A Development Of The Automatic Device For Local Calibration Of The Helicopter Test Stand Based On Artificial Neural Network

In order to describe the periodicity movement degree of blade attack angle, three control angles which were named collective pitch,lognitudinal periodic pitch and transverse periodic pitch were used by helicopter full periodic pitching control system. And they were also used by helicopter rotor model test to show the test states of rotor model. For the sake of improving the control precision and veracity of rotor model test, it's important to obtain the transform relations between control angles and displacement of actuators of rotor control system.Former local calibration of rotor model control system has following shortcomings: the instrument of measure is laggard, the degree of automatization is low, the efficiency is poor, the controls is complicated, too many invalid data influence the result of the calibration, it is incompatible between precision and efficiency, so it is necessary that the automatic device for local calibration is developed.Depending on the condition of helicopter test stand in existence, the development solves the problems in measure, data acquisition and transfers, control, calculate parts. In the measure part, using attack angle transducer as measuring method, the speed and precision of measure is enhanced. The problem of how to acquire azimuth is solved through dynamic calibration. The effect of circumgyratetion is eliminated through attack angle transducer angular velocity revising. The problem of how to measure rotary signal is solved through slip ring.In the data acquisition and transfers part, the acquisition mode that have external clock and external trigger function at the same time insures the veracity of pitch angle measure. The technique of pipelining alignment insures the real-time of control angle measure. The technique of Winsock and multithreading insures the speediness of data transfers.In the control part, using the bypast result as swatch, artificial neural network control system is built, it is achieved that controls used control angle as independent variable when conversion expressions is unknown.In the calculate part, it obviates the complicated operation of founding zero point by splitting the nonstandard overdetermined equations and solving it with generalized inverse matrix.By ameliorating the each part, the automatic device for local calibration of helicopter test stand is developed, and finally obtained the aims of improving test efficiency , advance test abilities, and ameliorate system automation degree.