How do you perform robustness analysis for control loop tuning?

Robustness analysis ensures controller performs adequately despite process variations. Methods: sensitivity analysis (vary process parameters +/-20%, check stability margins), structured singular value (mu) analysis for multivariable systems, Monte Carlo simulation with parameter distributions, and H-infinity design for worst-case optimization. Key metrics: maximum sensitivity Ms (peak of sensitivity function, target <2), complementary sensitivity Mt, gain and phase margins vs parameter variations. Document acceptable parameter ranges for maintained stability. Lambda tuning inherently provides robustness by setting closed-loop bandwidth below process uncertainty.