Effect of aggregation methods on individual modes in reduced order power system models

The extraordinary size and complexity of modern power systems results in an intense computational burden in transient stability studies. Use of reduced-order power system models could reduce the amount of the computations required by those studies. Such reduced-order models are called dynamic equivalents and can be derived by coherency-based techniques. A coherency-based equivalent is derived by first identifying the coherent groups, and second, aggregating all the generators belonging to the same coherent group. Two well-known aggregation procedures, the inertial averaging and the singular perturbations, are evaluated with respect to how accurately they preserve each dominant mode of the unreduced-order model in the equivalent system. An analysis proves that both aggregation methods exhibit an overstiffening phenomenon in which the dominant modes are increased in frequency. However, the effects are less pronounced when the equivalent model is obtained by the singular perturbation aggregation method