High-fidelity modeling of redundant EMA and small angle displacement analysis under different controllers

An increasing number of applications of the electromechanical actuator (EMA) in the flight vehicle control system have required accurate dynamic models and control strategies. This paper investigates an EMA system for guided projectiles characterized by a high power-to-weight ratio, high efficiency, and high integration. First, the working principle of the EMA is analyzed, along with the structural composition of its mechanical and electrical components. A detailed nonlinear dynamic model of the EMA is established, and internal nonlinearities such as friction and backlash are investigated. Subsequently, four intelligent controllers, namely, the fuzzy adaptive (FA)-proportional integral derivative (PID), single neuron (SN)-PID, back propagation (BP)-PID, and radial basis function (RBF)-PID, are designed and implemented in the automatic position regulator of the three-closed-loop EMA control system. In addition, simulations are conducted to compare the response tracking performance of the EMA for small angle step and sine displacement under different controllers. The results indicate that with the RBF-PID controller, the rise time of the 1° step displacement response is ∼318.950 ms, representing reductions of about 0.626%, 0.674%, 0.795%, and 0.981% compared with the BP-PID, SN-PID, FA-PID, and conventional PID controllers, respectively. Moreover, with the RBF-PID controller, the amplitude error (AE) in tracking the 1°-1 Hz sine displacement is ∼3.65 × 10−3, representing reductions of about 3.865%, 5.276%, 7.454%, and 10.254% compared with the BP-PID, SN-PID, FA-PID, and conventional PID controllers, respectively. Finally, experiments validate the command transmission delay time. Results from small angle step and sine displacement tests demonstrate the rapidity, stability, and robustness of the artificial intelligence strategies.