摘要
为提高燃气轮机状态监测中传感器的有效性,提出了一种用小波分析的传感器故障诊断方法。该方法通过对传感器输出信号进行小波多分辨分析(MRA),利用其表征信号趋势和细节的特点,能对缓变故障和突变故障进行准确定位,同时根据故障突变点的Lipschitz指数和故障前后能量的变化特征实现对突变故障类型的判别。对燃气轮机排气温度实时数据进行的故障仿真,表明了该方法的有效性。
For the purpose of increasing the effectiveness of sensors, used for monitoring the service state of gas turbines, a fault diagnosing method, using sensors in connection with a kind of wavelet analysis, is being proposed. With this method, by Multi Resolution Analysing （MRA） the sensor＇s output signals, drifting incipient faults and abruptly occurring faults can both be identified by making use of a signal＇ s apparent tendency and its details. The type of abrupt fault concerned can simultaneously be distinguished by using Lipschitz＇ s index, according to the fault＇s point of sudden change, and also by noticing the change in energy, triggered off by the fault. Fault simulation studies, using real time data of gas turbine exhaust gas temperatures, indicate the effectiveness of the proposed method. Figs 6, table 1 and refs 7.
出处
《动力工程》
EI
CSCD
北大核心
2006年第2期245-248,299,共5页
Power Engineering
关键词
动力机械工程
燃气轮机
传感器
小波变换
多分辨分析(MRA)
故障诊断
power and mechanical engineering
gas turbine
sensor
wavelet transformation
multi resolution analysis
fault diagnosis
