Renewed Interest in Hinge Moment Models for Failure Detection and Isolation

Current advances in reconfiguration techniques are encouraging engineers to reevaluate the usefulness of hinge moment models for failure isolation and estimation of control effectiveness. Analytic redundancy is currently used to achieve high reliability with minimal redundant hardware in fault tolerant systems. The functional relationships are subject to modeling errors including nonlinearities, unknown dynamics, parameter variations, and noise corruptions. Aircraft hinge moment models are especially inaccurate due to the difficulty of accurately predicting the effect of aerodynamic flow field interactions from first principles. For these reasons hinge moment models are not often used for flight control actuator/surface redundancy management, particularly not for effectiveness estimation. Improvements in empirical modeling techniques have the potential to greatly improve the prediction capability of hinge moment residuals. Even an imperfect estimate of surface hinge moments would greatly reduce false alarms, provide a useful measurement of control surface effectiveness, and supply multiple simultaneous damage isolation information. A hinge moment model was developed for the direct lift flap of the Total In Flight Simulator from data taken during a recent flight test. A typical linear trim point hinge moment model did not account for the significant oscillations in differential pressure due to nonlinear effects. An Algorithm for Synthesis of Polynomial Networks was used to construct a nonlinear model of differential pressure. Using readily measurable flight variables, the new model residuals have a smaller mean square error. Thus, the estimate of hinge moment is more accurate, leading to better detection, isolation, and estimation capability.