Development of an Adaptive Algorithm Based on Mathematical Morphology for Event Detection in Distribution Networks with DG Penetration

Adaptive protection schemes have been designed to handle the issue of variability in current contribution from the main network side, adjusting to state changes such as the connection and disconnection of distributed generation with inverter interfaces (IIDG). These schemes enhance the reliability and effectiveness of the system by reducing the risk of unnecessary load disconnections caused by inappropriate relay trips in the absence of actual fault conditions. In response to these challenges, this paper proposes an advanced algorithm that acts as a complementary tool for adaptive overcurrent protections. This algorithm is essential for recognizing state changes in the network, such as IIDG disconnections, and adjusting the relay settings to the new operational state. Identifying these conditions is complicated because disconnections of IIDGs, as well as other events like high impedance faults (HIF) and load switching, share the same characteristic of an increase in current observed by the relay that does not reach the overcurrent fault detection threshold, making them difficult to differentiate. To overcome this challenge, an enhanced feature extraction algorithm based on mathematical morphology has been developed, utilizing an adaptive triangular structuring element (SE) that conforms to the waveform of the input signal. This algorithm is capable of effectively distinguishing between various events such as HIF, load switching, and IIDG disconnections, enabling appropriate actions to be taken.