基于核主元分析方法的燃气涡轮发动机燃烧系统实时监测与诊断

针对燃气涡轮发动机燃烧室状态监测方法不足,故障定位难和故障早期发现难的问题,以涡轮排气温度场周向数据为分析依据,通过研究燃气在涡轮通流部分的偏转规律,利用核主元分析(KPCA)方法对经过有效性处理后的温度场数据进行分析,并结合两台发动机的故障数据,分别对燃烧系统自身故障和热电偶传感器故障进行检测与识别,验证了排气温度场燃气偏转规律与核主元分析相结合的方法对燃烧系统故障和传感器故障进行诊断的有效性.结果表明:该方法能够将安装了环管式分布火焰筒的燃气涡轮发动机燃烧室的故障诊断定位层次从目前的燃烧室这个大部件提高到火焰筒级别的小部件. 

Real-time monitor and diagnosis of combustion system based on kernel principal component analysis

According to the shortage of gas turbine engine combustor condition monitoring method, fault location and incipient fault detection, the low-turbine exhaust circumferential temperature field data were utilized to study the deflection law of gas in turbine through flow region, and the normalized temperature field data were analyzed by taking advantage of kernel principal component analysis (KPCA) method. Based on two engines'' fault data, the fault detection and identification for the combustor fault and thermocouple sensor were developed respectively. The method of combining exhaust temperature field deflection law with kernel principal component analysis method to study combustor camber fault and thermocouple sensor fault was proved to be effective. Results show that, this method improves the fault orientation level from combustion level to camber level for gas turbine engine combustor with loop tube camber.