涡轴发动机燃烧发生器热力过程控制研究

Study on Thermal Process Control of Turbine Engine Combustion Generator

涡轴发动机由燃气发生器与动力涡轮组成,燃气发生器由于进口速度畸变和燃油脉动等因素的存在,可以看作内环含有滞后环节的串级时变系统。而对于被控对象参数未知且时变的情况,仅用Smith算法调节不能有效补偿参数变化对系统性能产生的不利影响,控制效果也极差。故在建立燃气发生器与动力涡轮模型的基础上,采用递推增广最小二乘法和降阶处理进行参数的在线辨识,采用梯度法对延迟时间进行在线的辨识,将辨识的模型利用两级Smith预估控制进行实时闭环控制。仿真结果与理论一致,证明了辨识算法的正确性和有效性。上述算法与两级Smith预估控制共同应用于双闭环控制时,极大改善系统的抗干扰性、实时性,并提高了涡轴发动机的稳定性及鲁棒性。

The turbo engine consists of a gas generator and a power turbine. The gas generator can be regarded as a cascade time - varying system with a hysteresis due to the presence of factors such as inlet speed distortion and fuel ripple. For the case where the parameters of the controlled object are unknown and time - varying, the Smith algorithm can not effectively compensate the adverse effects of the parameter change on the system performance, and the control effect is also very poor. Therefore, on the basis of the establishment of the gas generator and the power turbine model, the on - line identification of the parameters was carried out by recursive least squares method and degenerate process. The gradient method was used to identify the delay time and the identified model was Smith predictions for real -time closed- loop control. The simulation results were in agreement with the theory, and the correctness and validity of the identification algorithm were proved. When the algorithm was applied to the double closed - loop control, the anti - jamming and real - time of the system were greatly improved and the stability and robustness of the turboshaft engine were improved.