The Intelligent Forecast System (IFS) consists of three units: the main unit is a multilayer neural network1 the second unit is based on fuzzy logic, which is used to prepare the data1 and the third unit is based on a genetic algorithm and is used to determine the best set of data to be analyzed. The data were set up by using a strategy that classifies the terms of the series into parts with the same quantity of elements. Each part is called a window. The IFS was trained by using a group of windows, which is called a set of training windows. Next, the IFS was applied to solve several problems related to the evaluation of structural integrity. The IFS was applied as an identification strategy and forecast of parameters in a flexible structure. Additionally, by using the IFS we have developed a new strategy to analyze the temporary series obtained from vibrations of flexible structures with the distribution of variable mass. The results obtained in this research, using the IFS applied as an identification strategy and forecast of parameters in a flexible structure, show the effectiveness of the IFS.

Experimental results obtained from Artificial Neural Network (ANN), Fuzzy Set and GeneticAlgorithms of an Intelligent Forecast System that is currently being developed for prediction and analysis of time series.