Online Condition Monitoring of Electrolytic Capacitors in DC-DC Interleaved Boost Converters, Adopting a Model-Free Predictive Controller

Aluminum electrolytic capacitors (AECs) are widely used in the DC link of power converters, such as the DCDC interleaved boost converter (IBC), which offers many advantages over the conventional amplifier converter. It is well known that AECs failures represent the major cause for breakdown of power electronics equipment. Hence, the development of robust control strategies, that allow to evaluate the electrolytic capacitor state, while keeping the performance of the converter control strategies, even under the occurrence of a capacitor fault, gains further emphasis. For that purpose, modelbased predictive control (MBPC) applied to power electronics is a hot research topic that has been receiving the attention of researchers, given its advantages in terms of robustness, reduced tuning effort and, when applicable, fault tolerance. In turn, model-free predictive control (MFPC) is an underdeveloped research topic, that requires further efforts. To the best of the authors' knowledge, the literature does not report any MFPC strategy aimed at DC-DC converters. Accordingly, this paper proposes the combination of a novel MFPC strategy with the Short Time Fourier Transform (STFT) method, aiming to perform an online estimation of the ESR and C parameters, thus allowing a continuous evaluation of the electrolytic capacitor condition and, consequently, avoid a total system failure. The effectiveness of the proposed strategy is verified by several experimental tests.