Projeto de experimento para análise de vibração em aeronaves de asas rotativas

Abstract

This work aims to study the Design of Experiments methodology, using Latin Squares in rotorcraft vibration analysis. This approach is due to the main rotor effect in forward flight condition for initial employment as a maintenance predictive control tool and future developments of recording systems. So it is necessary to identify the frequencies of interest that are transmitted to the fuselage and to develop the experiment plan, data monitoring and analysis, applying a statistical model based on analysis of variance, assuming a 95% confidence interval in the results. The Design of Experiments itself covers three basic principles in order to ensure data trends exemption: randomization, replication and blocking. It is proposed a method for vibration analysis presenting steps to be followed allowing its development in a systematic sequence. In this context, the factors that influence the process are identified along with its range and it is employed a tool to eliminate the influence of unwanted factors, leaving only the known effect of speed. Then checks are made with the purpose to validate the data recorded leading to consistent analysis. The implementation of the proposed method is performed by a computer model, using the "R" software, and rating for the increase of vibration in Eurocopter AS-355 aircraft. It also brings a preliminary study for Sikorsky Sea King aircraft, showing the usage and the method flexibility in different aircraft characteristics, and its conclusion is interesting for future studies.