Improved efficacy behavioral modeling of microwave circuits through dimensionality reduction and fast global...

Area of Science:

• Electrical Engineering
• Computational Electromagnetics
• Microwave Engineering

Background:

• Behavioral models are crucial for efficient design in high-frequency electronics, replacing time-consuming electromagnetic (EM) analysis.
• Challenges include circuit nonlinearity, large parameter spaces, and ensuring model validity across wide parameter and frequency ranges.

Purpose of the Study:

• To present a cost-effective and reliable behavioral modeling method for microwave passive circuits.
• To address the limitations of traditional surrogate modeling techniques.

Main Methods:

• Utilizing a fast global sensitivity analysis (FGSA) to identify key design parameters and their impact.
• Establishing a reduced-dimensionality domain based on FGSA results.
• Constructing behavioral models using limited data within the reduced domain.

Main Results:

• The proposed method significantly reduces the data required for model construction.
• The developed models demonstrate high predictive accuracy and capture substantial circuit response variability.
• Demonstrated superiority over traditional modeling techniques through analysis of diverse microstrip circuits.

Conclusions:

• The innovative approach offers a dependable and efficient solution for behavioral modeling of microwave passives.
• The method is well-suited for practical design applications in high-frequency electronics.
• FGSA is effective in identifying critical design parameters for dimensionality reduction.