Parametrics diagnostics of single-GTE on the basis of neural network modeling workflows

Abstract

The problem of increasing the efficiency of the parametric diagnostic single-shaft gas turbine engine, based on neural network modeling of its workflows. Spend a selection of the type of neural network diagnostic model, as well as a study of the influence, the efficiency of diagnosis of the condition: the number of the analyzed NA models from which conducted the choice of optimal diagnostic model, the type of neural network model, the volume of training sample, the number and the list of controlled parameters of the engine. Recommendations on the use of software "Gasturb" and "Statistica Neural Networks".