Fault remediation for distributed photovoltaic (PV) system

Abstract

Distributed and decentralized power electronics-based rooftop and building integrated photovoltaic (PV) systems are vulnerable to a variety of faults which are difficult to scan and rectify manually. This paper presents design, analysis and experimental validation of online fault remediation strategies in those PV systems. The remediation analysis and design depend not only on a fault type but also on the power conversion architecture of a PV system. The remediation for common failure modes in a typical PV power conversion architecture is proposed. The remediation strategies include minimum additional hardware (e.g. a bypass power semiconductor switch) for power converter switch failures and soft sensors (i.e. digital twin estimators) for electrical sensor failures. To validate these strategies, a laboratory prototype of a 300 W fault-tolerant dc-dc buck-boost converter is built. The converter has a fault-tolerant power stage and it hosts fault diagnosis, prognosis, and reconfigurable control on an FPGA device. The experimental results show significant increase in system output power (thus dependability) after proposed fault remediation is applied amid faulted scenarios such as switch and sensor faults in a distributed power converter.