Model-based Fault Detection and Isolation in a MPPT BOOST converter for photovoltaic systems

In this paper an observer-based fault diagnosis system is proposed for a Maximum Power Point Tracker (MPPT) BOOST converter in photovoltaic (PV) systems. Open-and Short-circuit switch faults can be diagnosed by the Fault Detection and Isolation (FDI) algorithm suggested in this study. A decoupled subsystem from the load and PV currents is obtained for residual generation, which is guaranteed at the price of 2 measurements, namely, PV current and load voltage. Hence, the FDI system is insensitive to load changes and sudden irradiance drops. According to observability properties of this subsystem, the fault detection time can be assigned arbitrarily. In addition, operation of the FDI system in open-and closed-loop has been evaluated through a prototype of 350 W. Only the most common measurements employed into the Maximum Power Point searching techniques are required in the proposed FDI system. Finally, as an important result, the proposed FDI system can be applied over the most common PV applications due to the residual generation system is decoupled from load conditions. © 2016 IEEE.

VIVO