On-Line Fault Diagnosis of DC-Link Electrolytic Capacitors in Boost Converters Using the STFT Technique

Faults in aluminum electrolytic capacitors (AECs) are classified as the major cause for power electronics equipment breakdown, mainly due to AECs wear out through the vaporization of the electrolyte, as a result of both aging and temperature effects. The majority of manufacturers identify the end-of-life threshold of such capacitors as their equivalent series resistance (ESR) increases by 80% or their capacitance (C) decreases by 20%, relative to their initial values. Therefore, the online estimation of these two parameters can provide valuable and timely information that allows detecting any potential capacitor failure. Changes in C and ESR parameters influence the relationship between the capacitor voltage and current ripples. This ratio is dominated by C at low frequencies, and by ESR in the high-frequency range. The aim of this article is to propose an algorithm allowing a continuous estimation and tracking of both parameters (ESR and C). This algorithm should be simple, fast, and suitable for online implementation. Therefore, this article introduces the use of the short-time Fourier transform (STFT) technique to determine and track C and ESR values, starting from some always-existing harmonics at low and high frequencies. The STFT is based on applying the discrete Fourier transform algorithm on a sliding window, which makes it more appropriate for online implementation. The effectiveness of the proposed method is proved by simulation and laboratory experiments, using a boost dc-dc converter.

Automated summary

This article introduces an algorithm for fault detection in aluminum electrolytic capacitors, based on online estimation of equivalent series resistance (ESR) and capacitance (C) parameters using short-time Fourier transform (STFT) technique. The algorithm is simple, fast, and suitable for online implementation.