Robust Framework for Online Parameter Estimation of Dynamic Equivalent Models Using Measurements

The ever-increasing demand for electricity, the advent of microgrids and the increasing penetration of distributed generators has renewed the interest in dynamic equivalencing, due to to its importance on power system analysis and control applications. This paper introduces a robust measurement-based framework for the online derivation of dynamic equivalent models. First, events/disturbances suitable for the derivation of dynamic equivalent models are automatically detected. Next, signal processing techniques are applied to recover missing samples and to remove noisy components from measured data. To exclude unnecessary post-disturbance data, a fine-tuning technique of the signal window length is also proposed as a supplementary offline process. Finally, model parameters are estimated using nonlinear least-squares optimization. The performance of the proposed methodology is tested using artificially created signals, simulation results obtained from a modified benchmark distribution grid and measurements acquired from a laboratory-scale active distribution network.

Automated summary

This paper presents a measurement-based online framework for deriving dynamic equivalent models, which includes automatic detection of suitable events/disturbances, signal processing techniques, fine-tuning of signal window length, and parameter estimation through nonlinear least-squares optimization. The methodology's performance is tested using artificially created signals, simulation results, and measurements from a laboratory-scale active distribution network.