Abstract

Though various control schemes have been investigated numerically for vibration control of large-scale membrane structure, there has not presented any practical method to recognize dynamic pattern from sensing signals. This study proposed a novel cable-based dynamic pattern recognition method of large-scale membrane antenna, giving the first-time practical implementation of full-state reconstruction for feedback control. Cubic polynomial function is adopted to reconstruct the pattern based on the displacements and angles of corresponding cables, according to the way vibration goes. To further improve the accuracy and continuity, convolutional correction and Kalman filter are applied. The pattern recognitions are compared in both physical and modal way via free vibration. And the performance of reconstructed shape in close-loop control is verified based on linear quadratic regular (LQR) controller. Simulations show that the proposed pattern recognition method can reasonably reconstruct the out-of-plane shape, and the suppression of vibration is as good as that in full-state-available case. With correction algorithms applied, the control force signals become realizable and practical, whose continuity is notably improved. This method introduces a novel technique to achieve full-state feedback for the first time, making it possible to realize many excellent control methods in practice.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call