| The key problem of unmanned helicopters, which are usful in military and civilianoperations, lies in the flight control system development. Dynamic models are essentialfor flight control design, and analysis of characteristics and flying qualities. However,modeling unmanned helicopters are difficult because of dynamic uncertainties, etc.To solve these problems, modes partition method (MPM) is proposed in this paper:The mode identification stragety includes external characteristics modeling, modepartition model and basic model structure. Using the modes partition model, themodeling of uncertain dynamics and large amount of parameters are avoided.The basic model strucrue is made up of on-axis and off-axis models. The former areestablished from first-principles analysis of the main rotor and tail rotor system. Thelatter are depends on the inter-axes coherence function of frequency-sweep test data.With the Bell-Hiller stabilizer bar, yaw rate gyro and engine governor, the entirehelicopter behaves like SISO systems with minor coupling in trim flight. Therefore, theoff-axis models are only applied when they are obviously effective.Real-flight tests are carried out to obtain identification and verification test data. Thecoherence function shows that the obvious coupling occours in vertical maneuvers.Off-axis model are designed to compensate the heave coupling.Through GA and PSO, the dynamic models are identified by MPM, with themaximum cost function is2.9046. Time-domain verification tests are perfomed to checkthe modeling effectiveness. The predicted responses of the on-axis models show perfectagreements with the flight data. Meanwhile, the off-axis models produce minorimprovement. With the PI controller developed from the on-axis models, the error ofyaw angle during ordinary flight and maneuvers are less than±3°and±10°, respectively.MPM is capable of accurately and effectively capturing helicopter dynamics.Moreover, MPM might be helpful to improve flight control design and its accuracy. |
PDF Full Text Request