On-line Estimation of Electrolytic Capacitor Parameters using Recursive Least Squares and Sliding Mode Differentiators

Electrolytic capacitors may experience a gradual deterioration that results in an increase in the equivalent series resistance (ESR) and a decrease in capacitance over time. An indication of the condition of a capacitor may be obtained by the on-line monitoring of these two electrical properties. In this paper, an on-line capacitor parameter estimation scheme that can be used for degradation monitoring is proposed. The approach involves the recursive least squares (RLS) algorithm with forgetting factor and robust exact differentiators based on sliding mode concepts. This scheme allows for on-line estimation of the ESR and capacitance under noisy conditions. Computer simulation results involving a DC-DC converter, provide evidence of the effectiveness of the scheme proposed in this paper.