摘要
由于航空发动机气路中传感器的数目有限,令气路健康参数的估计结果带有不确定性,为此,我们提出基于简约卡尔曼滤波器的机载自适应模型.该模型通过矩阵变换降低健康参数矩阵的维数,把简约卡尔曼滤波器的估计偏差和方差的加权和作为最小化的目标,采用自适应遗传算法构造出一个较准确,能反映发动机性能的健康参数子集.从理论上进一步论证了简约卡尔曼滤波器的估计方法.对某型涡扇发动机气路部件进行性能估计作数值仿真,结果表明,基于简约卡尔曼滤波器的机载自适应模型方法适用于在包线内传感器个数少于健康参数条件下,有效估计出发动机气路健康性能.
Because of the limited number of sensors, the health estimation results for the gas-path in a turbo-fan engine are uncertain. Based on the contracted Kalman filter, a self tuning on-board model is proposed. By using a matrix transfor- mation, we reduce the dimensions of the health parameter matrix. The weighted sum of the estimation bias and the variance of the contracted Kalman filter is employed as the object of optimization; and a subset of precise health parameters reflect- ing the engine performance in operations is obtained by using the adaptive genetic algorithm. The self tuning on-board model based on contracted Kalman filter is further proved theoretically. The simulation of a turbo-fan engine shows that the method of contracted Kalman filter with self tuning on-board model effectively estimates the health parameters when the sensor number is less than the health parameter number in the operation range.
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2012年第12期1543-1550,共8页
Control Theory & Applications
基金
中航工业产学研工程资助项目(HCA1000103)
南京航空航天大学青年创新专项基金资助项目(NZ2012110)
