Model Order Reduction of a Single Machine Infinite Bus Power System Using the Hankel-Norm Model Reduction Algorithm

This study focuses on reducing the model order of a Single Machine Infinite Bus Power (SMIBP) system in order to simplify the system dynamics while preserving its essential stability characteristics. The primary objective is to assess the effectiveness of the Hankel-Norm Model Reduction (HNMR) algorithm in truncating the state-space representation from the full-order model (order 7) to lower orders (4, 5, and 6). The research methodology involves converting the transfer function of the SMIBP system into its state-space representation, applying the HNMR algorithm to compute the controllability and observability Gramians, constructing the descriptor operator, performing Singular Value Decomposition (SVD), and extracting the stable part via the Schur complement. Experimental results indicate that the reduced-order model of order 4 exhibits an error of approximately 7.47 and a Mean Square Error (MSE) of 71.69, whereas the order 5 model shows significant improvement with an error reduced to 1.43 and an MSE of only 7.84. The order 6 model, on the other hand, achieves extremely low errors, with and MSE values of merely 0.05 and 0.08, respectively. These findings confirm that increasing the reduced order substantially enhances the accuracy of the approximated model relative to the original system. In conclusion, the HNMR algorithm is validated as an effective method for reducing the order of SMIBP systems, with the order 6 reduced model representing the optimal choice for applications requiring high accuracy and preservation of the dynamic characteristics of the original system.