Methodological basics of building an intelligent system for making decisions on continued operation of aircraft

Abstract

In order to improve the existing and create new innovative systems for making and making management decisions about the further operation of aircraft that have reached the end of their resource indicators, a methodical approach to preparing decision options for their adoption by responsible persons is proposed. The peculiarities of the creation and functioning of the intelligent decision-making system in the real conditions of the operation of military aviation equipment are considered. On the example, using statistical mathematical models and methods of fuzzy logical analysis, it is shown the possibilities of increasing the sample volume in classification tasks, the peculiarities of forecasting changes in the technical state, the order and phasing of the preparation of decision options.

The statistical mathematical model involves the use of methods of cluster analysis and determination of the equation of the regression line. As a result of the cluster analysis, a list of aircraft of the main group is determined, which have anticipatory values of the current calendar terms of service and resource relative to the aircraft, which are being investigated for the possibility of continued operation. In the process of performing the cluster analysis, methods of fuzzy comparative logical analysis of the operation data of the aircraft of the main group with the data of the aircraft under investigation were used. Making a decision on the possibility of extending the resource indicators of the service life and flight time involves the determination of the angular coefficient of the lower limit of the linear regression of the current values of the service life and flight time, which have the airframe designs of the aircraft of the main group after their last control inspection. In the case of a small number of aircraft in the main group, the SMOTE method was used , which allows you to increase the number by adding conditional synthesized aircraft. The angular coefficient of the forecast line is calculated as the result of half the sum of the angular coefficient of the regression line and the experts' forecast line, regarding the possibility of continued exploitation.