汽轮机故障在线智能诊断系统  

Online Intelligent Diagnosis System about Turbine Fault

作　　者：周洪煜[1] 常立民[1] 丁玉栋[1] 周爱红[1] 

机构地区：[1]重庆大学动力工程学院,重庆400044

出　　处：《控制工程》2003年第6期492-494,514,共4页Control Engineering of China

摘　　要：叙述了汽轮机故障诊断系统的应用必要性。根据系统功能,采用了分布式上下位机网络结构,给出了硬件设备的组成,上下位机之间以RS 232协议通信。提出了系统诊断的总体结构和诊断流程;讨论了反向传播神经网络和专家系统这两种诊断算法的诊断原理与诊断流程;探讨了BP网络激活函数的改进;介绍了应用于汽轮机几种典型故障的情况;编制出了具有丰富功能的上位机诊断软件。在转子模拟实验台上进行的实验表明,该系统稳定可靠,有较好的使用价值。The application necessity of diagnosis system about turbine fault is introduced. According to the function distributed system is applied.Communication protocol between updown is RS232.Whole structure and procedure of system diagnosis is put forward.It is emphasized the application of two kinds of diagnosis methods:back propagation and expert system.Improvement of BP networks cell function is introduced.In the end,diagnosis software with enriched function is developed. The experiment proves that this system is reliable and practical.

关 键 词：汽轮机 转子 状态监测 单片机 故障诊断 在线智能诊断系统 神经网络 专家系统 人工智能 

分 类 号：TK267[动力工程及工程热物理—动力机械及工程] TP18[自动化与计算机技术—控制理论与控制工程]