A group of researchers at the University of Lisbon has developed an advanced machine learning algorithm that can successfully classify and predict potential inverter failures in utility-scale photovoltaic (PV) plants. By monitoring the inverter subsystems and analyzing data, the algorithm is able to detect when maximum and minimum values are reached, and it sends alarms to alert operators of potential failures.
The algorithm categorizes variables based on their historical values, enabling it to identify different types of failures. These include grid faults, grid overvoltage, temporary grid overvoltage, grid undervoltage, low voltage, temporary AC overcurrent, grid overfrequency, grid underfrequency, grid power failure, excessive stray current, supply grid fault, 10-minute grid overvoltage, output overload, and unbalanced load of grid device fault.
To test the effectiveness of the algorithm, the researchers examined two ground-mounted PV systems with capacities of 140 kW and 590 kW, both of which were equipped with inverters from German manufacturer SMA. The algorithm analyzed the variables of each inverter and successfully identified the types of failures experienced.
The algorithm utilizes tree-based models to analyze the data. These models use splitting rules to divide the feature space into smaller regions with similar response values, enabling accurate prediction and classification of failures.
The researchers highlight that the algorithm is not only capable of identifying inverter failures but also demonstrates the potential to analyze seasonal variations in these failures. This information can be extremely valuable for reliability analysis and maintenance planning.
In conclusion, the researchers suggest implementing measures to protect inverters from inrush and overcurrent by using clamp circuits connected to resonant capacitance in parallel. This approach can significantly improve the power-conversion efficiency of the inverters.
The development of this novel machine learning algorithm provides a promising solution for predicting and preventing inverter failures in solar power plants, ensuring the continued optimal performance of utility-scale PV installations.
The source of the article is from the blog combopop.com.br