Scholarlink

<p>Backpressure forming is essential for precision forming of thin-walled complex scrolls by applying a controllable counter-pressure to regulate inhomogeneous material flow. However, the process involves strong thermo-mechanical coupling effects, and there exists a highly nonlinear relationship between the numerous process parameters and the forming quality, posing significant challenges to achieving stable and consistent part quality. Therefore, reasonable setting of process parameters is crucial. This challenge becomes even more complex when further considering factors such as forming energy consumption and mold life. This study focuses on the backpressure forming process of an orbiting scroll. The scroll vane end face height difference, under-filling rate, and peak forming load were selected as optimization objectives, with billet temperature, die temperature, forming speed, friction coefficient, backpressure distance, and backpressure load as design variables. Taguchi-based numerical experiments were conducted, and correlation analysis revealed competitive relationships among the objectives. Given the limited experimental data, a particle swarm optimization (PSO) optimized backpropagation neural network (BPNN) model was constructed to establish satisfactory precision nonlinear mappings, demonstrating reasonable prediction performance within the investigated parameter range across all three objectives. Multi-objective optimization was then performed using the fast elitist non-dominated sorting genetic algorithm (NSGA-III), yielding a Pareto optimal solution set that reflects the trade-offs between the objectives. Through subjective and objective evaluation, the preferred process parameter combination was ultimately determined based on the entropy weighted Technique for Order Preference by Similarity to Ideal Solution method (TOPSIS). Compared to the initial process parameters, the optimal parameter combination significantly reduced the peak forming load from 725 kN to 521 kN with a decrease of 28.1%, greatly decreased height difference from 5.2&#xa0;mm to 0.2&#xa0;mm with a reduction of 96.2%, and simultaneously increased filling rate from 92.9% to 99.2% with an improvement of 6.8%. Acceptable quality scroll components were successfully fabricated using these parameters. The proposed integrated multi-objective optimization method effectively coordinates conflicting objectives in backpressure forming, improving forming quality and providing a potential framework for developing precision forming processes for similar complex components.</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

Multi-objective optimization for backpressure forming of thin-walled complex scrolls via a hybrid PSO-BPNN, NSGA-III, and TOPSIS framework

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年8月5日

查看全文
  • Feng liu,
  • Wanji Chen,
  • Yadong Li,
  • Jianjun Pang,
  • Mingming Wang,
  • Huimin Tao,
  • Xiaomei Guo

摘要

Backpressure forming is essential for precision forming of thin-walled complex scrolls by applying a controllable counter-pressure to regulate inhomogeneous material flow. However, the process involves strong thermo-mechanical coupling effects, and there exists a highly nonlinear relationship between the numerous process parameters and the forming quality, posing significant challenges to achieving stable and consistent part quality. Therefore, reasonable setting of process parameters is crucial. This challenge becomes even more complex when further considering factors such as forming energy consumption and mold life. This study focuses on the backpressure forming process of an orbiting scroll. The scroll vane end face height difference, under-filling rate, and peak forming load were selected as optimization objectives, with billet temperature, die temperature, forming speed, friction coefficient, backpressure distance, and backpressure load as design variables. Taguchi-based numerical experiments were conducted, and correlation analysis revealed competitive relationships among the objectives. Given the limited experimental data, a particle swarm optimization (PSO) optimized backpropagation neural network (BPNN) model was constructed to establish satisfactory precision nonlinear mappings, demonstrating reasonable prediction performance within the investigated parameter range across all three objectives. Multi-objective optimization was then performed using the fast elitist non-dominated sorting genetic algorithm (NSGA-III), yielding a Pareto optimal solution set that reflects the trade-offs between the objectives. Through subjective and objective evaluation, the preferred process parameter combination was ultimately determined based on the entropy weighted Technique for Order Preference by Similarity to Ideal Solution method (TOPSIS). Compared to the initial process parameters, the optimal parameter combination significantly reduced the peak forming load from 725 kN to 521 kN with a decrease of 28.1%, greatly decreased height difference from 5.2 mm to 0.2 mm with a reduction of 96.2%, and simultaneously increased filling rate from 92.9% to 99.2% with an improvement of 6.8%. Acceptable quality scroll components were successfully fabricated using these parameters. The proposed integrated multi-objective optimization method effectively coordinates conflicting objectives in backpressure forming, improving forming quality and providing a potential framework for developing precision forming processes for similar complex components.