<p>This study analyzed the particle collection characteristics of an electrostatic precipitator (ESP) based on three performance-influencing parameters: applied voltage, volumetric flow rate, and collection plate type. We employed the flat, tilt, half-tilt, and expand type collection plates. Compared with the flat type, the tilt and half-tilt type exhibited enhanced corona discharge. The half-tilt type exhibited a 9 % higher collection efficiency than the flat type at φ = 12 kV and Q = 0.15 m<sup>3</sup>/s. Additionally, the expand type exhibited the same electric-field trend as the flat type; however, it developed a separation region in the flow-velocity distribution. Consequently, its efficiency was higher under φ = 9 kV than that at φ = 12 kV because the separation region prevented the particles from attaching to the collection plates. Using the results, we developed an prediction model with a single hidden layer comprising 18 neurons. The predictions agreed well with the numerical simulation results, exhibiting errors within ±10 %.</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

Numerical study on the particle collection characteristics of electrostatic precipitators with different types of collection plates

  • Jeong Geun Gwon,
  • Do Hwan Kim,
  • Ho Yeon Choi,
  • Hoon Ki Choi,
  • Young Min Seo,
  • Seokho Kim,
  • Yong Gap Park

摘要

This study analyzed the particle collection characteristics of an electrostatic precipitator (ESP) based on three performance-influencing parameters: applied voltage, volumetric flow rate, and collection plate type. We employed the flat, tilt, half-tilt, and expand type collection plates. Compared with the flat type, the tilt and half-tilt type exhibited enhanced corona discharge. The half-tilt type exhibited a 9 % higher collection efficiency than the flat type at φ = 12 kV and Q = 0.15 m3/s. Additionally, the expand type exhibited the same electric-field trend as the flat type; however, it developed a separation region in the flow-velocity distribution. Consequently, its efficiency was higher under φ = 9 kV than that at φ = 12 kV because the separation region prevented the particles from attaching to the collection plates. Using the results, we developed an prediction model with a single hidden layer comprising 18 neurons. The predictions agreed well with the numerical simulation results, exhibiting errors within ±10 %.