基于飞行包线划分的航空发动机T-S模糊模型辨识

针对航空发动机在建立Takagi-Sugeno(T-S)模糊模型时运算耗时长和过分依赖学习数据的问题,提出了一种基于飞行包线划分的航空发动机T-S建模方法.通过飞行包线划分和标称点求取确定T-S模型的前件结构；计算各标称点的状态空间模型,将其作为T-S模型的后件；最后通过对航空发动机发参数据的机器学习完成对模型前件参数的辨识.仿真对比结果表明:该方法缩短了航空发动机T-S模糊模型的建模时间,并使得高压转子和低压转子转速的绝对误差分别小于0.25％,0.10％,保持了辨识精度.

Identification of T-S fuzzy model for aero-engine based on flight envelop division

For the problems of time-consuming calculation and data dependence in Takagi-Sugeno(T-S) fuzzy model of aero-engine,a new T-S modeling algorithm based on flight envelop division was proposed.The premise structure of fuzzy model was confirmed by dividing the flight envelop and selecting the nominal points.The state space model at each nominal point was regarded as the consequence of T-S model.Finally,the parameters of premise structure were identified by training.Simulation results show that the new algorithm shortens the modeling time.The absolute error of high-pressure rotor speed is less than 0.25%,and that of low-pressure rotor speed is less than 0.10%.