No abstract

Time series forecasting is an important task in various fields of science, like economy, engineering and other areas that use historical data to predict future problems. In this context, Artificial Neural Networks have shown promising results for this task, when compared with the traditional statistical techniques. Thus, this research aims to evaluate the performance of NARX-neural network (Nonlinear Autoregressive Model with Exogenous Input) for the purpose of performing load forecasting for very short-term data from distribution substations. The cross validation was applied to evaluate different topologies. It is important to mention that the data was obtained by measures done in Brazilian substations located at two different cities. The results show the contribution of the paper once it demonstrates the efficiency of the NARXneural network compared with Feedforward and Elman neural networks, which are widely used to predict times series.Index Terms-Artificial neural networks, NARX-neural network, load forecasting.

No abstract

The main purpose of this paper is to achieve a comparative analysis among Autoregressive Integrated Moving Average model, Artificial Neural Networks and Adaptive NeuroFuzzy System techniques for load demand forecasting in distribution substations. The system inputs are three load demand time series, which are composed by data measured at intervals of five minutes each, during seven days, from substations located at Andradina, Ubatuba and Votuporanga. Autoregressive Integrated Moving Average models with suitable results have been analyzed, whereas several input configurations and different architectures have been investigated for Artificial Neural Networks and Adaptive Neuro-Fuzzy System techniques aiming the forecasting of twelve further steps. The results showed the Artificial Neural Network based technique superiority for such forecasting, followed by Autoregressive Integrated MovingAverage model and Adaptive Neuro-Fuzzy approach. The load demand forecasting can minimize costs of energy generation as well as improve the electric power system safety.Index Terms--Autoregressive integrated moving average processes, feedforward neural networks, fuzzy systems, intelligent systems, load forecasting, time series.

No abstract