Prediction of vertical tail maneuver loads using backpropagation neural networks

Federal aviation regulations require that structures critical to the safe o peration of an aircraft must not fail within their expected lifetimes due t o damage caused by the repeated loads typical to its operations. A backprop agation neural network has been used to predict maneuver-induced strains in the vertical tail spar of a Cessna 172P. Linear accelerometer, angular acc elerometer, rate gyro, and strain gauge signals were collected during fligh ts using a portable data acquisition system for Dutch roll, roll, sideslip, level turn, and push-pull maneuvers. Sensor signals were filtered and used to train the network. The strains in the vertical tail spar were predicted successfully by the network to within 50 mu epsilon of their strain gauge values. This is an inexpensive and effective technique for collecting verti cal tail load spectra for small transport airplanes already in service wher e installation of strain gauges are impractical.