應用於直流馬達速度控制之灰預測器

Abstract

本篇論文主要在驗證灰預測器在雜訊預測以及應用於直流馬達速度控制上的效能。在此我們採取順滑模態控制來做直流馬達的速度控制，因其對消除匹配式雜訊有很好的效果；但是傳統的順滑模態控制需在匹配式雜訊的上限值已知的前提下才能有如此好的性能。然而，雜訊的上限值是很難準確的量測或估算出來的，而要是雜訊的上限值估算得不準的話，整個系統就會產生不穩定的現象。可想而知，若是每一點的雜訊大小都能預測得到的話，那我們就可以直接對雜訊作處理而不必理會整個雜訊的上限值了。在論文中我們以多種不同組合的雜訊來做灰預測器的效能評估，並且與線性迴歸的方法作比較，結果顯示灰預測器對於較低頻或較沒有劇烈變化的雜訊有蠻好的效果。同樣地，在應用到順滑模態控制方面的模擬結果也顯示出這樣的結合確實可以解決估算雜訊上限值的問題。

This thesis verifies the efficiencies of grey predictor on the prediction of the disturbances and the application to speed control of DC motor. Sliding-mode control is chosen here for speed control of DC motor because of its robustness of eliminating the matched disturbances. The traditional sliding-mode control has such property with the hypothesis that the upper bound of the disturbances is known. However, the upper bound of the disturbance is difficult to be measured or estimated exactly, and the system would be unstable if the upper bound is not well estimated. It is obvious that if the value of the disturbances could be predicted well, the upper bound would not be necessary information any more. In the thesis, the performances of grey predictor are evaluated with several distinct disturbances and compared with the performances of linear regression method. The simulation results show that grey predictor has well performances for the disturbances with lower frequency or less rapid variations, and it can solve the problem of estimating the upper bound of the disturbances.