Abstract:
Precise and fast estimation of the phase information of grid voltages is crucial for grid synchronization of various power electronic devices, which is a research trend in modern smart grid technology. However, the task of instantaneous phase estimation has become difficult as the grid voltages may contain harmonics, DC offset, frequency variations, and voltage unbalances due to an increase in renewable energy penetration to grid, and domestic and industrial non-linear loads. In this dissertation, a fast and accurate instantaneous phase estimation technique is developed to overcome the present limitations in this field.
This dissertation proposes a three-phase phase locked loop (PLL) algorithm relying on adaptive Clarke transformation (ACT) for tracking the phase angle of unbalanced grid voltages associated with harmonics and DC offset. A meticulously tuned band pass filter (BPF) is inserted in each phase to remove harmonics and DC offset. Two separate algorithms are proposed to estimate the amplitudes and the phase-angle deviations of three-phase voltages. Using the estimated amplitudes and phase angle deviations, a set of analytical expressions is derived for the coefficients of Clarke transformation (CT) matrix to make them adaptive under both amplitude and phase unbalances, which is named as adaptive CT (ACT). The ACT is capable of generating orthogonal signals from unbalanced three-phase voltages, whereas the conventional CT based on constant matrix fails to do so. After getting the orthogonal voltages, a conventional SRF-PLL is used to track the phase-angles of all three-phases. A phase-correction technique is also developed to make the PLL frequency adaptive without using any frequency feedback loop.