Scholarlink

Purpose <p>Hybrid foil magnetic bearings (HFMBs) are high-speed and high-precision bearings which use foils and electromagnetic fields to levitate a rotor. It is a nonlinear system and in previous works, linear controllers were used by researchers based on mathematical approximations. This linearization of a nonlinear system hinders its efficient performance, and the application of a nonlinear controller can help in achieving its full potential. However, expert knowledge and experience are necessary for designing such controllers. To eliminate the need for expert knowledge, this paper introduces the design of a self-tuned fuzzy logic controller which is able to achieve the stable operation of HFMB with the help of different meta-heuristic algorithms.</p> Methods <p>The controller algorithm involves the combination of Mamdani type-2 fuzzy logic control with Particle Swarm Optimization and Pattern Search algorithms. The application of the proposed methodology resulted in the generation of rules from a blank rule base along with the upgraded input and output membership functions.</p> Results <p>The validation of the present methodology against the two literature results shows reductions in vibration along the x-axis by 21.26% to 55.04% and along the y-axis by 59.80% to 61.07%. Further, the error bar analysis indicates the variation in the rotor trajectory is lesser than 0.8% of the nominal air gap. The findings indicate that the designed controller performs better than a manually tuned fuzzy logic controller in the presence of initial deviation (almost zero steady-state error) and rotor mass uncertainty (-67.22% overshoot), along with better computational efficiency.</p> Conclusion <p>The proposed framework facilitates the development of a self-tuned optimal controller for HFMB. Integration of Mamdani type-2&#xa0;fuzzy logic control with Particle Swarm Optimization and Pattern Search algorithms results in 30.87% better computational efficiency. Also, it is capable of handling rotor positions for both trained and untrained data.</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

Self-Tuned Optimum Control of Hybrid Foil Magnetic Bearing

  • 期刊论文

  • 发表日期：

    2025年7月10日

查看全文
  • Nisha Singh,
  • Praveen Kumar Agarwal

摘要

Purpose

Hybrid foil magnetic bearings (HFMBs) are high-speed and high-precision bearings which use foils and electromagnetic fields to levitate a rotor. It is a nonlinear system and in previous works, linear controllers were used by researchers based on mathematical approximations. This linearization of a nonlinear system hinders its efficient performance, and the application of a nonlinear controller can help in achieving its full potential. However, expert knowledge and experience are necessary for designing such controllers. To eliminate the need for expert knowledge, this paper introduces the design of a self-tuned fuzzy logic controller which is able to achieve the stable operation of HFMB with the help of different meta-heuristic algorithms.

Methods

The controller algorithm involves the combination of Mamdani type-2 fuzzy logic control with Particle Swarm Optimization and Pattern Search algorithms. The application of the proposed methodology resulted in the generation of rules from a blank rule base along with the upgraded input and output membership functions.

Results

The validation of the present methodology against the two literature results shows reductions in vibration along the x-axis by 21.26% to 55.04% and along the y-axis by 59.80% to 61.07%. Further, the error bar analysis indicates the variation in the rotor trajectory is lesser than 0.8% of the nominal air gap. The findings indicate that the designed controller performs better than a manually tuned fuzzy logic controller in the presence of initial deviation (almost zero steady-state error) and rotor mass uncertainty (-67.22% overshoot), along with better computational efficiency.

Conclusion

The proposed framework facilitates the development of a self-tuned optimal controller for HFMB. Integration of Mamdani type-2 fuzzy logic control with Particle Swarm Optimization and Pattern Search algorithms results in 30.87% better computational efficiency. Also, it is capable of handling rotor positions for both trained and untrained data.