High-Gain Observer-Based Fault Detection Scheme for Short-Circuit Switch Faults in Grid-Connected PV Systems by using Optimizers

This paper proposes a model-based fault diagnosis scheme for grid-connected photovoltaic systems with distributed maximum power point tracking (MPPT) strategy. Specifically, short-circuit switch faults in optimizers (boost dc/dc converters in series connection for MPPT purposes) are considered in this study. Depending on the application, SCFs can be more or less detrimental to the stability and integrity of systems. In systems with high penetration, disturbances in voltage or frequency can be generated that can, in the worst case, cause instabilities in the system. Hence, a fast fault detection strategy is mandatory. It is worth noting that, unlike other works reported to date, the control action of the optimizers presents singularities under this fault scenario, making the fault diagnosis task difficult. For this, a decoupled subsystem from i) dc/ac converter dynamics and ii) irradiance changes are obtained for fault and disturbance isolation purposes. Then, a dedicated high-gain observers bank is proposed for a residual generation. In this way, fault time detection is achieved in 8 switching periods. Finally, a numerical evaluation has been carried out to validate the ideas proposed in this paper. For this, three series-connected optimizers interconnected to a three-phase 220V 60 Hz grid through a neutral point clamped (NPC) inverter with a first-order L filter were considered in the simulation.