摘要
一台支持向量机器(SVM ) 近似底的内部模型控制(IMC ) 策略为同步发电机的蒸气 valving 控制被介绍。建议 SVM IMC 策略包括二主要部分：在内部模型结构的 SVM 近似反的控制器和无常赔偿。SVM 反的控制器经由泰勒扩大用一条输入产量近似途径直接被导出，并且没有进一步的联机训练，它通过非线性的系统鉴定被实现。而且，一个坚韧性过滤器在内部模型结构被用于无常赔偿。模拟为蒸气 valving 控制显示出 SVM IMC 策略的有效性。
A support vector machine （SVM） approximate-based internal model control （IMC） strategy is presented for the steam valving control of synchronous generators. The proposed SVM IMC strategy includes two main parts： SVM approximate inverse controller and uncertainty compensation in the internal model structure. The SVM inverse controller is derived directly using an input-output approximation approach via Taylor expansion, and it is implemented through nonlinear system identification without further online training. Furthermore, a robustness filter is used for uncertainty compensation in the internal model structure. Simulations show the effectiveness of the SVM IMC strategy for the steam valving control.
出处
《自动化学报》
EI
CSCD
北大核心
2008年第2期172-179,共8页
Acta Automatica Sinica
基金
Supported by National Natural Science Foundation of China(60375001,60775047)
关键词
内模控制
非线性控制
支持向量机
神经网络
逼近模型
Nonlinear control, internal model control （IMC）, support vector machines, neural networks, approximate models
