航空发动机控制系统执行机构参数在线估计方法

Online actuator parameter estimation method for aero-engine control system

针对航空发动机控制系统因为执行机构性能退化而导致发动机控制品质降低或者严重时威胁发动机运行安全的状况,开展执行机构状态参数在线估计方法研究。在航空发动机实际执行机构控制回路实测信号较少的情况下,提出一种组合参数在线估计方法,通过作动模式识别分类,基于无迹卡尔曼滤波以稳态模式输出估计电液伺服阀平衡电流,基于拟牛顿算法(BFGS)以动态模式输出估计执行机构增益和作动延迟时间,实现模型参数的在线更新,建立实时自适应执行机构模型。以某涡扇发动机导叶作动控制回路为对象进行仿真,结果表明:在只有作动位置单一参数可测的条件下,在不同作动状态下对执行机构控制回路的平衡电流估计误差优于±0.2 mA,执行机构增益估计误差优于±4%,作动延迟周期估计误差不超过1个控制周期,能够实时跟踪并较为准确地估计执行机构的工作状态,为航空发动机执行机构控制回路设计与故障诊断提供技术支撑。

In order to solve the problem of performance degradation that consequently reduces the engine control quality and even endangers the engine operation safety,a method of estimating the real-time state and performance variation trend for the actuator was proposed.Considering the measurable signals of practical aero-engine servo actuator were less than the performance parameters,a adaptive estimation method of combined state was put forward by means of recognition and classification of the actuation pattern,the balanced current of electro-hydraulic servo in the steady state was estimated by unscented Kalman filter,the actuation gain and actuation time delay in the dynamic state were estimated by the method of Broyden-Fletcher-Goldforb-Shanno(BFGS),the performance parameters were updated in real time,and the adaptive model of servo actuator was established.A vane servo actuator loop of turbofan engine was simulated.The simulation results showed that,when single servo parameter can be measured,the absolute error of balance current estimation was less than ±0.2 mA,the relative error of actuation gain estimation was less than ±4%,and the absolute error of actuation delay period estimation was less than one control period in different actuation states,and the adaptive model can estimate the state of actuator accurately and track the performance variation trend in real time,so the method can provide technical support of the control loop design and fault diagnosis for aero-engine servo actuator.