A harmonic model validation methodology for power generation units based on voltage-current phasor characteristic

This paper proposes a comprehensive methodology to validate the harmonic model of power generation units based on the phasor (not only the magnitude) of harmonic quantities. A harmonic model validation method states ‘‘whether or how close a model emulates the actual harmonic behavior of a power electronic device’’ by comparing the model results and reference data (e.g., measurement data). The validation methods differ mainly in the way of comparison. This paper defines a unique characteristic in a 3D diagram for each harmonic model. This characteristic is called ‘‘voltage-current phasor characteristic.’’ The voltage-current phasor characteristic can give us a visual view of the harmonic current phasor response of any model to various harmonic voltage phasors with different magnitudes and phase angles in just one diagram. The main idea of the proposed methodology is to compare the voltage-current phasor characteristic of the harmonic model with the measured harmonic voltage-current pairs at the point of connection of the power generation unit in the 3D diagram. The proposed methodology considers both the phase angle and magnitude of harmonic quantities and also their correlation. The particular advantage of this methodology is the interpretability of its results for the diagnosis of possible reasons for the invalidity in the models. In the case of having an invalid harmonic model, the proposed methodology could give the model developers some valuable hints for improving their models. For example, in the case of the Norton model, the proposed methodology, in contrast to other methods, can diagnose both phase angle and magnitude errors in the current source and admittance and distinguish between them. The practicability of the methodology is shown by employing some test-bench (photovoltaic power generation unit) and on-site (wind power generation unit) measurement campaigns for validating some linear and nonlinear harmonic models. This paper also suggests some criteria for evaluating the quality of measurement data as the model validation input and for defining an acceptable deviation of models from measurement data.