Non-parametric multiple inputs prediction model for magnetic field dependent complex modulus of magnetorheological...

Area of Science:

• Materials Science
• Polymer Science
• Machine Learning Applications

Background:

• Magnetorheological (MR) elastomers exhibit complex, nonlinear behavior, making accurate modulus prediction challenging.
• Traditional modeling struggles with the vast parameter space and nonlinearities inherent in MR elastomers.
• Machine learning offers a promising non-parametric approach for modeling complex material responses.

Purpose of the Study:

• To develop a novel machine learning platform for predicting the complex modulus of MR elastomers.
• To investigate the influence of fabrication parameters and operating conditions on MR elastomer properties.
• To establish a data-driven model that overcomes limitations of traditional parametric methods.

Main Methods:

• Utilized feedforward neural networks, including Extreme Learning Machines (ELMs) and Artificial Neural Networks (ANNs).
• Trained models using experimental data encompassing excitation frequency, magnetic flux density, particle weight percentage, and curing magnetic field.
• Developed a data-driven approach for predicting multiple input-dependent complex moduli.

Main Results:

• Achieved high prediction accuracy for the complex modulus, with an R-squared value of approximately 0.997 compared to experimental results.
• Demonstrated that the machine learning models accurately capture the complex modulus pattern and magnetorheological effect.
• Validated model performance using both learned and unlearned datasets across various curing conditions.

Conclusions:

• The proposed machine learning platform provides a highly accurate and reliable method for predicting MR elastomer complex modulus.
• The data-driven approach effectively models the nonlinear behavior and complex interdependencies of MR elastomers.
• This study highlights the potential of machine learning in advancing the understanding and application of advanced materials like MR elastomers.