Some Novel Approaches For Islanding Detection In Grid Connected Distributed Generation System

Abstract

The microgrid consists of a renewable energy-based system with its local load and a Point of Common Coupling (PCC) where the grid is connected through a circuit breaker. Here, the solar Photo Voltaic (PV) system is considered as the Distributed Generation (DG) system. To increase the productivity of the PV system, Super superimposed sliding Mode Control Perturb and Observe (SISMC-PO) is used. The productivity of SISMC-PO is also checked with grid and isolated mode under varying weather conditions, where it shows its superiority over Perturb and Observe (PandO). When the local load demand is matched with DG power generation Zero Power Mismatch (ZPM) occurs. During this condition, the DG with its local load gets separated from the grid and starts operating independently is known as the islanded mode of operation. The importance of the Islanding Detection Scheme (IDS) is that it can detect islanding accurately, so that, the mal-operation can be reduced. The mal-operation of CB may cause damage to electrical equipment and danger to the life of the field operator. The area where the IDS fails to detect the islanding is known as the Non-Detection Zone (NDZ) which normally occurs near ZPM. If the area of the NDZ is greater, then the accuracy of detection gets reduced. Hence, the main objective is to reduce the area of NDZ near to zero, so that, the IDS can separate accurately the Islanding Events (IE) from Non-Islanding Events (NIE). Various techniques are found in the literature for reducing the NDZ to achieve high- accuracy detection. A voltage ratio and Artificial Neural Network (ANN) method is implemented as an IDS where the performance of the ANN is found to be superior as compared to the voltage ratio method. It is seen that the ANN performance is good, but the detection time is more. To get faster detection, the meta-heuristic optimizing algorithm of Sine Cosine Hyperbolic Optimization Neural Network (SCHO-NN) is employed in the 12.47 kV 50 Hz test system. In this thesis, the Fast Fourier Transform (FFT) is used fo