RIPTIDE

Real-time Interactive Prototype Technology Integration/Development Environment

Mathematical Model Development / Improvement

RIPTIDE provides a readily available real time, pilot-in-the-loop, simulation environment for evaluating mathematical models as they are developed and / or improved. Any type of mathematical model, from simple linearized models to high-fidelity blade-element type models, can be used, the requirement being the ability to read from and write to shared memory.

A simple linear model, known as the Enhanced Stability Derivative (ESD) model, is part of the base RIPTIDE software package. This model provides the basic feel of helicopter flight at hover and in forward flight while giving the user the ability to modify the flight characteristics by simply changing the stability derivatives. Furthermore, the >derivatives can be changed in real time through a GUI specifically designed for this purpose.

In addition to ESD, a blade-element model of the UH-60 Black Hawk (Gen Hel) has also been implemented. The following sections on model validation and model off-axes response improvement apply to this more sophisticated model.

Model Validation

Within the RIPTIDE environment, a frequency-domain approach is used to evaluate model responses and to compare these responses with flight data. The approach basically mimics the process of frequency sweep testing of an actual aircraft. Either pilot- or computer-generated sweeps can be used. In the case of computer generated sweeps, white noise can be added to the fundamental signal to improve spectral content. White noise can also be specified as the input to the 3 remaining controls.

In batch mode the sweeps are input to the model, one axis at a time, and model responses recorded. The main difficulty with running frequency sweeps through an unpiloted (batch mode only) simulation model is maintaining attitude and airspeed close to initial trim. Given the long duration of a typical sweep (90-120 seconds), some additional control has to be provided. This is added in the form of low-gain rate and attitude feedback loops on roll, pitch, and yaw. These loops have no effect on the extracted dynamic response obtained from multi-input/multi-output spectral analysis, since the frequency responses are based on the total input to the mixer.

Model time histories generated above are processed through the Comprehensive Identification from FrEquency Responses (CIFER) tool to generate Bode plots for comparison with flight data.

Off-axis Response Improvement

Until recently, one of the major weaknesses of existing math models had been their inability to correctly predict the off-axis response of single-rotor helicopters. Existing models not only failed to give an accurate prediction of the phenomenon, but they even failed to correctly predict the direction of the off-axis response.

Modeling and analysis efforts at Ames Research Center and Princeton University have shown that the cross-coupling discrepancy is associated with an error in the phasing of the fixed-frame aerodynamic moments of the rotor. An aerodynamic phase lag correction is highly effective in improving the off-axis correlation without affecting the on-axis response. The aerodynamic phase lag empirically accounts for the integrated effect of two distinct physical phenomena on the off-axes response of helicopters. One contribution to this aero phase lag is the geometric distortion of the shed wake during pitch or roll motion of the aircraft/tip-path-plane combination , as was discussed above. The second contribution is associated with the effect of compressibility on the 2-d unsteady shed wake (Theodorsen) response. The resulting total aerodynamic phase lag is implemented in simulation via a first-order filter applied to the rotating frame aerodynamic components.

Some Model Improvement Results

• Model Responses with Off-axes Correction
  • On-axis response
  • Off-axis response