The prediction of nonlinear dynamic loads on helicopters from flight variables using artificial neural networksA method of indirectly monitoring component loads through common flight variables is proposed which requires an accurate model of the underlying nonlinear relationships. An artificial neural network (ANN) model learns relationships through exposure to a database of flight variable records and corresponding load histories from an instrumented military helicopter undergoing standard maneuvers. The ANN model, utilizing eight standard flight variables as inputs, is trained to predict normalized time-varying mean and oscillatory loads on two critical components over a range of seven maneuvers. Both interpolative and extrapolative capabilities are demonstrated with agreement between predicted and measured loads on the order of 90 percent to 95 percent. This work justifies pursuing the ANN method of predicting loads from flight variables.
Document ID
19930035234
Acquisition Source
Legacy CDMS
Document Type
Conference Paper
Authors
Cook, A. B. (NASA Langley Research Center Hampton, VA, United States)
Fuller, C. R. (NASA Langley Research Center Hampton, VA, United States)
O'Brien, W. F. (NASA Langley Research Center Hampton, VA, United States)
Cabell, R. H. (Virginia Polytechnic Inst. and State Univ. Blacksburg, United States)