Wide-Area Measurement System-Based Subspace Identification for Obtaining Linear Models to Centrally Coordinate Controllable Devices

The contribution of this paper is the application of subspace system identification techniques, to derive a low-order black-box state-space model of a power system with many controllable devices using global signals. This model is a multiinput, multioutput open system model describing the power oscillatory behavior of the power system. The input signals are the controllable setpoints of the controllable devices, the output signals are the speed of selected generators measured by a wide-area measurement system. This paper describes how to achieve and preprocess the data to use subspace techniques to estimate and validate to finally assign an accurate model. This new approach can be used directly to design a central coordinating controller for all of the relevant controllable devices, with the aim to increase the damping of the modes in the system. Previously presented methods use local measurements or output signals dependent on the actual operational point. The benefit of the presented method is that the used output signals are independent of the system state. This makes it possible to use state-feedback control to combine the controllable devices to coordinately damp the modes. The presented method is applied in the CIGRÉ Nordic 32-bus system including two HVDC links. The case study demonstrates that accurate low-order state-space models can be estimated and validated by using the described method to accurately model the system's power oscillatory behavior.