Black-boxing of Converter State-Space Models for Power System Eigenvalue Analysis

Abstract

State-space models are useful for a wide range of power system analyses relying on eigenvalue-based small-signal stability assessment. However, state-space models for detailed analysis of power converter dynamics are typically not provided to third-parties due to intellectual property (IP) concerns. This work illustrates how diagonalization of a state-space model will effectively obscure sensitive information about the system structure and parameters. Thus, diagonalization can be utilized as a safe method for providing black-boxed state-space models. The process is performed on an example system to clearly demonstrate that the information revealed in a diagonalized model is no more than what can be determined from system identification of a black-boxed time-domain model. A practical method of implementing this technique to provide linearized small-signal models across the full range of operating points in a compiled application is presented.