电厂冷凝器清洗机器人设计与控制方法研究.pdfVIP

35 　6 ( ) Vol.35, No.6 2 0 0 8 6 Journal of Hunan Universit (Natural Sciences) Jun.2 0 0 8 :1674-2974(2008)06-0031-05 张　辉 , 王耀南, 彭金柱 ( , 　410082) 　　　:提出了面向电厂冷凝器清洗的新方法, 它以高压水清洗为主、以 学药剂清洗 为辅, 通过智能清洗机器人实现电厂冷凝器的自动 在线清洗.首先设计了水下冷凝器清洗 机器人的体系结构, 给出了清洗机器人的机械与电气结构, 在此基础上, 为实现机器人的高 效清洗, 设计了机器人的分布式控制系统, 提出了水下作业环境下二关节机器人的控制方 法, 建立了模糊高斯基神经网络控制系统结构, 通过 BP 学习算法进行神经网络训练.最后, 仿真实验表明:提出的控制方法在动态性能和稳定性方面优于传统的控制方法, 非常适用于 水下环境的清洗控制. :机器人;体系结构;神经网络;污垢清洗 :TP24 　　　　　　　　　　　　　　　:A Research on the Design and Control Strateg of Cleaning Robot for Pow er Plant Condenser ZHANG Hui , WANG Yao-nan, PENG Jin-zhu (College of Electrical and Information Engineering, Hunan Univ, Changsha, Hunan 　410082, China) 　　Abstract:A new method for online steam condenser fouling cleaning w as proposed .In this method, high pressure cleaning was chosen as the main method, chemical cleaning was applied to complement it, and all this cleaning method w as achieved b a new autonomous cleaning robot.Firstl , the robot s stem structure was designed, which consisted of a mechanical and an electrical structure.Secondl , the distributed control s stem was used to improve the cleaning effi- cienc , the cleaning control strateg based on fuzz -guass neural network was presented, and the BP algorithm was adopt- ed to train the network weights.Then, simulation results show that the proposed control strateg has better robustness and d namic performance than the traditional fuzz control, as it can not onl reduce the phenomenon of chattering in ef- fect, but also has good robustness and d namic performance.T