| Electric load simulator (ELS) is an important test device of the hardware-in-the-loop simulation for testing flight control systems. The task of ELS is to simulate the aerodynamic hinge moment the servo object bears under actual conditions, in order to test the performance of the servo system. Based on a number of practical engineering projects, this thesis studies the key technologies and theoretical issues of electric load simulator. Then, some creative work is carried out and several novel methods are proposed.First of all, the mathematical model of electric and hydraulic system is compared, and the equivalence of the two systems on the macroscopic dynamics is demonstrated. The equivalent relationship between the physical meaning of electric and hydraulic parameters is summed up too. Based on the equivalent relationship between armature voltage and load flow, the ratio of forced load flow (or back electromotive force) and valve-controlled load flow (or armature voltage) is deduced. This ratio, called as the "Passive-Active Flow Ratio", plays a significant role in comprehensive performance comparison of electric and electro-hydraulic load simulators.Secondly, in order to obtain the analytical expressions of its dynamic response performance parameters, the electric load simulator transfer function is simplified. After that, the effect of the motor parameters on the stability and performance of the loading system is clarified. This provides a more specific theoretical basis for designing electric load simulator. On this basis, a complete mathematical model of the electric load simulator is established, and the adjustment methods for loading system stability margin and the "Double Ten" bandwidth is proposed.Thirdly, the Bode diagram of the frequency response characteristics for the position disturbance is drawn. Through analyzing the gain and slope of the frequency characteristics curve, it is found that the surplus torque for the electric load simulator is essentially superposed by the synthesis friction load, inertial load and elastic position load. After spectrum analysis of the position disturbance signals and surplus torque signal, it is able to get the actual dynamic response characteristics of the position disturbance. The consistency between the experimental results and simulation results of the mathematical model proves the correctness of the theoretical analysis.Thereafter, in view of the fact that the output torque of the electric load simulator is in compliance with linear superposition principle, this thesis proposes a vector matching relationship-based disturbance elimination strategy. And according to the periodic of the motion of the tested object, the single vector matching method (SVM) and the compound vector matching method (CVM) are proposed, which are collectively referred to as vector matching based position disturbance torque elimination method. SVM is mainly used to suppress the sinusoidal perturbation position, while CVM is capable of suppressing more universal periodic location disturbance. Experiments are performed to verify the effectiveness of SVM and CVM. Experimental results vividly demonstrate that the vector matching method can not only eliminate surplus torque caused by the position disturbance, but also compensate for the load errors caused by the loading system phase lag. The proposed method improves the accuracy of dynamic loading for load simulator in essence.Finally, the insulation couplings design successfully blocks bearing voltage coupling channel and enhances the signal to noise ratio of the sensor signal. After intensive study on the Widrow-Hoff network, an improved adaptive linear neural network (RLS-ADALINE) is proposed. It is proved that the filter designed based on RLS-ADALINE network has better performance. To meet the practical requirement of measuring the speed and acceleration, the thesis proposes a novel sinusoidal excitation method to increase the efficiency of the calibration and accuracy of the sensor measurement. |
PDF Full Text Request