Techniques and applications of inverter-based online grid impedance estimation
The extensive integration of the distributed generation units is transforming the existing conventional synchronous generator-based power system from a passive to an active mode. While this new shift enables more integration of renewable energy resources, several challenges are posed, such as changing the grid impedance characteristics, reversing power flow, and voltage rise/drop along with the feeders of distribution networks. This thesis focuses on the first challenge, i.e., the equivalent network impedance seen by each inverter at its point of common coupling (PCC). The consequences of changing the grid impedance characteristics have significant impacts on the stable operation of the grid-connected inverters. Hence, enabling the online estimation of the grid impedance will benefit the inverters in numerous ways such as designing adaptive control.
Therefore, the reported work in this thesis is divided into two distinct parts. The first part presents several online grid impedance estimation techniques (e.g., frequency-based estimation and power variations estimation) using grid-connected inverters. Then, the second part focuses on applications where the information of the estimated grid impedance is used effectively to improve overall system performance (e.g., stability analysis and accurate power sharing between inverters).
In the first part, an extensive comparison is presented initially to assess the accuracy of two online frequency-based grid impedance estimation techniques at the fundamental frequency. The two techniques are based on the 75 Hz inter-harmonics frequency injection and pseudo-random binary sequence (PRBS) injection techniques, respectively. This comparison study investigates the effects of several parameters (e.g., perturbation magnitude and duration, grid impedance variations, and nonideal grid voltage) and proposes solutions to enhance the estimation accuracy of both techniques for non-ideal grid voltage containing inter-harmonics and sub-harmonics. Then, a novel event-based impedance estimation technique is proposed to significantly reduce the occurrence of the required variations of the inverter’s output power, thus minimizing the impact on power quality. Finally, two grid impedance estimation techniques are developed to improve the estimation accuracy under unbalanced grid voltage conditions. Both estimation techniques are implemented into the control loop of grid-connected inverters equipped with a positive- and negative-sequence control (PNSC) strategy.
The second part of the thesis focuses on the information of the estimated line impedance seen by the inverters for three different applications, i.e., stability analysis of grid-connected inverters, accurate power sharing of both droop-based and voltage oscillated controlled (VOC)-based islanded AC microgrids. The first proposed application consists of a fast and accurate impedance-based stability analysis in the synchronous (dq) reference frame for a broad range of frequencies based on the grid impedance, where the parametric transfer function of grid impedance is obtained using a complex curve fitting algorithm of the triangular impulse frequency response. The second application is based on a novel technique to achieve accurate proportional load power sharing between inverters in droop-based islanded AC microgrid. The technique relies on the adaptive tuning of the virtual impedance of each inverter in order to mitigate the negative impacts of unbalanced load power sharing between inverters. The third application focuses on VOC-based islanded AC microgrids. First, a new adaptive tuning approach of the VOC parameters to restore the microgrid voltage and frequency was proposed. Second, an adaptive tuning approach of the virtual impedance of each inverter based on the estimation of feeder impedances is proposed. Hence, accurate load power sharing between the VOC-based inverters can be achieved even in the case of impedance mismatch of the inverters’ feeders.