超磁歪材アクチュエータを有する直動型サーボ弁のモデリングとゲインスケジューリングによるスプール位置制御

抄録

Because of the reasons of favorable features such as high driving force, fast response, fine environmental durability and big distortion about 10 times more than that of PZT, GMM (Giant Magnetrostrictive Material) is suitable for the actuation structure of servovalve. But, GMM distortion depends on the magnitude of coil current and the output-gain declines according to decreasing current magnitude. Then, the dead-zone appears in small current range. One of the important characteristics of servovalve is the fine tracking performance to the reference signal. So far, we attempted to apply linear controllers, such as PI or H^∞controller. However GMM tracking performance is inferior to the traditional driving device for reference signal around null area, and this flaw remains as an important matter.
In this paper, adopting LPV (Linear Parameter-Varying) system modeling in which we regard the magnitude of input signal as a scheduling parameter, we design a gain scheduling controller and attempt to solve the problem that is caused by the nonlinearities mentioned above. Since it is hard to implement the controller which uses the information on the varying rate of scheduling parameters, we propose a new approach for output-feedback controller design. First, we describe the plant modeling as an LPV system, then present an outline of the controller design synthesis without any information on the varying rate of scheduling parameter and the design process. Lastly, we show the usefulness of the proposed controller by experimental results.