Towards On-line PMU-based Model Calibration for Look-ahead Frequency Analysis

The awareness of power system real time conditions has been improved thanks to the availability of PMU data, and current WAMPAC schemes are becoming a reality in various systems to correct and control unstable system conditions. Foreseing upcoming unstable system conditions to take preventive instead of corrective actions is also desirable, but it requires detailed and realistic models of the system to perform look-ahead dynamic simulations. This paper describes a conceptual framework for a look-ahead frequency dynamics tool. Based on a reduced-order model of system frequency dynamics, PMU data of generation contingencies can be used to calibrate the model to represent reality. Towards a field implementation, an algorithm to recognize generation contingencies is presented. Results of the algorithm are presented using data of the Chilean systems to obtain numerical results.