Condition Parameter Estimation for Photovoltaic Buck Converters Based on Adaptive Model Observers

DC-DC power converters such as buck converters are susceptible to degradation and failure due to operating under conditions of electrical stress and variable power sources in power conversion applications, such as electric vehicles and renewable energy. Some key components such as electrolytic capacitors degrade over time due to evaporation of the electrolyte. In this paper, a model-observer based scheme is proposed to monitor the states of Buck converters and to estimate their component parameters, such as capacitance and inductance. First, a diagnosis observer is proposed, and the generated residual vectors are applied for fault detection and isolation. Second, component condition parameters, such as capacitance and inductance are reconstructed using another novel observer with adaptive feedback law. Additionally, the observer structures and their theoretical performance are analyzed and proven. In contrast to existing reliability approaches applied in buck converters, the proposed scheme performs online-estimation for key parameters. Finally, buck converters in conventional dc-dc step-down and photovoltaic applications are investigated to test and validate the effectiveness of the proposed scheme in both simulation and laboratory experiments. Results demonstrate the feasibility, performance, and superiority of the proposed component parameter estimation scheme.