Electric load forecasting and spark detection using artificial neural networks

In this paper, the neural network, one of the artificial intelligence scheme, is proposed to improve the conventional electric switchboard. The neural network used in this paper has two purposes. First, neural network detect electrical spark from broken insulation of electric wire by decrepitude or line damage in an electric wire of indoors. Second, autoregressive neural network predict the electrical load in month so that a management of energy can be administered effectively. We confirmed that this method has better performance than the conventional one.