Simultaneous online identification and localization of disturbances in power transmission systems
Within this survey an approach is presented for the simultaneous online identification and localization of grid disturbances in transmission power systems using different techniques for multivariate time series classification. For the generation of the training data dynamic simulations are performed using DIgSILENT ® PowerFactory combined with a Monte Carlo based initial state selection. Within this survey different classifiers are developed and compared with each other including dynamic time warping, support vector machines, shapelets, recurrent neural networks and random forests. The performance is evaluated in terms of classification accuracy and prediction time.