Estimation of short-time daily power loads for Hanoi city using time series and neural networks

Short-time power loads prediction is one of fundamental task in power system problems. There are different methods in realizing this task, however the prediction model is very different for every region. For each data set, we should provide the model selection and model's parameters estimation in order to get the acceptable accuracy. Hence, this paper presents an estimation of short-time daily power loads for the Hanoi city (Vietnam) by means of time series analysis, Kohonen self-organizing map and multi-layer perceptron neural networks. The obtained results proved that both methodology and tool are usable. Future work will focus on integrating a forecast module based on the present work in a vertual power system simulation.