Application of Kalman filters in model-based fault diagnosis of a DC-DC boost converter

Abstract

This paper illustrates how Kalman filters were used in a model-based fault diagnosis of a DC-DC boost converter. A time-averaging model was used with the Kalman filters to generate residual signals. Multiple signature faults were developed in fault scenarios to identify critical variations in the elements of a power converter using the adaptive estimation technique. Results show a very precise and accurate fault diagnosis of signature faults. The fault diagnosis shows a high performance in transients and against noise in the circuit.