Scholarlink

Background <p>Particle damping is a passive vibration control technique that relies on energy dissipation through inter-particle collisions and friction.</p> Purpose <p>This study investigates the application of particle dampers within a printed circuit board enclosure of a radar transmit-receive module.</p> Methods <p>The damper cavity's location was selected based on vibration-sensitive regions and space constraints within the enclosure. Experiments were conducted with various parameters, i.e., particle size and filling ratio. To fine-tune these parameters, two artificial neural network models—a backpropagation neural network and a hybrid radial basis function neural network—were employed and trained on the experimental data.</p> Results <p>The hybrid radial basis function model demonstrated better performance metrics than the backpropagation model, leading to its selection for response prediction. The predicted results indicated significant influence of the damper parameters on acceleration responses, aiding in determining optimal settings. The optimized configuration, with a 2.35&#xa0;mm particle size and a 92% filling ratio was very effective in response reduction at different test inputs, such as an 85% reduction for a 2&#xa0;g input.</p> Conclusion <p>The optimal results were re-verified with a new set of experimental tests, confirming the accuracy and practical utility of the neural network-based approach in optimizing particle damper design parameters for enclosures.</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

Fine-tuning of Experimentally Obtained Particle Damper Parameters Using Hybrid Radial Basis Function Neural Network for Vibration Mitigation of a PCB Enclosure

  • 期刊论文

  • 发表日期：

    2025年1月22日

查看全文
  • Sunil Kumar,
  • Anil Kumar

摘要

Background

Particle damping is a passive vibration control technique that relies on energy dissipation through inter-particle collisions and friction.

Purpose

This study investigates the application of particle dampers within a printed circuit board enclosure of a radar transmit-receive module.

Methods

The damper cavity's location was selected based on vibration-sensitive regions and space constraints within the enclosure. Experiments were conducted with various parameters, i.e., particle size and filling ratio. To fine-tune these parameters, two artificial neural network models—a backpropagation neural network and a hybrid radial basis function neural network—were employed and trained on the experimental data.

Results

The hybrid radial basis function model demonstrated better performance metrics than the backpropagation model, leading to its selection for response prediction. The predicted results indicated significant influence of the damper parameters on acceleration responses, aiding in determining optimal settings. The optimized configuration, with a 2.35 mm particle size and a 92% filling ratio was very effective in response reduction at different test inputs, such as an 85% reduction for a 2 g input.

Conclusion

The optimal results were re-verified with a new set of experimental tests, confirming the accuracy and practical utility of the neural network-based approach in optimizing particle damper design parameters for enclosures.