Author(s):
A. Vicente T. leite 
; Henrique J. A. Teixeira ; A. J. Marques Cardoso ; Rui Esteves Araújo
Date: 2007
Persistent ID: http://hdl.handle.net/10216/23965
Origin: Repositório Aberto da Universidade do Porto
Subject(s): Ciências tecnológicas; Engenharia; Engenharia electrotécnica
Description
Electrolytic capacitors are usually used in power electronic systems for smoothing, energy storage or filtering. They have the best overall performance for these objectives being a critical element in the design of these systems, in several applications, with different requirements. But, infortunately, in most cases, they are the most life-limiting device. The expected life of electrolytic capacitors depends on their internal temperature and is determined by the ratio of the electrolyte solution evaporation used in their fabrication. The deterioration caused by this evaporation is reflected in electrical parameters, mainly the equivalent series resistance (ESR). As the volume of the electrolyte decreases, ESR increases and capacitance decreases. Additionally, the increase in ESR has a positive feedback effect since it leads the temperature to increase and this in turn leads to further evaporation of the electrolyte leading to the ESR increase and so forth. This paper presents a simple ESR identification methodology for electrolytic capacitors condition monitoring in view of preventive maintenance to signal the need of maintenance and or replacement. The identification methodology is based on a simple continuous-time model and some recursive prediction
error methods, namely, Kalman filter, gradient and orgetting factor approaches. Simulation and
experimental data acquired from a step down converter were used with these algorithms. The identification methodology is also suitable for different power converter topologies.
