An adaptive fault identification scheme for DC microgrid using event based classification

Abstract

The recent high penetration of stationary renewable energy sources yielding DC output coupled with the increased interest in storage systems, electronic loads, machine drives and other components operating with DC input, require the development of new techniques for control and link integration. Faults in power distribution systems have high impact on continuous supply and quality service. If a contingency occurs in a system, instability and cascading failures blackout would be possible. Protection against such severe faults increases the transfer capabilities of the existing transmission systems. The literature survey shows various methods available for the fault identification and fault location. In the existing system each protection unit is able to autonomously identify the type of fault using the current derivative fault identification method. Then the event judgment is sent to other interconnected protection units through high level data communication. But in the existing system there is a limitation that it is possible to limit the load voltage drop up to 2.88%. A novel protective method is proposed to identify the fault, isolate the faulted area and restore the system quickly. The fault identification is based on voltage derivative method. In the voltage derivative protection method, the derivatives ∂V/∂t and ∂I/∂t are calculated from the dc voltages and currents measured locally at the relaying point to detect, locate and isolate the fault. The local voltage and current measurements are taken for each instance of time. It requires less data and doesn't require high speed communication and synchronization. The advantage of this method is that the first incident/reflected wave from the fault is used in the detection and, therefore, the response time is very fast, providing fault detection within 2-3ms than current derivative method. The expected results of the proposed system are to identify the fault, isolated the faulted area and restore the system quickly. The results of the proposed system are verified using MATLAB Simulink.