Probabilistic Neural Network Based Islanding Detection in Distributed Generation

Minimum non detection zone for islanding detection using an optimal Artificial Neural Network algorithm based on PSO

“…It has two positive superiorities. First, the acceptable performance of ANN is possible with a small number of inputs (in our case only two inputs; the power of ROCOF and ROCOAP) in contrast to [64,65], in which many inputs are utilized. Also, the inputs in the proposed method can be simply derived without needing to the complicated transformations such as wavelet transform, which are implemented in [62,63].…”

A Universal Islanding Detection Technique for Distributed Generation Using Pattern Recognition

“…This work proved that AI techniques require more variety in their training data in order to achieve higher accuracies. Again, a Probabilistic Neural Network islanding detection technique was proposed [26]. Multiple parameters were derived and used as input to the probabilistic neural network in order to detect islanding.…”

“…Inverter-based and Synchronous-based DGs [22][23][24][25][26]. Twenty one features are measured locally in order to include as many parameters as possible that could be affected by islanding in the system.…”

Impact of Distribuion Generation on Protection Coordination

“…Gaonkar et al [33] presented islanding detection technique by estimation of phase angle using phase -locked loop which easily detects islanding events under matching distributed generation and load power ratings. Samantaray et al [34] proposed a probabilistic neural network (PNN) based islanding detection techniques for d istributed generation considering mu ltip le parameters in order to secure the detection of islanding for any p ossible network topology. However, large training sets, tedious training process, and large number o f neurons are the several disadvantages of the neural network (NN) based schemes.…”