Scholarlink

The article conducts simulation and optimisation of the inlet flow path of the block of electronically controlled multi-way valves based on Fluent. The flow field characteristics of the flow path of the electronically controlled multi-way valve block are analysed by simulation, and the pressure loss of the valve block mainly occurs at the sudden change of the flow path structure and the outlet area of the fluid cyclone by combining the pressure field and velocity field distributions in the flow path of the valve block. Based on the simulation results, the optimisation objective is determined, and the Box-Behnken method is used to design a 5-factor, 3-level experimental scheme to evaluate the effects of different factors on the pressure and flow state of the flow path and optimise them. A BP neural network model is set up to fit the simulation parameters to obtain a better optimisation model, and a more optimal path structure is obtained by combining with a multi-objective genetic algorithm. The optimised path structure reduces the path pressure loss by 45.18%, while the turbulence intensity is effectively reduced by about 14.85%.

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

Flow Path Optimization of Valve Blocks by Box-Behnken Response Surface Method Combined with BP Neural Network Based on Fluent

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Xukai Wang,
  • Xiqing Zuo,
  • Peisong Tang,
  • Yong Wang,
  • Yuede Lu

摘要

The article conducts simulation and optimisation of the inlet flow path of the block of electronically controlled multi-way valves based on Fluent. The flow field characteristics of the flow path of the electronically controlled multi-way valve block are analysed by simulation, and the pressure loss of the valve block mainly occurs at the sudden change of the flow path structure and the outlet area of the fluid cyclone by combining the pressure field and velocity field distributions in the flow path of the valve block. Based on the simulation results, the optimisation objective is determined, and the Box-Behnken method is used to design a 5-factor, 3-level experimental scheme to evaluate the effects of different factors on the pressure and flow state of the flow path and optimise them. A BP neural network model is set up to fit the simulation parameters to obtain a better optimisation model, and a more optimal path structure is obtained by combining with a multi-objective genetic algorithm. The optimised path structure reduces the path pressure loss by 45.18%, while the turbulence intensity is effectively reduced by about 14.85%.