and geometric accuracy functions. The maximum machining speeds that correspond to electrode wire diameters of 0.2 mm, 0.25 mm, and 0.30 mm are 125 mm2/min, 144 mm2/min, and 168 mm2/min, respectively. The BPNN prediction and GA optimization system has a mean absolute percentage error (MAPE) of less than 20%. In the five-factor test, the uniform design reduces the number of experimental runs by 80% relative to the orthogonal design. Therefore, this study demonstrates highly satisfactory prediction and optimization results.