Estimation of Rotorcraft Limit Envelopes Using Neural Networks

Helicopter operation is constrained by complex limit envelopes that are difficult to predict and poorly enunciated to the pilot. The ability to sense and automatically avoid these limits can protect the rotorcraft from loss of control, avoid catastrophic failure, reduce the number of minor exceedances that degrade airframe fatigue life, and permit the pilot to safely fly the aircraft closer to the edge of the envelope. The development and operation of a neural network- based rotorcraft limit detection system that generates rotorcraft limit avoidance data using on-board measurements is discussed. The system combines manufacturer’s envelope limit data together with adaptively determined envelope limits on-board the aircraft. Both time-varying and time-invariant limits are included in the neural network. The operation of the limit detection neural network is demonstrated using data obtained from six-degree-of-freedom manned rotorcraft simulations.