Electric load forecasting is very important for planning and operation of power systems. Knowing the load in advance aids in electrical system performance, i.e. providing energy with quality, security and reliability. The traditional statistical methods available in the literature as for example ARIMA of Box&Jenkins are widely used for solving the prediction problems, however Artificial Neural Networks are surpassing them. There are several architectures of Neural Networks that can be used in many applications, nevertheless mixing neural networks with other artificial intelligences (fuzzy logic, genetic algorithms, etc.) results powerful architectures than conventional ones. This work aims to use two different techniques, such as Artificial Neural Networks (ANN) and Fuzzy Logic to develop a neuro-fuzzy network and apply to short-term load forecasting problem. There are many papers in the literature using the MATLAB toolboxes, however the main objective of this work is to develop the algorithm in a computational language such as Fortran, C++, or even MATLAB to implement the neuro-fuzzy network applying to short term load forecasting.
News published in Agência FAPESP Newsletter about the scholarship: