Sliding mode observer-based control for a series active filter

In active compensation systems applications, such as series active filters (SAF), the right electrical variables measurement is critical for the correct system performance. However, in three phase applications this implies the use of many sensors, as consequence more system sensibility to fail and deviations introduced by the sensors is common. Since the SAF is causal and observable it can be reproduced variables which dynamics depend on the SAF desired behavior. The main aim of this paper is to reduce the sensor stage, avoiding the errors introduced by them and obtaining a good performance of the SAF. In order to achieve this objective, it has been implemented and tested a sliding mode observer in a digital signal processor (DSP) board and the estimated variables have been used to control a 1KVA SAF prototype presenting a good dynamic response, reducing the number of sensors and increasing immunity to interferences in the sense stage.