Neural network research on validating experimental tilt-rotor performance

Results from a neural network study to validate full-scale XV-15 tilt-rotor experimental hover and forward flight data are presented. In the context of the present study, neural-network-based test data validation includes the following: test data representation, test data quality assessment (e.g., isolating "bad" test points), and finally, for outdoor hover measurements, wind correction procedure development. Two test databases, acquired during separate tests conducted at NASA Ames, were used. These two isolated XV-15 rotor test databases were obtained from tests in the 80- by 120-Foot Wind Tunnel and an outdoor hover test facility. Neural networks were successfully used to represent and assess the quality of full-scale tilt-rotor hover and forward flight performance test data. The neural networks accurately captured tilt-rotor performance at steady operating conditions and it was shown that the wind tunnel forward flight performance test data were generally of very high quality. Compared to existing momentum-theory based wind corrections to outdoor hover performance, the present neural-network-procedure-based corrections were better. The present wind corrections procedure, based on a well-trained neural network, captured physical trends present in the outdoor hover test data that had been missed by the existing momentum-theory method.