Estudo de identificação de sistemas via algoritmo de realização de auto-sistemas

Abstract

The study of mechanical vibrations has always been a very comprehensive area of engineering, since almost everything in our modern post-industrial revolution constitu- tion is relative there is a machine or structure movement, so such an area is discussed from mathematical constructs in literatures and their models, to the constitutions of new techniques of experimental analysis, often going through numerical and computational studies through the prior construction of a system model. The importance of identifying a system, be it its dynamic characteristics or its vibrational behavior, so called modal analysis, gives us space to remodel a system or even apply control techniques on it, this is very used in the aeronautical, automotive industries , in structural studies, among others. After all, the main idea is to avoid failure, either by modifying your model or by controlling it. With the evolution of the new technologies, it is not only possible to assist precise models and numerical simulations as it is possible to use experimental tech- niques, but in the academic field, using such technology may not be easily accessible due to the costs of materials, whether they are sensors, acquisition systems or even specific programs for analysis. This work extends juxtaposedly there is a technique developed for modal analysis (dynamic parameter analysis) an inexpensive way of doing this th- rough the use of low cost technologies, acquisition system and sensors, as well as the use of open source programming for the treatment computational. The technique used for dynamic identification and main scope of this work is the ERA-Eignsystem Realization Algorithm, a mathematical algorithm developed by a former engineer from Nasa Juang, JN (NASA Langley Research Center, Hampton, VA, United States), which was developed through computational mathematics for use in identifying dynamic parameters of flexible structures, and then extended to control systems. Thus, the mathematical development will be attributed to this work for the construction of the model identification algorithm, which can later be used for both modal analysis and systems control, as well as accessible experimental use.