Stability Robustness Against Real Parameter Uncertainty of DC Electrical Power Conversion Systems

DC power systems which employ switch-made power converters am known to be notoriously prone to instability because of the high degree of sensitivity to parameter and load variation. Design of such systems therefore requires stability robustness assurance in the face of uncertainty. Aerospace applications often require multiloop treatment. This paper addresses some of the robust control theories which can be applied to stability analysis of DC power systems. In the robust control context, robustness analysis is performed by computing the Multivariable Stability Margin (k/sub m/) or Structured Singular Value (/spl mu/). The exact calculation of k/sub m/ or /spl mu/ is difficult. Several approximation methods including the Perron eigenvalue, the musol, the Muliplier and the Convex Hull Method, are used and compared. A simplified circuit representing a typical DC-DC switch-mode power converter with supply and load is used for comparison. It is concluded that the Convez Hull Method offers a good approximation and computationally forgiven without much sacrifice of conservativeness.