Abstract Nowadays, increasing the need to use renewable energies such as solar and wind energy with the aim of producing cleaner energy, has led to making use of clean energy sources and connecting these sources to the traditional electrical energy system. The concept of microgrids has been proposed in order to achieve this goal. Microgrids are divided into three types depending on their voltage type Alternating Current (AC), Direct Current (DC) and a combination of both types known as hybrid microgrids. Daily increase in DC loads such as computers, mobile phones, telecommunication systems etc. as well as the DC nature of the production of renewable energy sources increases the development of DC microgrids. Another reason to use these networks is higher efficiency compared to AC microgrids. The reason for this issue is the use of cables with a lower cross-sectional area in these grids and the absence of skin effects in the DC cables. One of the serious challenges facing a DC microgrid is designing a suitable protection plan. The reason for this is the high rate of current changes in these networks. Like other power grids, DC microgrids may also encounter faults. Timely detection of these faults has great importance to maintain the performance of the network and to minimize possible damages. Among other problematic factors in fault detection in DC microgrids is that they can be pointed out that the level of short circuit in these networks is low. Also, there is no natural point of zero crossing of the fault detection. This has made the networks problematic. This thesis will begin by studing the the protection of DC microgrids and the challenges of these grids. In the following, a new design for the purpose of detection faults in these networks have been suggested by using intelligent electronic devices. The performance of the proposed method in the first stage is based on signal processing and using the dynamic mode decopomsition method which is implemented by sampling voltage and current components. In the second stage, the signals obtained in the first stage have been entered into the Hilbert transform and instantaneous frequency algorithms in order to detect the fault. In order to eva‎luate the capability of the proposed method, the results of this method by different studies and through simulation in Matlab/Simulink is provided. The results show that the proposed scheme is able to detect all types of faults in the presence of existing transient states and production uncertainty by disturbuted production sources and is able to detect different faults in different places in the DC microgrid. It is worth explaining, the proposed scheme in the case where white Gaussian noise on the sampled signals has been affected, it is able to detect faults properly. Also, the proposed method is able to detect all types of pole-to-ground and pole-to-pole faults with different impedances in DC microgrid lines.

امير حسيني , وحيده السادات صادقي

احسان آزاد فارساني , حميدرضا عبدالمحمدي