| 摘要: |
| 针对火电厂锅炉过热汽温对象，将神经网络辨识技术和自适应逆控制技术相结合，提出了一种过热汽温自适应逆控制方案。该方案首先利用RBF神经网络在线辨识被控对象模型获得其Jacobian信息，在此基础上利用对角回归神经网络（DRNN）在线辨识获得被控对象的逆模型作为控制器，串联在控制对象前面构成自适应逆控制系统。通过对超临界600 MW机组过热汽温对象进行仿真研究表明，该控制方案能很好地适应过热汽温对象特性的变化，并且可以有效克服对象的大惯性和非线性，获得良好的控制品质。 |
| 关键词: 锅炉 自适应逆控制 过热汽温 DRNN Jacobian信息 串级控制 神经网络 |
| DOI： |
| 分类号: |
| 基金项目:国家自然科学基金资助项目（60974005） |
|
| Adaptive inverse control of superheated steam temperature |
|
WANG Wanzhao, WANG Jie
|
|
School of Electrical Engineering，Zhengzhou University，Zhengzhou 450001，China
|
| Abstract: |
| A scheme combining neural network identification technology and adaptive inverse control technology is proposed for the control of boiler superheated steam temperature in fossil-fired power plant，which applies the RBF neural network to online identify the Jacobian information of the controlled object and then the ＤＲＮＮ（Diagonal Recurrent Neural Network） to online identify the inverse model of the controlled object. The identified inverse model，as a controller，is set before and connected in series with the controlled object to form an adaptive inverse control system. Simulation with the superheated steam temperature of supercritical 600 MW unit as the controlled object shows that，the proposed control scheme，with excellent control quality，adapts well to the variable properties of object and overcomes its large inertia and nonlinearity effectively. |
| Key words: boilers adaptive inverse control superheated steam temperature DRNN Jacobian information cascading control neural networks |
