An ADALINE-Based PLL for Single-Phase Grid Synchronization with Enhanced Dynamics and Harmonic Rejection

This paper introduces a novel adaptive neural phase-locked loop (AN-PLL) based on an adaptive linear neuron (ADALINE) for grid-connected converters synchronization. The suggested AN-PLL contains two main stages; a zero-crossing-based frequency detector (ZCFD) in series with an ADALINE-based fundamental component estimator (AN-FE). In the first stage, a ZCFD accurately tracks the grid frequency, even under severe waveform distortions, harmonic disturbances, or the presence of DC offsets. In the second stage, the ADALINE estimates the fundamental component of the grid voltage, providing, after convergence, both its amplitude and phase angle. The use of the ADALINE is motivated by its fast convergence, inherent filtering capability, and simplicity, as it requires tuning of only one learning parameter. In the proposal, ZCFD and AN-FE work together to form a robust, fast, and easy-to-implement synchronization method, resilient to various grid disturbances. Performance assessment of the AN-PLL is carried out by integrating it into a single-phase grid-connected photovoltaic system. Both simulation and experimental tests were conducted, covering distorted grid conditions, frequency and voltage variations, and real-time power injection. For clarity and benchmarking, the developed AN-PLL is compared to an advanced enhanced PLL (EPLL), a widely recognized robust method in the literature. The proposed PLL effectively reduces estimation error, ensures fast dynamic response, and maintains low total harmonic distortion, even under grid distortions. Its strong tracking and rapid adaptation highlight its suitability for robust power electronic applications.