Classification of nonlinear loads based on artificial neural networks

In this paper we used artificial neural networks for classification of nonlinear loads. The special attention is payed to this kind of loads since they are the most difficult to register using existing monitoring methods. Carefully chosen parameters are first measured and calculated, and these values are then used as a training set for a neural network. We also considered situations when different processes are run on a PC, and we showed that this does not influence our method, since it works properly in these situations also.