Electric power systems load forecasting: a survey

This work reviews the latest works on load forecasting, classifying them according to presented methods and models, as statistical, intelligent systems, neural networks and fuzzy logic. As there are many different models and methods, we have studied the principal ones considering classical statistical and modern methods like neural networks and fuzzy logic. We emphasize the importance of load forecasting in modern operation centers, and conclude that nowadays the most used forecasting methods are the neural networks and fuzzy logic, but the statistical ones continue being used although to a lesser extent.