Flight Data Reduction via Rough Set

Abstract

It is of great importance that the maintenance of the aircraft is carried out in accordance with the procedures in order to ensure a safe flight. Monitoring of the health conditions of the engine from the aircraft components, preventive maintenance methods to take measures to prevent the occurrence of major failures are applied. The data recorded during the flight in the airplanes are evaluated on the ground and the engine maintenance is done as well as the engine evaluation of the data during the flight, and there are also some studies that determine the maintenance methods. The parameters which can give information about the health conditions of the engine are evaluated together with the parameter groups that affect them and the possible fault conditions are emphasized. One of the parameters used in decision making in order to monitor engine health conditions is exhaust gas temperature (EGT) value. In this study, some properties in the data are reduced by using the rough set method on the aircraft data which is considered as a pre-process in data mining methods. In this study, data sets of various qualities were obtained. After the obtained datasets are separated according to certain criteria, the fault prediction operation will be carried out. With this study, the preliminary stage of the fault prediction step to be performed in the next step has been realized.