Power system model order reduction based on dominant modes in modal analysis

Investigations of small-signal stability of electrical power systems are usually performed by modal and time-domain analysis using descriptive differential equations of machines. Thereby, the modal analysis enables the identification of all dominant states of the linearized power system model, which are associated with electromechanical interactions and power oscillations. For reasons of simplification and due to computation capabilities, a reduction of the model order is often advantageous. This can be done by neglecting different physical effects within the machines, retaining the topology and the modal characteristics of the power system regarding its electromechanical behavior. This paper presents a modal analysis based approach for power system model order reduction. In order to reproduce the small-signal behavior of power systems accurately, damping and synchronizing parameters are introduced within the reduced-order model. Subsequently, each coefficient is determined analytically for a single machine infinite bus system and using the particle swarm optimization method for a multi-machine system. The deviation to the electromechanical reference modes of the higherorder power system model constitutes the objective function to be minimized. The results confirm that the approach is suitable for the aspired goals and fulfills the accuracy requirements.