汽轮机通流部分在线故障诊断系统开发

介绍了一个汽轮机在线诊断系统，系统以VC 6.0作为前台开发工具，结合MS SQL Server作为后台数据库服务器，以隶属度函数具有自适应学习能力的模糊神经网络作为故障诊断模型，通过隶属度函数的在线自适应学习，在一定程度上能够反映汽轮机组热力参数的波动性，可以避免由于隶属度函数选取不当引起的误诊断和漏诊断的缺陷。该诊断方法不仅可以诊断出故障的类型，还可以诊断出故障的严重程度。经过电厂实际运行的检验，该系统运行稳定、结果正确，对提高电厂的安全运行起到了重要作用。

DEVELOPMENT OF ON-LINE FAULT DIAGNOSIS SYSTEM OF STEAM TURBINE FLOW PASSAGE

An on-line diagnosis system of steam turbine is developed by using VC++ 6.0 as its development tool and MS SQL Server as its database server. As for diagnosis model, the fuzzy neural network based on membership function with self-adaptation is put forward. To some extent, the fluctuation of the thermal parameter in steam turbine can be reflected by self-adaptation, so that the error and missing of the results caused by incorrect membership function can be avoided effectively. This model can not only diagnose fault types, but also indicate the fault degree. The practical test in power plant shows it runs stably as well as correctly and plays an important part in the safe operation of power plant.