A novel technique for online resonance frequencies monitoring based on wavelet transform for grid-connected solar inverters

Abstract

Photovoltaic systems are often connected to the main grid through multiple paralleled inverters to support and to reinforce the power generation into the grid. This interconnection can be affected by resonance phenomena that may be introduced by the excitation of passive filters and lead either to small oscillations that can deteriorate the power quality or large oscillations that can affect the whole system reliability. Therefore, early detection of such oscillations is essential to keep a resilient operation. This paper proposes a novel algorithm for an online resonance frequency monitoring based on wavelet transform (RFM-WT) that can perform fast detection of undesired resonance frequencies. The RFM-WT is performed through two steps. The first step consists in decomposing each inverter output current into different signals using the wavelet transform. This allows separating the different frequencies that exist in the inverter output current. The second step allows estimating the frequencies from the signals obtained in the first step. Accordingly, a comprehensive approach describing the design of the proposed monitoring algorithm is presented. Its effectiveness is compared to the classical PLL algorithm. Unlike the PLL algorithm, the proposed technique is easy to implement and doesn't require regulators’ parameters design and configuration. The simulations and experiments proved that the algorithm can accurately locate the resonance frequencies in inverters output currents. Thus, its robustness was validated through load variation and transition from grid- connected to islanded mode.