Detection and Identification of Physical Fault in DC Microgrid Systems Using Geometric Control Theory

Nowadays, DC microgrids have become more popular than AC microgrid systems due to their high efficiency, as they can reduce the requirement of the conversion stages. Also, they can integrate with renewable energy sources easily. Additionally, DC microgrids are more stable, reliable, and less complex in design. Nevertheless, DC microgrid fault detection and identification present a significant challenge due to the limited research available in this field. Out of the several faults in the DC microgrid (like pole to pole, pole to ground, and arc fault), arc fault is a more serious fault because Arc discharge raises the temperature, which can harm the circuit and increase the risk of a fire. The DC microgrid system's ability to operate safely will be impacted when a High Impedance Series Arc Fault arises, so to provide protection to the line from the arc fault, we need to detect and identify the fault accurately. However, because the fault current generated in the DC microgrid by an arc fault is nonlinear, it is difficult to detect the fault in a DC network using conventional techniques, such as CBs or fuses. In order to detect and diagnose the major arc fault more effectively and precisely, a new approach is presented here that builds an observer model using the Geometric Control Theory's idea of subspaces, such as unobservable and invariant subspaces, among others. This paper aims to design an observer that will generate residual signatures for each of the faults and show zero results for a consecutive fault.