Формирование базы данных характеристик отрывных диффузоров камер сгорания ГТД с целью использования технологий машинного обучения

Abstract

Preparation of high-quality data is a key stage of training artificial intelligence models, directly affecting their accuracy and efficiency. The paper considers the method of database generation for designing a gas turbine engine combustion chamber blow-off diffuser using numerical simulation and optimization in Ansys Fluent. To reduce the calculation time, the number of geometric parameters is reduced from five to three by calculating one opening angle and introducing the assumption of equality of areas in the cross section. The proposed approach allows us to accelerate the creation of training data sets and expand the application of AI in the design of gas turbine engines.