Active Compensation of Reflector Antenna Structural Deformation Based on Deep BP-Neural Network

The active reflector is a method used to compensate for the surface deformation of the reflector due to gravity and other factors, which will lead to the degradation of the electromagnetic performance of the antenna. The time-invariant surface deformation of large antenna caused by wind cannot be compensated quickly and effectively. A method based on deep BP neural network (DBPNN) is proposed to quickly obtain the adjustment of active reflector actuators. The mapping relationship between the deformation of the primary reflector and the adjustment of thousands of actuators is established. In order to ensure the unity of data dimension, the relation matrix of the actuators and panels is established. By processing the deformation data of the primary reflector, the deformation and position characteristics of the panels are obtained, and the network parameters are trained by the Adam optimizer. The results show that the proposed method can obtain thousands of actuator adjustments at different elevation angles very quickly under the premise of meeting the accuracy requirements. Compared with the traditional numerical algorithm, this method is suitable for compensating the loss performance at all elevation angles, and the time of calculation is reduced significantly.