Abstract
The increasing photovoltaic (PV) installations and their integration with the utilities have complexed the operation of the power system network making them vulnerable to various faults and abnormalities. The traditional methods developed to handle this problem are aimed to explore the ability of PV inverter to operate in standalone (SA) mode when there are predictable grid side abnormalities or scheduled maintenances. In this article, a grid condition monitoring based transition control approach is developed using machine learning algorithm and a hybrid control strategy. This article is motivated at handling the intentional and unintentional islanding conditions by operating the PV system in both grid-connected (GC) and SA modes. The switching between the controllers is performed by the central controller ensuring a smooth transition and continuous power delivery to the load. For validating the claims, numerical simulations and experimental analysis are carried out with a 4 kWp GC PV system. The results depicted fast grid condition monitoring under 20 ms and smooth transition without any transients or harmonics.
Original language | English |
---|---|
Pages (from-to) | 2332-2340 |
Journal | IEEE Transactions on Industry Applications |
DOIs | |
Publication status | Published - Mar 2022 |
Externally published | Yes |
Keywords
- condition monitoring
- distributed power generation
- invertors
- learning (artificial intelligence)
- photovoltaic power systems
- power distribution faults
- power grids