Identify EMI Contributors in Circuit Simulation using Morris Sensitivity Analysis

Electromagnetic interference (EMI) analysis is essential to optimize circuit designs to meet regulatory standards, especially as switching speeds in modern power electronics continue to increase. However, identifying which circuit components contribute most to EMI remains challenging. Methods like Monte Carlo simulations require thousands of runs, making them impractical for early-stage design. In this work, we demonstrate that Morris global sensitivity analysis (GSA) provides a computationally efficient alternative for screening EMI contributors. We apply Morris GSA to a buck converter simulated in LTspice with 39 variable circuit components, using the frequency-domain EMI spectrum as the output parameter. We found that only 12 optimized trajectories (480 simulations) are sufficient to obtain stable sensitivity rankings in order to identify dominant contributors to conducted emissions (differential- and common-mode) across frequency. Although Morris GSA cannot resolve higher-order interactions, our findings suggest that Morris GSA can support EMC engineers to quickly prioritize impactful design changes early in the development process.